HIV-Infected African Americans are Willing to Participate in HIV Treatment Trials

Abstract

Background

Despite the high prevalence rates of HIV infection in the African-American community, African Americans remain underrepresented in HIV treatment trials.

Objective

(1) To develop a questionnaire that measures attitudes and concerns about HIV treatment trials among HIV-infected African Americans at a university-based clinic. (2) To determine actual participation rates and willingness to participate in future HIV treatment trials among HIV-infected African Americans at a university-based clinic.

Design

Questionnaire development and cross-sectional survey.

Setting, Participants, and Measurements

In a sample of 200 HIV-infected African-American adults receiving medical care at the Pittsburgh AIDS Center for Treatment (a university-based ambulatory clinic), we assessed research participation rates and willingness to participate in future HIV treatment trials, trust in the medical profession, sociodemographic characteristics, attitudes, and concerns about HIV treatment trials.

Main Outcome Measures

Research participation rates and willingness to participate in future HIV treatment trials.

Results

Only 57% of survey respondents had ever been asked to participate in an HIV treatment trial but 86% of those asked said yes. Prior research participation was significantly related to willingness to participate in future HIV treatment trials (P = .001). Contrary to previous studies, neither trust/distrust in the medical profession nor beliefs about the dishonesty of researchers was associated with research participation rates or willingness to participate in future HIV treatment trials.

Conclusions

Having never been asked to participate in research is a major barrier to the participation of HIV-infected African Americans in HIV treatment trials. African Americans who seek medical care for HIV infection should be asked to participate in HIV treatment trials.

Introduction

In the United States, the HIV epidemic disproportionately affects the African-American population. In 2004, African Americans accounted for 49% of the estimated 42,524 AIDS cases diagnosed; this rate was 10 times that for whites.¹ Despite these disproportionate numbers, African Americans remain underrepresented in HIV treatment trials.^2–5 In a 1995 study, among 4,604 HIV-infected individuals, African Americans were significantly less likely to be in trials than all other races or ethnicities.⁶ As of October 1999, African Americans accounted for only 23% of the adults enrolled in HIV treatment trials.⁷ Recent data from the HIV Cost and Services Utilization Study confirmed that HIV-infected African Americans are less likely to have access to experimental treatment or to participate in research than whites.³

Numerous studies exploring the barriers to African Americans’ participation in medical research concluded that compared to whites, African Americans are more distrustful of researchers and less willing to participate in research.^8–10 However, a study comparing African Americans who participate in research with those who do not showed that research participants had more positive perceptions of clinical research.¹¹ We believe that within-culture comparisons are critical for developing effective mechanisms to increase the recruitment and retention of African Americans in research. Therefore, we developed a survey to compare the attitudes and characteristics of HIV-infected African Americans who have and have not participated in HIV trials. Given that previous studies document African Americans’ distrust of research, we were particularly interested in whether research distrust influences participation rates in HIV trials.

Methods

Questionnaire Development

Before constructing the questionnaire, we reviewed existing instruments used in past research and searched the following terms in the MEDLINE database: African-American human subjects, attitudes towards research, (HIV) clinical trials, African Americans distrust of research, research participants survey instrument, and research participants questionnaire. We found no psychometrically validated surveys that assess research participants’ trust of physician-researchers or their attitudes about HIV trials. However, previous studies identified the following variables that influenced the participation of African Americans and other minority groups in research: knowledge of research,^5,10 African American (dis)trust of researchers and the medical community,^7–10 economic group membership,⁷ altruism,⁷ religiosity,⁷ knowledge of informed consent,¹⁰ low socioeconomic status,¹² speaking a primary language other than English,¹² personal inconvenience,¹² and fear of medical procedures.¹² We incorporated these variables into our questionnaire, often using the same language found in previous studies’ instruments. One of the authors (GS), a psychometrician, reworded some questions and rating scales to increase consistency across the questions from the multiple sources and reduce ambiguity where it existed. We included questions about sociodemographic characteristics, insurance status, social support networks, drug and alcohol use, prior research participation, and willingness to participate in future HIV trials.

