Behavioral and Clinical Characteristics of Persons Living with Diagnosed HIV Infection—Medical Monitoring Project, United States, 2019 Cycle: Technical Notes and Appendix

Technical Notes

Population of Inference

For the 2019 Medical Monitoring Project (MMP) data collection cycle (data collected June 1, 2019–May 31, 2020), the population of inference was adults with diagnosed HIV (aged ≥18 years) living in the United States.

A total of 23 areas were funded to conduct data collection for the 2019 cycle: California (including the separately funded jurisdictions of Los Angeles County and San Francisco), Delaware, Florida, Georgia, Illinois (including the separately funded jurisdiction of Chicago), Indiana, Michigan, Mississippi, New Jersey, New York (including the separately funded jurisdiction of New York City), North Carolina, Oregon, Pennsylvania (including the separately funded jurisdiction of Philadelphia), Puerto Rico, Texas (including the separately funded jurisdiction of Houston), Virginia, and Washington.

Data Collection

Persons with diagnosed HIV were sampled for MMP using data from the National HIV Surveillance System (NHSS). Sampled persons were recruited to participate by mail, by telephone, or in person. To be eligible for MMP, the person had to be, as of December 31, 2018: living with diagnosed HIV infection, aged ≥18 years, and residing in an MMP project area. The participant eligibility criteria were the same in all participating project areas.

A trained interviewer conducted either a telephone interview or an in-person interview. English and Spanish versions of the questionnaire were used in the 2019 cycle (June 2019–May 2020). Persons who agreed to participate were interviewed over the telephone or in a private location (e.g., at home or in a clinic). The interview (approximately 45 minutes) included questions about demographics, health care use, met and unmet needs for ancillary services, sexual behavior, depression and anxiety, gynecologic and reproductive history (females only), drug and alcohol use, and use of prevention services. Participants were given a token of appreciation of no more than $50 in cash or the equivalent for participation; tokens differed by project area according to local considerations.

After the interview, MMP staff abstracted clinical data from the medical records of participants at the health care facility identified by the participant as their most frequent source of HIV care. Abstracted information included diagnoses of AIDS-defining conditions, prescription of antiretroviral therapy (ART) medications, laboratory results, and health care use in the 24 months before the interview.

Appendix

Methods

The Medical Monitoring Project (MMP) uses a stratified, 2-stage sampling design. States were sampled first, with probability proportional to size (PPS). All 50 states, the District of Columbia, and Puerto Rico (defined as primary sampling units [PSUs]) were eligible for selection.

From these 52 PSUs, 20 were selected by using PPS sampling based on AIDS prevalence at the end of 2002. According to the PPS sampling method, states with a higher AIDS prevalence had a higher probability of selection, and those with a lower AIDS prevalence had a lower probability of selection [1]. Six municipal jurisdictions receive separate funding for HIV surveillance (Chicago, Illinois; Houston, Texas; Los Angeles County, California; New York City, New York; Philadelphia, Pennsylvania; and San Francisco, California); these areas were included with the state for first-stage sampling and constituted a city-state unit. If a state included a city with independent HIV surveillance authority (e.g., Texas, which includes Houston), selection of the state included selection of the city (i.e., city-state units were selected together).

In 2004, 19 states (including the 6 separately funded areas within those states) and Puerto Rico were selected from the 52 PSUs, resulting in 26 MMP project areas. Because of funding constraints for the 2009 data collection cycle, 3 project areas (Maryland, Massachusetts, and South Carolina) were randomly selected to discontinue participation in MMP, and the total number of MMP areas was reduced to 23.

An analysis carried out in 2014 found that the original measure of size with which states were originally sampled (i.e., AIDS prevalence in 2002) was still a reasonable proxy for the distribution of HIV prevalence in 2010 (the most recent year for which prevalence estimates were available at the time). The selected sample of states was still sufficiently representative of the population of persons with diagnosed HIV; consequently, selecting a new sample for the 2015 and subsequent data collection cycles was unwarranted. In addition, the change in the sampling frame and the availability of national totals from the National HIV Surveillance System (NHSS) presented new options for calibrating weights, further lessening the need for any adjustments to the sample of states.

