Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Feb 1.
Published in final edited form as: AIDS Care. 2019 Sep 16;32(2):175–181. doi: 10.1080/09540121.2019.1668522

Predicting HIV Testing in Low Threshold Community Contexts among Young African American Women Living in the Southern United States

Jalie A Tucker 1,*, Susan D Chandler 1, JeeWon Cheong 1
PMCID: PMC6937398  NIHMSID: NIHMS1540656  PMID: 31526051

Abstract

African American women in the Southern United States have disproportionately high HIV infection rates, and increasing HIV testing is a prevention priority in communities of color. Research suggests that optimal conditions for promoting testing involve reaching out to community members and offering free tests in private, supportive contexts with minimal delays to receiving results. These conditions were implemented with young African American women (N = 223, M age = 20.4 years) living in disadvantaged areas of a Southern U.S. city to identify participant characteristics associated with declining and accepting testing in this low threshold context. Participants were recruited using Respondent Driven Sampling, a peer-driven method useful for recruiting young adults in community settings. Structured field interviews assessed personal and social network characteristics, sexual practices, substance use, and behavioral impulsivity (assessed by a delay discounting task). A free HIV test was then offered. Participants were free to accept or decline testing, which was the outcome variable. Testing was accepted by 69%, which exceeded the national lifetime test rate for this population by 7.4% (p < .05). All were sero-negative. Test refusal (31%) was associated with poorer educational performance, higher delay discounting rates indicative of greater impulsivity, less social network encouragement to use birth control (ps < .05), and lower engagement in sexual risk behaviors (p < .10). Substance involvement did not differ as a function of test choice. Thus, low threshold community testing promoted acceptance among the majority of the priority population, but a sizeable minority with specific characteristics likely need additional incentives to fulfill the prevention potential of “know your status” through HIV testing.

Keywords: HIV/AIDS, community HIV testing, African-American women, young adults, choice architecture, delay discounting

Introduction

For the U.S. “High Impact Prevention” program to succeed (U.S. Centers for Disease Control & Prevention [CDC], 2016), individuals’ knowledge of their sero-status through HIV testing is fundamental (Kay, Batey, & Mugavero, 2016). Improving testing rates is particularly important among young African American women living in the Southern United States where new HIV/AIDS cases continue to rise among women of color (CDC, 2016, 2017). However, clinic-based HIV testing is limited in reach and impact with this population because of racial and economic inequities and infrastructure barriers to clinical services in disadvantaged Southern communities (Reif, Pence, Hall et al., 2015). Moreover, younger women often do not have access to a regular physician or preventive health care, despite improvements since the implementation of the Patient Protection and Affordable Care Act (Transamerica Center for Health Studies, 2018).

For these reasons, community-based HIV testing programs that involve active outreach are promising alternatives to clinic-based testing for younger adults and have received some favorable empirical attention (e.g., Johns, Bauermeister, & Zimmerman, 2010; Murray, Hussen, & Toldedo, 2018), but further work is needed. Key issues in developing effective community testing programs for young African American women, our target group of interest, involve how best to reach them in the community and then offer a testing context that facilitates test acceptance. With respect to reaching them, peer social networks are highly valued among adolescents and young adults and influence many health behaviors, including HIV-relevant behaviors (Davies, Cheong, Lewis, Simpson, Chandler, & Tucker, 2014; Hahm, Kolaczyk, Jang, Swenson, & Bhindarwala, 2012). Thus, community outreach using peer social networks should be an effective engagement approach.

