Posters - Detailed Abstracts

1. Findings from the 2010 National Survey of Residential Care Facilities (NSRCF)

Christine Caffrey, Lauren Harris-Kojetin, Eunice Park-Lee, Manisha Sengupta, and Abigail Moss, National Center for Health Statistics; Emily Rosenoff, Assistant Secretary for Planning and Evaluation

Background: People living in state-regulated residential care facilities (RCFs)–such as residents of assisted living communities–receive housing and supportive services because they cannot live independently, but generally do not require the skilled level of care provided by nursing homes. The ability to provide a comprehensive picture of the long-term care industry has been hampered by a lack of data on RCFs.

Objective: Using data from the first nationally representative survey of RCFs with 4 or more beds, this poster presents national estimates of these RCF facilities and residents by select characteristics.

Methods: Resident data from the 2010 NSRCF were used for these analyses. Differences among subgroups were evaluated using chi-square and t-tests. All significance tests were two-sided using p < 0.05 as the level of significance. Data analyses were performed using the statistical package SAS-callable SUDAAN version 10.0 (9).

Results: In 2010, there were 31,100 facilities and 733,300 residents. RCF residents were mostly female, Non-Hispanic white, aged 85 years and older, and had a median length of stay of about 22 months. For about two-tenths of residents—or 137,700 people—Medicaid paid for at least some of their long-term care services provided by the RCF. Almost four-tenths of all residential care residents received assistance with 3 or more ADL limitations and over four-tenths had Alzheimer’s disease or other dementias.

Conclusions: These findings suggest a vulnerable population with a high burden of functional and cognitive impairment. This brief profile of residential care residents may provide useful information to policymakers, providers, and consumer advocates as they plan for the future long-term care needs of older as well as younger adults. In addition, these findings serve as baseline national estimates as researchers continue to track the growth and changes in the residential care industry.

2. Determinants of Emergency Room Visits in Seniors: Population-based Analysis

Joseph Finkelstein MD, PhD, Eunme Cha MPH

Background: Limited number of studies had explored the determinants of emergency room (ER) utilization by elderly patients on a population-based level.

Objective: In this study we assessed demographic, clinical, and behavioral determinants of ER visits in the elderly population.

Methods: Based on the National Health Interview Survey (NHIS) for the year 2010, a logistic regression model was run to assess the predictors of ER visits in elderly population. The number of ER visits during the past 12 months was dichotomized to ‘Yes’ or ‘No’. Demographic factors included age, sex, race, marital status, education, income level, US regions, insurance status, and total number of people in the household. Clinical determinants included presence of chronic conditions, Activities of Daily Living (ADL), and overall health status compared with 12 months. The behavioral determinants consisted of current physical activity, smoking and drinking status.

Results: The prevalence of ER utilization was 23.7% in the elderly population (>=65 years). Among the demographic factors, the oldest age group (>=80) was 1.5 time more likely to visit ER (95% CI 1.1-2.2, P=0.004) then the younger age group (65-69 yrs). The participants without high school education were 1.5 times more likely to visit ER (95% CI 1.0-2.1, P=0.04) than those with college education (>12 yrs). Among the clinical determinants, those with chronic conditions were 1.9 times (95% CI 1.2-3.0, p=0.009) more likely to visit the ER than those without co-morbidities. Those who needed assistance with ADL were 2.2 times (95% CI 1.3-3.9, P=0.004) more likely to visit the ER. Among the behavioral determinants, those who didn’t exercise showed 30% (95% CI 1.0-1.7, P=0.03) increased chance of going to the ER.

Conclusions: The predictors of ER visits in elderly population included the oldest age group, chronic conditions, no physical activity, no high school diploma, and assistance with ADL.

3. Trends in Hospitalization Associated with Atrial Fibrillation among the Elderly in the US

Adebayo Arowolaju II RN, MSHA; Richard F. Gillum, MD; Departments of Sociology and Medicine, Howard University

Background: Recent studies of death with a diagnosis of atrial fibrillation (AF) within the US showed a decline in the number and rate of death since 1999.

Objectives: We tested the hypothesis that the trend in hospitalization with AF would be similar to that for mortality in the US.

Methods: We examined data for 1999-2009 from the US National Hospital Discharge Survey for discharges with ICD-9-CM code 427.3 among up to seven coded diagnoses. We computed All-listed diagnosis numbers and rates per 100,000 for discharges among the civilian population.

Results: In 2009, the estimated number of all-listed diagnoses of AF in the US was 2,643,000; 1,330,000 (50%) in females, 2,155,000 (82%) at ages 65+ years and 467,000 (18%) first-listed diagnoses. At age 65+ years, the number increased from 2,049,000 in 1999 to 2,573,000 in 2005, then declined to 2,155,000 in 2009 (10% higher than 1999). The rate per 100,000 increased 17% from 5,984 in 1999 to 6,994 in 2005, then decreased 22% to 5,445 in 2009, 9% lower than 1999. The average annual growth rate was 3.12% from 1999 to 2005 compared to -5.00% from 2005 to 2009.

Conclusions: In contrast to steadily decreasing rates of death with AF, rates of diagnoses of AF at hospital discharge increased from 1999 to 2005, and then declined slightly between 2005 and 2009. Further research is needed to assess explanations for the decline in rates and discordance between 1999 and 2005 in rates of mortality and hospitalization.

4. Older Adult Access to Vision Rehabilitation Care in the U.S.

Lori L. Grover¹, Kevin D. Frick^2
1 Ophthalmology, Johns Hopkins University Wilmer Eye Institute, Baltimore, MD; ² Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD.

Purpose: Access to care is a concept that includes geographic accessibility to providers and facilities. A key access measure for patients with chronic vision impairment (VI) involves eye care professionals providing vision rehabilitation (VR) care. We lack data on where VI patients and VR clinicians are geographically located and if access to VR care is problematic. We aimed to create a VR access measure for U.S. older adults with chronic VI.

Methods: Data were analyzed to assess ratio of older adult Medicare population (65+ years) to U.S. VR clinicians (optometrists or ophthalmologists). Clinician data were compiled from national/state professional eye care organizations (AOA, AAO, AAOph) and aggregated at state level. Population data were drawn from the 2005-2010 Behavioral Risk Factor Surveillance System (BRFSS). VI prevalence by state using 2010 census data was used to calculate number of older adults with VI per VR physician and state-level potential access.

Results: A national VR clinician census was identified (n=987) and stratified by state. Self-reported VI population data in 18 states were counted; providers within those states were identified (n=446). Access for older adults to VR care varied across states, and identified clinicians within these states were primarily optometrists vs. ophthalmologists with overall provider ratio of 32:1 (national ratio 31:1). NY, OH, IA, and TX had highest population access; WV, AR, AL and NM the lowest.

Conclusions: The relationship between clinician access and improved health outcomes is well documented. However, limitations exist. VR clinicians may see patients from other states, there are no standards to compare to as what constitutes ‘adequate’ access, and higher access does not ensure uptake of VR care or account for potential barriers that may arise. This measure will further assist investigation into factors impacting care, realized access, and quality of care measures.

5. Socioeconomic predictors of mortality risk for adults with intellectual disability

Scott D. Landes

Background: Socio-economic and healthcare disparities have been identified by researchers as likely causes of increased morbidity and premature mortality for persons with intellectual disability. However, despite this insight, mortality research on this population has not explored the effect of income, education, and healthcare access on mortality risk.

Objective: The aim of the present study was to investigate whether education, income level, and healthcare access limitations had differential mortality risks for adults with intellectual disability living in the United States.

Methods: Data was analyzed from the 1997 to 2003 (n=1201) National Health Interview Survey linked with the National Health Interview Survey Linked Mortality Files (103 deaths) that match death certificate records from the National Death Index through 2006. Survival models estimated the effects of education, income level, and healthcare access limitations, as well as demographic and health related covariates, on mortality risk.

Results: While increased age and poor self-rated health increased the likelihood of mortality, education had a quadratic effect on mortality risk. Mortality risk for adults with intellectual disability decreased as level of education increased up to High School level, but then increased for post-High School education.

Conclusions: Similar to research on the general population of persons without intellectual disability, results reflect the benefit of education on mortality risk for persons with intellectual disability. However, for those adults with intellectual disability included in the 1997 to 2003 National Health Interview Survey, post-High School education appears to have diminishing salutory returns on mortality risk. These findings, coupled with the predictive effect of self-rated health, suggest possible benefits for increased public health education campaigns focused on adults with intellectual disability.

6. Vitamin Status Profile and Cognitive Function in a National Sample of Older Americans

A Naqvi, RF Gillum, J Kwagyan,TO Obisesan, Howard University College of Medicine, Washington, DC,

Background: In observational studies of vitamins and cognitive function, examination of nutrient profiles rather than single nutrients is gaining favor. We examined the association of a previously-reported vitamin status profile with memory in a large, national sample using objective serum measures of vitamin status.

Methods: In a national sample of 3128 persons aged 60 years and over examined in 1991-94, 2308 had complete data on serum folate (FOL), vitamin B12 (B12), beta-carotene (BC), vitamins C (C), D(D), E(E) and the Boston story-recall test (BST). A vitamin score (VS) was computed giving 1 point each for FOL, B12, C, D, E and BC > median (i.e., VS= 0, 1, 2, 3, 4, 5, or 6). Dummy variables were created to indicate low (0-2), medium (3-4) and high (5-6) VS for regression analyses.

Results: VS was 0 in 5.0%, 1 in 13.7%, 2 in 18.0%, 3 in 19.9%, 4 in 16.8%, 5 in 15.9% and 6 in 11.9%. Mean BST score showed an increase from VS=0 to VS=6 (i.e. 7.7, 7.4, 7.8, 8.4, 8.0, 8.6, 8.5). In a linear regression model adjusting for age, a low VS was significantly associated with lower BST score (coefficient = -1.13, p<0.001) compared to a high score and a test for trend was significant (p<0.001). After adjusting for age, sex, race and years of education, a low VS remained significantly associated with lower BST score (coefficient -0.58, p=0.006) compared to a high VS and a test for trend was significant (p=0.01). No interaction of VS with age or sex was seen.

Conclusion: VS was positively associated with BST, a measure of logical memory.

7. Trends in Hospitalization Associated with Alzheimer’s Disease in the US.

Ashley Pinette, MS; Richard F. Gillum, MD, MS; Thomas O. Obisesan, MD, MPH; Division of Geriatrics, Department of Medicine, Howard University College of Medicine

Background: Recent studies of trends in death with dementia within the US showed marked increases in the number and rate of death since 1999. We tested the hypothesis that the trend in hospitalization with Alzheimer’s disease (AD) would be similar to that for mortality in the US.

Methods: Examination of data for 1999-2009 from the US National Hospital Discharge Survey was done to find discharges with ICD-9-CM code 331.0 (AD) and codes 290.0-290.43 which correspond to Alzheimer’s disease related dementia (ADRD) among up to seven coded diagnoses. Discharge numbers and rates per 100,000 civilian population were computed.

Results: In 2009, the estimated number of all-listed diagnoses of AD in the US was 372,000; of which 259,000 (70%) were females, 363,000 (98%) were ages 65+ years, and 55,000 (14%) were first-listed diagnoses. The number of diagnoses with ADRD was 174,000. At age 65+ years, the AD rate per 100,000 was 917; the rate showed an increase of 1.9% per year in 1999-2004 and then a subsequent decrease of 4% per year in 2004-2009. The ADRD rate per 100,000 was 409; this rate showed decreases of 3% per year in 1999-2004 and 1% per year in 2004-2009.

Conclusions: In contrast to steadily increasing rates of death with dementia and AD between 1999 and 2009, rates of hospital discharge with AD increased from 1999 to 2004 and then decreased from 2004-2009 and rates of ADRD decreased from 1999 to 2009. More research is needed to assess explanations for the discordant trends in mortality and hospitalization.

8. Racial Variations in Office-Based Medical Care for Mental Health

Grishma Patel, MPH

Background: The U.S. Surgeon General’s report titled “Mental Health: Culture, Race, and Ethnicity–A Supplement to Mental Health” illustrated the racial and ethnic disparities found in mental health treatment.

Objective: This exploratory study evaluates the extent to which office-based treatment of mental health visits vary by patients’ race.

Methods: The 2006 – 2009 National Ambulatory Medical Care Survey (NAMCS) was used (n=123,192). Logistic regression analysis was conducted to evaluate the impact of the patients’ race on mental health visits, rate of diagnosis, type of treatment received (psychotherapy or other mental health counseling), and rate of psychotropic medication prescription, while controlling for age; gender; source of payment; geographic region; urban-rural classification; and percent poverty, median household income, and percent of adults with bachelor’s degree or higher in the patient’s zip code.

Results: It was determined that minorities have lower odds than their white counterparts to have a mental health visit (OR=0.6, 95% CI: 0.5-0.8). Within mental health visits, the provision of treatments and rate of diagnosis were not statistically different, after controlling for predisposing and enabling factors.

Conclusions: There is a continued need to develop innovative strategies to identify and address racial variations in the treatment of mental health.

9. Neighborhood Disorder and Mental Health among Baltimore City Residents

Hollisa Rosa¹, Adam J. Milam, MHS², Alexandra Duncan, DrPH, MPH, CPH², C. Debra M. Furr-Holden, PhD^2
1Johns Hopkins University; ² Johns Hopkins Bloomberg School of Public Health, Dept. of Mental Health

Background: While many factors impact mental health, environmental effects are understudied. Neighborhood disorder is essential for understanding mental health outcomes in communities.

Objective: This study examines the relationship between neighborhood disorder and mental health outcomes in a sample of Baltimore City residents.

Methods: Data from 711 Baltimore City participants were drawn from the Windows to Health Study (WtoH). The WtoH combines an in-person household interview on health status, and violence, alcohol and other drug exposure with a block face assessment of neighborhood disorder using the Neighborhood Inventory for Environmental Typology instrument. Binary logistic regression was used to assess the relationship between depression, emotional health, and neighborhood disorder.

Results: In the unadjusted regression model, alcohol and other drugs (AOD) had an independent association with depression (OR=1.11, p<0.05) and emotional health (OR=1.15, p<0.01). However, after adjusting for demographic characteristics of race, sex, gender, educational attainment, and employment status, results show that neighborhood disorder (AOD and violence) is not statistically significantly associated with mental health outcomes (depression and emotional health). In addition, findings suggest that the association between neighborhood disorder and depression is confounded by race, employment, and education.

Conclusions: These results suggest characteristics of the physical environment are not significant predictors of mental health status when demographic characteristics are considered. Future studies should expand covariates to include information on residents’ mental health status to increase the study scope and provide information for other possible factors associated with mental health status.

10. How Have Access and Cost Barriers to Mental Health Changed Over the Last Decade Among Insurance Groups? Findings from the Integrated Health Interview Series (IHIS)

Kathleen Rowan

Background: Recent years have seen important changes in the financing of mental health services. While changes in Medicaid eligibility and parity legislation have potentially increased access to care, the use of managed care for mental health services, increased cost sharing, and the decline in employer sponsored insurance may have reduced access. We update and extend previous research and discuss implications.

Objective: We address two research questions: 1) How has access to care for those with serious psychological distress changed over time within each insurance type (public, private, none)? 2) How have cost barriers to care changed over time within each insurance type?

Methods: Data come from the Integrated Health Interview Series. A mental health problem is defined as any serious psychological distress (SPD) measured by the Kessler-6 scale. Access is defined as visiting any mental health professional in the past 12 months. A cost barrier defined as forgoing mental health care in the past 12 months due to cost. Insurance is categorized as public (Medicare, Medicaid, other State programs), private, and no insurance. We limit the analysis to adults age 18 to 65 and use logistic regressions to test for trends in access and cost barriers for each insurance type between 1999 and 2010.

Results: Trends in insurance coverage for those with SPD paralleled those without SPD; statistically significant differences were observed for public coverage, with more persons without SPD gaining public coverage. Findings show no increase in access for any insurance type, but rising cost barriers among the uninsured and privately insured.

Conclusions: The increase in cost barriers for those with private insurance suggests that such plans leave many without needed care and are not sufficient benchmark plans for the Exchanges. The rise in un-insurance among those with SPD is particularly adverse because of the greater need for services.

11. Correlates of Psychological Distress in U.S. Adults during 2009-2010

Cory Blackwell

Background: Mental health disparities continue to exist in the United States, and remain a cause for concern as they may exemplify historical injustice, inequality, and even de facto racism, sexism, and/or classism. Individuals feeling such injustice and inequality may be more likely to be experience poorer mental health.

Objective: In order to better understand these disparities, I examine the joint associations between mental health and a number of explanatory variables, including demographic characteristics, socioeconomic status, and various health status indicators.

Methods: Using 2009-2010 National Health Interview Survey data, I utilize logistic regression to analyze the mental health status of adults by a number of characteristics including (but not limited to) age, sex, race/ethnicity, educational attainment, marital status, general health status, poverty status, and leisure-time physical activity. Mental health is measured by indicators of some or serious psychological distress (relative to none or little) and whether respondents felt that their emotional state interfered with their life in the past 30 days.

Results: Initial results indicate that women, non-Hispanic whites, adults without a college degree, unemployed adults, and especially those in fair or poor health were more likely to be in serious psychological distress. For example, the odds of being in serious psychological distress for adults in fair or poor health were 6.8 times the odds of being in such distress for adults in good, very good, or excellent health.

Conclusions: Given the strong relationship between one’s general health status and how he/she feels emotionally, doctors should be aware that sicker people may be more depressed or unhappy, and that these feelings can interfere with their lives on a daily basis. Such serious distress could even interfere with a patient’s ability to recover from illness. Medical doctors should work more closely with mental health professionals in treating patients.

12. The importance of financial health in predicting mental and physical health outcomes

Jessica L. Sloan & Daniel J. Kruger

Background: Numerous studies document the inverse relationship between socioeconomic status and health. Previous research has also shown that unemployment and job insecurity are associated with adverse health outcomes. However, these studies typically utilize samples of individuals seeking mental health care and do not account for retired individuals or those experiencing economic hardships despite stable employment. The recent economic recession provides an opportunity to examine the relationship between financial adversity and health outcomes.

Objective: We examine the relationships between subjective financial health, declines in subjective financial health, and mental and physical health outcomes with a demographically representative county-wide survey conducted in Spring 2009.

Methods: We assessed the relationships with linear regressions for 733 adult participants. We included general health and mental health items from the Behavioral Risk Factor Surveillance System (BRFSS), Brief Symptom Inventory (BSI-18) depression subscale, and Perceived Stress Scale as outcome variables. We used the item “My financial situation is much worse this year than it was in the previous year” and a 6-item subjective financial health scale (Cronbach alpha = .836) as predictors, along with socio-demographic variables.

Results: The degree to which an individual’s financial situation declined over the past year was associated with worse self-reported general and mental health, increased number of days that poor general and mental health interfered with daily activities in the past month, as well as higher levels of self-reported stress and depressive symptoms. These relationships were independent of education, income, age, sex, and minority status. Subjective financial health was the strongest predictor of mental and physical health outcomes.

Conclusion: Declines in financial status are associated with adverse self-reported physical and mental health independent of traditional demographic and socioeconomic indicators. These effects were observed in a community experiencing decades of economic hardship in addition to the recent economic crisis.

13. Predictors of discharge planning in criminal justice settings: Implications for reentry of offenders with co-occurring substance use and mental health disorders

Hiroki Toi, MSW; Doctoral Student, University of Connecticut School of Social Work

Background: Offenders with co-occurring disorders (COD) present significant challenges to both criminal justice and community-based treatment settings. While concern for this population is increasing, little is known about discharge planning for the population and factors that affect the planning.

Objective: Research question for the study is: what are the predictors of the degree of discharge planning for offender with COD in the criminal justice system?

Methods: Secondary dataset using a sample of administrators of correction, probation, and outpatient substance abuse treatment programs (n = 147) from the National Criminal Justice Treatment Practices Survey of Co-Occurring Substance Use and Mental Disorder Treatment Services in Criminal Justice Settings is used for the analysis. Medication is defined as provision of medication upon release. Community support is assessed by support from communities and collaboration with external groups. Resources are assessed by follow-up services and staff positions in discharge planning. Standard multiple regression was used to test the five-factor model.

Results: Preliminary analyses were conducted to ensure no violation of the assumptions of normality, linearity, multicollinearity and homoscedasticity. After entering the five factors, the total variance explained by the model as a whole was 46.2%, F (5, 113) = 19.42, p < .001 with a 95% confidence interval. The result indicated that three factors were significant (medication beta = .43, p < .001; staff positions beta = .31, p < .01; collaboration with external groups beta = .16, p < .05), whereas two factors, support from communities and follow-up services, were not significant in the model.

Conclusions: Although prior studies showed that community support and resources were positive predictors of offender’s access to treatment services, the results suggested that medication was the best predictor for discharge planning over other predictors. A critical examination of implications for reentry of offenders with COD will be discussed.

14. Prevalence of hazardous and harmful drinking patterns among older US adults: data from the 2005-2008 NHANES.

Sandra R. Wilson,¹ Sarah B. Knowles,¹ Qiwen Huang,¹ Arlene Fink^2,3
1 Department of Health Services Research, Palo Alto Medical Foundation Research Institute, Palo Alto, CA; ² David Geffen School of Medicine, School of Public Health, University of California at Los Angeles, Los Angeles, CA; ³ Langley Research Institute, Pacific Palisades, CA.