At the second stage, persons with a reported diagnosis in NHSS were sampled after the selection of the states. The sampling frame was the national case surveillance data set containing records submitted to the Centers for Disease Control and Prevention (CDC) as of December 31, 2018. Using NHSS data, the initial national frame dataset was created for persons who were alive, had diagnosed HIV infection, 18 years or older, and living in the United States, the District of Columbia, or Puerto Rico on the sampling date (December 31, 2018). Each case was assigned to a surveillance jurisdiction based on the most recently reported residence in NHSS. These addresses primarily came from case report forms and HIV-related laboratory reports. From this initial national frame, CDC staff drew simple random samples for the 23 project areas; project area staff then linked their samples to local case surveillance systems and extracted contact information for use in locating sampled persons, whom they then attempted to recruit.

Eligibility and Response Classifications

Persons were eligible for participation if, as of the sampling date, they had received a diagnosis of HIV, were aged ≥18 years, alive, and a resident of an MMP project area. Sampled persons were presumed to be eligible based on their information in NHSS unless data from another source contradicted this status. Persons were classified into 4 categories: (1) eligible respondents, (2) contacted nonrespondents, (3) nonrespondents who were not contacted, and (4) ineligible persons. These categories were used in calculating final response rates and contact rates following standard formulas [2].

Weighting

Overview

For the 2019 MMP cycle, sets of weights were produced nationally, for the city-state combinations, and for each project area. This report presents national weighted data and, thus, represents all adults with diagnosed HIV infection living in the United States. Nationally, data were weighted based on known probabilities of selection at the state or jurisdiction level and person level and then adjusted for multiplicity and nonresponse. After adjusting for nonresponse, the weights were poststratified to population totals from the NHSS frame. Extreme weights were trimmed, and the weights were adjusted to the same population totals.

For the weighting process, an updated sampling frame was obtained from NHSS data approximately a year and a half after sampling, during which time additional information reported to NHSS may have become available for sampled persons and additional diagnoses may have been reported. This updated sampling frame added records that would have been eligible if their information had been reported to NHSS on the date the initial sample was drawn; primarily, these were diagnoses that occurred during the year prior to the MMP sampling date (for the 2019 cycle, December 31, 2018). Additionally, some persons were found to have had multiple records at the time of sampling that were later identified as duplicate records. In some cases, updated information indicated that a person originally judged eligible and included on the original frame was ineligible. The updated sampling frame data also provided descriptive information for all sampled persons regardless of response and were the source of data used for nonresponse analysis and weighting.

Adjustments for Unequal Selection Probabilities

The first step in the computation of weights was the calculation of base weights that reflect the sampling design probabilities. The base weight for each sampled person incorporates both the probability of selecting a project area, and the probability of selecting a person within a project area. A person who was sampled from one jurisdiction but lived in another area at the time of sampling retained the original base weight. Prior to weighting, such cross-jurisdictional records were grouped with their project area of residence at the time of sampling. This moving of records did not affect the national weights, but did affect the project area weight totals, increasing some while decreasing others.

Adjustments for Multiplicity

A multiplicity factor was applied to the person weight for persons with records found to be present more than once after the original frame was compared to the updated sampling frame. This factor, which accounts for some persons’ multiple opportunities for being sampled, was capped at 2.0 and was applicable for only 48 persons.

Adjustments for Nonresponse

A nonresponse adjustment factor was applied to the multiplicity-adjusted base weight based on an analysis of nonresponse. In 2019, updated sampling frame data provided descriptive information about all sampled persons, which was used to assess how these characteristics were associated with nonresponse. The potential predictors of nonresponse were: race/ethnicity, men who have sex with men (MSM) HIV transmission category, HIV/AIDS disease stage, disease progression measured by most recent viral load test reported to NHSS, time since HIV diagnosis, age of most recent contact information, the person’s frequency of receipt of HIV care (as indicated by NHSS records), movement to a different MMP jurisdiction since the time of sampling, non-U.S. birthplace, sex at birth, and age at sampling date. The nonresponse analysis followed a 2-step process. First, a bivariate analysis was conducted to determine which characteristics were potential predictors of nonresponse; then, a multivariate analysis using the significant characteristics from the bivariate analysis was conducted to identify independent predictors of nonresponse. Three significant predictors from this multivariate analysis were used to create weighting classes for the national data. In 2019, the significant predictors of nonresponse were: the person’s frequency of receipt of HIV care (as indicated by NHSS records), sex at birth, and disease progression measured by the most recent viral load test reported to NHSS. Within weighting classes, the adjustment factor for nonresponse was the ratio of the sum of the multiplicity-adjusted base weights for eligible sampled cases to the sum of these weights for eligible respondents. The multiplicity adjusted weight within each nonresponse weighting class was then multiplied by the nonresponse adjustment factor to produce the nonresponse adjusted weights.