With respect to creating contexts conducive to HIV testing, we were guided by “choice architecture” approaches that seek to promote beneficial behavior and reduce problem behavior by crafting decision-making contexts in ways that promote good choices without restricting freedom of choice (Loewenstein, Brennan, & Volpp, 2007; Thaler & Sustein, 2008). A basic strategy is to accept, rather than attempt to remediate, biased decision-making, such as the normative tendency to devalue or discount delayed rewards in favor more immediate rewards, and to create decision-making contexts that use the biases to increase the likelihood of healthier choices (Tucker, Chandler, & Cheong, 2017). In applying choice architecture to promote HIV testing, which is a form of health-related help-seeking, we considered the established range of influences on service utilization for health and behavioral health problems, namely, social network responses, stigma, access to services, monetary and time costs of services, and the extent to which individuals’ decision-making is sensitive to shorter versus longer-term consequences (Evangeli, Paddy, & Wrote, 2016; Mechanic, 1986; Morrisey, 1993; Tucker, Simpson, & Khodneva, 2010). Particular attention was given to the fact that young people tend to be relatively impulsive and to heavily discount future consequences, including but not limited to those related to health (Madden & Bickel, 2010).

Taken together, these concepts and findings suggest that optimal conditions for promoting HIV testing include active community outreach using peer networks and offering free testing in private, supportive contexts with minimal delays to receiving results. The present study implemented these conditions with young African American women living in economically disadvantaged areas of a Southern U.S. city. We sought to identify personal, behavioral, and social network characteristics of participants for whom this low threshold context was and was not sufficient for test acceptance. We predicted that acceptance would be associated with social network encouragement of HIV risk reduction and reproductive health behaviors and with greater sensitivity to future (delayed) outcomes when making decisions. We explored associations among sexual behaviors, substance use, and testing decisions because prior reports are mixed (Decker et al., 2015; Evangeli et al., 2016; Johns et al., 2010).

Materials and Methods

Participant Recruitment and Characteristics

African American women ages 15-25 years (N = 223, M age = 20.4 years, SD = 2.5) were recruited as a supplemental sample to a larger parent study of sexual and other health risk behaviors in African American emerging adults living in economically disadvantaged areas of a Southern U.S. city (Tucker, Simpson, Chandler, et al., 2016). As in the parent study, the supplemental sample was recruited using Respondent Driven Sampling (RDS) (Heckathorn, 1997; Johnston & Sabin, 2010), a chain referral method that corrects limitations of snowball sampling while maintaining benefits of peer-driven access to community groups. U.S. Census data from 2000 were used to identify disadvantaged neighborhoods for sample recruitment that had high percentages of young African Americans and household incomes below federal poverty guidelines. In addition to age, race, and gender eligibility criteria, participants were required to have telephone access, no plans to move within the next six months, lifetime misuse of any substance, and any sex with a partner in the past 90 days.

Research staff similar in age, gender, and/or race to the target sample recruited initial “seed” participants who met eligibility criteria in person. Seeds were trained to recruit peers “like you,” who then recruited peers in an iterative process (excluding blood relatives). Seeds and recruits received three time-limited coupons to distribute to peers that provided a toll free study number to call if they wished to enroll, which provided freedom of choice to participate. All participants received $30 for an initial 1.5-hour data collection interview and $15 for each peer recruit who enrolled in the study (up to $45). The sample included 81 seeds and 142 recruits who recruited an average of 1.76 peers (SD = 1.90; range = 0 to 7 recruitment waves); seeds directly recruited 49 peers, who in turn recruited 93 peers. Recruitment chains were allowed to continue until they naturally stopped.

Procedures

The research received university Institutional Review Board approval and a federal Certificate of Confidentiality and adhered to STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines for cross-sectional studies (Von Elm et al., 2007). Supplemental and parent (Tucker et al., 2016) study procedures were the same, with one exception: At the end of 1.5-hour data collection interviews conducted in safe, private community locations, participants were offered free voluntary HIV counseling and testing (OraQuick) with guaranteed referral for confirmatory testing and HIV care if results were positive. Test choice was the outcome variable.

In addition to demographic characteristics, predictors of test choice were derived from four measures:

An adapted Youth Risk Behavior Surveillance System Questionnaire (CDC, 2009) assessed variables associated with HIV risk and reproductive health, including age of first intercourse, condom use during last sexual intercourse, substance use before last intercourse, ≥ 2 sexual partners during last 90 days, sex with an injection drug user, birth control use, and sex in exchange for food, shelter, money, or drugs. A binary sex risk variable was created for analysis (1 = ≥ 1 risk variable; 0 = none). Binary variables also were created for any prior HIV testing (1 = yes; 0 = no) and sexually transmitted infection (1 = any STI; 0 = none).