Background: Even moderate alcohol consumption may be risky in older adults with certain co-morbidities, medication use, and functional limitations. However, national estimates of such risk are presently unavailable.

Objective: To estimate the prevalence of hazardous and harmful alcohol use in U.S. adults aged 65 years and older and the relative odds of at-risk and harmful drinking by sex, age, and race/ethnicity.

Methods: Reported alcohol consumption of 2005-2008 National Health and Nutrition Examination Survey respondents was considered in the context of the individual’s co-morbidities, medication use, and physical and psychological functional status to classify their consumption as Healthwise, Hazardous (at risk of alcohol-related health problems), or Harmful (already experiencing alcohol-related health problems), using the classification algorithm of the Alcohol-Related Problems Survey (ARPS), a screening tool for older adults. Weighted prevalence estimates and adjusted prevalence odds ratios were calculated using survey statistical procedures.

Results: Among seniors who reported drinking, 57% drank 2-4 times per month or less and 63.5% consumed less than one drink on the days they drank. However, in relation to their co-morbidities, medication use, and functional status, the alcohol consumption patterns of 53.3% (nearly 9 million persons) were classified as Harmful (37.4%, 95% Confidence Interval [CI]: 34.9%, 40.0%) or Hazardous (15.9%, 95% CI: 13.4%, 18.4%). Men had more than twice the odds of women of being Hazardous or Harmful drinkers (Odds Ratio [OR]=2.14, 95% CI: 1.77, 2.60). Compared to non-Hispanic whites, non-Hispanic blacks were at increased odds of risky drinking (OR=1.49, 95% CI: 1.02, 2.17) and even greater odds of Harmful drinking (OR=1.83, 95% CI: 1.03, 3.26).

Conclusions: Most Americans aged 65 and older drink less than the current recommended maxima, yet substantial proportions drink in a manner that is harmful considering their co-morbidities, medication use, and functional status. Alcohol misuse constitutes a major but neglected public health problem.

15. The Association between Receipt of an Antidepressant Prescription and Weight Status

Devin Sanders MSCR, Kisha Hill MSCR, Andrew Muzyk PharmD, Melissa Johnson PharmD, MHS

Background: Studies have suggested an association between obesity and depression. Additionally, antidepressants can cause weight gain, which may influence prescribing of these medications.

Objective: To determine (A) if there was an association between an antidepressant prescription and weight status and (B) if there was an association between weight status and a diagnosis of depression or anxiety.

Methods: This retrospective, cross-sectional study sampled 28,603 adults (≥18 years of age) that presented to clinics participating in the National Ambulatory Medical Care Survey (NAMCS) from 2006-2008 and prescribed an antidepressant. Weight status was determined by Body Mass Index (BMI).

Chi square analysis was used to test the association between weight classification and (a) the proportion of visits prescribed an antidepressant and (b) the proportion of visits with a diagnosis of anxiety/depression. The normal weight category was the referent group for all analyses. Multivariate logistic regression was conducted to control for factors such as age, sex, race, and ethnicity.

Results: There was an association between receipt of an antidepressant prescription and weight status for underweight subjects (OR 1.74 (95% CI 1.19-2.55) and obese subjects (OR 1.18 (95% CI 1.04-1.35) but not overweight individuals (OR 1.01 (95% CI 0.88-1.15) on chi-square analysis. In multivariate analysis, there was an association between prescription for an antidepressant and weight classification (normal vs. not normal), when controlling for age, sex, race, and ethnicity. In secondary chi square analysis, there was an association between obese status and diagnosis of anxiety/depression. Only age and race were predictors of anxiety/depression diagnosis.

Conclusions: We observed a U-shaped association between receipt of an antidepressant prescription and weight status, with underweight and obese populations being more likely to receive an antidepressant. Thus, it appears anxiety/depression is being addressed in subjects with these weight classifications.

90. Premigration Harm and Depression: Findings from the New Immigrant Survey, 2003

Michelle Montgomery, MPH student
Elizabeth Kelvin, Assistant Professor, CUNY School of Public Health at Hunter College, Epidemiology & Biostatistics

Background: Studies examining refugees from conflict areas have found that persecution in the place of origin is a risk factor for depression. Few studies have looked at the association between mental health and the experience of premigration harm due to race, gender or religion in the general population of US immigrants.

Objective: To examine how the experience of premigration harm relates to depression in general and depression specifically related to the legalization process among immigrants to the USA.

Methods: We used data from the New Immigrant Survey (NIS) baseline questionnaire, which was administered to a random sample of adults receiving legal permanent residency in the US between May and November 2003 (n=8,573). Logistic regression was performed to assess the association of premigration harm with general depression and depression related to the legalization process. Potential confounders examined included gender, educational status, region of origin, and visa type (refugee/asylee/parolee versus other).

Results: In multivariate analysis, premigration harm was a predictor of general depression of borderline statistical significance (odds ratio [OR], 1.33; 95% CI [confidence interval] 0.98-1.80, p=0.068) and a significant predictor of depression related to the legalization process (OR, 1.45; 95% CI 1.11-1.89, p=0.007). Those with a refugee/parolee/asylee visa had significantly higher odds of reporting depression (OR, 1.47; 95% CI 1.10-1.97, p=0.009). There was no evidence that the association between premigration harm and depression was modified by visa type.

Conclusions: Our results suggest that persecution in one’s country of origin is a risk factor for depression related to the legalization process. The association did not differ between those on refugee/asylee/parole visas versus other visa types. Overall, these findings support existing research on the negative effects of persecution on mental health. Future policies should consider the benefits of screening all immigrants for premigration harm irrespective of refugee status to better identify depression risk.

16. Did Changes in the U.S. Centers for Disease Control´s Recommendations for HIV Testing Change the Characteristics of Adults Tested for HIV?: A Quantitative Analysis of Persons Tested under Three Policy Regimes, Based on Data from the U.S. National Health Interview Survey, 1988-2010

Ryan Moltz (student), Christine Gille Kunitz (student), and Miriam L. King; Minnesota Population Center, University of Minnesota

Background: CDC has provided increasingly inclusive recommendations for HIV testing. In Regime 1 (1988-1992), CDC emphasized testing persons “most likely to be infected.” In Regime 2 (1993-2005), CDC extended testing guidelines to “hospitalized patients and persons obtaining health care as out-patients in acute-care settings.” In Regime 3 (2006-present), CDC recommended “routine voluntary HIV screening as a normal part of medical practice.”

Objective: We evaluate whether HIV-tested adults differed significantly (95% confidence level) under the 3 HIV Testing Regimes.

Methods: Using National Health Interview Survey data for 1988-2010, we classified adults who ever had an HIV test as being tested under Regime 1, 2, or 3, based on year of last HIV test (N=103,844). We used multinomial logistic regression to determine how persons tested in Regimes 1 and 3 differed from those tested in Regime 2. Using descriptive statistics, we evaluated differences across Regimes in reasons for testing, and whether persons with doctor visit or hospitalization in past year had been recently tested.

Results: Across Regimes, characteristics of those tested changed (in race/ethnicity, social class, access to health care and health status). Those who evaluated their risk of contracting HIV as “medium/high” were 31 percent more likely to be tested in the earliest period, and 18 percent less likely to be tested in the third period, than in the middle period. Disadvantaged groups (women, African-Americans, poor) gained greater access to testing, and routine medical care increasingly predominated as the reason for testing. However, most of those receiving routine care in the past year were not tested.

Conclusions: Changes between 1988 and 2010 in the type of persons tested and reasons for testing are largely consistent with CDC recommendations. However, although most people tested are now in the health care system, most people in the health care system are still not tested.

17. Community Health Center Effects on County Level Mortality Rates in U.S. Medically Underserved Areas, 2000 – 2008

Joseph Benitez, MPH; Seijoung Kim, PhD; Richard Barrett, PhD; Division of Health Policy and Administration, School of Public Health, University of Illinois at Chicago

Background: In 2010, the passage of the Affordable Care Act (ACA) called for a massive expansion of the Health Resources and Services Administration (HRSA) Community Health Center (CHC) program. CHCs fill considerable gaps in access to healthcare services for over 4 million people living in medically underserved areas (MUA) across the U.S. In 2000 there were 982 medically underserved U.S. counties that were without a CHC, of which 223(23%) received a CHC between 2001 and 2008.

Objective: CHC effects on local infant mortality rates (MR) as well as prenatal care services is known and positive; however, CHC effects on community level MR and otherwise preventable deaths have been underexplored—the primary objective of this study is to evaluate the effects of new CHCs on county level MRs across time.

Methods: The major sources of data include age-adjusted county mortality data from the CDC Wonder website and administrative data from the HRSA data warehouse on MUAs and CHCs. OLS regression models with robust clustered standard errors were used to estimate the treatment effect. Methods to estimate this causal effect include: first differencing (FD) and Difference-in-Difference (DiD).

Results: In the FD and DiD models, establishing a new CHC was, though not statistically significant, correlated with a reduction in the county MRs except for diabetes related mortality. Model statistics for the DiD estimations ranged from F=7.5(p<0.0001, R-sq=0.1059) for Cancer to F=100.96(p<0.0001, R-sq=0.3436) for Heart Disease.

Conclusions: While CHCs improve access to care among the underserved, we cannot say with confidence that past expansions in the program caused reductions in county MRs. Trends observed in our analysis of past CHC expansions suggest there may be a minor causal effect on the outcome, it should not be expected that the new expansion in CHCs will result in significant years of preventable lives lost.

18. Secondhand Smoke on the Johns Hopkins University Homewood Campus.

Jackie Ferguson, Erika Avila-Tang, Lisa Hepp, Patrick Breysee.

Background: Exposure to secondhand smoke (SHS) has been directly linked to numerous adverse health outcomes.

Objectives: The study objectives were to 1) assess the effect of the current smoking policy on the Homewood campus SHS exposure inside and outside of non-smoking buildings and 2) to investigate the ability of SHS to seep from outdoor locations where smoking is allowed to indoor non-smoking locations.

Methods: Data was collected from three study sites in three seasons in 2011. The sites included the Student Health and Wellness Center (summer), two freshmen dormitories (fall), and the library and an academic hall (winter). SHS concentrations in indoor and outdoor locations were assessed by measuring air nicotine and particulate matter 2.5μ (PM2.5) using passive air nicotine monitors and a TSI SidePak AM510 Personal Aerosol Monitor, respectively. Monitors were strategically placed to evaluate the seepage pathways of SHS.

Results: We found that despite the presence of an indoor smoking ban, air nicotine was detected, on average, in 46% of the indoor locations studied (87.5% of dorm rooms monitored, 40% of indoor library locations, and 14% of academic indoor locations). Only one indoor location in theacademic hall tested positive for airborne nicotine. Yet, every outdoor location monitored was positive for nicotine. PM2.5 was found in all locations, with higher concentrations in locations with closer proximities to outdoor smokers. Simultaneous testing of PM2.5 showed real-time seepage of outdoor PM2.5 to indoor air through doorways.

Conclusions: Despite the enforcement of the indoor non-smoking policy, the ability to smoke directly outside of buildings causes smoke to seep through entryways into the building. As result, during their years study at the Homewood campus, Hopkins students are potentially exposed to SHS. We recommend to the Johns Hopkins University to make Homewood a smoke-free campus to protect their students from SHS exposure.

19. Secondhand Smoke on the Johns Hopkins University Homewood Campus.

Marcella Thompson

Background: Lead, mercury and polychlorinated biphenyls (PCBs) are neurotoxicants with intergenerational health consequences from maternal body burden and gestational exposures. Little is known about multiple chemical exposures among childbearing-aged women.

Objectives: 1. What is the percentage of childbearing-aged women who have all three xenobiotics at or above the median in their blood/serum? 2. What is the most common binary chemical combination based on body burden? 3. Which women were disproportionately burdened based on vulnerability factors?

Methods: This descriptive study involved analysis of existing data from the National Health and Nutrition Examination Survey (NHANES), a national probability sample. The outcome of interest was based on evidence of the biological uptake of lead and mercury in blood and select PCBs in serum of these women aged 16 to 49 of diverse races and ethnicities who were living in the U.S. 1999 to 2004. The scientific literature guided the selection of 62 measures of vulnerability (susceptibility- and exposure-related attributes, socioeconomic factors and race-ethnicity).

Results: More than one fifth of childbearing-aged women had all three xenobiotic levels at or above the median. Lead-PCBs was the most common binary chemical combination. The best-fit logistic regression model without interactions contained 12 variables (R=.25). Four risk factors were notable (p≤ 0.05). An exponential relationship was demonstrated with increasing age. Any fish consumption in past 30 days more than doubled the odds. Heavy alcohol consumption or binge drinking increased risk. A history of breastfeeding children lowered their risk as did occasional alcohol consumption.

Conclusions: These findings are the first description of multiple chemical exposures among U.S. childbearing-aged women. This study further supports increasing age, fish consumption and heavy alcohol consumption as significant risk factors. Prior history of breastfeeding lowered their risk. Some evidence was found of increased body burden among minority women but this relationship was not stable.

20. The 2010 Occupational Health Supplement to the National Health Interview Survey (NHIS-OHS)

Sara Luckhaupt and Taylor Shocke

Background: Integration of occupational health into general disease surveillance systems supplements data available from the Bureau of Labor Statistics Survey of Occupational Illness and Injury, workers’ compensation data, and other “traditional” occupational health surveillance systems.

Objective: To describe the prevalence of work-related health conditions and exposures to potential psychological and physical occupational hazards in the US working population.

Methods: The National Institute for Occupational Safety and Health (NIOSH) sponsored an occupational health supplement (OHS) to the 2010 National Health Interview Survey (NHIS). Sample adults who had worked at least part of the 12 months preceding their interviews (N=17,524) received the OHS questions. We calculated prevalence rates of key work-related health conditions and exposures among these current/recent workers, stratified into 8 broad industry sectors of employment, as defined by the National Occupational Research Agenda.

Results: The 12-month prevalence of dermatitis varied from a low of 5.5% among workers in the Agriculture, Forestry, & Fishing sector to a high of 11.9% among workers in the Healthcare and Social Assistance sector. The prevalence of current carpal tunnel syndrome varied from a low of 1.9% among Transportation, Warehousing, & Utilities to a high of 5.6% among Agriculture, Forestry, & Fishing. Non-standard work arrangements varied from a low of 7.8% among Mining to a high of 44.1% among Construction. Job insecurity varied from a low of 22.8% among Agriculture, Forestry, & Fishing to a high of 54.7% among Construction. Workplace exposure to secondhand smoke among non-smokers varied from a low of 6.9% among Agriculture, Forestry, & Fishing to a high of 28.4% among Mining.

Conclusions: The prevalence rates of work-related health conditions and exposures to potential occupational hazards vary widely by industry of employment. Collecting this type of data in population-based surveys facilitates the development of evidence-based interventions to reduce work-related illnesses and injuries among specific groups of workers.

21. Patterns of Mortality in New Hampshire by Industry and Occupation

Mary Butow¹, M.S., Michael Laviolette², Ph.D., M.P.H., JoAnne Miles², M.P.H., Sai Cherala², M.D., M.P.H, David Kriebel¹, Sc.D., Pamela Schumacher³, Karla Armenti², Sc.D.
¹ University of Massachusetts-Lowell; ² New Hampshire Department of Health and Human Services; ³ National Institute for Occupational Safety and Health

Background: As the employment landscape in New Hampshire (NH) has changed from an agricultural to a service economy, the need has arisen to study possible associations between emergent technologies (e.g. use of chemicals) and occupational mortality.

Objective: This study seeks to determine patterns of mortality among occupations and industries in NH by the following means: (1) quantitatively by identifying the distribution, and excess of, mortality for NH residents by occupation and industrial sector, and (2) qualitatively by comparing results to the first comprehensive study of NH mortality by occupation and industry conducted in 1986 for data from 1975-1985.

Methods: Occupation, industry, age, date and cause of death information was abstracted from the death records of all NH resident decedents (aged 15 years or older) for a 5 year period (2002-2005 & 2007). The National Institute for Occupational Safety and Health (NIOSH) coded the occupation and industry text to Census 2000 Occupational and Industry codes. Standardized proportionate mortality ratios (SPMR’s) were calculated for each combination of occupation/industry and selected ICD-113 cause of death. Statistically significant results based on 95% confidence intervals were determined and qualitatively compared to proportionate mortality ratios from the historical study.

Results: Selected key findings include elevated risk of ovarian cancer in elementary and middle school teachers (SPMR = 2.80, 95% CI = 1.93-3.95), all other and unspecified malignant neoplasms in ship & boat builders (2.35, 1.60-3.36), and accidental poisoning and exposure to noxious substances for workers in the construction industry (1.45, 1.20-1.73).

Conclusions: Death certificate analysis generated promising leads for further occupational health studies. Future projects include both intervention-based (e.g. poison control and awareness) and research-based (e.g. geographic assessment of excess risk in ovarian cancer) work.

22. Paid Sick Leave and Nonfatal Occupational Injury*

Abay Asfaw, Regina Pana-Cryan, and Roger Rosa(Centers for Disease Control and Prevention, National Institute for Occupational Safety and Health)

Objectives. Despite the advantages of paid sick leave (PSL) for both employees and employers, the number of private sector employees who have access to it remains low. For example, between 1996 and 1998 nearly 90% of workers in state and local governments had access to PSL compared to only 45% of workers in the private sector. Additional empirical evidence about the advantages and costs of PSL would help inform employer decisions about offering or expanding PSL to employees. We examined the association between worker access to PSL and the incidence of nonfatal occupational injuries in the U.S. from the employer’s perspective. We also examined this association in different industries and occupations.

Methods. We developed a theoretical framework to examine the business value of offering PSL. The 2005-2008 National Health Interview Survey was used to empirically test the hypothesis that offering PSL is associated with reduced incidence of occupational injuries. We estimated a multivariate logit model using data on 38,000 working adults.

Results. Holding all other variables constant, the likelihood of workers with access to PSL to be injured was 28% (95% CI = 0.52-0.99) lower than that of workers without access to PSL. The association between the availability of PSL and the incidence of occupational injury varied across sectors and occupations, with the greatest differences occurring in high-risk sectors and occupations. For instance, ceteris paribus, a construction worker without access to PSL had a 21% higher predicted probability of suffering a nonfatal occupational injury than a construction worker with access to PSL.

Conclusions. These findings suggest that, similar to other investments in worker safety and health, introducing or expanding PSL programs might help businesses to reduce the incidence of nonfatal occupational injuries, particularly in high-risk sectors and occupations.

* The paper is accepted for publication by the American Journal of Public Health

23. Characteristics of Vulnerable Users’ Deaths in Brazilian Road Safety 10 Cities (RS-10)

Veralice Gonçalves¹, Birchie Erin Seiffert², Tanara Souza¹, Aruna Chandran², Flávio Pechansky^1
1 Center for Drug and Alcohol Research, Federal University of Rio Grande do Sul, Brazil
² International Injury Research Unit, Johns Hopkins Bloomberg School of Public Health

Background: In Brazil since 2006, deaths among vulnerable road users (VRUs) have made up more than half of all traffic-related deaths, with 51.1% of all deaths in 2010 occurring among motorcyclists, pedestrians, and bicyclists. We investigated mortality rates among motorcyclists and pedestrians in five state capitals targeted by a global road safety initiative.

Objective: To evaluate road traffic fatalities among VRUs in five Brazilian cities from 2000 to 2010, describing demographic trends among victims.

Methods: Mortality data from 2000 to 2010 for the five cities were obtained from the Ministry of Health. Road traffic deaths were compiled using ICD-10 codes (V01-V89), subcategorized by VRU type and disaggregated by gender, age, and city.

Results: The five cities considered represent 18.4% of total deaths occurred in capitals from 2000 to 2010. Mortality rates varied from 25.0 per 100,000 (Belo Horizonte) to 42 (Palmas), all above the national rate of 21.5. Motorcyclist deaths increased by 288.4%, while pedestrian deaths fell by 5.6% over the decade. The highest pedestrian death rate was observed in Belo Horizonte (9.7 per 100,000). The highest rate of deaths among motorcyclists occurred in Palmas with 13.9 deaths per 100,000. Disaggregated by age, VRUs between 61 and 80 died at the highest rates.

Conclusions: Given high proportion of road traffic fatalities that occur among VRUs, policy implications can be gleaned from the results. Brazil’s transit authorities vary from between states, so analysis of the burden of VRU deaths between capitals can be helpful in targeting interventions towards appropriate risk groups.

24. Chronic obstructive pulmonary disease, hospital visits, and comorbidities – National Survey of Residential Care Facilities, 2010.

Wheaton AG, Croft JB

Background: National prevalence estimates of chronic obstructive pulmonary disease (COPD) and other chronic diseases are often based on data from surveys that include only civilian noninstitutionalized adults. Residents of long-term care facilities are excluded from these surveys. An estimated 733,000 individuals nationwide lived in residential care facilities (RCFs) in 2010.

Objective: To characterize the prevalence of COPD among residents RCFs in the United States, as well as to compare patterns of hospital visits and comorbidities to residents without COPD.

Methods: Data regarding demographics, chronic health conditions, and hospital stays in the previous 12 months for 8,094 residents of RCFs were obtained from interviews with RCF staff in the 2010 National Survey of Residential Care Facilities (NSRCF). COPD included responses of emphysema, chronic bronchitis, or COPD when asked about residents’ medical history. Weighted estimates were obtained using SAS-callable SUDAAN.