To measure trust, we included The Trust in the Medical Profession (TIMP) scale, an 11-item scale with a single factor structure, good internal consistency, and good response variability (Cronbach alpha = .89).¹³ The scale uses a 5-point Likert response of “strongly agree” to “strongly disagree” and the responses are averaged to form a composite score that ranges from 1 to 5.¹²

To measure religiousness, we included items from the Multidimensional Measurement of Religiousness/Spirituality for Use in Health Research.¹⁴ The final questionnaire was a one-time only, self-administered, anonymous questionnaire with items from content areas that were of importance in previous research. Except for the TIMP, the intent was not to use or develop composite measures in specific content area but to look at ratings of specific items extracted from other scales.

After developing a draft of the questionnaire, we asked content experts if it accurately represented the concepts and the full breadth of information we were trying to assess. We asked members of the Community Research Advisory Board, a group housed within the Center for Minority Health at the University of Pittsburgh Graduate School of Public Health, to review the questionnaire. We revised the questionnaire in response to their suggestions.

To further ensure that the questionnaire covered topics relevant to the target population, we conducted a focus group consisting of 12 HIV-infected African-American patient volunteers from the Pittsburgh AIDS Center for Treatment (PACT). The focus group was 40% female, all were aged 25–44, and a third had participated in HIV trials. Focus group participants were asked to review and critique the contents of the revised questionnaire. Based on their recommendations, we clarified and reworded ambiguous questions and response options, and revised the visual format of the questionnaire. Finally, we asked 5 PACT patient volunteers to complete the 56-item questionnaire to ascertain the time needed to complete it. The time varied from 20 to 45 minutes; this variation was attributed to varying education levels and reading ability. A copy of the final questionnaire is included in the Appendix.

Methods

Overview

We invited all HIV-infected African-American adults receiving medical care at the PACT to complete the newly developed questionnaire. The University of Pittsburgh Institutional Review Board (IRB) approved this study. Because the survey instrument was an anonymous questionnaire, the IRB waived a written informed consent document.

Setting, Participants, and Survey Administration

The University of Pittsburgh Medical Center (UPMC) houses the PACT, a facility that provides comprehensive medical care to HIV-infected patients from diverse racial and socioeconomic backgrounds, regardless of their medical insurance status. UPMC also houses the NIH-funded Pittsburgh AIDS Clinical Trials Unit (ACTU), which conducts clinical trials of antiretroviral therapies and vaccines. Individuals registered at PACT are recruited into ongoing trials at the ACTU.

In December 2002, 897 patients were registered at PACT; 679 (76%) were men, 597 (66%) were white, and 286 (32%) were African-American. This is comparable to HIV prevalence rates among African Americans in Southwestern Pennsylvania (36%),¹⁵ but lower than prevalence rates in Pennsylvania (51%) and nationally (40%).¹⁶ All African-American patients at PACT, aged 18 years or older, were eligible for our study. The only exclusion criterion was clinical history of cognitive dysfunction as reported by the medical assistant.

Beginning in July 2003, every African-American patient presenting for a scheduled clinic visit was invited to participate in the study by the medical assistant. Patients who expressed interest were then approached by the Principal Investigator, screened for eligibility criteria, presented with the study introductory script, and asked to complete the study questionnaire. For five participants who had problems reading the questionnaire, the Principal Investigator read each question and the participant independently filled in the appropriate option on a separate questionnaire. Each participant was given a $10 gift certificate after completing the questionnaire.

Data Analysis

We obtained descriptive statistics for (1) respondents who had ever been asked to participate in HIV trials, (2) those who agreed to participate, (3) those who enrolled, and (4) those who completed a trial. In addition, we obtained frequency distributions for the responses to the question about willingness to participate in future HIV trials. Using chi-square tests and multivariable logistic regression, we compared independent variables among the following groups: (1) those who were never asked to participate in HIV trials versus those who were asked; (2) those who agreed to participate versus those who refused; (3) those who agreed to participate and enrolled in a study versus those who agreed but never enrolled in a study; and (4) those who enrolled and completed a study versus those who enrolled but did not complete a study. Any independent variable with a P value of ≤0.15 was included in the multivariable analysis. Using backward, stepwise, logistic regression, we arrived at a predictor model for each comparison group with variables showing statistical significance at P < .05. We also compared independent variables among those who expressed willingness, uncertainty, and unwillingness to participate in future HIV trials.

Stata 7.0. (Stata Corporation, College Station, Tex) was used for all analyses. A P value of ≤0.05 was considered statistically significant.