The Alcohol, Smoking and Substance Involvement Screening Test (ASSIST v 3.0; World Health Organization, 2010) assessed lifetime and past 90-day alcohol, illicit, and non-medical prescription drug use and yielded a Global Continuum of Risk score (range = 0 to 280) that was used for analysis. Higher scores indicate greater substance involvement.

Delay discounting of future outcomes in favor of more immediate rewards was assessed using a computerized hypothetical money choice task (Richards, Zhang, Mitchell, & de Wit, 1999). Participants made repeated choices between a smaller money amount available immediately and a larger amount available at 5 delays (1, 2, 30, 180, and 365 days from now; e.g., $2 now or $10 in 30 days; $50 now or $100 in 180 days). An equivalence point was determined for each delay, estimating the amount of immediate money that was subjectively judged equivalent to the larger later amount. These equivalence points were used to derive a discount rate (log k) for analysis that modeled the nonlinear trajectory of changes in the devaluation of future rewards as a function of delay to receipt. Higher values indicate more immediate reward preferences. It was not necessary to incentivize choices because hypothetical and real money ranging from small to large amounts generate equivalent delay discounting rates in diverse populations, including among emerging adults (e.g., Dixon, Lik, Green, & Myerson, 2013; Johnson & Bickel, 2002; cf. Madden & Bickel, 2010).

Using an expanded Norbeck Social Support Questionnaire (Norbeck, Lindsey, & Carrieri, 1981; Tucker, Cheong, Chandler, Crawford, & Simpson, 2015), participants listed up to 12 network members by first name and relationship and then rated the extent to which each member encouraged health behaviors, including safer sex practices and birth control use (1 = not at all and 5 = a great deal). Member feedback ratings for each health behavior were averaged across network members to yield scores for analysis that reflected overall network encouragement to practice safer use and use birth control, which adjusted for individual differences in network size (M = 7.43 members, SD = 2.83).

Data Analysis

Analyses were conducted using SAS 9.4. First, participants who did and did not accept HIV testing were compared on demographic characteristics, HIV risk variables related to substance involvement and sexual practices, delay discounting, and social network feedback (see Table 1). To increase power for analysis, seeds and recruits were combined because the only difference found on eligibility criteria was that seeds were about a year older (M = 20.08) than recruits (M = 21.06; t(1, 221) = 2.88, p < 0.005). Second, variables of conceptual significance (sex risk indicators, substance involvement, log k, network feedback) and/or that differed significantly among HIV testing groups in the bivariate comparisons were included together in a logistic regression analysis to evaluate predictors of HIV testing decisions. A dichotomous age variable (age < 21 years or age ≥ 21 years) also was included to control for any effects of age (e.g., on parent status; access to alcohol, tobacco, and adult entertainment venues; differences in impulsivity). Of 223 participants, 33 had missing values on predictor and outcome variables and 14 had invalid response patterns on the discounting task and were excluded in the final logistic regression (n = 176). Finally, an exploratory analysis evaluated whether our testing context improved testing rates over the national prevalence for young African American women. A z-test compared the proportion of study participants who accepted testing to the unweighted proportion of lifetime HIV testing prevalence in the representative U.S. National Health Interview Survey (National Center for Health Statistics, 2017) matched for age, gender, race, and data collection period (2012-2014).

Table 1.