Results: Overall prevalence of any COPD was 12.4%. COPD was more common among men and divorced/separated individuals and less common among residents aged ≥85 years. Compared to residents without COPD, at least one visit to an emergency department (42.5% vs. 33.4%) or an overnight hospital stay (31.6% vs. 22.7%) in the previous 12 months were more prevalent (p<0.05) among residents with COPD. Less than 3% of residents with COPD had no comorbidities. Among residents with COPD, several comorbidities were more common than among those without COPD (arthritis (33.7% vs. 26.2%), depression (31.5% vs. 26.8%), chronic heart failure (26.4% vs. 11.3%), diabetes (25.3% vs. 16.0%), coronary heart disease (18.0% vs. 12.6%), and asthma (14.5% vs. 2.7%)), but not Alzheimer’s disease (33.7% vs. 43.0%).

Conclusions: COPD is associated with more emergency department visits and overnight hospital stays, as well as various chronic diseases, among RCF residents. The NSRCF is a valuable source of health data on this understudied population.

25. Chronic Obstructive Pulmonary Disease and Movement Limitations in Adults – National Health Interview Survey (NHIS), 2008-2010

Kosacz N, Wheaton A, Croft J.

Background: Over 12 million Americans currently live with chronic obstructive pulmonary disease (COPD). It is a significant cause of hospitalization and negatively affects quality of life. Little is known about how COPD contributes to functional impairment among adults in the general population.

Objective: To examine associations between COPD and movement limitations.

Methods: We analyzed data for 36,897 adults aged ≥ 45 years who participated in the National Health Interview Survey (2008-2010). COPD status was based on self-report of emphysema diagnosis (ever) or current chronic bronchitis (previous 12 months). Participants were asked about difficulty performing several physical tasks without special equipment. Adjusted prevalence ratios (PR) and 95% confidence interval (CI) for COPD-associated movement limitations were calculated with SUDAAN using logistic regression that included sex, age, race/ethnicity, smoking status, and body mass index.

Results: Prevalence of COPD among adults aged ≥ 45 years was 8.2%. Compared to those without COPD, more adults with COPD reported each movement difficulty: lifting/carrying 10 lbs (PR= 2.7; 95% CI: 2.48, 3.27), grasping small objects (PR=2.4; 95% CI: 2.11, 2.74), pushing large objects (PR=2.5; 95% CI:2.33, 2.73), reaching overhead (PR=2.5; 95% CI: 2.26, 2.84), sitting 2 hours (PR=2.4; 95% CI: 2.14, 2.64), standing 2 hours (PR=2.2; 95% CI: 2.05, 2.32), stooping/kneeling (PR=2.0; 95% CI: 1.89, 2.14), and walking a quarter mile (PR=2.4; 95% CI: 2.23, 2.96). The most commonly mentioned causes of movement limitations among participants with COPD included arthritis (44.6%), lung/breathing problems (33.3%), and back/neck problems (31.7%).

Conclusions: All movement limitations were more common among adults with COPD than among those without COPD. More research is needed to elucidate the extent to which COPD may cause and exacerbate movement limitations.

26. Clustering and Determinants of COPD in Central Appalachia

Nicole Johnson, Timothy S. Hare, and Chad Wells

Background: We explore the regional clustering of chronic obstructive pulmonary disease (COPD) mortality and associated social, economic, and environmental factors across central Appalachia. Identification and characterization of the diverse set of factors contributing to COPD clusters is necessitated by the complex interaction of determinants.

Objective: What are the relative contributions of specific factors contributing to COPD mortality and how do the effects of these factors vary across the study area?

Methods: This is an ecological study that uses county-level aggregated data for a wide variety of factors to explore local or neighborhood effects on COPD mortality. We analyze these data using a combination of regression, spatial regression, and geographically weighted regression (GWR). Spatial regression controls for the process of geographical diffusion and minimizes spatial error in the regression coefficients and significance. GWR evaluates the potential for geographical variation in the effects of different factors on COPD mortality.

Results: Our analysis shows that rates of tobacco use and proximity to coal mining do not adequately explain the presence of clustered high rates of COPD in some areas of the region. In addition, the effects of some factors vary in different parts of central Appalachia.

Conclusions: Tobacco use and proximity to coal mining are primary determinants of COPD mortality in some areas, but not in others. In particular, other factors have greater effects in the poorest areas of central Appalachia.

27. Trends in Hospitalization with Pulmonary Hypertension

Orlando Valle, MS; Alem Mehari, MD; Howard University, Washington, DC

Background: Pulmonary hypertension, which is listed on hospital records and death certificates as either primary (i.e., idiopathic) pulmonary hypertension or pulmonary hypertension secondary to another underlying condition or disease, is considered rare amongst the U.S. population. Few reports have been published regarding surveillance data for this debilitating and often fatal condition.

Objective: To assess trends in death and hospitalization with pulmonary hypertension in the US.

Methods: For the years 1999-2008, we examined multiple causes of death mortality data from the National Vital Statistics System (NVSS) and hospital discharge data from the National Hospital Discharge Survey (NHDS) for 1999–2009. Because pulmonary hypertension might be reported secondary to other diseases, this report presents data for pulmonary hypertension as any contributing cause of death or any-listed hospital diagnosis.

Results: Since 1999, the number of deaths and hospitalizations, as well as death rates and hospitalization rates has increased for pulmonary hypertension, particularly among women and older adults. During 1999 death rates were higher for men than women; however, by 2002, no difference in rate was observed because of increasing death rates among women and declining death rates among men; after 2003 higher death rates were observed for women. Death rates throughout the reporting period 1990-2008 have been higher for blacks than for whites. In addition, a higher rate of hospitalization was observed in the Northeast than in other regions of the United States.

Conclusions: Increases in rates of mortality from and hospitalization for pulmonary hypertension were observed in the US in 1999-2009.

28. Analysis of NAMCS and NHAMCS data for chronic fatigue syndrome (CFS), 2000-2009

Jaeyong Bae§^1,2, Jin-Mann S. Lin¹
§Presenting author
¹ Chronic Viral Diseases Branch, Division of High-Consequence Pathogens and Pathology, National Center for Emerging and Zoonotic Infectious Diseases, Centers for Disease Control and Prevention. 1600 Clifton Rd. NE, Mailstop A-30, Atlanta, Georgia 30333
² Department of Health Policy and Management, Rollins School of Public Health, Emory University, Atlanta, Georgia

Background: CFS is a disabling illness with substantial economic burden and functional impairment comparable to heart disease and multiple sclerosis. Many patients with CFS do not receive appropriate healthcare, partially due to rule-out-conditions and physicians’ lack of knowledge/attitudes/beliefs about CFS.

Objective: Surveys such as National Ambulatory Medical Care Survey (NAMCS) and National Hospital Ambulatory Medical Care Survey (NHAMCS) could be useful in profiling demographics, symptom complaints and healthcare for CFS patients and their co-morbidities.

Methods: We examined 9.06 billion (weighted) adult visits from 2000-2009 NAMCS/NHAMCS data. CFS-related visits were identified by ICD-9-CM code, 780.71, up to tertiary diagnosis.

Results: We estimated 2.9 million (95% CI: 1,844,391 – 3,944,931) CFS-related visits during 2000-2009, with no statistical evidence for an increasing trend across years. Patients with CFS-related visits were mostly aged 40-49 years (31.44%), female (66.07%), white (86.95%), metropolitan/urban residents (92.05%), and insured (87.26%). Apart from fatigue, the top four symptoms were muscle/joint pain (14.62%), headache (3.85%), post-exertional malaise (2.31%), and sore throat (2.31%). About 68% of CFS patients had co-morbidities, including depression (37.29%), arthritis (25.42%), and hypertension (23.73%). Physicians spent more time for CFS-related visits than non-CFS visits (26.96 vs. 21.29 minutes). The most common counseling provided to CFS patients was diet/nutrition (11.30%), stress management (8.47%), exercise (6.09%), smoking cessation (5.08%), and weight reduction (3.09%). Stress management was more prevalent in CFS-related visits than non-CFS visits.

Conclusion: CFS-related visits did not increase from 2000 to 2009. A high proportion of CFS patients had one or more co-morbid conditions, adding to its complexity. While physicians spent more time with CFS patients, a low proportion of CFS patients received counseling for diet/nutrition, exercise, and weight reduction given their complex illness. The NAMCS/NHAMCS data are useful in evaluating co-morbidities, healthcare utilization and quality indicators (e.g. preventive counseling) for healthcare among CFS patients.

29. Rising Incidence and Rates of Opioid Medication Prescription for Regional Body Pain in the Emergency Department

Ming-Chih J. Kao¹, Lyly Cao Minh¹, Grace Y. Huang², Jeffrey K. Teraoka³, Matthew Smuck^1
1 Department of Orthopaedics, Stanford University Medical School, Palo Alto, CA
² Department of Orthopaedics, University of California San Francisco Medical School, San Francisco, CA
³ Department of Physical Medicine and Rehabilitation, Palo Alto VA, Palo Alto, CA

Background: Opioid medications have come under much scrutiny in recent years. In the Emergency Department (ED) setting, we previously demonstrated increasing low back pain patients from year-to-year, who are increasing more likely to receive opioid medication prescriptions.

Objective: To describe trends in the prescription of opiate medications for regional body pain in EDs in the United States.

Methods: National Hospital and Ambulatory Medical Center Survey (NHAMCS) 1996-2008 data was obtained from the National Center for Health Statistics. Data was analyzed using SAS version 9.2 and R. For variance estimation, masked “ultimate cluster” weights were used for 2002-06. Domain analysis was employed in subset analyses. We estimated the incidence of regional body pain as chief complaint in United States EDs for various regional body pains: axial (headache, neck, back, low back), upper extremity (shoulder, arm, elbow, wrist, hand), lower extremity (hip, leg, knee, ankle, foot); as well as rates of opioid medication prescription for each regional body pain.

Results: From 1996 to 2008, there is a progressive year-to-year increase in most regional body pain complaints, including the shoulder, hand, hip, knee, leg, foot, head, back, and low back (2-9% per year, p<0.01), in particular headache, back, and leg pains. Except for the elbow and the head, opioid medications have demonstrated a generally rising rates of prescription (5-11% per year, p<0.05). The largest increases were for the hip (10% per year) and the neck (11% per year).

Conclusions: We described a rising incidence of regional body pain in the ED. The corresponding increasing trend of opioid medication prescriptions suggests that opioid medications may be relied on as a temporizing measure in the ED. Given the progressively important role of the ED in healthcare access and delivery, judicious pain management strategies are warranted.

30. Labor Force Status among U.S. Adults 45-64 Years with and without Arthritis, 2002-2010

KA Theis, L Murphy (Centers for Disease Control and Prevention, Atlanta, GA, USA)

Background: Arthritis is a leading cause of work disability. Working during preretirement (ages 45-64) is critical as earning potential is typically at its highest and income loss can lead to poverty in retirement.

Objective: To examine labor force status (LFS) of US adults ages 45-64 by arthritis status, 2002-2010.

Methods: Data: 2002-2010 National Health Interview Survey (annual multistage probability survey by in-person interview designed to represent US civilian, noninstitutionalized population; mean response rate 68.6%). Arthritis: “Have you ever been told by a doctor or other health professional that you have some form of arthritis, rheumatoid arthritis, gout, lupus, or fibromyalgia?” (sensitivity=77%). LFS: employed, unemployed (seeking work), retired, disabled. Weighted proportions and prevalence ratios (PR) with 95% confidence intervals (CI) estimated in SAS 9.2.

Results: In all years, the percent of people with arthritis (PWA) who were employed was lower than people without (2010: 53.9% [CI=51.5-56.2] vs 70.2% [CI=68.9-71.5]). For both groups, the percent employed decreased by 6.6% over the study period. In 2002 and 2010, PWA were 20% less likely to be employed than those without (PR=0.8; CI=0.7-0.8). For both groups, unemployment increased >3% over the study period. Retirement was consistently higher among PWA (2010: 10.4% [CI=9.1-11.8] vs 6.5% [CI=5.8-7.2]); in 2010, PWA were >2 times more likely to be retired (PR=2.4; CI=2.1-2.7). Each year, disability was >3 times higher among PWA (2010: 21.0% [CI=19.1-22.9] vs 6.4% [CI=5.6-7.1]). In 2010, there was a slight decrease in unemployment and early retirement among PWA in 2010 which may be attributable to a corresponding increase in disability.

Conclusion: Overall, employment declined and unemployment increased from 2002-2010. Preretirement PWA had less favorable LFS than those without. Retirement and disability were higher in PWA and suggest important income loss. PWA may need targeted interventions to improve work retention and delay retirement.

32. Increasing Congestive Heart Failure Hospitalization from 2000 to 2010 for those under 65: A Cause for Concern?

Shaleah Levant, M.P.H. and Margaret Jean Hall, Ph.D.

Background: Between 2000 and 2010, an estimated one million patients were hospitalized annually for congestive heart failure (CHF). While many of these patients were 65 and over, a growing percentage were under age 65. Previous research has focused primarily on CHF inpatients 65 and over. Little is known about younger patients with this very serious condition and potentially preventable hospitalizations.

Objectives: To fill the information gap about inpatients under age 65 hospitalized for CHF, a condition associated with older patients; to assess whether CHF hospitalizations decrease for this age group in view of efforts to prevent these hospitalizations by promoting timely and appropriate ambulatory care.

Methods: Using data from the National Hospital Discharge Survey, the study compares demographic and clinical characteristics, and hospital utilization of under age 65 CHF patients hospitalized in 2000 and 2010, as well as trends from 2000-2010. Differences among the subgroups were evaluated with two-tailed t-tests; a weighted least squares regression method was used to test the significance of time trends (p<0.05).

Results:

• The rate of CHF hospitalizations for those under 65 increased significantly from 2000-2010, from 9.4 to 10.8 per 10,000 population (15% increase).
• In 2000, 23% of CHF hospitalizations were for those under 65; in 2010, it was significantly higher at 29%.
• In 2000, the percentage of CHF hospitalizations for males and females under 65 were equal; by 2010, the percentage of males (54%) was significantly higher than that of females (46%).
• In 2000, 30% of the under 65 CHF inpatients were on Medicare; the share with Medicare in 2010 was 34%. Medicare eligibility was likely due to disability.

Conclusions: Hospital utilization data should continue to be tracked to evaluate whether preventive care and coordination of care provisions in recent healthcare reform legislation leads to a decrease in CHF hospitalizations in the future.

33. Evaluating the Feasibility of Using Four-Digit Social Security Numbers as National Death Index Search Criteria

Jesse Bassich and Donna Miller, MS

Objective: The National Center for Health Statistics (NCHS) regularly links NCHS survey respondents to the National Death Index (NDI) to ascertain death and cause of death information. Personal identifiers available from both the survey and the NDI, including Social Security number (SSN), are utilized during the linkage. Recently the National Health Interview Survey began collecting the last four digits of the SSN instead of the full nine digits. However, the NDI matching algorithm does not accept partial SSNs. The purpose of this project was to evaluate the feasibility of using the last four digits of SSN to find unique NDI matches.

Methods: Data Strings for NDI death years 1999-2008 were created by combining four and nine digit SSNs with other personal identifiers used by the NDI matching algorithm, including combinations of first name, last name, and date of birth (e.g. JaneDoe033044). Using these Strings, we tabulated the number of times each String occurred in the NDI: once, twice, and more than twice.

Results: Data Strings created with month, day, and year of birth, sex, and nine digit SSN were unique for 99.99% of records; using four digit SSNs, strings were unique for 96.75% of records. Some, but not all, other Strings considered led to similar results. For example, while Strings based on month and year of birth, sex and nine digit SSN, were unique for 99% of records; when four digit SSN was used, strings were unique for 45% of records.

Conclusion: Data Strings based on four and nine digit SSNs could lead to comparable tabulations, depending on included personal identifiers. These results support the development of matching criteria using the last four digits of the SSN.

34. A Statistical Method to Improve Truthful Responses to Surveys of Sensitive Information: A Quantitative Optional Unrelated-Question Randomized Response Model

Anna Tuck, Tracy Spears
Faculty Mentors: Sat Gupta, Mary Crowe
University of North Carolina at Greensboro

Background: Obtaining accurate survey responses is essential to public health research. Despite assurances of anonymity, subjects often give untruthful responses when information is requested regarding illegal or highly stigmatized behaviors. Greenberg’s Unrelated-Question Randomized Response Technique has been shown to improve accuracy in these situations.

Objective: We propose a generalization of the Unrelated-Question Randomized Response model to estimate the extent of a sensitive behavior, as well as the proportion of subjects who consider the question sensitive (known as the sensitivity level) for questions requiring a quantitative response, while taking into account that a question may be sensitive to one person, but not to another.

Methods: The mean of the innocuous behavior is assumed known. Estimators are derived for both the mean prevalence of the sensitive behavior and the sensitivity level, and their theoretical properties are examined (expected value, bias, sampling distribution). A split sample approach is utilized to solve for two unknowns. Computer simulations are conducted using SAS System for Statistical Analysis, for a range of values for sensitivity level, sample size, and mean prevalence of sensitive and innocuous behaviors. Simulation results are compared to theoretical results.

Results: The properties of the estimator for the mean prevalence of the sensitive behavior can easily be shown to be unbiased with asymptotically normal distribution by Central Limit Theorem, and is confirmed by computer simulation. The properties of the estimator for Sensitivity Level are problematic due to the presence of a random variable its denominator. Simulations, however, reveal an approximately normal distribution and accurate estimations of the sensitivity level.

Conclusions: Theoretical and simulation results indicate that this model accurately estimates mean prevalence of sensitive behaviors. Estimation of the mean prevalence and sensitivity level of a sensitive behavior among undergraduate students will serve as a field test of this model, planed for fall of 2012.

35. A comparison of Hispanic Population From the National Health Interview Survey and the Behavioral Risk Factor Surveillance System

Julia Lee, PhD Candidate, U. of Michigan

Background: A recent report comparing the national estimates from the National Health Interview Survey (NHIS) and the Behavioral Risk Factor Surveillance System (BRFSS) indicated noticeable discrepancy on estimates, such as general health status, among Hispanics, despite identical question wording.

Objective: The purpose of this study was to compare characteristics of Hispanic population from the NHIS and the BRFSS on comparable demographic and background variables.

Methods: The characteristics of Hispanic population, including gender, age, race, education attainment, job class, income, and health care access, were investigated between two surveys. Subgroup analyses were conducted on those who were interviewed by Spanish and by English. The characteristics of Hispanic population from both surveys were further compared to that of the American Community Survey which served as the gold standard for the characteristics of U.S. Hispanic population in regard to their demographic characteristics.

Results: Age, sex, education attainment, and work class were found to be signicantly different. This population between the two surveys represent samples from different socio-demographic segments. This difference can not be explained by interview language and mode of data collection. Hispanic population from NHIS is more similar to that of the ACS. Hispanic population from BRFSS consisted higher proportion of female, older age, lower education attainment, and people not in labor force.

Conclusions: National estimates from the NHIS and BRFSS on Hispanic population are not comparable. Cautions need to be taken when using BRFSS to guide National policy and program decisions, especially for policies with high relevance to Hispanic population. Alternative sampling strategies should be considered to obtain representative respondents on Hispanics.

36. Access to health care among Hispanic subgroups

Gulnur Freeman, M.P.A., Patricia F. Adams, and Robin A. Cohen, Ph.D.

Background: Access to health care plays an important role in the quality of health care. Although an effort has been made to eliminate inequality in health and health care, disparities in access to care continue to exist in the United States. Previous studies have indicated that Hispanic persons were more likely to experience a lack of access to health care than non-Hispanic white or non-Hispanic black persons.

Objective: This study describes differences in access to health care and health care utilization among adults in Hispanic subgroups aged 18-64 in the United States. For comparison purposes, estimates will also be presented for non-Hispanic white and non-Hispanic black adults aged 18-64.

Methods: This study uses data from the Family Core and Sample Adult components of the 2009-2011 National Health Interview Survey (NHIS). The 2009-2011 data were combined in the analysis to improve the precision and reliability of the estimates. Access measures examined include health insurance, usual source of care, emergency room utilization in the past 12 months, and unmet medical needs due to cost in the past 12 months.

Results: A preliminary analysis based on earlier NHIS data (2008-2010) found that among Hispanic subgroups, Puerto Rican adults were more likely to have a usual source of health care and to have visited a hospital emergency room on two or more occasions in the past year than Mexican, Cuban, or Central or South American Hispanic adults. Central or South American (40%) and Mexican (36%) adults were more likely to be uninsured than Cuban (28%) or Puerto Rican (16%) adults.

Conclusions: Disparities in access to health care for adults aged 18-64 exist among Hispanic subgroups. The NHIS can play an important role in monitoring changes in health care access of Puerto Rican, Cuban, Mexican, and Central or South American Hispanic adults living in the United States.

37. Racial disparities in outcomes of spinal surgery for lumbar stenosis

Jacob H. Bagley, BS¹, Shivanand P. Lad, MD, PhD¹, Ranjith Babu, MS¹, Beatrice Ugiliweneza, PhD, MSPH³,
Carlos A. Bagley, MD¹, Oren N. Gottfried, MD¹, Robert E. Isaacs, MD¹,
Chirag G. Patil, MD², Maxwell Boakye, MD³

¹ Division of Neurosurgery, Department of Surgery, Duke University Medical Center
² Center for Neurosurgical Outcomes Research, Maxine Dunitz Neurosurgical Institute, Department of Neurosurgery, Cedars-Sinai Medical Center
³ Department of Neurosurgery, University of Louisville

Background: Racial inequalities have been described in the outcomes of a variety of surgical procedures. Lumbar stenosis is among the most common indications for spine surgery, but there has been minimal investigation of racial disparities in complications and costs of lumbar laminectomies and fusions.