Poststratification and Trimming

Poststratification methods ensure that weighted totals sum to known population totals and, therefore, minimize the potential for biases due to nonresponse and noncoverage. However, poststratification can also add additional variance to the weights. Thus, trimming procedures are used to control weight variability and reduce its impact on survey variances. MMP used an iterative approach that combines poststratification and trimming so that trimmed weights retain their variance-reducing features after poststratification and ensures that poststratified weights add up to known population totals.

The nonresponse adjusted weights were first poststratified to population totals from the updated sampling frame. The poststratification cells were defined by crossing sex at birth, race/ethnicity, and age group. Nationally, there were 32 poststratification cells. Poststratification adjustments were performed within each poststratification cell so that the weighted sum was preserved in each cell. To reduce additional variance added to poststratified weights, cells were collapsed and the need for weight trimming was evaluated. Poststratified cells were collapsed when cells had 2 or fewer respondents or had an extreme adjustment factor (≥ 1.75). The need for trimming was then assessed. If the design effect due to weighting (measured as 1 + CV², where CV is the coefficient of variation of the weights) had exceeded 1.75, we would have capped the weights at the median weight plus 4 times the interquartile range of the weights; where trimming occurred, the weights were redistributed and poststratified again to the population totals. However, no trimming was needed for the national weights.

Design Variables

Nationally, design variables indicating strata and cluster membership for each participating person accounting for the sample design were created. Many project areas were sampled with certainty because of higher AIDS prevalence, and each of these was defined as its own stratum. Elsewhere, strata were created by grouping 2 to 3 project areas (PSUs in the stratified PPS design) that had similar selection probabilities. Among the 23 project areas, 14 were sampled with certainty. The 14 certainty project areas each represent a stratum, and each person within the stratum is a cluster. The remaining 9 noncertainty project areas were grouped to create strata, and each noncertainty project area was a cluster within the stratum. Multiple project areas within certainty states were effectively substrata, and each project area remained its own stratum. For local estimates, variance estimation was conditional on the initial sampling of states as PSUs, meaning that this stage of sampling was ignored. Participants were treated as having come from a simple random sample with replacement, although the various adjustment factors induced unequal weights.

Definitions

Sociodemographic Characteristics

• Gender: Categories were male, female, and transgender. Participants were classified as transgender if reported sex at birth and current gender as reported by the participant were not the same or if the participant answered “transgender” to the interview question regarding self-identified
• Health insurance or coverage for care or medications: Participants were asked whether they had health insurance or coverage for care or medications (including antiretroviral [ART] medications) during the 12 months before interview. Responses to these questions were combined and categorized as private health insurance, Medicaid, Medicare, Ryan White HIV/AIDS Program, Tricare/CHAMPUS and Veterans Administration coverage, insurance classified as other public health insurance, and unknown insurance. Participants could select more than 1 response for health insurance or coverage for care or medications.

• Federal poverty guidelines: Participants were asked about their combined monthly or yearly household income (in U.S.$) from all sources during the 12 months before interview. The number of persons meeting the current federal poverty threshold was determined by using the U.S. Department of Health and Human Services poverty guidelines that corresponded to the calendar year for which income was asked. These guidelines are issued yearly for the 48 contiguous states and Washington, D.C., and are an indicator used for determining eligibility for many federal and state programs. The 2018 guidelines [3] were used for participants interviewed in 2019, and the 2019 guidelines [4] were used for persons interviewed in 2020. Because the poverty guidelines are not defined for the territory of Puerto Rico, the guidelines for the contiguous states and Washington, D.C., were used for this jurisdiction. Participants were asked to specify the range of their income, and household income was assumed to be the midpoint of the income range.