Results

All 228 (80% of all possible black clinic patients) HIV-infected black (African, African-American, and Caribbean-American) patients who presented to PACT from July to November 2003 were invited to join the study. Of those patients, 203 (89%) who were African Americans were recruited, 2 were ineligible because of dementia, and 1 declined. Of the 200 respondents, 3 failed to complete the TIMP scale; 197 respondents were included in all of the analyses. Subanalyses of the group of patients who required assistance with completing the survey was not done because of the relatively small number of patients⁵ in this group.

Figure 1 provides a breakdown of the subject flow and research participation rates among the respondents, illustrating the four comparison groups. Of 197 survey respondents, only 114 (57%) had ever been asked to participate in an HIV trial, 98 (86%) agreed. Eighty (82%) of those who agreed enrolled, 69 (86%) of whom completed a trial.

Figure 1.

Breakdown of the subject flow and research participation rates.

Table 1 shows the sociodemographic characteristics of the respondents. Fifty-six percent were in the 25–44 age category, 78% had known about their HIV infection for >2 years, and 61% had been registered at PACT for >2 years. The mean TIMP score was 3.44 (halfway between undecided and agree, or slightly positive trust). The observed coefficient alpha on the TIMP for our subjects was 0.73 (95% confidence interval [CI] 0.67–0.78).

Table 1.

Sociodemographic Characteristics of 197 HIV-Infected African-American Participants

Characteristics	Percentage
Gender
Male	56
Age (y)
<25	5
25–44	56
45–64	36
≥65	3
Insurance status*
Medicaid	44
Medicare	22
Private	23
None	15
Education
Some high school	12
High school graduate	25
Trade school, college, or higher	63
Income
<$10,000	33
$10,000–$19,999	16
$20,000–$29,999	6
$30,000–$39,999	4
≥$40,000	3
Unknown	38
Time aware of HIV infection
<1 y	7
1–2 y	10
>2 y	78
Unknown	5
Time registered at the clinic
<1 y	20
1–2 y	19
>2 y	61
AIDS classification based on CD4 count
CD4 < 200 cells/μL	21
CD4 ≥ 200 cells/μL	45
Unknown	34

*Some patients had more than one type of insurance coverage.

In the regression analyses, after controlling for being registered at PACT for >2 years, those asked to participate in a trial were significantly more likely to be employed and less likely to be negative about research participation; those who enrolled were significantly more likely to receive social support from their spiritual leader and less likely to use marijuana. Table 2 lists unadjusted odds ratios (OR) from the bivariate analyses and adjusted OR from the multivariable analyses; only statistically significant variables from the bivariate analyses are reported.

Table 2.

Significant Predictors of Being Asked, Agreeing, Enrolling, and Completing HIV Treatment Trials

	Unadjusted		Adjusted
	Odds ratio	95% confidence interval	Odds ratio	95% confidence interval
ASKED: Compared with patients who were never asked to participate in an HIV treatment trial (n = 84), patients who were asked to participate (n = 113)*
More likely to have known of their HIV infection for more than 2 y	2.9	1.3–6.4
More likely to have been a patient registered at the clinic for more than 2 y	3.5	1.9–6.3	3.1	1.4–7.0
More likely to have acquired at least a trade school education or some college education	1.9	1.3–2.9
More likely to be engaged in some form of employment	2.9	1.6–5.3	3.2	1.1–9.3
More likely to have private insurance	2.2	1.1–4.5
Less likely to be negative about research participation	0.4	0.2–0.8	0.3	0.1–0.7
Less likely to believe that research participants must remain in a study until the protocol is completed	0.4	0.2–0.9
AGREEING: Compared with patients who agreed to participate (n = 97), patients who refused to participate in an HIV treatment trial (n = 16)†
Less likely to believe that research doctors are required to disclose risks to research participants	0.2	0.6–0.9
Less likely to believe that research participants can leave a study at any time during the protocol	0.2	0.1–1.0	0.1	0.0–0.6
Less likely to acknowledge support from God or a higher power in dealing with their disease	0.2	0.1–1.0
Less likely to be married or in a committed relationship	0.1	0.0–0.7
Less likely to have tried to cut down on alcohol use at some point in their life	0.2	0.1–0.9	0.1	0.0–0.6
ENROLLING: Compared with patients who agreed to participate but did not enroll in an HIV treatment trial (n = 18), patients who agreed and enrolled (n = 79)‡
More likely to have been a patient registered at the clinic for more than 2 y	5.4	1.9–15.3	6.7	2.0–22.4
More likely to receive support in dealing with their HIV disease from their minister or spiritual leader	5.0	1.5–16.5	4.0	1.1–14.8
More likely to receive support in dealing with their HIV disease from their HIV physician	4.0	1.1–15.0
More likely to receive support in dealing with their HIV disease from their friends	3.4	1.1–10.4
Less likely to question whether God exists	0.4	0.1–1.1
Less likely to interpret life events without relying on God	0.2	0.1–0.6
Less likely to use marijuana	0.3	0.1–0.8	0.3	0.1–0.9
More likely to believe research doctors are required to identify HIV treatment trials as a form of research	9.0	1.5–54.2
More likely to believe that research doctors are required to disclose risks to research participants	3.7	0.9–15.1
COMPLETING: Compared with patients who enrolled and completed the trial (n = 68)§, patients who enrolled but did not complete an HIV treatment trial (n = 11)
Less likely to believe that HIV medical research tests whether new medications will benefit patients	0.2	0.0–0.8	0.2	0.0–0.8
Less likely to believe that the government has rules to ensure that medical research is done properly	0.3	0.1–1.2