Objective: To evaluate racial disparities in outcomes of lumbar stenosis surgery.

Methods: We analyzed the Medicaid dataset of the Reuter’s MarketScan database between 2000 and 2009. This dataset encompasses more than 63 million patients in more than 45 health plans and government organizations in the United States. African American and non-Hispanic white patients who underwent laminectomy or fusion for lumbar stenosis with at least two years post-operative data were included. We examined the effect of race on the rate of re-operations, complications and the charges associated with surgery.

Results: African American patients in the Medicaid database were at no higher risk for re-operation in the two years following index procedures than white patients (7.14% vs. 7.89%. p = 0.65). However, we did find that African American patients were more likely to experience post-operative complications within 90 days of their procedures, even after adjusting for length of hospital stay, comorbidities, gender, and age (aOR = 1.6, 95% CI: 1.11 – 2.32 p < 0.05). White patients had a significantly shorter length of stay (3 vs. 5 days, p < 0.0001) and accrued fewer hospital-related charges ($16,148 vs. $24,267, p < 0.0001). African Americans, despite having more comorbidities in our sample, were prescribed fewer medications in the two years following index procedures (47 vs. 71 prescriptions, p < 0.0001), and had fewer medication charges ($5,297 vs. $8,450, p = 0.002).

Conclusion: At the national level, there are a number of racial disparities in the rate of complications, length of stay and charges after surgery for lumbar spinal stenosis.

38. Gender and Racial Disparities in Opioid Medication Prescription for Regional Body Pain in the Emergency Department

Ming-Chih J. Kao¹, Lyly Cao Minh¹, Grace Y. Huang², Salar Deldar¹, Jeffrey K. Teraoka³, Matthew Smuck^1
1 Department of Orthopaedics, Stanford University Medical School, Palo Alto, CA
² Department of Orthopaedics, University of California San Francisco Medical School, San Francisco, CA
³ Department of Physical Medicine and Rehabilitation, Palo Alto VA, Palo Alto, CA

Background: Opioid medications have come under much scrutiny in recent years. In the Emergency Department (ED) setting, we demonstrated increasing low regional body pain complaints from year-to-year, who are increasing more likely to receive opioid medication prescriptions.

Objective: To describe differences in the rate of opiate medication (OM) prescription by race and gender for regional body pain in Emergency Departments (EDs) in the United States (US)

Methods: National Hospital and Ambulatory Medical Center Survey 1996-2008 data were analyzed using SAS 9.2 and R. For variance estimation, masked “ultimate cluster” weights were used for 2002-06. Domain analysis was used in subset analyses. Weighted logistic regression was used to adjust for demographic factors. Rates of OM prescription for various regional body pains: axial (headache, neck, back, low back), upper extremity (shoulder, arm, elbow, wrist, hand), lower extremity (hip, leg, knee, ankle, foot) by race and gender.

Results: Controlling for age and gender, African-Americans were less likely to receive OM for head (OR=0.43, p<0.001), back (OR=0.45, p<0.001), hip (OR=0.46, p=0.01), and arm pain (OR=0.47, p=0.003) compared to white male controls. There appears to be a trend toward normalization of this disparity. Controlling for age and gender, Hispanics were less likely to receive OM for head (OR=0.41, p<0.001) and back pain (OR=0.55, p=0.001). Controlling for age and race, females were less likely to receive OM for headache (OR=0.64, p=0.002) but more likely to receive OM for back pain (OR=1.48, p<0.001), a trend that is normalizing at 4% per year.

Conclusions: We describe 13-year trends on persistent race and gender disparities in US EDs showing statistically significantly different rates of OM prescription for African-Americans, Hispanics, and females that are manifest in pain complaints of certain body regions.

39. Quantitative Analyses on Health Leaders’ Perceptions Towards Health Promotion in a Visible Minotor Populations

Amy Ng

Background: Although health leaders in Alberta have promoted primary health care intervention, ethnic minorities respond to health promotion activities slowly. For instance, the high incident rates of colorectal cancer among ethnic minority women are evidenced; and the low colorectal screening rates of ethnic minority are found. The current study is to examine health leaders’ perception towards health promotion in visible minority populations from a quantitative viewpoint.

Objectives: To test a primary health care intervention satisfaction in promoting health in visible minority populations by comparing the frequency of health professionals’ participation into health care practices among visible minority people to the degree of interest in serving ethnic clients, cross-cultural competency, the level of strategic health care plan, and the effect of health promotion in diversity. The qualitative analysis of health leaders’ inputs is implemented to explain the satisfaction of primary care promotion among ethnic minority women groups.

Method: In a survey, 25 health professionals were recruited for phone interviews on various occasions in order to gather their inputs. Statistical Package for the Social Sciences (SPSS) was used to manage and analyze the research data.

Results: The results indicated that the co-relation factor is 32.85% between the degree of interest in serving diverse clients and the frequency of serving diverse health clients. The frequency of serving among visible minorities is co-related with the degree of cross-caring competence in terms of 19.5%. The high level of cross-cultural competence in health professionals is related to the health professional research informants’ negative perception on effect of health promotion in minority groups, which is the co-relation factor is -41%.

Conclusion: The health promotion in visible minority groups appears to be dominated by the influence of cross-cultural caring practices that perpetuate a client-centered focus of service delivery. Health promotion in cross-cultural issue is worthwhile for a further research study.

40. Parental Investment, Child Health Formation and Racial Differences

‘Muriel’ Zheng Fang

Background: Cross-sectional researches in social and medical sciences suggest that early life circumstances have long-term consequences for adult physical, cognitive and emotional health. This has given rise to the theory of developmental plasticity that physical health is most sensitive to inputs received during early years. Evaluation of this hypothesis demands a study that accounts for the dynamic and cumulative nature of health formation process, which is lacking in the literature.

Objective: To evaluate the developmental plasticity hypothesis in child physical health, and to explain how racial health disparity from birth persists through childhood.

Methods: This paper estimates a value-added health production function with time varying rates of return to investment. It uses anthropometric measurements and limiting physical conditions as proxies for child health, and general demographic as well as age-appropriate parenting measurements to proxy for parental investment, from the Children of National Longitudinal Survey of Youth 1979’ (CNLSY79). It uses a multiple-indicator multiple-cause (MIMIC) model in which concurrent measurements act as instrumental variables.

Results: The results find that health increasingly is dependent on lagged health, and that investment rates of return during the prenatal and infancy periods are higher than subsequent periods. I also explore racial differences in the production function and find that rates of return to investment are suppressed for black children.

Conclusions: The paper gives empirical support to developmental plasticity in child physical health. The suppressed investment return in black children contributes to an understanding of how racial health disparity persists through childhood.

41. Beyond Battering: Intimate Partner Violence and Preventive Screening Behaviors among Women

Monique J. Brown, MPH, PhD student¹; Kate L. Lapane, PhD¹; Sherry Weitzen, PhD^2
1 Department of Epidemiology and Community Health, Virginia Commonwealth University
² Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel

Background: Intimate partner violence (IPV) is defined as physical, sexual or psychological harm that can be perpetrated by a former or current spouse. Data from a national survey indicated that 76% of all rapes and physical assaults against women in the US are perpetrated by a former or current intimate partner. IPV has been linked to sexually transmitted infections and chronic diseases. Racial/ethnic disparities have been seen in help-seeking behaviors among abused women. Barriers to healthcare for ethnic minority women have also been identified.

Objectives: To determine the association between IPV and preventive screening behaviors, including HIV testing, cervical cytology (Pap smears), colorectal cancer screening, cholesterol testing, breast examination and mammography among women; and to explore the racial/ethnic disparities concomitant to this association using data from the Behavioral Risk Factor Surveillance System (BRFSS) survey.

Methods: We used data from eight states that collected IPV information in the 2006 or 2007 BRFSS survey (n=30,182). Multivariable logistic regression models provided adjusted estimates of odds ratios (aOR) and 95% confidence intervals (CI).

Results: Approximately one in four women reported a history of IPV. Relative to those who did not report a history of IPV, IPV victims were twice as likely to get tested for HIV (aOR: 2.35; 95% CI: 2.07 to 2.66), had cervical cytology tests (aOR: 2.08; 95% CI: 1.30 to 3.34) or had a breast exam (aOR: 1.76; 95% CI: 1.38 to 2.25). However, IPV victims were 23% less likely to ever have a mammogram (aOR: 0.77; 95% CI: 0.64-0.94). Variations in associations between IPV and screening behaviors were observed by race/ethnicity.

Conclusion: IPV victims seem to be vigilant about certain screening practices but seem to be less vigilant about others. Some ethnic minority women seem to be less likely to have certain screenings compared to non-Hispanic White women.

42. Child bullies: Characteristics of young school-aged children who bully

LaJeana Howie MPH, Rosa Avila MSPH, Patricia Pastor PhD

Background: Bullying is intentional, repeated negative behavior reflecting a power disparity between victim and perpetrator (bully). Prior research has found that bullying others during childhood is associated with problematic behaviors in adulthood. The goal for many youth anti-bully efforts is to stop bullying before it starts, but national data have not been used to examine the factors associated with bullying in young school-aged children.

Objective: To examine selected characteristics of young-school aged U.S. children who bully other children.

Methods: Parent-reported data were analyzed for 33,047 children 6-12 years old from the 2007 National Survey of Children’s Health, a nationally representative random-digit-dial telephone survey. Bivariate analyses and logistic regression models were used to examine associations between child behaviors (positive social skills and problematic behaviors), parental stress, neighborhood characteristics and bullying, after adjusting for socio-economic and demographic characteristics. Children were categorized as bullies if they were cruel/mean to others. Parental stress was defined as often feeling aggravated by their child. Neighborhood characteristics were defined by perceived neighborhood safety and support, and school safety.

Results: Approximately 15% of young school-aged children were bullies. Compared to children who do not bully, bullies were more likely to be non-Hispanic black or Hispanic, lived in poor households, and resided in unsafe and unsupportive neighborhoods. Child bullies often lacked positive social skills and frequently exhibited problem behaviors. Compared to non-bullies, bullies were also more likely to have mothers with sub-optimal mental health and have a parent who often felt stressed. After adjustment, bullying remained associated with living in an unsafe neighborhood (Odds ratio [OR]=1.27; Confidence Interval [CI]=1.02-1.58), having a stressed parent (OR=3.73; CI=2.99-4.65), and having a mother with sub-optimal mental health (OR=1.87;CI=1.56-2.24).

Conclusions: This study found that negative child behaviors, unsafe neighborhoods, parental stress, and sub-optimal maternal mental health are associated with bullying despite socio-economic/demographic differences.

43. Teen pregnancy in the United States is higher when the nativity of the mother is inside the US

Jagit Teji

Background: As the latest vital statistics have shown that teen pregnancy is dropping in the US while there are sporadic studies implicating that teen pregnancy in the US is less likely when the nativity of the mother is outside the 50 states of the US.

Objective: To test the hypothesis that teen pregnancy is lower in the mothers whose nativity is outside the 50 states of the US.

Methods: VSS data from the NCHS was analyzed for the years 1995 thru 2002. Logistic regression analysis was performed with the dependent variable teen pregnancy, TP, with maternal place of birth, MPB as independent variable. The confounding variables were maternal race, gestational age, GA, birth weight, BWT, maternal conditions such as diabetes, hypertension, pregnancy induced hypertension, maternal exposure to alcohol and tobacco. Also considered variables were prenatal care , plurality, and gender. Stata 12.0 was utilized for statistical analysis.

Results: Out of over 31 million records for births during the period 1995 thru 2002 over 25 million births had usable data. Over all, births to teen mothers were significantly more likely if their nativity was within the 50 states of the US , OR 2.49 cf (95%) 2.48-2.50 while controlling for other confounders. Every ethnicity and race was affected by the maternal nativity.

Conclusion:

1. We accept the hypothesis that teen pregnancy is lower when the maternal nativity is outside the US.
2. Reason for higher teen pregnancy rates when mother’s nativity is from the US is concerning and needs to be decreased.

44. Pre-pregnancy body mass index (BMI) and intra-abdominal visceral fat accumulation in the first trimester

Anjel Vahratian ^1,2, Marjorie Treadwell ², Christina Carroll ^2
1 Data Analysis and Quality Assurance Branch, Division of Health Interview Statistics, National Center for Health Statistics,
² Department of Obstetrics and Gynecology, School of Medicine, University of Michigan at Ann Arbor

Background: Body mass index (BMI) is used as a proxy measure for body fat or adiposity in both pregnant and non-pregnant populations. In pregnant women, pre-pregnancy BMI is often ascertained based on maternal self-reported weight and measured height at the first prenatal visit. However, this simple and inexpensive assessment of obesity may not be the most appropriate proxy for risk during pregnancy.

Objective: The objective of this study was to examine the relationship between pre-pregnancy BMI and an ultrasound measure of intra-abdominal visceral fat accumulation during the first trimester of pregnancy.

Methods: Data from 48 women enrolled in a single-center longitudinal study of maternal obesity and metabolic dysfunction in pregnancy were used for this analysis. The minimal thickness of subcutaneous fat and the maximum thickness of preperitoneal fat were recorded directly from the ultrasound screen with electronic calipers at 8-12 weeks gestation, and the abdominal wall fat index (AFI) was calculated as the ratio of the thickness of preperitoneal to subcutaneous fat (P/S ratio). An analysis of variance and correlation coefficients were used to test for significant differences across BMI subgroups. P-values of < 0.05 were considered statistically significant.

Results: There was a statistically significant, positive association between pre-pregnancy BMI and the thickness of the a) preperitoneal fat layer and the b) subcutaneous fat layer (p<0.001). However, there was no difference in the AFI between the normal, overweight, and obese women.

Conclusions: Intra-abdominal visceral fat accumulation during the first trimester of pregnancy is associated with maternal self-reported pre-pregnancy BMI. This information may be useful in the assessment of maternal risk for adverse maternal and fetal health outcomes during pregnancy.

45. Gestational Weight Gain in Underweight Women and Associations with Infant Death.

Regina R. Davis; S. Hofferth; E. Shenassa; University of Maryland College Park, School of Public Health

Background: More than 28,000 babies die before their first birthday in the United States each year. Programmatic and Policy focus on prematurity and birth weight stem largely from their known relationship to infant mortality and morbidity. A large body of literature exists linking poor gestational weight gain to prematurity and low birth weight, but few studies have examined infant death as an important pregnancy outcome of inadequate gestational weight gain. As a measure of healthy gestational weight gain, the Institute of Medicine (IOM) published guidelines which provide a recommended weight gain for each category of pre-pregnancy Body Mass Index (BMI).

Objective: The association between gestational weight gain and infant mortality among U.S. infants were examined through two research questions: Is gestational weight gain associated with the risk of infant mortality? Does the effect of gestational weight gain on infant mortality depend upon pre-pregnancy BMI?

Methods: Using data from the Pregnancy Risk Assessment Monitoring System in 35 states and New York City, we investigated the association between the 2009 IOM weight gain guidelines and infant death by maternal pre-pregnancy BMI among 160,111 women who delivered a singleton infant from 2004-2008. Descriptive and logistic regression analyses were used to assess the risk of infant mortality associated with inadequate gestational weight gain compared to normal weight gain.

Results: Infants born to underweight women with inadequate gestational weight gain had odds of infant death that were 6.37 times (p<0.0001, 95% Confidence Interval: 2.60, 15.64) the odds for infants born to normal weight women with weight gain within the recommended range. Increased odds remained after adjustment for gestational age, low birth weight, parity, maternal age, maternal education, prenatal care, maternal race, marital status, diabetes, hypertension, tobacco use, and alcohol consumption.

Conclusion: There is a significant association between inadequate gestational weight gain and infant death.

46. Identifying factors related to the implausible gestational ages using mixture models

Alena Maze; Guangyu Zhang; Jennifer D Parker; Nathaniel Schenker; National Center for Health Statistics, CDC

Background: Birth weight and gestational age are two important measurements of infant health and maturity. For each gestational age, birth weights of fetuses are usually assumed to follow a normal distribution. However, in the US Birth data at early reported gestational ages, the birth weight distribution is often skewed to the right or even bimodal, appearing as a combination of two normal curves. Errors in gestational age that are calculated from the mother’s reported last menstrual period (LMP) are believed to be the source of this bimodal distribution.

Objective: We sought to identify maternal characteristics associated with the implausible gestational ages calculated from LMP.

Methods: The data were taken from the 2004-2009 United States birth data for gestational weeks 21-34. We fit Gaussian mixture models to identify factors related to the probability of implausible gestational ages.

Results: The trimester prenatal care, the number of prenatal visits, infant’s live birth order, maternal age, race, Hispanic status, educational level, marital status, and tobacco use were significantly associated with the probability of implausible gestational ages using the univariate analysis. Multivariate analysis revealed the trimester prenatal care, the number of prenatal visits, infant’s live birth order, mother’s education, age, and race to be significant predictors of implausible gestational ages.

Conclusion: Women who had no or late prenatal care (after 1st trimester), more prenatal visits, less education, were younger, and were Hispanic were more likely to report implausible gestational ages than their counterparts.

47. The Association between the Medical Home and Receipt of Standardized Developmental Screening among US Children 5 years and younger

Nicole Richmond, MPH (1, 3), Tri Tran, MD, MPH (2, 3), Susan Berry, MD, MPH (1, 3)
(1) Louisiana Office of Public Health, CSHS Program; (2) Louisiana Office of Public Health, MCH Program; (3) LSUHSC, School of Medicine, Department of Pediatrics

Background: The American Academy of Pediatrics’ best practice recommendation for early childhood pediatric medical care is use of Standardized Developmental Screenings (SDS), which improves early detection and treatment for developmental disabilities, and is an integral service of the Medical Home (MH).

Objective: Is there an association between having a MH and receipt of SDS among US children 5 years and younger.

Methods: The 2007 National Survey of Children’s Health data were used. PDS, MH, and Special Health Care Needs variables were created using codes provided by the National Center for Health Statistics. Child’s race, ethnicity, age, and sex, federal poverty level, language, insurance status, family structure, highest education, urban/rural, and Census Bureau Region (CBR) were selected. Simple and multiple logistic regression were conducted. Alpha was set to 0.05. SAS-Callable SUDAAN v. 9.3 was used.

Results: Approximately 80.8% (22,270) of US children 5 and younger met study inclusion criteria, and represent more than 19 million children nationwide. Only 19.5% (95% CI: 18.3-20.8) received SDS, and 63.0% (61.3-64.6) have a MH. Child’s sex had a differential effect on the association between having a MH and SDS receipt; controlling for insurance, age, and CBR females exhibited an association (1.48, 1.1-1.8), whereas among males the MH did not increase rates for SDS receipt (1.1, 0.8-1.3).

Conclusions: National data shows that the MH does not substantially increase the odds for SDS. And any marginal influence by the MH is differentially experienced according to the child’s sex. The substantially large proportion of children aged 5 years and younger who did not receive an SDS in the past year signifies a considerable need to implement a multifaceted approach to increase SDS rates, since receipt of SDS can increase the likelihood that a child will receive proper medical, social, and educational services during his/her critical periods of development.

48. Does home visiting improve birth outcomes? Findings from the 2007-2008 Virginia PRAMS

Monisha Shah, MPH Candidate, Kristin Austin, MPH, Marilyn Wenner, Gandarvaka Gray, MPH, Derek Chapman, PhD
Virginia Department of Health, Virginia Commonwealth University, Tulane University School of Public Health and Tropical Medicine

Background: Home visiting programs have gained momentum with the recent appropriation of federal funds for states to implement evidence-based maternal and child health home visiting programs. Virginia will be implementing several evidence-based models and evaluating the effectiveness of home visiting. However, currently, the effects of having a home visit by a health care worker or trained lay outreach worker are unknown at the state level.

Study Question: Does participation in a home visiting program during pregnancy improve birth outcomes?

Methods: Using cross-sectional data from the 2007-2008 Virginia Pregnancy Risk Assessment Monitoring System (PRAMS), several maternal and infant health outcomes were examined. Women who reported receiving a home visit by a healthcare worker or nurse during pregnancy on the PRAMS survey were assigned into the Home Visiting (HV) participation group. Women who did not have a home visit, but had risk factors similar to participants, were assigned to the comparison group. Multivariate logistic regression was used to measure the association between HV and the selected birth outcomes of interest.

Results: Controlling for socioeconomic status and selected maternal characteristics, women who had a home visit by a healthcare worker or nurse during pregnancy showed a 87.5% reduction in the risk of having a low birth weight infant [0.13, 95% CI: (0.02, 0.78)], a 4.5 times greater odds of initiating breastfeeding [AOR: 4.50 95% CI: (1.05, 19.54)], and a 38.6 times greater odds of using contraceptives postpartum [AOR: 38.55, 95% CI: (3.14, 473.21)] than women in the comparison group.

Conclusions: Receiving home visiting services during pregnancy increased breastfeeding initiation and postpartum contraceptive use and reduced the risk of low birth weight.