Clinical Characteristics

• CDC stage of disease classification for HIV infection: Defined according to CDC’s 2014 revised surveillance case definition for HIV infection [5]. Information from NHSS was used to determine the most advanced HIV disease stage ever reached by participants.

Use of Health Care Services

• Outpatient HIV medical care: Defined as documentation of any of the following: encounter with an HIV care provider, viral load test result, CD4 test result, HIV resistance test or tropism assay, ART prescription, pneumocystis pneumonia (PCP) prophylaxis, or Mycobacterium avium complex (MAC) prophylaxis. All were measured through documentation in the person’s medical record; an encounter with an HIV care provider was also measured based on interview self-report. Persons were considered to be retained in care if they had 2 elements of outpatient HIV care at least 90 days apart in each 12-month period reviewed.
• ART prescription: Defined as a prescription in the medical record, during the 12 months before interview, of any of the following medications: abacavir, amprenavir, atazanavir, bictegravir, cobicistat, darunavir, delavirdine, didanosine, dolutegravir, doravirine, efavirenz, elvitegravir, emtricitabine, enfuvirtide, etravirine, fosamprenavir, ibalizumab, indinavir, lamivudine, lopinavir/ritonavir, maraviroc, nelfinavir, nevirapine, raltegravir, rilpivirine, ritonavir, saquinavir, stavudine, tenofovir alafenamide, tenofovir disoproxil fumarate, tipranavir, or zidovudine. Persons with no medical record abstraction were considered to have no documentation of ART prescription.
• PCP prophylaxis: Defined as documentation in the medical record that prophylaxis for PCP was prescribed among persons with a CD4 count of < 200 cells/µL in the 12 months before interview [6]. Persons prescribed regimens typically given as PCP prophylaxis (trimethoprim-sulfamethoxazole, dapsone with or without pyrimethamine and leucovorin, aerosolized pentamidine, and atovaquone) were not presumptively categorized as having received PCP prophylaxis unless it was specifically stated in the medical record that prescription of these medications was for PCP prophylaxis or no length of time was specified for the course of treatment.
• Influenza vaccination: Participants were asked whether they had received seasonal influenza vaccine during the 12 months before
• Neisseria gonorrhoeae testing: Defined as documentation in the medical record, during the 12 months before interview, of a result from culture, Gram stain, enzyme immunoassay (EIA), nucleic acid amplification test (NAAT), or nucleic acid probe performed on a specimen from any anatomical site for screening or diagnostic purposes.
• Chlamydia trachomatis testing: Defined as documentation in the medical record, during the 12 months before interview, of a result from culture, direct fluorescent antibody (DFA), EIA or enzyme-linked immunoassay (ELISA), NAAT, or nucleic acid probe performed on a specimen from any anatomical site for screening or diagnostic purposes.

• Syphilis testing: Defined as documentation in the medical record, during the 12 months before interview, of a result from nontreponemal serologic tests (rapid plasma reagin [RPR], Venereal Disease Research Laboratory [VDRL]), treponemal serologic tests (Treponema pallidum hemagglutination assay [TPHA],  T. pallidum particle agglutination [TP-PA], microhemagglutination assay for antibodies to T. pallidum [MHA-TP], chemiluminescence immunoassay [CIA], fluorescent treponemal antibody absorption [FTA- ABS] tests), polymerase chain reactions (PCR), or dark-field microscopy performed for screening or diagnostic purposes.

Self-reported ART Medication Use and Adherence

• ART adherence: Participants were asked about their adherence to ART in the 30 days before interview using questions from a 3-item scale developed by Wilson and colleagues [7]. Participants were asked about how many days they missed at least 1 dose of their HIV medicines, how often they took their HIV medicines in the way they were supposed to, and how good a job they did at taking their HIV medicines in the way they were supposed to during the 30 days before interview.

