Adherence and Retention in Clinical Trials: A Community-Based Approach

Abstract

Background

The Community Health Advisor (CHA) model has been widely used to recruit rural and low-income, mostly African American women into clinical and behavioral research studies. However, little is known about its effectiveness in promoting retention and adherence of such women in clinical trials.

Methods

The Community-based Retention Intervention Study (CRIS) evaluated the effectiveness of a community-based intervention strategy using the CHA model and the empowerment theory to improve the retention and adherence of minority and low-income women in clinical trials. The research strategy included the training and use of the volunteer CHAs as research partners. The target population included women participating in the University of Alabama at Birmingham (UAB) clinical site of the ASCUS-LSIL Triage Study (ALTS), a multicenter, randomized clinical trial. Two communities in Jefferson County, Alabama, matched according to population demographics, were identified and randomly assigned to either intervention or control group. Thirty community volunteers were recruited to be CHAs and to implement the intervention with the ALTS trial participants. A total of 632 ALTS participants agreed to participate in the project: 359 in the intervention group with CHA care, and 273 in the control group with standard care.

Results

Adherence rates for scheduled clinic visits were significantly higher in the intervention group (80%) compared to the control group (65%; p < 0.0001).

Conclusion

Results indicate that volunteer CHAs can be trained to serve as research partners and be effective in improving the retention and adherence of minority and low-income women in clinical trials.

Introduction

To ensure retention and adherence of minority and underrepresented women in clinical trials, researchers have had to: 1) expand theoretical frameworks to include minorities and women; 2) reduce barriers to accessibility and participation; and 3) improve trials’ acceptability by reducing fear and mistrust of clinical procedures and experiences.^1–4 Studies addressing the conceptual and structural barriers ^5–9 suggest that strategies to retain African American women may necessitate labor-intensive population-based approaches.

The Community Health Advisor (CHA) model has been widely used in Southern states to recruit rural low-income, mostly African American women into clinical and behavioral research studies, and projects report overall positive results. ^10–16 The model appears particularly promising in addressing the health needs of low-income women. However, little is known about the effectiveness of the CHA model in promoting retention and adherence of African American women in clinical trials.

The Community-based Retention Intervention Study (CRIS) evaluated the effectiveness of a community-based intervention strategy to enhance the retention and adherence of African American and low-income women in clinical cancer research and to test the CHA model as a research partnership. Evaluating the effectiveness of this strategy included examining both the adherence of the research participants and the CHAs. We hypothesized that both retention in a clinical trial and adherence to scheduled appointments would be higher among research participants who were paired with a CHA, or natural helper.

Methods

Design of the CRIS Study

The CRIS project collaborated with the University of Alabama at Birmingham (UAB) clinical site of the ASCUS-LSIL Triage Study (ALTS), a multi-center clinical trial funded by the National Cancer Institute to evaluate the optimal clinical management of low-grade cervical cytologic abnormalities. ALTS participants were randomized to three management strategies: 1) immediate colposcopy; 2) human papilloma virus (HPV) DNA testing, which triaged to colposcopy only participants with oncogenic HPV type; and 3) conservative management followed with serial Pap smears and colposcopy if Pap smear progressed to high grade. All participants, regardless of condition, were asked to attend follow-up clinic visits every six months over a period of two years (a total of four visits). Three of the visits were for a Pap smear, HPV test, and a cervigram; the exit visit was for colposcopy ± LEEP as indicated.¹⁷

The ALTS trial targeted women residing in Jefferson County, Alabama. When the CRIS ancillary study began, some ALTS participants had started their series of visits; altogether 40% of all study visits had been made. The geographic area for CRIS included two low-income communities in Jefferson County, Alabama, matched according to population demographics. A total of 632 ALTS participants residing in the communities identified for CRIS consented to participate in the CRIS study. Following the consent, the two CRIS communities were randomly assigned to a treatment or a control group. Participants in the intervention group (N=359) received CHA-supported retention activities (four types of communication and support), in addition to the reminder calls, cards, and retention incentives offered by the ALTS trial ($20 at each visit and $100 and a gift bag at final visit). Participants in the control group (N=273) received only the reminder calls, cards, and retention incentives offered by the ALTS trial.

The CRIS retention strategy incorporated the CHA model¹⁸ and an adaptation of the empowerment model¹⁹ to retain participants in the ALTS trial for four follow-up visits. The CHA model allows community members to serve as a link between providers and consumers for improved health.¹⁸ The empowerment model is based on an educational theory that promotes both personal and social change. It encourages researchers to allow community members to identify their critical issues, and then to initiate a dialogue about designing and implementing interventions that address these critical issues.¹⁹ The model was used to create awareness about cervical cancer and cancer clinical trials in two ways: 1) by developing the knowledge, communication abilities, and the capacity of CHAs; and 2) by motivating CHAs to help change participants’ behavior and follow study protocols for improved health. CHAs in the CRIS study served as a link between study participants and study investigators and staff and were considered research partners.