Public Health Implications: These findings demonstrate that home visiting services have positive effects on the health and development of women and infants.

49. Utilization of Well Child Care and Effective Care Coordination: An analysis of the 2007 National Survey of Children’s Health data.

Adejoke Ogunrinde

Background: The need for an improved delivery of primary care services is increasingly being noted as important in efforts towards improving access to health care and having a primary care provider who coordinates and assures this quality care for children have also been identified as a key factor in health care. Research shows that guidance given during a well-child care (WCC) visit can change parenting practices and affect choice to breastfeed.

Research Question: Is there an association between utilization of WCC and effective care coordination? What other factors affect the utilization of WCC?

Methods: Data used was collected by telephone interview on 60,845 children 0-10 years using parental reports from the National Survey of Children’s Health 2007. A chi-square analysis was used to describe the relationship between care coordination and utilization of WCC. Logistic Regression was performed to examine the association between the utilization of WCC visits, coordination of care and other control variables.

Results: Having all components of care coordination is associated with utilization of WCC visits; Black non-Hispanics, Hispanics were 49%, 17% of time respectively more likely to utilize WCC visits in the US. The poorer the family, the less likely to get WCC, those with public health insurance are 4 times more likely to attend WCC than the uninsured and 95% odd greater than those with private insurance who are 3 times more likely to attend.

Conclusion: There is significant association between effective care coordination and WCC visits and additional research needs to be conducted to understand other factors that may influence utilization such as poverty levels and insurance coverage.

50. Ease of Use of Community-Based Services among Children with Special Health Care Needs

Ngai Kwan, Myra Rosen-Reynoso, & Susan Foley, National Center for Ease of Use of Community-Based Services, Institute for Community Inclusion, University of Massachusetts Boston, Boston, MA

Background: In response to the growing population of children with special health care needs (CSHCN), the National Center for Ease of Use of Community-Based Services has attempted to operationalize one of the six outcome indicators -ease of use of community based services among CSHCN.

Objective: The objective of this study is to explore factors associated with the outcome5, ease of use(EU) among CSHCN.

Methods: Data of a weighted sample of 9,813,214 CSHCN from 2009-10 National Survey of Children with Special Health Care Needs (NS-CSHCN) was analyzed. The dependent variable was dichotomized as meeting outcome5(EU) where families must experience no difficulties or delays in getting services, and be only sometimes or never frustrated in efforts to get services for CSHCN. Incorporating the complex sample design structure, bi-variate and multivariate analyses were performed.

Results: Bi-variate results showed that failing to meet outcome5 is associated with minority status, low socioeconomic measures, and improper insurance coverage. Results from a logistic regression further illustrated that, controlled for covariates, Hispanics (OR =.80; 95%CI=.68-.92) and Other races (OR =.82; 95%CI=.69-.96) were less likely to report EU than Whites. Wealthier families were also more likely to report EU comparing to families with less financial resources. Relative to CSHCN with private insurance only , those with public insurance only(OR =.76; 95%CI=.67-.87); or with both private and public insurance(OR =.82; 95%CI=.69-.97); or with other insurance only(OR =.67; 95%CI=.52-.91); or with no insurance (OR =.19; 95%CI=.14-.24)were all less likely to report EU.

Conclusions: Thirty-five percent CSHCN families reported services were not easy to use. The situation is further aggravated when the CSHCN are in ethnic minority groups with less financial resources and inadequately insured. Efforts need to be made to provide a comprehensive, easy-to-use, and family-centered system of care for these CSHCN and their families.

51. Congenital Anomalies and Childhood Cancer in Oklahoma

Amanda E. Janitz¹, Barbara R. Neas¹, Janis E. Campbell¹, Sharon Vaz², Anne Pate^2
1 Department of Biostatistics and Epidemiology, College of Public Health
² Oklahoma State Department of Health

Introduction: Several data-linkage studies found an association between congenital anomalies and childhood cancer. However, this linkage has not been done in Oklahoma using the existing birth defects and cancer registries. This linkage can generate hypotheses about common pathways of congenital anomalies and cancer.

Objective: We aim to understand the process of data-linkage between these registries and determine if there is an association between congenital anomalies and childhood cancer.

Methods: We will collect several variables, including characteristics of the child at birth, parental characteristics at the child’s birth, and information on the cancer and congenital anomaly. We will then link the databases, calculate descriptive statistics and assess the relationship between congenital anomalies and childhood cancer.

Results: In Oklahoma, the age-adjusted incidence rate for all childhood cancers was 161.9 cases per million from 1999-2008, with the highest incidence in children under 1 year. The most common types were Leukemia, Brain/Central Nervous System tumors, and Lymphoma. Major congenital anomalies occurred in 3% of infants and caused 20% of deaths in infants under 1 year. The most common congenital anomalies among live births were cleft lip with and without cleft palate (13.8 cases/10,000) and Down syndrome (12.4 cases/10,000) from 2002-2006. We will present our findings from more in-depth analyses of the registries and the process of linking these databases.

Discussion: Because this is the first linkage of this nature in Oklahoma, the process of working with different departments within the Oklahoma State Department of Health is important. This study will provide an opportunity to work with the health department and contribute to the literature regarding the association between congenital anomalies and childhood cancer.

52. Understanding Food Deserts: A Comparison of Food Insecure Neighborhoods in the Southeastern U.S.

Christopher Langeler

Background: Eliminating health disparities has become a focus of local, state, and national efforts to improve population health. One way health disparities manifest is through unequal access to affordable healthy food, particularly in low-income “food desert” neighborhoods. Lack of availability of adequate nutritious food has been associated with numerous negative physical and mental health outcomes and constitutes an ongoing obstacle to achieving health equality in the U.S.

Objective: The purpose of this study is to compare and contrast data collected through assessments of individual, family, and community-level data nine different neighborhoods in four major cities in the Southeastern U.S. (Nashville, TN; Memphis, TN; Atlanta, GA; St. Louis, MO).

Methods: The “Food Procurement Survey” (an unpublished instrument) was used to assess food insecure neighborhoods in National Mobile Market’s partner cities by local staff and trained community volunteers. Participants were obtained through convenience or snowball sampling due to resource constraints of community partners. Primary analytic techniques include ANOVA, Pearson correlations, and descriptive analyses.

Results: Preliminary findings suggest that food procurement practices, barriers to healthy food procurement, and demographic characteristics vary significantly between different “food desert” neighborhoods. This variance was found both between cities and among neighborhoods within cities. Final data collection for one of our cities (St. Louis) is pending and will be incorporated into our analysis by the end of June 2012, which will lead to more sophisticated analyses and detailed results.

Conclusion: Results indicate that food insecure neighborhoods, sometimes called “food deserts” vary greatly and need to be uniquely assessed and approached in order to maximize the effectiveness of interventions. Treating all “food deserts” similarly will result in poorly adapted initiatives that may not adequately address barriers and issues specific to a given neighborhoods.

53. The Impact of Pizza on the American Diet

Donna G. Rhodes, Meghan E. Adler, John C. Clemens, Alanna J. Moshfegh

Background: Pizza, a popular food in the United States, has a considerable impact on nutrient intakes in the U.S. population. Previous nationally representative dietary data have shown pizza consumption impacts total daily energy and sodium intakes of Americans.

Objective: Estimate food and nutrient intakes in the U.S. population and report differences for those individuals who reported pizza vs. non-reporters.

Methods: Consumption of pizza was assessed using nationally representative one-day dietary data from What We Eat in America (WWEIA), NHANES 2007-2008 for individuals 2 years and over (n=8529). Nutrient intakes of pizza reporters (PRs), individuals who reported pizza as defined in WWEIA Food Categories 2007-2008 at least once, were compared to non-reporters.

Results: Overall, PRs comprised 13% of the population. For PRs, the contribution (mean ± SE) of pizza to total daily energy was 27 ± 0.8% and mean energy from pizza was 662 ± 21 kcal/d. Total daily energy intake, from all foods was 2435 ± 39 kcal for PRs and 2018 ± 29 kcal for non-reporters. Total daily sodium intake, from all foods was 3805 ± 74 mg for PRs and 3262 ± 58 mg for non-reporters. Pizza contributed 37% of saturated fat, 35% of sodium, 36% of calcium and 53% of lycopene to total daily intakes of PRs. Results will be updated with What We Eat in America, NHANES 2009-2010 data.

Conclusions: Pizza provided a substantial contribution to total daily nutrient intakes for the U.S. population. This study emphasizes the sizable impact a single food can have on the dietary intake of Americans.

54. Food Intakes Converted to Retail Commodities Database: A Novel Approach to Convert Foods Consumed by Americans Back to Retail Commodities

Shanthy A. Bowman¹, Carrie L. Martin¹, Jennifer L. Carlson¹, James E. Friday¹, John Clemens¹, Alanna J. Moshfegh¹, and Biing-Hwan Lin²
U. S. Department of Agriculture, ¹Agricultural Research Service, Beltsville, MD and ²Economic Research Service, Washington, D.C.

Background: At present, there are no national databases that convert foods consumed in the national dietary surveys to food commodities at the retail level. The study objectives are: (1) to develop a Food Intakes Converted to Retail Commodities Database (FICRCD) that converts foods eaten in What We Eat In America (WWEIA), the dietary component of the National Health and Nutrition Examination Survey (NHANES), to retail commodities, and (2) to determine changes in selected retail commodity consumption over the survey periods.

Methods: There are two major steps in FICRCD development: (1) foods consumed in the surveys are separated into their ingredients that are assigned to an appropriate retail commodity category, and (2) conversion factors are applied to adjust for food preparation and processing losses to convert foods consumed back to their commodity levels. There are eight major retail-level commodity categories in FICRCD: Dairy Products; Fats and Oils; Fruits; Grains; Meat, Poultry, Fish and Eggs; Nuts; Caloric Sweeteners; and Vegetables. Because each of these has several components, FICRCD has a total of 65 retail-level commodities.

Results: Mean consumption of selected retail commodities among persons 2 years and older in the two survey periods 1994-98 and 2003-04, respectively are: fluid milk (266g vs. 265g), low fat and skim milk (89g vs. 81g), total fruit (322g vs. 313g), potatoes (117g vs. 103g), tomatoes (93g vs. 111g), and total vegetables (371g vs. 357g).

Conclusions: FICRCD is useful for nutrition and agricultural policy development since it can be used to compute national estimates of retail commodity use by Americans and to identify disparities among socio-economic groups. Further, it can be linked with economic data to determine food cost. FICRCD for subsequent WWEIA, NHANES surveys are currently under development. Study funded by ARS & ERS, USDA.

55. Retooling of the USDA Food Patterns Equivalents Database: Methodology

Shanthy Bowman, Miyuki Shimizu, James Friday, Rachel Condie, John Clemens, Brian Barrows, Alanna Moshfegh. U.S. Department of Agriculture, Agricultural Research Service, Beltsville, MD

Background: The Dietary Guidelines for Americans, 2010 (DGA 2010) provide recommendations on how much to eat from the USDA Food Patterns Groups such as Grains, Dairy, Fruits, Vegetables, and Protein Foods. The USDA Food Patterns Equivalents Database (FPED), formerly known as the MyPyramid Equivalents Database (MPED), converts foods consumed in What We Eat In America (WWEIA), the dietary component of the National Health and Nutrition Examination Survey (NHANES), into cup equivalents of dairy, fruits, and vegetables; ounce equivalents of protein foods and grains; teaspoon equivalents of added sugars; gram equivalents of discretionary solid fats and oils, and the number of alcoholic drinks.

Objective: To retool the FPED process to simplify the FPED development methodology and to apply consistent decisions across similar foods.

Methods: The retooling steps include developing a 100-gram FPED for the Food and Nutrient Database for Dietary Studies (FNDDS) ingredients and applying it to compute the final FPED for FNDDS foods; consolidating the weights of 1 cup of fruits and vegetables such that similar types of fruits and vegetables will have the same weights; estimating added sugar equivalents from the total sugar values of the foods that are defined as added sugars; and the use of 16 grams of flour as the basis for defining one ounce equivalent of grains for grain products made of flour and using 28.34 grams as one ounce equivalent of grains for breakfast cereals and grains.

Results: A comparison of the FPED and MPED applied to WWEIA, NHANES 2003-04 showed no appreciable differences between the national mean intakes of the 32 FPED/MPED components estimated using the two databases. Development of FPEDs for WWEIA, NHANES 2005-2008 is ongoing.

Conclusions: The retooling simplifies the FPED development process and reduces the time requirement without affecting the quality of the database.

56. Importance of Ethnic Foods in What We Eat in America (WWEIA), NHANES

Deirdra N. Chester, PhD, RD; Carrie L. Martin, MS, RD; Grace Omolewa-Tomobi, BS; Joseph D. Goldman, MA; Meghan Adler, MS, RD; Alanna J. Moshfegh, MS, RD
U.S. Department of Agriculture, Agricultural Research Service, Beltsville Human Nutrition Research Center, Beltsville, MD

Background: The availability and consumption of ethnic foods has increased among the U.S. population. There is a critical need to capture the ingredients and nutrient content of ethnic foods in nutrient databases. The Food and Nutrient Database for Dietary Studies (FNDDS) is used to analyze dietary intakes from WWEIA, the dietary component of NHANES. Every two years, the FNDDS is updated to reflect the changing marketplace and dietary trends of reported intakes from WWEIA, NHANES survey participants. Asians will be oversampled in NHANES, 2011 – 2014, which further enhances the need to capture ethnic foods in FNDDS.

Objectives: (1) To categorize ethnic foods available in FNDDS, (2) report consumption of ethnic foods in WWEIA, NHANES, 2005 – 2008, and (3) to evaluate and update ethnic recipes in FNDDS.

Methods: Estimates were calculated using WWEIA, NHANES, 2005 – 2008 day 1 dietary intake data for adults 20 years+ (n=10,480). Ethnic foods were identified and grouped into two categories; Asian-type and Hispanic-type foods.

Results: Over 800 ethnic foods were identified in FNDDS. The percentage of individuals reporting Asian-type and Hispanic-type foods was 31% and 24%, respectively. Those identifying their race/ethnicity as Other than Non-Hispanic (NH) White, NH Black, or Hispanic reported consuming the most Asian-type foods (57%) followed by NH Whites, NH Blacks, and Hispanics, 31%, 26%, and 24%, respectively. Hispanics reported consuming the most Hispanic-type foods (66%) followed by NH Whites, NH Blacks and Others, approximately 18% each.

Conclusions: Ethnic foods are being consumed by a large portion of the population. FNDDS is continuously updated with new foods and recipes to support WWEIA, NHANES dietary analyses. With oversampling of Asians for NHANES 2011-2014, emphasis is being placed on assuring that Asian-type ethnic foods reported by survey respondents reflect what is available in the U.S. food supply.

57. A comparison of field methods to assess body fat of minority elementary school children

T.W. Johnson¹, A.Y. Spurlock¹, J.S. Parrot², M.F. Cousins¹, J.W. Dunn,
G. Avery¹ and B. Hamilton^1
1 Troy University School of Nursing, Troy, Alabama, USA;
² University of Medicine and Dentistry of New Jersey, USA.

Background: School-based interventions (SBIs) that address pediatric obesity frequently rely on field-friendly surrogate markers to assess body composition. The purpose of this study was to compare two of these methods in a population disproportionately affected by incidence of pediatric overweight and obesity.

Methods: Adiposity of elementary school children (n=319) from a rural, underserved area was assessed by skinfolds and bioelectrical impedance analysis (BIA). Represented grades were kindergarten through third grade, and the students were predominantly African- and Mexican-American.

Results: The correlation of skinfolds to BIA was r=0.822 (P<0.001). The mean difference between BIA and
skinfolds methods was -4.06% (95% CI -3.46 to -4.65); 95% of individuals in the sample had a difference in body fat measurements of 6.78% to -14.9% indicating skinfolds underestimated percent body fat relative to BIA in this comparison.

Conclusions: This study demonstrated the Tanita BIA SC-331C model provided a reasonable estimate of percent body fat in a group of younger African- and Mexican-American school children. The mean difference between measurements was not clinically relevant since it was only 4%. However, 95 % of individual differences fell within a band from 6.78% to -14.9% indicating that the difference between skinfolds and BIA could be clinically meaningful for an individual child. Given the limitations of BMI and skinfolds to assess effectiveness of SBIs to reduce incidence of pediatric overweight and obesity, BIA derived from this Tanita
model represents a reasonable, practical approach in this population.

58. Factors Associated with Physical Activity Participation among Children

Ngai Kwan, Institute for Community Inclusion, University of Massachusetts Boston, Boston, MA and Anita N. Lee, Department of Health and Physical Education, Eastern Connecticut State University, Willimantic, CT

Background: Previous research suggested that participation in regular physical activity (PA) at young age plays a positive role in physical development and overall well-being for children.

Objective: The objective of this study was to explore the factors (Intrapersonal, Interpersonal and Environmental) associated with PA participation among children by the modified Theory of Planned Behavior.

Methods: Data from 2007 National Survey of Children Health (NSCH) was used with 60,365 children aged 6-17 as the final sample. PA participation was dichotomized as meeting current recommendations or not (McKay et al., 2007). A binary logistic regression was used to determine the association among determinants and PA participation among children.

Results: Controlled for other variables, the Odds Ratios (OR) estimation for male (OR = 1.75; 95%CI: 1.68-1.81) suggested that the expected odds of being active increased by about 75% for boys compared to girls. Hispanics (OR = 0.877; 95%CI:0.82-0.94) and Blacks (OR = 0.90 ; 95%CI: 0.84-0.96) are less likely to actively engage in PA compared to Whites. Children who spend more than 3 hours watching television or using computers are less likely to participate in PA (OR = 0.73; 95%CI: 0.71-0.76). Moreover, children with active parents (OR = 1.97; 95%CI: 1.90-2.04); living in a family with two parents (OR = 1.05; 95%CI: 1.01-1.10) ; feeling safe about the neighborhood community (OR = 1.13 ; 95%CI: 1.07-1.20); and living close to a playground (OR = 1.11; 95%CI: 1.05-1.16) are more likely to be physically active.

Conclusion: Existing PA programs for children should be evaluated to improve services for girls and children in minority groups . Promoting an active lifestyle among family members and providing a safe as well as an accessible environment are also crucial to encourage PA participation among children.

59. Calcium, potassium and fiber intake in yogurt consumers

Nielsen, SJ; Tang, Y

Background: Numerous studies have examined dairy consumption and associated nutrient intake profiles. Few of these studies focus specifically on yogurt consumption or its relationship to nutrient intake.

Objective: To analyze trends in the percent of adults consuming yogurt and to compare the mean intake of calcium, potassium and fiber of yogurt consumers to dairy non yogurt consumers and non-dairy consumers.

Design: Data from adults, ages 18 years and older, in 5 cycles of the National Health and Nutrition Examination Survey (NHANES); 1999-2000, 2001-2002, 2003-2004, 2005-2006, and 2007-2008. Sample sizes for the survey cycles ranged from 4529 to 5265. Percent of yogurt, non-yogurt dairy and non-dairy consumers were calculated. Mean nutrient intake for calcium, potassium and fiber were calculated for these 3 groups of consumers. Trends in the percentage of the population consuming yogurt, dairy (non yogurt) or consuming neither on a given day were examined. Differences in mean nutrient intake between the groups were examined using a t-test with a significance level of 0.05

Results: In 2007-2008 6.7% of US adults consumed yogurt, up from 3.7% in 1999-2000 (p<.05). Yogurt consumers had higher mean intakes of calcium, potassium and fiber than all other adults (p<.05).

Conclusion: Between 1999-2000 and 2007-2008, yogurt consumption almost doubled among US adults. Yogurt consumers had higher mean intakes of calcium, potassium and fiber, three of the four nutrients highlighted by the Dietary Guidelines for Americans 2010 as shortfall nutrients. Substituting a yogurt for another snack item may help to increase people’s intake of calcium and potassium.

60. Total vitamin and mineral intakes of the U.S. population

Donna G. Rhodes, Meghan E. Adler, John C. Clemens, Alanna J. Moshfegh

Background: The widespread use of dietary supplements (DS) can contribute substantially to vitamin and mineral intakes of the U.S. population. Beginning with National Health and Nutrition Examination Survey (NHANES) 2007-2008, intake estimates from food, beverages and dietary supplements can be easily combined since data are now collected using similar methods.

Objective: Estimate total vitamin and mineral intakes of the U.S. population using DS data in conjunction with food intake from What We Eat in America, NHANES.

Methods: The 2007-2008 sample included nationally representative one-day dietary intake data from 8421 males and non-pregnant/lactating females, aged 2 years and older. Intake from foods and beverages was collected using the 5-step U.S. Department of Agriculture Automated Multiple-Pass Method. The DS component, administered after the dietary recall, collected data on usage of all vitamins, minerals, herbals, and non-prescription antacids. Nutrients intakes for 22 vitamins and minerals were analyzed from food and DS, and then combined for total nutrient intake.

Results: Overall, 37% reported taking at least one DS; prevalence was highest (70%) in females 60+ years and lowest (16%) in adolescents aged 12-19 years. Use of DS containing calcium, reported by 27% of the sample, was most common. Calcium intake (mean ± SE) was 142 ± 9 mg/d from all multi- and/or single-ingredient DS and 946 ± 20 mg/d from food. Intake of vitamin D (µg/d), reported by 26% of the sample, was 4.1 ± 0.5 and 4.6 ± 0.1 from DS and food, respectively. For vitamin B12, intakes from DS (25.1 ± 4.7 µg/d) were much higher than intakes from food (5.2 ± 0.1 µg/d). Results will be updated to include data from NHANES 2009-2010.