Demographic characteristics and descriptive statistics

Demographic Characteristics Accepted
HIV test
(n = 154)		 Declined
HIV test
(n = 69)		 Total
sample		 Test statistic
Frequency (%)/Mean (SD)
 High school completed/in progressa 150 (97.4) 63 (91.3) 213 (95.5) χ2 (1) = 4.14*
 Average gradesb
  Mostly As 32 (20.8) 8 (11.6) 40 (17.9) χ2 (1) = 2.73+
  Mostly Bs 98 (63.6) 35 (50.7) 133 (59.6) χ2 (1) = 3.30+
  Mostly Cs or lower 24 (15.6) 26 (37.7) 50 (22.4) χ2 (1) = 13.38***
 Employedc 86 (55.8) 38 (55.1) 124 (55.6) ns
 Receipt of public assistance 114 (74.0) 49 (71.0) 163 (73.1) ns
 Married 14 (9.1) 11 (15.9) 25 (11.2) ns
 Have children 37 (24.0) 26 (37.7) 63 (28.3) χ2 (1) = 4.38*
 Age in years 20.4 (2.3) 20.5 (2.8) 20.4 (2.5) ns
Health Risk Behaviors
 Substance used
  Alcoholic beverages 139 (90.3) 64 (92.8) 203 (91.0) ns
  Tobacco products 92 (59.7) 39 (56.5) 131 (58.7) ns
  Illicit drugs 100 (64.9) 39 (56.5) 139 (62.3) ns
 Sexual history
  Sexual risk behaviors (yes/no)e 112 (72.7) 47 (68.1) 159 (71.3) ns
  Any STI (yes/no) 39 (29.3) 17 (31.5) 56 (30.0) ns
  Prior HIV testing (yes/no) 114 (76.0) 58 (87.9) 172 (80.0) χ2 (1) = 3.99*
 ASSIST global continuum of substance risk scored 21.0 (17.3) 19.6 (20.2) 20.5 (18.3) ns
 Delay discount rate (log k)f −3.4 (2.0) −2.6 (1.8) −3.2 (2.0) t(179) = −2.65**
Network Characteristics
 Network encouragement of safer sex 4.0 (1.8) 4.0 (2.5) 4.0 (2.1) ns
 Network encouragement of birth control 3.6 (1.7) 2.9 (1.7) 3.4 (1.7) t(221) = 2.69**
Open in a new tab

Note. Descriptive statistics are frequencies (percentages) for categorical variables and means (standard deviations) for continuous variables based on participants who provided data for a given variable. Comparisons among participants who accepted or declined HIV testing are chi-square tests for dichotomous variables and t-tests for continuous variables. ASSIST = Alcohol, Smoking and Substance Involvement Screening Test. STI = sexually transmitted infection.

a

Participants currently in high school or received high school or higher education.

b

Average grades in the last 2 years in school (1 = mostly As; 2 = mostly Bs; 3 = mostly Cs; 4 = mostly Ds; 5 = mostly Fs).

c

Employment resulting in at least weekly pay.

d

Use of specific substances and global risk scores were based on ASSIST reports of lifetime substance use; the Global Continuum of Risk scale also assessed lifetime network concerns about substance use, failed quit attempts, and injection drug use.

e

Binary variable of endorsing any of seven sexual risk factors (see text; 0 or ≥ 1 factor).

f

Log k was based on 181 participants, excluding 42 with invalid response patterns or missing data on the delay discounting task.

+

p < .10,

*

p < .05,

**

p < .01,

***

p < .001.

Results

As summarized in Table 1, consistent with study recruitment goals, the sample as a whole was economically disadvantaged and reported HIV risk behaviors; i.e., 73% lived in households receiving government assistance, 44% were unemployed, 71% reported one or more sexual risk factors (e.g., no condom use during last sex, 41%; ≥ 2 sexual partners in last 90 days, 19%), and 78% reported substance use in the last 90 days. Furthermore, 28% were mothers; mothers were older (M = 22.05 years) than participants without children (M = 19.81; χ2(1) = 34.9, p < .0001, and 78% of mothers were unmarried.

Of the 223 participants, 69.1% accepted testing, and 30.9% declined. No participant tested positive. The sample proportion that accepted testing was significantly higher than the national lifetime testing rate of 61.7% for African American women ages 18-25 (z = 2.07, p < .05; 95% CI: .004, .143). This is a 7.4% increase in test acceptance compared to the matched national sample.