Depression and Substance Use

• Depression: Participants were asked questions from the Patient Health Questionnaire (PHQ-8), an 8-item scale used to measure frequency of depressed mood in the preceding 2 weeks [8]. The PHQ-8 has the following question: “Over the last 2 weeks, how often have you been bothered by any of the following problems?” The respondent is then asked about the following problems: (1) little interest or pleasure in doing things (anhedonia); (2) feeling down, depressed, or hopeless; (3) trouble falling/staying asleep, or sleeping too much; (4) feeling tired or having little energy; (5) poor appetite or overeating; (6) feeling bad about yourself or that you are a failure or have let yourself or your family down; (7) trouble concentrating on things, such as reading the newspaper or watching television; and (8) moving or speaking so slowly that other people could have noticed, or being fidgety or restless or moving around a lot more than usual. Response categories were “not at all,” “several days,” “more than half the days,” and “nearly every day” with points (0–3) assigned to each response category, respectively. The PHQ-8 responses were scored by using 2 methods. Method 1: an algorithm involving criteria from the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV-TR) [9], for diagnosing major depression was used to classify adults with diagnosed HIV as having major depression, other depression, or no depression. To meet the criteria for major depression, a participant must have experienced 5 or more symptoms at least “more than half the days,” and one of the symptoms must be anhedonia or feelings of hopelessness. For other depression, a participant must have experienced 2 to 4 symptoms at least “more than half the days,” and one of the symptoms must be anhedonia or feelings of hopelessness. Method 2: scores for each response category were summed to produce a total score between 0 and 24 points. Current depression of moderate or severe intensity was defined as a total score of ≥10.
• Anxiety: Participants were asked questions from the Generalized Anxiety Disorder Scale (GAD-7), a 7- item scale used to screen for and measure the severity of generalized anxiety disorder [10]. The GAD-7 has the following question: “Over the last 2 weeks, how often have you been bothered by any of the following problems?” The respondent was then asked about the following problems: (1) feeling nervous, anxious, or on edge; (2) not being able to stop or control worrying; (3) worrying too much about different things; (4) trouble relaxing; (5) being so restless that it is hard to sit still; (6) becoming easily annoyed or irritable; and (7) feeling afraid as if something awful might happen. Responses were scored according to criteria from the DSM-IV-TR [9]. Response categories were “not at all,” “several days,” “more than half the days,” and “nearly every day,” with points (0–3) assigned to each response category, respectively. Scores for each response category were summed to produce a total score between 0 and 21 points. “Severe anxiety” was defined as having a score of ≥15; “moderate anxiety” was defined as having a score of 10–14; and “mild anxiety” was defined as having a score of 5–9.
• Alcohol use: Participants were asked about alcohol use during the 30 days and the 12 months before interview. A drink was defined as 12 ounces of beer, a 5-ounce glass of wine, or a 1.5-ounce shot of liquor.
• Binge drinking: Defined as ≥ 5 drinks in a single sitting for men and ≥ 4 drinks in a single sitting for women in the past 30 days.

Sexual Behavior

• Prevention modalities: Reported behaviors that decrease the likelihood of HIV transmission to a sexual partner, including:
  • Sex while having sustained viral suppression: Vaginal or anal sex and the person’s HIV viral loads were documented in the medical record as < 200 copies/mL at every measure in the past 12 months before interview.
  • Condom-protected sex: Condoms were consistently used with at least 1 vaginal or anal sex partner.
  • Condomless sex with a partner on preexposure prophylaxis (PrEP): At least 1 condomless-sex partner with an HIV-negative status was on PrEP use was only measured among the 5 most recent partners and was reported by the partner with HIV.
  • Sex with a partner with HIV: Vaginal or anal sex with at least 1 partner with HIV.
• High-risk sex: Vaginal or anal sex with at least 1 partner with an HIV-negative or unknown status while not having sustained viral suppression, when a condom was not used, and the partner was not known to be taking PrEP.

Met and Unmet Needs for Ancillary Care Services

Ancillary care services were defined as services that support retention in routine HIV medical care and viral suppression, such as HIV case management, dental care, and mental health services. Ancillary care services were grouped into three categories: HIV support services, non-HIV medical services, and subsistence services. HIV support services included: HIV case management, medicine through ADAP, adherence support services, HIV peer group support, and patient navigation services. Non-HIV medical services included: dental care, mental health services, drug or alcohol counseling or treatment, and domestic violence services. Subsistence services included: Supplemental Nutrition Assistance Program (SNAP) or Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), transportation assistance, meal or food services, and shelter or housing services.

• Met need: Defined as an ancillary service received during the 12 months before interview.
• Unmet need: Defined as an ancillary service that the participant reported as needed, but not received, during the 12 months before interview.

