Trends in Health Insurance and Access to Care from 1997-2008

Print
Related Pages
On This Page

These tables present estimates from the National Health Interview Survey on trends in health insurance and access to care from 1997-2008 for persons 18-64 years of age. Included are tables showing the numbers and percentages of people with different types of health insurance coverage and with no coverage; annual state-specific estimates for the 20 largest states by insurance type and age group; and quarterly estimates of unmet need for medical care and delay of medical care due to cost by age group, employment status, and insurance coverage.

Tables 1-4 present trends in the percentage of persons 18-64 years of age with different types of health insurance coverage and with no coverage by quarter and employment status. Estimates were derived from 12 years of the National Health Interview Survey (NHIS) over the period 1997 to June 2008. Annual estimates are presented as well. Note that the standard errors associated with quarterly estimates for the unemployed are relatively large, so caution should be used in making comparisons over time as apparent differences my be statistically insignificant.

Table 5 provides annual state-specific estimates for the 20 largest states in terms of population as well as national estimates from 1997 – 2008 by insurance type and age group for persons under 65 years of age. The original table produced included data through 2007. This table was updated recently to include full year 2008 estimates. Although the relative standard errors are generally less than 30%, they are still quite large for children under 18 in some states, especially for the measures of public coverage and uninsured. Large swings in rates – especially for children with public coverage – are often statistically insignificant.

Tables 6-14 provide quarterly estimates of unmet need for and delay of medical care due to cost for persons under 65 years of age by age group, employment status, and insurance coverage from 1997-June 2008. Annual estimates are presented as well. Note that the standard errors associated with quarterly estimates for the unemployed are relatively large, so caution should be used in making comparisons over time, as apparent differences may be statistically insignificant.

Technical information on the data source and definitions of NHIS coverage categories are provided below.

Data Source and Methods

The National Health Interview Survey (NHIS) is the source of data for tables 1-14. NHIS is a continuous multistage probability sample survey of the civilian noninstitutionalized population of the United States. It is a multipurpose health survey conducted by the Centers for Disease Control and Prevention’s (CDC) National Center for Health Statistics (NCHS). NHIS interviewers are from the U.S. Census Bureau. Information on basic health topics is collected for all household members, by proxy from one family member if necessary (all members of the household 17 years of age and over who are home at the time of the interview are invited to participate and respond for themselves). More information is available on the National Health Interview Survey homepage.

The sample size of the NHIS varied over the years 1997-2008. The sample size of persons under age 65 ranged from 91,275 to 67,065.

Household response rates for the period ranged between 86-90%. In the tables, unknown values (responses coded as “refused”, “not ascertained”, or “don’t know”) were not counted in the denominators when calculating estimates. The item nonresponse rate for the health insurance items and employment items used in the tables are about 1%. Item nonresponse rates for the unmet medical needs questions are around 0.2%.

Estimation procedures

Estimates were calculated using the NHIS survey weights, which are calibrated to census totals of the U.S. civilian noninstitutionalized population by sex, age, and race or ethnicity. The 1997 to 2002 weights were derived from the 1990 census-based population estimates. Starting with 2003, weights were derived from 2000 census-based population estimates.

Point estimates and estimates of their variances were calculated using SUDAAN software to account for the complex sample design of NHIS. The Taylor series linearization method was chosen for variance estimation.

State-specific estimates are presented for 20 states. The Taylor series linearization method was chosen for estimation of standard errors for the 10 states with the largest sample sizes (California, Florida, Illinois, Michigan, New Jersey, New York, North Carolina, Ohio, Pennsylvania, and Texas). Due to small sample sizes and limitations in the NHIS design, estimated standard errors for other states could be unstable or negatively biased. For these states, an estimated design effect was used to calculate standard errors. For this report the design effect is the ratio of the true standard error of an estimated percentage (taking the complex sample design into account) to the standard error of the estimated percentage from a simple random sample (SRS) based on the same number of persons.

The average design effect for each health insurance measure and domain was estimated by averaging the design effect derived from SUDAAN-based estimates of standard errors for the ten largest states. For this report, standard errors for 10 states were estimated by multiplying the SRS standard error by the average design effect for each health insurance measure and domain. The average design effects ranged from 1.33 to 2.40.

Definitions of NHIS coverage categories

NHIS questions about health insurance are point in time, and the estimates in tables 1-14 pertain to coverage status at the time of interview.

Private coverage includes comprehensive health care coverage obtained through an employer, purchased directly, or obtained through any other means. It excludes plans that pay for only one type of service such as accidents or dental care. Private coverage also includes persons covered thru TRICARE.

Public coverage includes persons who reported having Medicaid, Children’s Health Insurance Program (CHIP), state-sponsored or other government-sponsored health plan, Medicare (disability), Veterans Administration (VA) and CHAMP-VA. A small number of persons were covered by both public and private plans and were included in both categories. For 2008, this category also includes persons covered by TRICARE, therefore caution should be used in comparing these estimates with prior years.

The uninsured category includes persons who have no private health insurance, Medicaid, military coverage, Medicare, CHIP, or a state-sponsored health plan or other government program. A person was also defined as uninsured if he or she had only Indian Health Service coverage or only had a private plan that paid for one type of service such as accidents or dental care.

Tables