Table 1 presents the bivariate comparisons between participants who did and did not accept HIV testing. Significant personal characteristics associated with test acceptance included high school completion/in progress, higher average grades during the last two school years, and lower frequency of motherhood. As predicted, lower delay discount rates indicative of greater self-control and greater social network feedback to use birth control also were associated with test acceptance. However, with one exception, the test choice groups did not differ significantly with respect to substance use, sexual risk indicators, or social network feedback encouraging safer sex. Participants who had not been tested for HIV before were more likely to accept testing.

Table 2 presents the results of the logistic regression analysis, which included all significant variables in the bivariate comparisons, the non-significant variables of sexual risk behaviors and substance involvement because of their emphasis in HIV testing research, and age (< 21 or ≥ 21 years). Test acceptance was predicted by higher school achievement (p < .05), lower behavioral impulsivity (p < .05), and higher network encouragement to use birth control (p < .01). Also consistent with the bivariate findings, substance use was not a significant predictor. But in the multivariable model, higher frequency of endorsement of sexual risk behaviors was marginally associated with test acceptance (p < .10).

Table 2.

Associations among HIV testing decisions and personal, social, and risk behavior characteristics: Logistic regression results

Predictors B (SE) OR (95% CI)
Agea −0.57 (0.41) 0.57 (0.24, 1.27)
High completed/in progress 0.56 (1.03) 1.74 (0.24, 12.95)
Have children (yes/no) −0.27 (0.44) 0.76 (0.32, 1.81)
Average gradesb −0.75 (0.29)** 0.47 (0.27, 0.84)
Sexual risk behaviors (yes/no)c 0.80 (0.42)+ 2.22 (0.97, 5.09)
Prior HIV test (yes/no) −0.40 (0.56) 0.67 (0.22, 2.02)
Network encouragement of birth controle 0.28 (0.11)* 1.32 (1.06, 1.63)
ASSIST global continuum of substance risk scored 0.01 (0.01) 1.01 (0.99, 1.03)
Delay discounting log kf −0.22 (0.11)* 0.81 (0.65, 0.99)
Open in a new tab

Note. Odds ratio (OR) with 95% confidence intervals (CI) for binary testing decision (accept = 1; reject = 0).

a

Dichotomized as age < 21 years or ≥ 21 years

b

Average grades in the last 2 years in school (1 = mostly As; 2 = mostly Bs; 3 = mostly Cs; 4 = mostly Ds; 5 = mostly Fs).

c

Binary variable reflecting endorsement of any of seven sexual risk factors (0 or ≥ 1 factor; see text).

d

ASSIST = Alcohol, Smoking and Substance Involvement Screening Test; higher Global Continuum of Risk scores indicate greater substance involvement.

e

Encouragement of safer sex and use of birth control (1 = not at all to 5 = a great deal) averaged over ratings for all network members.

f

Log-transformed to reduce skewness.

+

p < .10;

*

p < .05;

**

p = .01.

Discussion

HIV testing was offered to young African American women living in disadvantaged urban communities under conditions highly conducive to accepting it: The test was free, confidential, offered on the spot after interviewers had established rapport with participants, and results were available with minimal delay (20-30 minutes) with assurance of immediate referral for medical care if indicated. These conditions guided by choice architecture resulted in test acceptance by almost 70% of participants. The sample acceptance rate for this single testing opportunity was significantly higher than the cumulative lifetime U.S. testing prevalence for young African American women (National Center for Health Statistics, 2017). Furthermore, our direct observation of test choices lends confidence to study findings (cf. Fan, Fife, Cox, Cox, & Zimet, 2018) and contrasts with many earlier studies that relied on retrospective reports of testing over variable or indeterminate intervals that are susceptible to recall and other reporting biases (Evangeli et al., 2016). Overall, the present approach appears promising for increasing testing rates among young African American women if implemented on a larger scale.