Conclusions: To accurately assess vitamin and mineral intakes of the U.S. population, nutrient intakes form DS need to be included with nutrient intakes from foods.

61. When America Eats

Lois C. Steinfeldt and John C. Clemens
USDA, ARS, Beltsville Human Nutrition Research Center, Beltsville, MD 20705

Background: Both differences and similarities exist in the mean daily intake of energy and other nutrients between population subgroups. Similarly, population subgroups exhibit a variety of food consumption patterns.

Objective: Are there significant differences in nutrient intake by time of day and weekdays compared to weekends between population subgroups?

Methods: This analysis was conducted using What We Eat in American, National Health and Nutrition Examination Survey 2007-2008 day 1 dietary intake data for males and females ages 20 and over. Intakes on Monday to Friday were classified as weekdays and Saturday and Sunday as weekends. The intakes were divided into 24 hourly periods. All statistical tests were done at p < .001. Standard errors were estimated using SUDAAN 10.0.1.

Results: Males age 20 and over have significantly higher mean daily energy intake on weekends (2655 kcal ± 40) compared to weekdays (2480 ± 33). Their mean energy intake by hour of the day shows differences in the morning and early afternoon with significantly lower intake at 6 am (23 ± 3 for weekends compared to 60 ± 6 for weekdays) and significantly higher intakes at 9 am (138 ± 12 for weekends compared to 82 ± 7 for weekdays) and 2 pm (149 ± 9 for weekends compared to 99 ± 8 for weekdays). There are no other significant differences. Females age 20 and over do not have significantly different energy intake on weekends compared to weekdays, but do show significantly higher intake at 3 pm (104 ± 8 for weekends compared to 68 ± 4 for weekdays).

Conclusions: There are differences in the patterns of mean energy intake by time of day between weekends and weekdays for both males and females. This is true whether or not there are differences in mean daily energy intake.

62. Barriers to Healthy Eating: A Focus Group Study among Women at a Transitional Housing Facility

Heather Brotsos

Background: Poor nutrition is a serious health concern across the United States. The obesity epidemic has become a central issue in the national public health discourse in recent years. American obesity rates have more than doubled from “15% in 1976-1980 to 31% in 1999-2002” (Kim 432).

Objective: To understand eating habits and food perspectives as they relate to healthy eating, nutrition and cooking among women in a transitional housing facility.

Methods: Focus groups were held at a transitional housing facility: one with staff and a second with residents. Questions focused on the eating habits of residents and staff, food sources and budget of residents, definitions of a well-balanced diet and healthy eating, as well as barriers to achieving a healthy diet. Results were compiled, systematically coded and analyzed using a grounded theory approach.

Results: Five key themes were uncovered as they relate to these concepts: an individual’s family and culture, finances, peer group, time and effort requirement, and knowledge and skills.

Conclusions: Through this analysis, the most critical barrier that emerged was the inability to translate a desire to change into action. Participants had a high degree of self-awareness with regard to the food choices they were making and the consequences of those choices on their health in both the short and long term. Knowledge and cooking skills were not primary barriers for many. For those whom this was, the barrier generally appeared to be one that has only emerged within their generation.

Three key themes emerged that inhibit this behavior change: a perception that healthy food is always more expensive than unhealthy food, a belief that a meal is not complete without a meat component and a lower level of motivation when caring for oneself, compared to others.

These factors must be considered in development of effective intervention strategies.

63. Obesity and Mortality: Are the Risks Declining? Evidence from a Meta-Analysis of Eighteen Prospective Studies in U.S.

Primary Author: Tapan Mehta; Department of Biostatistics, University of Alabama at Birmingham
Co-Authors: Kevin R. Fontaine, Ph.D. Scott W. Keith, Ph.D. Nicholas M. Pajewski, Ph.D. Gustavo de los Campos, Ph.D. Sai Santosh Bangalore, M.S. Alfred Bartolucci, Ph.D. David B. Allison, Ph.D.

Background: Previous analyses of the National Health and Nutrition Examination Survey have suggested that the mortality rate associated with obesity has decreased over calendar time in the U.S. population, potentially due to improvements in the medical care of obesity and its associated diseases.

Objective: To evaluate whether this reduction in the detrimental association of obesity with mortality rate over calendar time is supported by data from an additional 18 U.S. prospective studies.

Methods: Eighteen U.S. based prospective studies with at least two assessments of body mass index (BMI, kg/m2) at different calendar years (waves) were acquired from National Heart Lung and Blood Institute and the Inter-University Consortium for Political and Social Research data repositories, and through contact with investigators. Individual-level data for 2 waves were extracted from each of these studies with a total sample size of 215,243 and 55,460 deaths. Log hazard ratios were estimated by BMI category (i.e., underweight : <18.5, normal weight: 18.5 to <25, overweight: 25 to <30, grade 1 obesity: 30 to <35, grade 2-3 obesity: 35), separately for each study wave using a Cox regression, after adjusting for smoking, sex and race. These log hazard ratios were then pooled as a response in a weighted random effects meta-regression to test for interactions between BMI categories and calendar time after controlling for length of follow-up.

Results: From 1960 to 1990, the deleterious associations of overweight and grade 1 obesity decreased by 14.9% (p=0.01) and 10.5% (p=0.19) respectively. The deleterious association of grade 2-3 obesity in the same period increased by 25.1% (p=0.06).

Conclusions: Our results suggest that there were no consistent reductions in the deleterious associations of obesity (grade 1 or 2-3) with mortality rate. There is, however, clear and compelling evidence that the deleterious association of overweight has declined over calendar time.

64. The Impact of Physical Activity and a Healthy Eating Diet on Post-Heart Attack Mortality: Analysis from the Third National Health and Nutrition Examination Survey (NHANES III) and Linked Mortality Files

Steven Lee
The George Washington University

Background: The Third National Health and Nutrition Examination Survey (NHANES-III, 1988-1994) estimated the 3.4% of the adult population who have experienced a heart attack. A follow-up through the end of 2006 showed that of the samples who had died since the initial survey, a majority (55.1%) had died as a result of heart disease. Such statistics reaffirm the need to investigate self-control factors that may be protective against post-heart attack mortality.

Objective: The study sought to examine the impact of two self-control factors – Physical Activity Level (PAL) and Healthy Eating Diet (HED) – on post-heart attack mortality.

Methods: This study selected 938 adult samples with history of heart attack from the NHANES-III cohort. PAL and HED data as well as data of several confounding factors (e.g., Age-at-initial-survey, Diabetes, Blood Pressure/Cholesterol, Tobacco Use, etc…) were obtained from NHANES-III. Follow-up heart-disease-related mortality and post-survey-survival-time were obtained from the NHANES-III-Linked Mortality File. Several competing weighted Cox Proportional Hazards models with the post-survey-survival-time as the outcome variable and the research and confounding factors as explanatory variables were tested. The adjusted PAL and HED Hazard Ratios (HR, relative risk) for heart-linked mortality were each estimated by holding the age-at-initial-survey and the other research and confounding variables constant and by taking into account the design effect of the NHANES-III using SUDAAN software.

Results: With regard to PAL, the two subgroups with the least PAL had post-heart attack mortality risks that were 87% greater (HR=1.87, p <0.05) and 122% greater (HR=2.22, p <0.05) than did that of the subgroup with the highest PAL, respectively. The overall quality of individuals’ HED appeared to be neutral (HR=1, p >0.05) in this regard.

Conclusions: The results indicate that adults with a history of heart attack would benefit from greater physical activity. Unexpectedly, a healthy eating diet played no significant role.

65. Diabetes on Navajoland: Meat, Marvels and Mysteries

Primary Author: Kelly Gneiting, MAT Geography; Co-Author: Alan French, Bachelor of Science Nursing, Registered Nurse, Diabetes Management (DM) Educator

Background: We will be researching the prevalence of diabetes within the eastern portion of the Navajo Nation—Northern Apache County, AZ and Western McKinley County, NM (henceforth TMCR).

Objective: Why is diabetes so prevalent among Navajos living 60 miles from Tsehootsooi Medical Center (TMC, located in Fort Defiance, AZ), yet nowhere else on the Navajo Nation Reservation or among Native American groups elsewhere?

Methods: Include:

1. A comparison of TMCR’s diagnosed diabetes rates with the other counties in the U.S., and especially the 100+ counties with a high proportion of Native American population.
2. Maps and X-Y Scatterplots showcasing the Pearson and R-squared Correlation Coefficients for Percent Native American population and Diabetes, in relation to #1 above.
3. The variables of Inactivity and Obesity are included in further determining cause-effect relationships. (Data sources for #1 -#3 are “County-Level Estimates of Diagnosed Diabetes”, from CDC, and county-level ethnicity population proportions from the 2010 U.S. Census)
4. A comparison of TMCR’s mortality rates AND ranks attributable to diabetes with the Navajo Nation’s 6-county region, the state of Arizona, and the United States, and by Native American/Non-Native American ethnicity and Age Groups. (Data source: CDC)

Results: Rates of diagnosed diabetes at TMCR are unexplainably high, according to various measures including those related to the Pearson Correlation Coefficient. In addition, detailed study of the #4 Method reveals comparatively sky-high ranks and rates of mortality being caused by diabetes at TMCR.

Conclusions: No conclusive evidence exists that suggests Navajos at TMCR are predisposed, by virtue of genetics (Nature), to the diagnosis of diabetes. This would indicate that an outside factor (Nurture) is the primary contributor of the disease. 31 charts/tables/maps/figures collaborates these conclusions. Extremely high diabetes rates at TMCR—some of the highest in the United States—remains a mystery.

66. Declining Rates of Hospitalization for Lower Extremity Conditions Among Adults Aged ≥20 Years with Diabetes – United States, 1988-2009

YanFeng Li, Nilka Ríos Burrows, Linda S. Geiss

Background: Hospitalization rates for nontraumatic lower extremity amputation (LEA) have declined in the U.S. diabetic population.

Objective: We examined whether hospitalization rates for lower extremity conditions (LEC) that are major risk factors for LEA, including peripheral arterial disease (PAD), foot ulcer/inflammation/infection, and neuropathy, have also declined in U.S. adults with diabetes.

Methods: Diabetes-related hospital discharges were identified in the National Hospital Discharge Survey (NHDS) as those discharges having an ICD-9-CM diagnosis code of 250. Hospital discharges with LEC (ICD-9-CM codes of 250.7, 440.2, 442.3, 443.8-443.9, or 444.22 for PAD; 454, 707.1, 680.6-680.7, 681.1, 682.6-682.7, 711.05-711.07, 730.05-730.07, 730.15-730.17, 730.25-730.27, 730.35-730.37, 730.85-730.87, 730.95-730.97, or 785.4 for ulcer/inflammation/infection; and 337.1, 357.2, 355, 358.1, 713.5, 094, or 250.6 for neuropathy) from the NHDS served as the numerator for rates, and the U.S. diabetic population derived from the National Health Interview Survey served as the denominator. Rates were age-adjusted based on the 2000 U.S. standard population. Joinpoint regression was used to calculate an annual percentage change (APC) and to assess trends.

Results: From 1988 to 2009, the age-adjusted LEC discharge rate per 1,000 diabetic population decreased by 2.2% per year from 21.4 to 12.4 (P<0.01). Similarly, significant declines were observed among people aged 45-64 and 65-74 years and among men and women. However, the LEC rate changed little among people aged 20-44 years throughout study period (APC=-0.1%, P=0.82). For people aged 75 and older, the LEC rate changed little from 1988 to 1996, but then declined significantly from 1996 to 2009 (APC=-5.9%, P<0.01).

Conclusions: National hospitalization rates for LEC in the diabetic population have decreased, with declines occurring in most demographic groups examined. The observed decline is similar to the decline in LEA, suggesting that these changes might be attributed to improvements in the incidence or treatment of conditions that underlie LEA.

67. Cardiovascular Risk Factor (CVD-RF) Management in Diabetics in the United States (US). Still an Elusive Target?

Talluri SK¹, Siddegowda Bangalore B¹, Talluri J¹, Marella PC², Besur S¹;
¹ Department of Internal Medicine, McLaren, Flint, Michigan ² Department of Internal Medicine, Banner Estrella Medical Center, Phoenix AZ.

Introduction: Individuals with diabetes have twice the risk of myocardial infarction and stroke as the general population. Although the incidence of CVD events in patients with diabetes has declined over the past decade, many studies have shown that CVD-RF control has been inadequate.

Objective: Our study objective is to re-examine achievement of the of American Diabetic Association (ADA) goals for CVD-RF modification in the United States (US).

Methods: We conducted a secondary data analysis of a retrospective cohort of adult diabetic patients included in the 2008 NHANES. We excluded those not seeing a doctor for their diabetes. The primary outcome was to measure the proportion of US diabetic patients that met ADA diabetes care guidelines on control of blood sugar, blood pressure, cholesterol and albuminuria and secondarily examine gender differences. We compared adherence to guidelines among men and women using Chi-square test.

Results: There were 759 adults with diabetes in our sample (projected to 18.5 million in the US). We included 577 adults with diabetes (projected to 14.6 million in the US). The goal of hemoglobin A1C < 7 was met in 49% (95% CI 39%-59%), LDL was < 100 in 62% (95% CI 54%-69%), triglycerides < 150 in 54% (95% CI 46%-62%), blood pressure < 130/80 in 51% (95% CI 41%-62%), HDL goal of > 40 in men and > 50 in women was met in 60% (95% CI 50%-69%), urine albumin creatinine ratio < 30 in 59% (95% CI 51%-66%). Men were more likely to meet guideline goal achievement only for LDL, odds ratio 2.7 (95% CI 1.4-5.3, P =0.006), and HDL, odds ratio 2.18 (95% CI 1.3-3.5, P =0.003).

Discussion: Treatment goals for CVD-RF were met in slightly more than half of diabetic patients. Men are more likely to meet LDL and HDL goals compared to women.

68. Profile on Emergency Department Visits among Adults with Diabetes, US 2009

Jing Wang, Linda S Geiss

Background: Prior research estimated that the number of diabetes related Emergency Department (ED) visits doubled from 1997 to 2007.

Objective: To better understand ED utilization by people with diabetes, we profiled ED visits among adults aged 18+ years with diabetes and compared to those without diabetes in the U.S.

Methods: ED visits were identified in 2009 National Hospital Ambulatory Medical Care Survey. Diabetes status was determined by a checkbox on diabetes or ICD-9-CM code of 250 in any of three diagnoses. To calculate ED visit rates, we used as denominators the number of adults with and without diabetes estimated from 2009 National Health Interview Survey.

Results: In 2009, there were 10.6 million ED visits by adults with diabetes, representing 10.3% of all adult ED visits. ED visits by diabetic adults were more urgent based on triage and resulted in higher hospital admission rates (p<0.01). Overall age-adjusted ED visit rates were 54.8% (+4.5%) for diabetic adults and 45.6% (+3.1%) for non-diabetic adults (p=0.09). Older adults (age ≥75 years) and non-Hispanic blacks had higher ED rates regardless of diabetes status. Among persons aged 18-44 years or women, diabetic adults had notably higher ED rates than their non-diabetic counterparts. ED rates for ischemic heart disease and heart failure were higher for adults with diabetes than for adults without (all p<0.05).

Conclusions: ED visits among diabetic adults tend to be more complicated and more likely to involve cardiovascular disease, suggesting the importance of preventing diabetes complications to reduce ED burden. Higher ED utilizations by diabetic adults aged 18-44 years and women warrants further investigation.

69. Variants in or near PNPLA3, GCKR and PPP1R3B associated with ultrasound-defined hepatic steatosis (fatty liver) in NHANES III

Ruben Hernaez^1,2,3, Jody McLean⁴, Mariana Lazo^1,3, Frederick L. Brancati^1,3,5, Joel N. Hirschhorn^6,7,8, Genetics of obesity-related liver disease (GOLD) Consortium, Thutrang Nguyen^6,7, Ihab R. Kamel⁹, Susanne Bonekamp⁹, Mark S. Eberhardt¹⁰, Jeanne M Clark^1,3,5, Wen Hong Linda Kao^3,5, Elizabeth K. Speliotes^7,11

1) Department of Medicine, Johns Hopkins School of Medicine, Baltimore, MD.
2) Department of Medicine, Washington Hospital Center/Georgetown University Hospital, Washington, DC;
3) Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD;
4) NOVA Research Company, Bethesda, MD;
5) Welch Center for Prevention, Epidemiology and Clinical Research, Johns Hopkins University, Baltimore;
6) Divisions of Endocrinology and Genetics and Program in Genomics, Children’s Hospital, Boston, MA;
7) Broad Institute, Cambridge, MA;
8) Department of Genetics, Harvard Medical School, Boston, MA;
9) Department of Radiology, Johns Hopkins School of Medicine, Baltimore, MD;
10) National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, MD
11) Department of Internal Medicine, Division of Gastroenterology, and Center for Computational Medicine and Bioinformatics, University of Michigan, Ann Arbor, MI

Background: A genome-wide association study identified five genetic variants associated with computerized tomography measured hepatic steatosis (HS) in individuals of European ancestry.

Objective: Test whether these variants associated with ultrasound-measured HS in participants (20-74 yrs) in the third National Health and Nutrition Examination Survey (NHANES III Phase-2).

Methods: Genetic and HS data was obtained from NHANES III, a nationally representative population based survey of non-institutionalized US residents. HS was defined as any steatosis observed by ultrasound. We tested whether HS was associated with rs738409 (PNPLA3), rs2228603 (NCAN), rs12137855 (LYPLAL1), rs780094 (GCKR), and rs4240624 (PPP1R3B) with an additive genetic model in population-specific logistic regression analyses adjusted for age, age2, gender, and alcohol use. Using population-weighted regressions, we estimated overall association across three ancestries, non-Hispanic White (NHW), non-Hispanic Black (NHB), Mexican-American (MA).

Results: Of the 4,804 adults [1,825 NHWs, 1,442 NHBs, and 1,537 MAs] 57.2% were female; mean age at exam was 43.2 years; and the weighted HS prevalence was 37.3%. The G allele of rs738409 (PNPLA3) and T allele of rs780094 (GCKR) were associated with HS with high alanine aminotransferase (ALT) (OR: 1.36, p=0.01 and 1.30, p=0.03 respectively); the A allele of rs4240624 (PPP1R3B) and the T allele of rs2228603 (NCAN) with HS (OR 1.33, p=0.0129 and 1.40, p=0.04 respectively); and the variants at PNPLA3 and NCAN with ALT across all three ancestries. Using SNP testing, the following showed significant associations in race/ethnicity (for NHW: PNPLA3, NCAN, GCKR and PPP1R3B, for MA PNPLA3; none for NHB).

Conclusions: This study validated the association between rs738409 (PNPLA3), rs780094 (GCKR), and rs4240624 (PPP1R3B) and HS with or without elevated ALT across three ancestries for the first time in NHANES III. Some but not all associations between variants in NCAN, LYPLAL1, GCKR, PPP1R3B and HS with and without elevated ALT were significant within sub-populations.

70. Visualizing the Data: Using visual analytics software to paint a picture of community need

Catherine Rains

Background: An urban pediatric hospital provides benefit to the surrounding community through free medical and health education programs. Services include asthma education and treatment, dental exams and restorative care, and a mobile medical clinic providing vaccinations, growth analysis, and screening for blood lead level, scoliosis, iron deficiency anemia, vision and hearing impairment. Health education services include home and vehicle safety education, car seat installation and pedestrian safety education.

Objective: As part of the hospital’s community health needs assessment and subsequent strategic planning, the hospital sought to visualize where services were being provided by hospital outreach compared to other agencies in the area.

Methods: Internet search engines and referrals from hospital partners were used to identify agencies with similar services. Agencies were called via phone to verify services offered.

Tableau software was used to convert the latitude and longitude of locations of community services provided by the hospital and other agencies into a map to visualize areas of highest need and pinpoint areas that have a gap in health services. The map of services was underlain with median household income and population from United States Census data, a native feature in Tableau.

Results: The results show that hospital outreach programs are unique in the community and are currently being offered in locations of high need. Additionally, the project has shown that areas of expansion or partnership include services to rural and suburban areas that are greater than 25 miles away from the hospital.

Conclusion: This project has helped to justify current and future funding for hospital outreach programs. The data generated from this project along with public health data will be used to determine areas that have the greatest need of services and will be used to enhance the hospital’s community health needs assessment.

71. Racial Differences in Refusal of Surgery for Colorectal Cancer in the United States

Talluri SK, Besur S Talluri J
Department of Internal Medicine, McLaren, Flint, Michigan.

Background: Colorectal Cancer (CRC) is the second leading cause of cancer death in the United States. Surgery is the most common treatment for all stages of CRC. The racial variations in surgical treatment for CRC persist even after adjustment for disease stage at presentation, lower socioeconomic status and lower educational levels. Refusal of surgery in CRC patients as a reason for disparity in survival among different races has not been studied.

Objective: The primary objective of our study is to determine racial differences in refusal of recommended surgery in CRC patients; the secondary objective is to investigate other predictors of refusal.

Methods: We analyzed a retrospective cohort of CRC patients included in the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) database from 1988 to 2008. We included Caucasian and African -American patients with primary CRC and age greater than 65 years. We excluded patients in whom surgery was not recommended, surgical status was unknown and reason for not performing surgery was unknown.