Centers for Disease Control and Prevention National Indicators

Measures in this section are used by CDC for national monitoring and evaluation purposes.

• Homelessness among persons receiving HIV care: Defined as living on the street, in a shelter, in a single- room–occupancy hotel, or in a car at any time during the 12 months before interview among persons who received any outpatient HIV medical care in the 12 months before interview.
• HIV stigma: HIV stigma since HIV diagnosis was defined as the median score on a 10-item scale ranging from 0 (no stigma) to 100 (high stigma) that measures 4 dimensions of HIV stigma: personalized stigma since HIV diagnosis, current disclosure concerns, current negative self-image, and current perceived public attitudes about people living with HIV [11]. HIV stigma during the past 12 months was defined as the median score on a 10-item scale ranging from 0 (no stigma) to 100 (high stigma) that measures 4 dimensions of HIV stigma during the past 12 months: personalized stigma during the past 12 months, current disclosure concerns, current negative self-image, and current perceived public attitudes about people living with HIV.
• High-risk sex: See “Sexual Behavior” section.

Ethics Statement

In accordance with guidelines for defining public health research [12], CDC has determined MMP is public health surveillance used for disease control, program, or policy purposes. Local institutional review board approval was obtained at participating states and territories when required. Informed consent was obtained from all interviewed participants.

References

1. Frankel MR, McNaghten A, Shapiro MF, et A probability sample for monitoring the HIV-infected population in care in the U.S. and in selected states. Open AIDS J 2012;6:67–76. doi:10.2174/1874613601206010067external icon
2. The American Association for Public Opinion Standard Definitions: Final dispositions of case codes and outcome rates for surveys. 9th ed. https://www.aapor.org/Standards-Ethics/Standard-Definitions-(1).aspxexternal icon. Revised 2016. Accessed June 23, 2021.
3. U.S. Department of Health and Human Services. Computations for the 2018 poverty guidelines. https://aspe.hhs.gov/2018-poverty-guidelines-computations-pageexternal icon. Published January 2018. Accessed June 23, 2021.
4. U.S. Department of Health and Human Services. Computations for the 2019 poverty guidelines. https://aspe.hhs.gov/computations-2019-annual-update-hhs-poverty-guidelines-48-contiguous-states-and-district-columbiaexternal icon. Published January 2019. Accessed June 23, 2021.
5. CDC [Selik RM, Mokotoff ED, Branson B, Owen SM, Whitmore S, Hall HI]. Revised surveillance case definition for HIV infection—United States, 2014. MMWR 2014;63(RR-03):1–10. https://www.cdc.gov/mmwr/indrr_2014.html. Accessed June 23, 2021.
6. Panel on Antiretroviral Guidelines for Adults and Guidelines for the use of antiretroviral agents in adults and adolescents with HIV. https://clinicalinfo.hiv.gov/en/guidelines/adult-and-adolescent-arv/external icon. Updated June 3, 2021. Accessed June 23, 2021.
7. Wilson IB, Lee Y, Michaud J, Fowler FJ Jr, Rogers Validation of a new three-item self-report measure for medication adherence. AIDS Behav 2016;20(11):2700–2708. doi:10.1007/s10461-016-1406-xexternal icon
8. Kroenke K, Strine TW, Spitzer RL, et The PHQ-8 as a measure of current depression in the general population. J Affect Disord 2009;114(1–3):163–173. doi:10.1016/j.jad.2008.06.026external icon
9. American Psychiatric Diagnostic and Statistical Manual of Mental Disorders: DSM-IV-TR. 4th ed. Washington, DC: American Psychiatric Association; 2000.
10. Spitzer RL, Kroenke K, Williams JB, Löwe A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med 2006;166(10):1092–1097. doi:10.1001/archinte.166.10.1092external icon
11. Wright K, Naar-King S, Lam P, Templin T, Frey Stigma scale revised: reliability and validity of a brief measure of stigma for HIV+ youth. J Adolesc Health 2007;40(1):96–98. doi:10.1016/j.jadohealth.2006.08.001external icon
12. CDC. Distinguishing public health research and public health nonresearch. https://www.cdc.gov/os/integrity/docs/cdc-policy-distinguishing-public-health-research-nonresearch.pdfpdf icon. Published July 2010. Accessed June 23, 2021.