Even under these favorable conditions, about 31% declined testing and had characteristics that may help identify when additional incentives for test acceptance are needed. Compared to test acceptors, refusers had relatively poorer school performance and higher discount rates indicative of greater impulsivity. Refusers also had less social network encouragement to use birth control, which, like HIV testing, is a protective reproductive health behavior. However, network encouragement of safer sex and substance involvement did not differentiate testing decisions. In the multivariable model only, there was a marginal tendency for test refusal to be predicted by lower endorsement of sexual risk behaviors. These largely negative findings replicate other studies using similar samples (e.g., Decker et al., 2015; Jones et al., 2010) and suggest that participants’ testing decisions are in line with their more general patterns of choice that vary in the degree of sensitivity to future outcomes. This supports expanding the scope of targeted testing programs beyond young adults known to be engaging in substance use and risky sex.

The study has limitations. First, except for the computerized discounting task, risk predictors were based on verbal reports of sensitive behaviors during structured interviews. This helped establish rapport considered important to promote HIV testing, but verbal reports can be biased. To facilitate accuracy, confidentiality was protected, validated measures were used, and participants and interviewers were of similar age, gender, and race. Second, the sample was drawn from a particular race/ethnicity group living in a southern city, and results may not generalize to other populations. Third, due to unexpected funding cuts during data collection, the sample size was smaller than originally planned and insufficient to apply RDS analysis procedures to check and adjust for possible recruitment bias due to chain referrals (Johnston & Sabin, 2010). Checks conducted on the larger parent sample showed no evidence of bias (Tucker et al., 2016). Similar results would be expected for the supplemental sample given that the same methods were used to recruit participants of the same race and age range in the same neighborhoods.

With these qualifications, the study supported the utility of peer-driven recruitment and consideration of social and economic factors when crafting appealing testing programs for younger women who are not seeking clinical care. When implemented in community settings, the present testing context promoted test acceptance among the majority of participants. Among the minority who refused testing, participant characteristics were identified that may guide provision of additional incentives to promote testing. This could entail having peers, pharmacists, or community health workers disseminate free or low-cost home test kits that increase convenience and confidentiality (Meyerson, Carter, Lawrence, et al., 2016; Tobin, Edwards, Flath, Lee, Tormohlen, & Gaydos, 2018; cf. Estem, Catania, & Klausner, 2016) and incentivizing HIV testing with commodities of value to young people, such as money, club admission, or free drinks (Murray, Hussen, Toldedo, et al., 2018). Reaching young African American women with agreeable HIV testing options is a vital initial step for the U.S. “High Impact Prevention” program to succeed (CDC, 2016) in Southern communities of color.

Acknowledgements

This research was supported in part by the U.S. Centers for Disease Control and Prevention [cooperative agreement #5U48DP001915] awarded to the UAB Prevention Research Center/Center for the Study of Community Health and by the UAB Center for AIDS Research CFAR, an NIH funded program [P30 AI027767] made possible by the following institutes: NIAID, NCI, NICHD, NHLBI, NIDA, NIMH, NIA, NIDDK, NIGMS, NIMHD, FIC, and OAR. Study design and data collection were conducted when all authors were affiliated with the Department of Health Behavior, University of Alabama at Birmingham (UAB), Birmingham, AL and the UAB Prevention Research Center/Center for the Study of Community Health and the UAB Center for AIDS Research. Portions of the research were presented at the annual meeting of the American Psychological Association, August 2016, Denver, CO. The authors thank Cathy A. Simpson, Ph.D. and Ms. Julie Hope for contributing to measurement development and Michael J. Mugavero, M.D., M.H.Sc. for arranging clinic-based HIV testing through the UAB 1917 Clinic.

Footnotes

Declaration of Interests: None.