Results: We included 160,412 patients with CRC. The majority of the study population was Caucasian (92%). African-Americans presented with advanced metastatic disease as compared to Caucasians (16% vs. 12.6%). Most patients underwent recommended surgery in both racial groups (Caucasian (98.6%) and African-Americans (97.2%). African-Americans were more likely to refuse surgery (2.8% vs. 1.4%) compared to Caucasians (P<0.001). The odds-ratios (ORs) and corresponding 95% confidence intervals (CIs) for patient characteristics that independently predicted patient refusal of surgery for CRC are: African-American race (OR=2.3, CI=2.05-2.6); age (OR=1.12, CI=1.11-1.28); male gender (OR=1.28, CI=1.17-1.40); marital status (OR=0.52, CI=0.47-0.58). As expected, mean survival time was significantly lower in those who refused surgery (3 yrs) as compared to those who had surgery (14.3 yrs) (P <0.001).

Conclusion: African-Americans are twice as likely to refuse recommended surgery for colorectal cancer as compared to Caucasians.

72. Racial/Ethnic Disparities in Healthcare Receipt among Male Cancer Survivors: Findings from the National Health Interview Survey, 2006-2010.

Nynikka R. A. Palmer, MPH, DrPH¹; Ann M. Geiger, PhD²; Tisha M. Felder, MSW, PhD^3,4; L. Douglas Case, PhD⁵; Lingyi Lu, MS⁵; Kathryn E. Weaver, PhD, MPH^1
1 Department of Social Science and Health Policy, Division of Public Health Sciences, Wake Forest School of Medicine, Winston-Salem, North Carolina
² Department of Epidemiology and Prevention, Division of Public Health Sciences, Wake Forest School of Medicine, Winston-Salem, North Carolina
³ Department of Clinical Pharmacy and Outcomes Sciences, South Carolina College of Pharmacy, University of South Carolina, Columbia
⁴ South Carolina Cancer Prevention and Control Program, Arnold School of Public Health, University of South Carolina, Columbia
⁵ Department of Biostatistical Sciences, Division of Public Health Sciences, Wake Forest School of Medicine, Winston-Salem, North Carolina

Background: Men, especially minorities, underutilize healthcare services. This may be particularly problematic after cancer diagnosis and treatment.

Objective: We examined racial/ethnic disparities in healthcare receipt among a nationally representative sample of U.S. male cancer survivors.

Methods: Using data from the 2006-2010 National Health Interview Survey, we identified a cohort of men aged 18 years and older who reported ever being told by a healthcare provider that they had cancer. Four self-reported measures assessed healthcare receipt: primary care visit in past year, specialist visit in past year, flu vaccination in past year, and ever having the pneumococcal vaccination. We employed hierarchical logistic regression modeling, stratified by age (younger: <65 versus older: 65+), to examine racial/ethnic disparities in healthcare receipt. We adjusted for race/ethnicity, age, marital status, education, health insurance, usual place for care, comorbidities, health status, functional limitations, time since diagnosis, and number of cancer diagnoses.

Results: In adjusted models, younger African-American survivors were almost twice as likely as non-Hispanic Whites to not see a specialist (odds ratio [OR] 1.92, 95% confidence interval [CI] 1.14-3.24). Among older survivors, African-Americans and Hispanics were approximately twice as likely as non-Hispanic Whites to not see a specialist (OR 1.78, CI 1.19-2.68 and OR 2.09, CI 1.18-3.70, respectively), not receive the flu vaccine (OR 2.21, CI 1.45-3.37 and OR 2.20, CI 1.21-4.01, respectively) and not receive the pneumococcal vaccine (OR 2.24, CI 1.54-3.24 and OR 3.10, CI 1.75-5.51, respectively). There were no statistically significant differences found for primary care receipt in either age group.

Conclusions: Racial/ethnic disparities in healthcare receipt were apparent among older survivors, despite them having access to Medicare. This may have implications for their health and receipt of follow-up care if minority male survivors are less likely to see specialists, including oncologists, and get basic preventive care.

73. Using Achievable Benchmarks and Time to Benchmark to Highlight Disparities in Healthcare Services: Cancer Case Study.

William Freeman, MPH; Barbara Barton, MPH; Atlang Mompe
Agency for Healthcare Research and Quality, Center for Quality Improvement and Patient Safety

Background: In 2010, the National Healthcare Quality and Disparities Reports (NHQR-DR) began presenting achievable benchmarks based on the performance by top States for many measures. In conjunction with benchmarks, the NHQR-DR began calculating the time for different subgroups to achieve the benchmarks, providing a new way to view disparities in healthcare quality.

Objective: To present time to benchmark for NHQR-DR measures related to the quality of cancer care.

Methods: Cancer care measures presented in the 2011 NHQR-DR were analyzed using achievable benchmarks and time (years) to benchmark. The achievable benchmarks are averages of the most recent data year estimates for the top 10% of reporting States. Time to benchmark for each racial/ethnic, gender, income and insurance subpopulation was determined by first calculating the annual rate of change by regression. Then the number of years to reach the benchmark was extrapolated based on the annual rate of change for each subpopulations.

Results:

• Preventative: The 2008 top 5 State achievable benchmark for cervical cancer screening was 89% for women ages 21-65. Women with private insurance have achieved the benchmark, while women in other insurance groups were moving away from it.
• Treatment: The 2008 top 5 State achievable benchmark for women treated with breast-conserving surgery who received radiation therapy within one year of diagnosis was 93%. High-income women could achieve this benchmark in about 7 years, but it will take all other income groups over 20 years.
• Outcome: The 2008 top 5 State achievable benchmark for people age 50 and over who have colorectal cancer diagnosed at advanced stage was 67 per 100,000 population. Women have already achieved the benchmark, while men would take 4 years.

Discussion/Conclusions: The time needed to reach a benchmark poignantly highlights the large disparities that currently exist, illustrating the importance of reducing disparities and setting realistic goals.

74. A Population-Based Survey of Celiac Disease in the United States

Talluri SK, Besur S, Talluri J
Department of Internal Medicine, McLaren-MSU Internal Medicine Residency Program, Flint, MI

Background: Celiac disease is a gastrointestinal disorder triggered by the ingestion of gluten-containing food products. Gluten triggers an immunological reaction in the small intestine of susceptible patients and causes malabsorption of vitamins and minerals. A previous study of healthy blood donors found the serological prevalence based on anti-endomysial antibodies to be 1:250. The serological prevalence of celiac disease in the United States based on the presence of anti-tissue transglutaminase antibodies and anti-endomysial antibodies was not studied in the general US population.

Objective: The primary objective of our study is to estimate the disease prevalence as well as serologic prevalence of celiac disease in the United States.

Methods: We conducted a secondary data analysis of a cross-sectional sample of patients included in NHANES (2009-2010). We included all adult (18 years and older) patients included in NHANES (2009-2010). We calculated the prevalence of celiac disease. We estimated serological prevalence by identifying the individuals with either the presence of anti-endomysial or anti-transglutaminase antibodies. We analyzed the data with SPSS complex samples 19.

Results: We included 6360 adults of NHANES 2009-2010 in this study (projected to 226 million in the US). The number of adults with celiac disease was 14 (projected to 396,968 in the US). Adults with celiac disease were predominantly women (63%), non-Hispanic White (66%), and had health insurance (82%). The estimated prevalence (95% CI) of celiac disease was 0.1% (0.1%-0.2%). The serological prevalence (95% CI) of celiac disease was 0.9% (0.7%-1.3%). The proportion of adults (95%CI) with celiac disease who eat a gluten-free diet was 52.1% (27.5%-75.7%).

Conclusion: The serological prevalence of celiac disease in the United States was much higher than prevalence of celiac disease. Slightly more than half of adults with celiac disease adhere to a gluten-free diet.

75. Women’s Access to and Use of Hormonal Contraceptives, 1995 – 2006-2010

Jo Jones, Ph.D., and Kimberly Daniels, Ph.D.
National Survey of Family Growth
National Center for Health Statistics

Background: Five decades ago the oral contraceptive pill and the inert IUD were introduced in the U.S. The pill remained the only hormonal contraceptive available for 30 years. In 1990, the 5-year implant was approved by the FDA, followed by a 3- month injectable in 1992, a 1-month injectable in 2000, a low-dose, progestin-only IUD, the contraceptive patch and ring in 2001, and a 3-year implant in 2006. These contraceptive methods are considerably more effective at preventing unintended pregnancy, compared with traditional methods, e.g., withdrawal or the condom. Given the relatively recent introduction of these highly effective contraceptive methods, descriptive work on women’s use of these methods across demographic and community characteristics is essential information for policy makers interested in reducing unplanned pregnancies – about half of all pregnancies.

Objective: To describe the introduction of highly effective contraceptive methods; the current use of the pill, the IUD, and other hormonal methods, by selected socio-demographic and contextual variables; and to identify groups of women who may be underutilizing these methods and, thus, putting themselves at increased risk of an unintended pregnancy.

Methods: Using data from the 2006-2010 National Survey of Family Growth, we use bivariate and multivariate analysis to describe differences in use of the pill, IUD, and other hormonal methods across socio-demographic characteristics, such as education, income, insurance, and Hispanic origin and race, as well as metropolitan residence and U.S. region.

Results: Differential patterns of method use are found across race and ethnicity, age, marital status, parity, education, insurance coverage, and region of residence in the use of the pill, the IUD, and other hormonal methods of contraception.

Conclusion: At any point in time, a woman’s choice of contraception is influenced by her individual characteristics, the context in which she lives which creates differential access, and her reproductive life-course stage.

76. The Weakening Link: The Sequencing of First Sex, First Marriage, and First Births among Women in the United States: National Survey of Family Growth, 1988-2010

Gladys M. Martinez; Casey E. Copen
National Center for Health Statistics (NCHS), CDC

Background: Family life in the United States has undergone rapid changes as noted by decreased age at first sexual intercourse, delayed or forgone marriage and increases in nonmarital childbearing. The increase in nonmarital births, as well as an increase in the age at first marriage, suggests the linkage between marriage and childbearing in the U.S. is weakening or may have been replaced by a new ‘normative’ sequence. This poster explores racial differences and trends over time in the sequencing of first sex, first marriage, and first births. We focus not only on the marital status of the first birth but also on the different paths women experienced from the time they became sexually active.

Methods: Data from the 1988, 1995, 2002 and 2006-2010 National Survey of Family Growth are used to examine the sequencing of first vaginal intercourse, first marriages and first births and how these patterns have changed over time, focusing on race differentials.

Results: For all women, the average age at first marriage was younger than the average age at first birth in 1988. By contrast, in 2006-2010 the age at first birth was younger than the age at first marriage, indicating that women are now more likely to have a first birth before marriage. The difference between the average age at first sex, age at first marriage and age at first birth widened for white, Hispanic and black women between 1988 and 2006-2010.

Conclusions or Significance: This poster showcases the marriage and fertility histories in the NSFG which are an ideal resource to examine the sequencing of these important life transitions among American women and how these patterns have changed over time.

77. Ambulatory Health Care Visit by Patients with Parkinson Disease in the United States 2005 -2008

Wenyu “Andy” Sun, Anbesaw W. Selassie, Paul B. Pritchard
Medical University of South Carolina, Department of Neurosciences

Background: To our knowledge, there is very little information on the patterns of ambulatory health care visit among patients with Parkinson disease in the United States.

Objective: To describe ambulatory health care visits by patients with Parkinson disease in the United States during 2005-2008.

Methods: Patients visits attributed to Parkinson disease (both primary and secondary diagnoses) in the National Ambulatory Medical Care Survey (NAMCS) and National Hospital Ambulatory Medical Care Survey (NHAMCS) from 2005 to 2008 severed as the study basis for epidemiological analysis. The unit of analysis was the visit. All estimates have been adjusted using weights provided by the National Center for Health Statistics (NCHS) to account for the multi-stage sampling design of the NAMCS and the NHAMCS. PD patient visit was defined by ICD-9-CM coding of 332.

Results: There was an estimated 6.9 million PD patient visit to the physicians’ offices and hospital outpatient departments during 2005 – 2008. Top 3 specialties being visited were neurology (48%), Internal medicine (20%), General/Family practice (13%). 53% of all PD visits were made by males. 94% of all PD visits were made by the white. 99% of all PD visits were made by patients 45 years and above, 75 years and over age group account for half of all PD visits. 64% of all PD visits were paid by Medicare, private insurance covered 25% and Medicaid covered 7% of all PD visits. Top co-morbidities for PD visits included hypertension 38%, Diabetes 20%, Depression 16%, Arthritis 16%, Hyperlipidemia 16%, cerebravascular disease 11%, ischemic heart disease 11%, cancer 5% of all PD visits.

Conclusions: Health care providers and policy-makers should be aware of this pattern in order to develop strategies that lead to better ambulatory health care for PD patients. Further investigation to explore the relationship between PD and co-morbidities is warranted.

78. Prevalence of Helminth parasites from vertebrates at Maharashtra region, India.

Dr. Hemlata Wankhede

Background: Study of helminth parasites from vertebrates which are consumed by many people as a main source of food. I.e. fishes, mammals and birds affecting the human health hence the biological study of helminth parasites are essential to give the awareness for proper diagnosis, treatment and preventation of parasitic infection in medical field.

Objective: The parasites constitute major health problem in livestock and in human throughout the world. For example approximately 30% of the world population is infected with helminth parasites. The incidences of many parasitic diseases viz. Schistosomiasis, Hydatidiasis, Ascarasis, and Taeniasis have increased in the recent years.

Methods: The cestodes parasites and trematodes were flattened and preserved in 4% formalin,stained with Harris haematoxyline . Nematode parasites preserved in 10% glycerine jelly,and glycerine alcohol . Prevalence is calculated by following Formula. Prevalence of infection = (Infected host / Total host examined) X 100

Results: The cestode parasites collected were Circumoncobothrium (Shinde, 1968), Cotugnia(Diamare, 1893) Monozeia (Blanchard, 1891)from the host viz. Mastuscembelus armatus, Gallus gallus domesticus and Capra hircus. The trematode parasites collected were Azygia (Looss, 1899) from the host Clarias batrachus. The nematode parasites collected were Bunostomum (Railliet, 1902) and Heterakis (Dujardin, 1845) from the host Capra hircus and Bandicota bengalensis

Conclusions: The prevalence of cestode parasites Circumoncobothrium from the host Mastuscembelus armatus were (32.66%) in rainy season, (61.83%) in winter and (73.81%) in summer. Cotugnia from the host Gallus gallus domesticus infection (31.66%) in rainy season (65.12%) in winter and (68.48%) in summer. Moniezia from the host Capra hircus were (49.11%) in rainy season, (69.66%) in winter and (67.03%) in summer. The prevalence of Trematode parasites Azygia from the host Clarias batrachus were (27.10%) in rainy season, (25.84%) in winter and (18.05%) in summer. The prevalence of Nematode parasites Bunostomum from the host Capra hircus were (22.65%) in rainy season, (43.20%) in winter and (30.62%), in summer Heterakis from the host Bandicota bengalensis were (22.51%) in rainy season, (39.95%) in winter and (23.48%) in summer.

79. Increase in mortality rates from Clostridium difficile Infection: United States, 1999-2009

Jiaquan Xu, MSB/DVS/NCHS/CDC

Background: Clostridium difficile (often called C. difficile) is an anaerobic, Gram-positive, spore-forming bacillus that can cause symptoms ranging from diarrhea to life-threatening inflammation of the small intestine and colon (Clostridium difficile Infection, CDI). The majority of CDI are typically associated with use of systemic antibiotics and most commonly affect older adults in hospitals or in long term care facilities. The mortality from CDI increased dramatically since 1999 in United States.

Objective: To find out the CDI mortality trend by age, race and sex from 1999 to 2009 in the United States.

Methods: Data represented in this report are based on information from all resident death certificates filed in the 50 states and District of Columbia. The statistics information is compiled in a national database through the Vital Statistics Cooperative Program for years 1999-2009. The cause code A047 for CDI is classified in accordance with the 10th revision of the International Classification of Diseases (ICD-10).

Results: Mortality rate for CDI increased by 7.3 times from 1999 through 2009. More than 90 percent of the CDI deaths occurred to persons aged 65 years or older. Men and women are at the similar risk of dying from CDI. Majority of CDI deaths occurred in hospital (80%). Mortality rates for CDI were higher for non-Hispanic white than for non-Hispanic black and Hispanic population.

Conclusions: CDI is a rising cause of death for persons aged 65 years and over. It was the 18th leading cause of death in 2009 moving from 29th leading cause of death in 1999 for people aged 65 years and over.

80. Joint modeling of multivariate longitudinal measurements and competing risks survival data: application to the study on older Mexican Americans

Ping Yao
School of Nursing and Health Studies, Northern Illinois University

Background: Research on the longitudinal association between depression and mortality in elderly population remains inclusive. The main issue with longitudinal outcomes is caused by non-ignorable missing values due to death and dropout process. Joint modeling of longitudinal and competing risk model can be conducted to fit this gap.

Objective: To use a joint modeling approach to examine the association between longitudinal depressive symptoms and mortality under controlling non-ignorable dropout.

Methods: We propose a joint model that consists of a linear mixed sub-model for the longitudinal outcome, and a proportional hazards sub-model for the competing risks of death and dropout process, linked together by shared random effects. The proposed method is applied to a data set, from Hispanic Established Population for Epidemiological Study of the Elderly (HEPESE) that consisted of repeated measurements of Center of Epidemiological Studies Depression Scale (CES-D) and time to death and dropout for 3050 Mexican-Americans. The results are adjusted by demographic, social-economic and health relative variables such as sex, age, body mass index(BMI), education, marriage status, household income, chronic disease, daily activities, smoking status, and prescription medicine.

Results: overall the total mortality rate was 30.59% in older Mexican-Americans. CES-D at baseline was significant risk factor of mortality at 8 year follow-up with and without adjusting for baseline covariates. The joint analysis indicated that the intercept (estimated baseline of CES-D ) and slope (change of CES-D) of linear trajectory of CES-D are associated with mortality. But for drop-out, the joint analysis indicated that the intercept is significantly risk factor and slope is not.

Conclusion: Joint modeling of competing risks with longitudinal outcomes fits the depression and mortality data well. Baseline CES-D is significant predictive of mortality and dropout, which indicates higher CES-D at baseline is associated with higher hazards of death and dropout

81. Analyzing the Determinants of Health Care Provider Choice During Recessionary Periods

Kelley L. Cullen, PhD
Assistant Professor, Economics
Eastern Washington University

Background: The U.S. economy was in recession from March 2001 to November 2001 and then again from December 2007 to June 2009 resulting in significant increases in unemployment with many Americans finding themselves without health insurance. The loss of employer-provided health insurance causes growth in demand for government-provided health coverage that may also result in individuals making greater use of emergency rooms (ERs), failing to use preventative care, and waiting until an illness worsens before seeking care.

Objective: U.S. data collected from the National Health Interview Survey (NHIS) from 2000 to 2009 is used to examine how individuals change their choice of primary care provider during recessions.

Methods: A nested multinomial logit model is used to estimate choice of primary care provider. The choices range from clinics, doctor’s offices, ERs, to outpatient services at a hospital. The survey provides data on demographic characteristics such as age, gender, family size, employment, and type of insurance that were used as explanatory variables. Since the data was collected quarterly, a panel data approach to test for fixed and random effects is utilized.

Results: Ultimately, this research shows that the loss of employment and lack of health insurance places greater demand on ERs, outpatient services, and clinics as a source of primary care. However, education and improvement in health status serve to lessen demand for these services in favor of either a doctor’s office or utilizing more than one choice for primary health care.

Conclusions: During recessions, households used clinics and outpatient services less frequently. Increased levels of employment and employer-provided health care would cause a significant drop in the demand for ERs and clinics. Additionally, overall improvement in health status including an emphasis on fitness or health education would also reduce the likelihood of households utilizing ERs or clinics for primary care.

82. Comparison of selected health care utilization measures for young adults aged 19-25 years before and after implementation of the Affordable Care Act

Whitney K. Kirzinger, MPH, Robin A. Cohen, Ph.D.

Background: An early provision of the 2010 Affordable Care Act (ACA) extended dependent health care coverage up to age 26. National Health Interview Survey (NHIS) data has shown that more adults are now covered by private health insurance, following the implementation of ACA, but data has not fully been analyzed investigating changes in health care utilization.

Objective: This analysis will examine changes in health care utilization of adults aged 19-25 by a variety of sociodemographic characteristics prior to and following the implementation of the ACA.

Methods: This study uses data from the Family Core and Sample Adult components of the 2009-2011 NHIS. The 2009-2010 data were combined in the analysis to improve the precision and reliability of the estimates. Measures examined include emergency room utilization, utilization of general health care providers, and unmet medical needs due to cost. Sociodemographic characteristics including sex, race, ethnicity and poverty status were incorporated into the analysis.

Results: Prior to the implementation of the ACA, 21% of males and 29% of females aged 19-25 had an emergency room visit in the last 12 months, 43% of males and 58% of females saw a general doctor in the last 12 months, and 60% of males and 76% of females had a usual place for health care. 15% of young adults aged 19-25 delayed medical care due to cost and 12% did not get medical care because they couldn’t afford it. Preliminary 2011 results show changes in multiple utilization measures following the ACA extension of dependent coverage. These estimates will be compared to 2011 data when final data is released.