A P value of ≤.05 was considered statistically significant in the multivariable analysis.

The main reasons for participating were to help find a cure for AIDS (cited by 77%), to help the HIV community (53%), to get a new and effective experimental drug (47%), and to help the black HIV community (46%). Only 16% of the group reported financial compensation as a motivation for research participation.

*Independent variables with a P value of ≤.15 in the univariate analysis were included in the multivariable analysis.

^†In the group of 16 patients who refused to participate, the 2 most common reasons for refusal were time constraints (cited by 36% of patients) and fear of adverse effects of medications (cited by 29%).

^‡In the group of 18 patients who agreed to participate but did not enroll, the most common reason for failure to enroll was time constraints (cited by 56%).

^§In the group of 68 patients who completed an HIV treatment trial, 55 (81%) said they "felt good" about contributing to research.

Trust in the medical profession did not play a significant role; there were no statistically significant differences in levels of trust in any of the four analyses.

Finally, we asked survey respondents about their willingness to participate in future HIV trials. Eighty-three percent of prior research participants would be willing to participate in future trials. Among all respondents, 67% were willing to participate in future trials, whereas 23% were uncertain and 10% were unwilling. The reasons for willingness to participate in future trials were to help find a cure for AIDS (86%), to help the HIV community (65%) and to obtain new and effective experimental HIV medication(47%). Reasons for unwillingness and uncertainty about participation were fear of adverse effects of medications (45%), concern about the general risks of research (24%), and time constraints (21%). We assigned those uncertain about future participation to the unwilling group; in multivariable analysis the only significant factor related to willingness to participate in future research was prior research participation (adjusted OR = .6; 95% CI 1.9–11). There was no statistically significant difference in levels of trust in the medical profession between those willing, unsure, and unwilling to participate in future HIV trials. (F_2,194 = 1.15, P = .3) Only 13% of those unwilling or uncertain were concerned about the honesty of researchers.

Discussion

Multiple studies document African Americans’ low participation rates in HIV trials.^3–6 However, when we surveyed HIV-infected African-American patients receiving medical care at PACT (a university-based clinic), we found that only 57% had ever been asked to participate in an HIV trial. In fact, when asked, African Americans seemed very willing to participate in a trial. Thus, having never been asked was the major barrier to participation. Contrary to findings in previous studies, we found that neither distrust in the medical profession nor beliefs about dishonesty of researchers was significantly associated with actual participation rates or willingness to participate in future HIV trials.

These findings are consistent with recent publications that African Americans are willing to participate in research: A national study of veterans attending outpatient facilities reported no race differences in self-reported research participation.¹⁷ Wendler et al.¹⁸ performed a systematic review of 20 health research studies that reported consent rates by race or ethnicity. They also found that when invited, African Americans were not less likely and occasionally were more likely to consent to research participation compared to non-Hispanic whites.

Although the bivariate analyses show that both having been diagnosed with HIV infection for >2 years and being a patient at PACT for >2 years were significantly associated with being asked to participate in a trial, the multivariable analyses revealed that having been a patient at PACT for >2 years was the significant determinant. After adjusting for being registered at PACT for >2 years, patients asked to participate in trials were less likely to be negative about research and less likely to believe that research participants must remain in a study until the protocol is completed. This suggests that public education about research may improve participation rates.