References

  1. Davies S Cheong, Lewis T, Simpson C, Chandler S, & Tucker J (2014). Sexual risk typologies among urban African American youth in the southern United States: Relationships with early parenthood and STI outcomes. Sexually Transmitted Infections, 90(6), 475–477. [DOI] [PubMed] [Google Scholar]
  2. Decker MR, Rodney R, Chung S-E, Jenning JM, Ellen JM, & Sherman SG (2015). HIV testing among youth in a high-risk city: prevalence, predictors, and gender differences. AIDS Care, 27(5), 555–560. 10.1080/09540121.2014.986048 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Dixon MR, Lik NMK, Green L, & Myerson J (2013). Delay discounting of hypothetical and real money: The effect of holding reinforcement rate constant. Journal of Applied Behavior Analysis, 46, 512–517. [DOI] [PubMed] [Google Scholar]
  4. Estem KS, Catania J, & Klausner JD (2016). HIV self-testing: A review of current implementation and fidelity. Current HIV/AIDS Reports, 13, 107–115. [DOI] [PubMed] [Google Scholar]
  5. Evangeli M, Pady K, & Wrote ML (2016). Which psychological factors are related to HIV testing? A quantitative systematic review. AIDS and Behavior, 20, 880–918. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Fan H, Fife KH, Cox D, Cox AD, & Zimet GD (2018). Behavior and health beliefs as predictors of HIV testing among women: A prospective study of observed HIV testing. Aids Care, 30(8), 1062–1069. 10.1080/09540121.2018.1442555 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Hahm HC, Kolaczyk E, Jang J, Swenson T, & Bhindarwala AM (2012). Binge drinking trajectories from adolescence to young adulthood: The effects of peer social network. Substance Use & Misuse, 47(6), 745–756. doi:10.3109/10826084.2012.666313. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Heckathorn D (1997). Respondent-driven sampling: A new approach to the study of hidden populations. Social Problems, 44, 174–99. [Google Scholar]
  9. Johns MM, Bauermeister JA, & Zimmerman MA (2010). Individual and neighborhood correlates of HIV testing among African American youth transitioning from adolescence into young adulthood. AIDS Education and Prevention, 22, 509–522. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Johnson MW, & Bickel WK (2002). Within-subject comparison of real and hypothetical money rewards in delay discounting. Journal of the Experimental Analysis of Behavior, 77, 129–146. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Johnston LG, & Sabin K (2010). Sampling hard-to-reach populations with respondent driven sampling. Methodological Innovations Online, 5, 38–48. [Google Scholar]
  12. Kay ES, Batey DS, & Mugavero MJ (2016). The HIV treatment cascade and care continuum: Updates, goals, and recommendations for the future. AIDS Research and Therapy, 13, 35. eCollection [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Loewenstein G, Brennan T, & Volpp K (2007). Asymmetric paternalism to improve health behaviors. Journal of the American Medical Association, 298, 2415–17. [DOI] [PubMed] [Google Scholar]
  14. Madden GJ, & Bickel WK (2010). Impulsivity: The Behavioral and Neurological Science of Discounting. Washington, DC: American Psychological Association. [Google Scholar]
  15. Mechanic D (1986). Illness behavior: An overview In: McVugh S & Vallis TM, eds. Illness Behavior: A Multidisciplinary Model. New York: Plenum Press; (pp.101–109) [Google Scholar]
  16. Meyerson BE, Carter G, & Lawrence C, et al. (2016). Expanding HIV testing in African American communities through community-based distribution of home-test vouchers AIDS Patient Care and STDs, 30, 141–145. [DOI] [PubMed] [Google Scholar]
  17. Morrisey MA (1992). Price Sensitivity in Health Care: Implications for Health Care Policy. Washington, DC: National Federation of Independent Business Foundation. [Google Scholar]
  18. Murray A, Hussen SA, Toldedo L, Thomas-Seaton L Gillespie S, Graves C. … Camacho-Gonzalez AF. (2018). Optimizing community-based HIV testing and linkage to care for young persons in metropolitan Atlanta. AIDS Patient Care and STDs, 32, 234–40. [DOI] [PubMed] [Google Scholar]