Conclusions: Along with health insurance coverage, access to and utilization of health care is changing for young adults aged 19-25. The NHIS can play an important role in monitoring these changes in the United States.

83. Living Longer? Living Healthier? Estimates of Healthy Life Expectancy in Colorado

Kirk A. Bol, MSPH
Statistician, Colorado Department of Public Health and Environment

Background: Out of Colorado’s Public Health Improvement Plan of 2009 came the recommendation to ‘Ensure a comprehensive set of public health indictors’. Included in these indicators are measures of health status, namely quality of life and life expectancy. One metric used to combine these two concepts is healthy life expectancy.

Objective: It is the intent of this work to share estimates of life- and healthy life expectancy for Colorado and selected sub-populations.

Methods: Following previously published guidelines for calculating life- and healthy life expectancy, estimates were computed (with 95% confidence intervals) for all of Colorado, as well as for various demographic groups, including sex, racial/ethnic, and sub-state /regional populations. Data utilized included Colorado vital statistics and estimates of general population health derived from the Behavioral Risk Factor Surveillance System and the Colorado Child Health Survey.

Results: From 1990 to 2010, life expectancy at birth among Colorado residents rose from 77.2 years to 80.4. Despite this increase, disparities remained between male and female in 2010 (78.3 and 82.4, respectively); as well as across white/non-Hispanic, white/Hispanic, and black/African American populations (80.5, 80.3, and 76.9). Similar relationships were seen when examining healthy life expectancy at birth, which for the combined years 2008 to 2010 in Colorado was 71.0 years: Males had a lower healthy life expectancy than females (69.4 and 72.7), while healthy life expectancy among white/non-Hispanic residents was higher than that for white/Hispanic and black/African American residents (73.0, 61.1, and 64.0).

Conclusions: As life expectancy has increased, so must the implications of this transition be examined. Of particular concern are the growing proportion of the population in elderly age, the desire to maximize the proportion of one’s life spent in good health, and persistent disparities. These considerations are integral in the prioritization and planning of Colorado’s present and future public health efforts.

84. Developing Vital Records Standards for Electronic Health Record Systems (EHR-S)

Donna Hoyert, Joyce Martin, Hetty Khan, Michelle Williamson

Background: Many data items required by birth and death certificates and fetal death reports are captured in hospital medical records. The vital statistics and standards communities have collaborated to build upon more than a century’s effort in standardization to encourage Electronic Health Records systems (EHR-S) to capture medical and health information in a format that is consistent with the vital statistics requirements. The goal is to reduce duplicative entries by hospital personnel, and improve the timeliness and quality vital statistics data.

Objective: Can vital records requirements for the collection and transmission of data for birth and death certificates, and fetal death reports be represented in national standards for electronic health record systems?

Methods: CDC/NCHS has collaborated with the National Association of Public Health Statistics and Information Systems (NAPHSIS) and other vital records stakeholders to develop vital records standards that are supported by the standards development organization (SDO), Health Level Seven International (HL7) and the standards organization, Integrating the Healthcare Enterprise (IHE).

Results: A description of these standards will be provided:

• Health Level Seven International (HL7)
  • V2.5.1 Vital Records Birth and Fetal Death Reporting Implementation Guide
  • V2.5.1 Vital Records Death Reporting Implementation Guide
  • Vital Records Domain Analysis Model
  • Vital Records Functional Profile
• Integrating the Healthcare Enterprise (IHE)
  • IHE Birth and Fetal Death Reporting Technical Supplement.

Conclusion: The representation of vital records requirements in EHR-S standards has been successful. Based on fiscal year 2012 funding, CDC/NCHS is planning to pilot test in several states interoperability between EHR and Vital Record systems utilizing these HL7 and IHE standards.

85. Which is the fragile sex? Understanding sex differences in mortality rates with a life history framework

Daniel J. Kruger

Background: Sex differences in mortality rates stem from interconnected genetic, physiological, behavioral, and social causes that are best understood in an evolutionary life history framework. Males in many species are selected for higher levels of risk-taking than females, which enhance reproductive success at the expense of health and longevity. Human males have higher mortality rates for most causes of death across the life span, especially in young adulthood when mating competition is most intense. The degree of male competition will be affected by the availability and distribution of resources in the environment. In areas where there is high potential or actual variability and skew in male resource holding or status, comparatively more risky male behavioral strategies may lead to more excess male mortality.

Objective: Examine the relationship between economic and reproductive inequality and excess male mortality (above the rate for females) across nations.

Methods: A stepwise linear regression examined the relationship between the degree of inequality in economic and reproductive outcomes and sex differences in mortality rates across 70 nations. We obtained mortality data from the World Health Organization’s Mortality Database, the Gini coefficient indicating economic inequality from the Central Intelligence Agency World Factbook, and the degree of polygyny from Kanazawa and Still (1999).

Results: Areas with greater economic inequality and greater male reproductive inequality, consistent with a greater intensity of male competition, had greater sex differences in mortality. The degree of economic inequality and degree of polygyny independently predicted sex differences in mortality rates, accounting for 53% of the variance.

Conclusion: Sex differences shaped by natural and sexual selection interact with the current environment to yield predictable mortality patterns across the lifespan. The degree of excess male mortality is influenced by social conditions corresponding with the intensity of male competition.

86. A Picture of LGBT Health: Using Nationally Representative Data to Develop Objectives for Healthy People 2020

J. Michael Ryan, Anjani Chandra, & Emily DeCoster

Background: There is a limited amount of data available at the national level on the health status of lesbian, gay, bisexual and transgender (LGBT) populations. Methodological and administrative challenges contribute to the limited amount of data on LGBT populations, including a lack of standardized survey questions for the collection of reliable information on sexual identity among diverse respondent groups and relatively small samples of LGBT respondents in those surveys which currently collect sexual identity data.

Objective: The objective will be to explore how to use nationally representative data to develop measurable objectives for the Healthy People 2020 LGBT Health topic area.

Methods: The release of 2006-2010 data from the National Survey on Family Growth (NSFG) provides an opportunity to review initial findings for lesbian gay and bisexual persons related to a number of health issues that have been identified in non-nationally representative studies. In addition, the QDRL has been doing research using cognitive interviews to test the development of a new sexual identity question for the NHIS.

Results: Results from initial analysis of the NSFG data will be presented along with the new sexual identity question tested by the Questionnaire Design Research Laboratory (QDRL).

Conclusions or Significance: This poster will include an overview of the preliminary findings from the NSFG data, along with a description of the ongoing work with the NHIS related to the collection of nationally representative data on LGBT populations.

87. Changes in age-specific death rates from six leading causes of death in fifty US states and the District of Columbia, 1970-2008

Rong Wei, Melonie Heron and Meena Khare
National Center for Health Statistics, Hyattsville, MD 20782

Background: Mortality rates from the leading causes of death are important tools for monitoring the health status of populations. The National Center for Health Statistics has well-documented US mortality data collected through the National Vital Statistics System since 1968.

Objective: Data from 1970 to 2008 are used to demonstrate variation in death rates from six leading causes of death: heart disease, cancer, stroke, accidents, chronic obstructive pulmonary disease (COPD) and diabetes over 39 years in 50 states and the District of Columbia (DC).

Methods: Death rates are computed by single years of age, so differences over time can be explored across age-, cause of death- and state-specific groups. A locally weighted regression is applied to smooth the rates in two dimensions: age and year of death. Death rates for 50 states and DC are mapped to explore geographic distributions. Correlations among causes of death and changes in death patterns across states are studied statistically and graphically.

Results: This study shows that the age distribution of the leading causes varies over time, and the time trend of death rates varies across states. Interestingly, while most states follow similar patterns over time, some states have unique trends over time. Relative to other states, Alaska has lower death rates from heart disease and diabetes but higher rates from accidents; Louisiana persistently has the highest death rates from diabetes; Nevada and Wyoming have higher death rates from accidents and COPD, but lower rates from diabetes; DC has higher death rates from cancer, heart disease and diabetes, but lower rates from COPD; Hawaii has lower deaths from all six leading causes.

Conclusions: The study focuses on cause-specific mortality over time, rather than on all-cause mortality. The results demonstrate variations in health conditions and deaths in fifty US states and DC over the last four decades.

88. Implications of the ICD 10 Transition for Public Health Surveillance: Challenges, Opportunities and Lessons Learned

Donna Pickett, Prachi Mehta, Lea Trujillo, and Marella Bradway

Background: The U.S. health care industry will undergo a major transition when it moves from ICD-9 CM to ICD-10 CM/PCS proposed on October 1, 2014. This transition will have a significant impact on public health surveillance systems and activities which are largely secondary users of ICD-9 coded data. For example, public health officials rely on the receipt of ICD-9-CM coded data to conduct regular disease surveillance activities, from cancer case finding lists, to injury surveillance, to real time public health situation awareness. It is imperative that public health agencies prepare their “systems, processes and people” for a seamless transition to ICD-10 coded data to ensure continuity of surveillance services and functions.
Objective: The audience will be able to describe the impact that the transition from ICD-9 CM to ICD-10 CM and ICD-10 PCS will have on public health epidemiology and surveillance, and the audience will understand key differences between ICD-9 CM and ICD-10 code sets.

Methods: In 2011, Centers for Disease Control & Prevention (CDC) Information Resources Governance Council formed the ICD-10 Transition Workgroup to facilitate agency wide ICD-10 transition efforts. The Workgroup has engaged in several activities including conducting a needs assessment survey to ascertain program readiness and identify challenges, needs and gaps; develop training webinars; and compile/organize resources to assist programs with their technical needs.

Results: CDC’s activities in preparing for this transition will be presented, including findings from the needs assessment survey. Resources developed relevant to secondary data users and lessons learned from CDC’s transition efforts will also be shared.

Conclusions: ICD-10 incorporates greater specificity and clinical detail but also increased complexity, which requires additional effort to make the data meaningful. Public health entities will be able to leverage the work done within CDC as they transition their own public health activities to ICD-10 code sets.

89. Has Text Messaging Increased Participant Compliance on NHANES?

Tatiana Nwankwo, MS.

Background: The National Health and Nutrition Examination Survey (NHANES) has experienced a decline in fasting compliance in the Oral Glucose Tolerance Test (OGTT) and in reporting Sexually Transmitted Disease (STD) results since 1999. Hence, there is a need for improved communication with NHANES participants and text messaging technology holds promise for this application.

Objective: To reduce fasting noncompliance by sending out a text message and to improve the communication of STD test results by text messaging a reminder to call-in for results.

Methods: Starting in May 2011 two text messages were sent to participants aged 12 and older who were assigned to the morning session the day before the examination for the OGTT component. For STD reporting, participants who agreed to STD testing would receive a text message if they failed to call for test results within 38 days after their exam.

Results: There were 1611 sample participants who were eligible to fast for the OGTT. Three hundred twenty six (48%) text messages were sent to sample participants. Of those who were sent a text message, there was a 97% compliance rate compared to a 93% compliance rate for those who were not sent a text message. In the STD reporting, there were 1163 participants who gave permission to text. Nine hundred ten participants failed to call within 38 days and of those 855 were sent a text message. To date, 395 (46%) participants called in.

Conclusion: Preliminary findings suggest that text messaging provided a simple and efficient “call-to-action” to reach sample participants with no complaints reported. The short-term data indicates that the OGTT component had a decrease in non-compliance while the STD reporting stayed steady. Consequently, NHANES has opted to expand the use of text messaging to the Physical Activity Monitor (PAM) component.

31. Prevalence of Multiple Chronic Conditions among U.S. Adults: Data from the National Health Interview Survey, 2010

Jeannine S. Schiller, MPH; Brian W. Ward, PhD

Background: Ameliorating chronic conditions has long been a priority in the United States; however, the increase of individuals with two or more conditions, or multiple chronic conditions (MCC), has added a layer of complexity with which to contend.

Objective: The objective of this study was to present the prevalence of MCC by selected demographic characteristics, and the prevalence of the most common MCC dyads and triads.

Methods: Data from the 2010 National Health Interview Survey (NHIS) were used to study the U.S. adult civilian noninstitutionalized population 18 years and over (n=27,157). To determine number of chronic conditions, respondents were asked if they had ever been told by a doctor or health professional that they had: hypertension, coronary heart disease, stroke, diabetes, cancer, arthritis, hepatitis, chronic obstructive pulmonary disease, asthma (with a follow-up question regarding current asthma), or weak or failing kidneys during the past 12 months. Descriptive estimates were generated and significance tests were used to test for statistically significant differences.

Results: Fourteen percent of adults had two of the previously-listed chronic conditions, 7% had three, and 5% had four or more. Significant findings (p<.05) include that prevalence of MCC increased with age, was higher among women than men, and that non-Hispanic white and non-Hispanic black adults had a higher prevalence of MCC than Hispanic adults. The most common dyad of MCC for men and women was arthritis and hypertension. The most common triad was arthritis, hypertension, and diabetes.

Conclusions: The findings of this study contribute valuable information to the field of MCC research. The NHIS can be used to identify population subgroups most likely to have MCC, and this collection of information may be useful in helping clinicians develop prevention strategies in order to reduce the prevalence of these conditions, and consequently reduce health care costs among this population.

Vitamin D Intakes and Status among US Children Aged 1 to 18 Years: Do Obese and Racially/Ethnically Diverse Youth Need More Vitamin D?

Lauren E Au MS, RD; Gail T Rogers MA; Susan S Harris DSc; Johanna T Dwyer DSc, RD; Paul F Jacques DSc; and Jennifer M Sacheck PhD

Background: Obesity and race/ethnicity are linked to poor 25-hydroxyvitamin D (25OHD) levels, however little information exists concerning whether either modifies the relationship between vitamin D intake and 25OHD.

Objective: The objective of this study was to determine whether adiposity and race/ethnicity modified the association between vitamin D intake and serum 25OHD in a nationally representative sample of US children/adolescents aged 1-18 years.

Methods: 2,608 children/adolescents in the 2005-2006 NHANES were examined. Vitamin D intake was estimated from foods and dietary supplements. Race/ethnicity was categorized as Non-Hispanic White, Non-Hispanic Black, Mexican American, or Other. Weight status was dichotomized to healthy weight (BMI > 5th and < 85th percentile-for-age) and overweight/obese (BMI > 85th percentile-for-age). Multivariate logistic regression was used to assess whether the odds of inadequate 25OHD (<20 ng/mL) for a given vitamin D intake is modified by adiposity or race/ethnicity.

Results: The likelihood of vitamin D inadequacy was significantly greater for children who consumed less than the Estimated Average Requirement (EAR) for vitamin D intake (400 IU/d) than those who met or exceeded it (OR: 2.7; 95% CI: 2.0-3.6). This association was not modified by adiposity (P=0.67). However, it was modified by race/ethnicity (P=0.009). Specifically, stratified analyses showed that NH Black children who met the EAR were 5.8 times more likely to have adequate 25OHD than NH Black children who did not meet it (95% CI: 3.3-10.2), but this association was much smaller among NH White children (OR: 2.2; 95% CI: 1.1-4.4), and Mexican American children (OR: 1.8; 95% CI: 1.2-2.8).

Conclusions: Adiposity did not modify the association of vitamin D intake with 25OHD in the study population. However, these findings show that NH Black children are more likely to be inadequate in 25OHD than other children when consuming less than 400 IU/d, the current EAR.

Association of special health care needs and parental mental health and parenting stress on sedentary lifestyles in children

Janet Gingold, MD

Background: Children with special health care needs (CSHCN) may be at increased risk of obesity and other co-morbidities associated with sedentary lifestyles. Parents’ ability to promote active lifestyles in children may be affected by parental mental health (MH) and ability to cope with demands of parenting.

Objective: We aimed to compare the prevalence of sedentary lifestyle among CSHNC with and without emotional, behavioral and developmental conditions (EBD) and non-CSHCN, and to determine whether parental MH/stress problems were associated with increased prevalence of sedentary lifestyle.

Methods: The 2007 National Survey of Children’s Health provided data from parent interviews about 64,076 children aged 6-17. Sedentary lifestyle was defined as exercising less than six days/week and having more than 120 minutes/weekday of screen-based leisure activity. Children were categorized as CSHCN without EBD, CSHCN with EBD or non-CSHCN based on the CSHCN Screener, need for counseling, and the presence of a listed EBD diagnosis. Parents were categorized as having MH/stress problems based on self-rated mental health and 4 items about parenting stress. Multiple logistic regression was used to calculate odds ratios adjusted for age, gender, race/ethnicity and respondent’s education using SAS survey procedures to account for the complex survey design.

Results: Prevalence of sedentary lifestyle was 37.8% for CSHCN with EBD, 34.9% for CSHCN without EBD and 32.4% for non-CSHCN. Children of parents with MH/stress problems had increased likelihood of sedentary lifestyle (39.9%; 95%CI, 37.6-42.2) compared with those without MH/stress problems (31.5%; 95%CI, 30.5-32.5). After adjusting for demographic factors and parental MH/stress problems, the likelihood of sedentary lifestyle was higher among CSHCN without EBD (AOR 1.167; 95%CI, 1.011-1.347) and among CSHCN with EBD (AOR 1.198; 95%CI, 1.047-1.372) than among non-CSHCN.

Conclusion: Special health care needs and parental MH/stress problems are associated with small increases in the likelihood of sedentary lifestyles in children.

Urban and Rural Disparities in Tobacco Use

Mingyang Shan; Elizabeth A. Lancet, MPH; Zach Jump, MA
American Lung Association, Epidemiology and Statistics Unit

Background: Differences in geographic location may play a role in the incidence and prevalence of tobacco use.

Objective: This analysis aims to examine predictors of tobacco use among rural vs. urban areas in the United States.

Methods: We analyzed data from the Health and Human Services’ National Survey on Drug Use and Health. Binary and multiple logistical regression were used to estimate odd ratios (OR) and 95% confidence intervals (CI) in examining the relationship between tobacco use (cigarette smoking and smokeless tobacco) and geographic region (urban and rural) before and after controlling for selected demographic characteristics.

Results: Although crude estimates showed differences in cigarette usage between rural and urban areas, the disparity was not significant after controlling for selected demographic variables; smokeless tobacco use was 2.03 times more likely among rural individuals (95% OR CI: 1.68-2.46). Smoking cigarettes was 1.27 times more common amongst men than women (95% OR CI: 1.18-1.38), while smokeless tobacco use was 24.25 times more prevalent in men (95% OR CI: 17.52-33.57). Significant rural/urban vs. age (P=0.013) and rural/urban vs. race (P=0.0006) interactions were found for smokeless tobacco use. Family income and education were inversely related to cigarette smoking in both rural and urban areas as expected; however, non-linear trends were seen among smokeless tobacco use. Smoking was more frequent among women who were not pregnant compared to women who were pregnant in urban areas (OR=3.49, 95% CI: 2.56-4.71), but this difference was not significant in rural areas (OR=1.80, 95% CI: 0.71-4.48). Rural women who were pregnant smoked at approximately the same rate as urban women who were not pregnant (OR=1.70, 95% CI: 0.59-4.85).

Conclusion: Smokeless tobacco use was higher in rural than urban areas across multiple demographic characteristics. The non-linear trend between socioeconomic status and smokeless tobacco use was unexpected and further evaluation is warranted. Lastly, smoking interventions to assist rural women in quitting, especially during pregnancy, are desperately needed.

Generalizing Observational Study Results: Applying Propensity Score Methods to Complex Surveys Association of special health care needs and parental mental health and parenting stress on sedentary lifestyles in children

Eva H. DuGoff¹, MPP
Elizabeth A. Stuart^2,3, PhD
Megan Schuler³, MS
¹ Department of Health Policy & Management, Johns Hopkins Bloomberg School of Public Health
² Department of Biostatistics, Johns Hopkins Bloomberg School of Public Health
³ Department of Mental Health, Johns Hopkins Bloomberg School of Public Health

Background: Randomized controlled trials (RCT) are the gold standard for establishing treatment efficacy but do not always provide evidence of real-world effectiveness. Nationally representative survey data offer important data for effectiveness studies but a challenge is adequately controlling for confounding. Propensity score methods are often used to address confounding, yet few guidelines exist on how to implement these methods with complex survey data to maintain representativeness of the data. We performed a limited systematic review and found wide variability in propensity score methodology used with complex survey data, suggesting misunderstanding of these methods.

Objectives: This paper provides a tutorial for using propensity score methods with complex survey data, highlighting trade-offs of various methods and discussing correct interpretation of estimated effects.

Methods: We first describe different propensity score methods (matching, subclassification and weighting) and their trade-offs in the context of complex surveys. We then illustrate these methods in a cross-sectional analysis of data from the 2008 Household Component of the Medical Expenditure Panel Survey that compares health care spending among adults who report a generalist versus a specialist as their usual source of care. We implemented several different propensity score methods and varied whether the outcome model was run with or without survey weights.

Results: The choice of propensity score methods determines the interpretation of the results and whether the results are generalizable to the original survey population. We present both Average Treatment Effect on the Treated (ATT) and Average Treatment Effect (ATE) estimates and clarify whether the resulting estimates are for the survey sample or the full population. Incorrectly applying survey weights to matched samples consistently resulted in larger point estimates and smaller, statistically significant p-values.

Conclusions: This paper discusses methodological choices researchers should consider when using propensity score matching methods with complex survey data and provides a checklist to guide analysis.