We postulate several reasons for the finding that only 57% of respondents had ever been asked to participate in an HIV trial. First, physicians and researchers may have preconceived notions that African Americans are likely to refuse to participate in research. Assuming that African Americans are not interested in research, clinicians may not raise the issue. A similar dynamic was found in studies of organ donation where health care providers perceiving black families were less willing to donate, were requested for organ donation less frequently.¹⁹ While organ donation and research participation are different topics with different emotional impact, it is consistent with the broader communication theory that suggests that a speaker’s preconceptions influence interactions. For example, white researchers may be uneasy talking to African Americans about research assuming (incorrectly) that patients will react with anger or distrust given past abuses. The relationship between the health care practitioner’s ethnicity and whether patients were asked to participate in trials could not be assessed in this study because all of the practitioners at PACT are white. Further research into investigators’ ethnicity, beliefs, and attitudes is needed. Second, medical researchers simply may not know how to effectively communicate with African Americans about research participation particularly in light of the history of research abuses on the African-American population. Third, researchers may believe that African Americans may not be reliable research participants and thus may be reluctant to enroll them in research. For example, Shavers-Hornaday et al.²⁰ report that researchers believe that it is more difficult to obtain African-American patients’ compliance and that African Americans have higher attrition rates. Fourth, some researchers believe that they are more likely to encounter difficulties in contacting patients from minority populations or in financing effective recruitment efforts that target these populations.^20–22 Fifth, it is also possible that some patients were not asked to participate simply because they were ineligible for ongoing trials. Ineligibility may be because of lack of requisite clinical diagnoses and ongoing risky behaviors. However, given that the general policy at PACT is to ask all patients if they are interested in research and then to determine eligibility, this is less likely to be a reason.

Several reasons might explain our finding that 86% of African Americans asked to participate in HIV trials agreed to do so. First, it may be attributable to the unique clinical setting in which the survey was done. PACT is a university-affiliated clinic that provides comprehensive clinical and social services. Second, PACT employs an African-American peer advocate who addresses patients’ concerns and complaints. Third, PACT employs a resident research coordinator who acts as a liaison between the clinic and the ACTU; familiar-face coordinators, who ask patients about research participation in an environment that caters to their medical and social needs, might positively influence research participation rates. A study of HIV-infected patients beginning HIV primary care at Boston City Hospital reported that when a research associate provided patients with information about the purpose, role, and availability of HIV clinical trials, and addressed their concerns about these trials, the overall rates of participation improved and there was no longer a significant difference between the rates of participation by whites and by people of color.²³

Our study found that trust/distrust in the medical profession was not associated with research participation rates and willingness to participate in future trials. This is contrary to reports by Sengupta et al.⁷ that African Americans’ distrust about research institutions was the strongest inverse predictor of willingness to participate in AIDS clinical trials. The reasons for our distinctive findings are not clear. It may be that HIV-infected African Americans who seek medical care at a university-based clinic are less distrustful of the medical profession. Ninety percent of survey respondents indicated that their PACT physician provides considerable support in coping with their disease. El–Sadr and Capps⁴ have noted that African Americans’ fear of exploitation in research and their distrust of researchers can be overcome if trust develops between the primary care provider and the patient. Similarly, Kass et al.²⁴ found in a survey of 1,900 research participants that the patients’ trust in their physicians was instrumental in the participants’ decision to enroll in a research study.

The major limitation of this study is that of selection bias. Because the study population is limited to African Americans at a single university-based clinic, the findings may not be generalizable to other clinical settings. Although the distinctiveness of some of our findings may also be due in part to the desire of survey participants to provide socially desirable responses to questions, we believe this effect is minimal in our study, not only because the questionnaire was anonymous but also because participants reported considerable amounts of alcohol and substance abuse. There is evidence to suggest that when individuals self-report extensive drug use, the information provided is likely to be valid.²⁵

Future studies are needed to assess the beliefs of African Americans in other clinical settings. These studies should assess whether there is a relationship between the clinical services received, satisfaction with these services, and willingness to participate in research. Studies should also assess the relationship between how African Americans are asked about research and their willingness to participate in research.

Conclusion

This study demonstrates that HIV-infected African Americans seeking medical care at a university-based clinic are likely to participate in HIV trials if asked. This study suggests that HIV-infected African Americans who seek medical care for their HIV infection should be asked to participate in HIV trials.

Appendix