  19. National Center for Health Statistics (2017). Early release of selected estimates based on data from the National Health Interview Survey, January – March 2017 Human immunodeficiency virus testing. Hyattsville, MD: National Center for Health Statistics, 69–74. [Google Scholar]
  20. Norbeck JA, Lindsey AM, & Carrieri VL (1981). The development of an instrument to measure social support. Nursing Research, 30, 264–269. [PubMed] [Google Scholar]
  21. Reif S, Pence BW, Hall I, Hu X, Whetten K, & Wilson E (2015). HIV diagnoses, prevalence and outcomes in nine southern states. Journal of Community Health, 40, 642–651. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Richards JB, Zhang L, Mitchell SH, & de Wit H (1999). Delay or probability discounting in a model of impulsive behavior: Effect of alcohol. Journal of the Experimental Analysis of Behavior, 71, 121–143. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Thaler RH, & Sunstein CR (2008). Nudge: Improving Decisions about Health, Wealth, and Happiness. New York: Penguin Books. [Google Scholar]
  24. Tobin K, Edwards C, Flath N, Lee A, Tormohlen K, & Gaydos CA (2018). Acceptability and feasibility of a Peer Mentor program to train young Black men who have sex with men to promote HIV and STI home-testing to their social network members. AIDS Care, 30(7), 896–902. 10.1080/09540121.2018.1442553 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Transamerica Center for Health Studies (2018). Sixth Annual Nationwide TCHS Consumers Healthcare Survey: Stressed Out: Americans and Healthcare. https://www.transamericacenterforhealthstudies.org/docs/default-source/research/tchs2018-healthcare-consumer-research-report.pdf
  26. Tucker JA, Chandler SD, & Cheong J (2017). Role of choice biases and choice architecture in behavioral economic strategies to reduce addictive behaviors In Heather N & Segal G (Eds.), Addiction and choice: Rethinking the relationship. Oxford, U.K: Oxford University Press; (pp. 346–364) [Google Scholar]
  27. Tucker JA, Cheong J, Chandler SD, Crawford MS, & Simpson CA (2015). Social networks and substance use among at-risk emerging adults living in disadvantaged urban areas in the southern United States: A cross-sectional naturalistic study. Addiction, 110, 1524–1532. [DOI] [PubMed] [Google Scholar]
  28. Tucker JA, Simpson CA, Chandler SD, Borch CA, Davies SL, Kerbawy SJS … Michael M (2016). Utility of Respondent Driven Sampling to reach disadvantaged emerging adults for assessment of substance use, weight, and sexual behaviors. Journal of Health Care for the Poor and Underserved, 27, 194–208. DOI: 10.1353/hpu.2016.0006 [DOI] [PubMed] [Google Scholar]
  29. Tucker JA, Simpson CA, & Khodneva Y (2010). The role of time and delay in health decision making In: Madden GJ & Bickel WK, eds. Impulsivity: The Behavioral and Neurological Science of Discounting. Washington, DC: APA Books; (pp. 243–272) [Google Scholar]
  30. U.S. Centers for Disease Control and Prevention (2017). HIV among African Americans. February 2017. Available from: https://www.cdc.gov/nchhstp/newsroom/docs/factsheets/cdc-hiv-aa-508.pdf.
  31. U.S. Centers for Disease Control and Prevention (2016). HIV in the Southern United States. May 2016. Available from: https://www.cdc.gov/hiv/pdf/policies/cdc-hiv-in-the-south-issue-brief.pdf.
  32. U.S. Centers for Disease Control and Prevention. Youth Risk Behavior Survey 2009. Available from: https://www.cdc.gov/healthyyouth/data/yrbs/questionnaires.htm
  33. Von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, & Vandenbroucke JP STROBE initiative. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies (2007). Epidemiology, 18, 800–804. [DOI] [PubMed] [Google Scholar]
  34. World Health Organization. The Alcohol, Smoking and Substance Involvement Screening Test (ASSIST) v3.0. Geneva, Switzerland: WHO Press, 2010. Available from: http://www.who.int/substance_abuse/activities/assist_test/en/. [Google Scholar]

RESOURCES