Identification and Selection of CHAs

The CHA training protocol^20,21 guided the recruitment and training process. Potential CHAs were recruited from churches, civic organizations, and through referrals made by community leaders (e.g., ministers, mayors, and city council members). Potential CHAs were to: 1) be highly regarded in their community; 2) be identified by well-respected leaders and peers in the community; 3) be viewed as non-judgmental leaders themselves; 4) identify with and support their community members; 5) reside in the ALTS intervention community; 6) have no plans to leave the community during the course of the study; 7) have similar demographics as the ALTS participants; 8) have access to reliable transportation; 9) be literate; and 10) consent to participate in the study. A total of 38 women were identified and recruited for the CHA training.

CHA Training and Maintenance Program

At the close of the orientation session, 30 women consented to be trained as a CHA and participate in the CRIS study; all 30 completed the training. The 20-hour training program consisted of two hours of training over the course of ten consecutive weeks. At the end of the ten-week training, a ceremony was held and each CHA received a graduation certificate and a $50 gift card.

During the orientation session (Session 1), CHAs were provided with an overview of clinical trials research, information about the ALTS trial (inclusion criteria and goals), and specific details about the CHAs’ role in the CRIS study. CHAs were expected to be research partners and fulfill four specific functions: they were to be marketers, mentors, motivators, and monitors. As marketers, they were expected to promote the benefits of research and the importance of participation of women in clinical trials by answering questions regarding clinical trials in general, the ALTS trial, and the CRIS project. Since the CHAs resided in the same community as the ALTS participants and had understanding of their life experiences, they were to mentor the younger women, helping them to overcome barriers and comply with the follow-up visits in the trial. As motivators, the CHAs were expected to increase adherence by making direct contact with ALTS trial participants through telephone, with reminder cards, or by visits when appropriate, to encourage women to make their scheduled appointments. As monitors, the CHAs were to document all interactions with their assigned ALTS cases.

Subsequent training sessions provided more details regarding 1) Pap smears, 2) clinical trials research, 3) specific information about the ALTS trial, 4) skills-building activities to prepare CHAs as motivators and mentors, and 5) training to prepare CHAs as monitors. CHAs also practiced a variety of problem-solving scenarios. Following the didactic training, each CHA participated in a site visit to the ALTS trial clinic. This visit allowed the CHAs to experience first-hand what an ALTS trial participant would undergo at a follow-up visit, with the exception of the examination. The simulation also allowed CHAs to review additional ALTS trial materials, meet the ALTS physicians and nurses, and become familiar with the clinic environment.

CRIS Intervention

Each trained CHA was assigned 8 to 12 ALTS trial participants who had consented to participate in CRIS. The CRIS study protocol required that CHAs make four types of intervention calls. The first contact was an introductory call where the CHA identified herself, promoted the CRIS study, discussed confidentiality, and reviewed the ALTS visit protocol. The second type of call was a reminder call: the CHA inquired about any barriers to adherence, assisted with problem-solving, reiterated the importance of making the visit, and reminded the participant to cancel and reschedule the appointment if unable to keep it. The third type of call was a follow-up call the day after a scheduled visit: the CHA was to provide positive feedback for a kept appointment and identify potential negative experiences that could impact completing the next visit. If the ALTS participant was a no-show, the CHA asked her to reschedule and call the CHA back with the new appointment time within 3 days. If this did not occur, the CHA would contact the woman and encourage her again to reschedule a visit. The fourth type of call was a miscellaneous call made to acknowledge special occasions such as birthdays or anniversaries, to inquire about special needs, or to respond to questions.

Each time a CHA contacted one of her assigned ALTS cases, she completed a contact form recording the time, date, mode of contact, and outcome of the contact. Throughout the CRIS project, monthly maintenance or debriefing meetings were held to support the CHAs. CHAs were encouraged to discuss their activities, successes, and difficulties retaining ALTS participants, to submit their contact forms, and to receive the next month’s schedules and updates for the trial participants assigned to them.

The CHAs were not employees of the university; they were volunteers who were compensated and rewarded for their contribution as partners in research.¹³ As travel compensation, CHAs received $10 for attending monthly maintenance sessions. As a reward for effort and success, CHAs received $5 each month for contacting ALTS trial participants, and $5 for each ALTS trial participant that attended their follow-up visit based on ALTS trial clinic records. Additionally, CHAs received non-monetary incentives, including special presentations on health-related topics requested by them.

Data Collection and Statistical Analysis

ALTS trial participants and CHAs signed an informed consent document approved by the Institutional Review Board at the University of Alabama at Birmingham when they agreed to participate in the CRIS study. At the end of the two-year project, CHAs completed an exit survey. Questions inquired about the CHA training, case load, role as a research partner, types of contacts made with the ALTS participants, and satisfaction with incentives. Data regarding the ALTS trial follow-up visits were collected from electronic appointment records provided by the ALTS trial research staff and investigators.

Descriptive statistics were used to compile the data for the CRIS project. Chi-square analyses compared retention and adherence rates for ALTS trial follow-up visits by CRIS study participants (intervention and control groups). The probability level used to determine statistical significance was p=.05.

Results

Community Health Advisors (CHAs)

Of the 30 African American women who completed the CHA training, 27 responded to the CHA exit survey. Respondents ranged in age from 42 to 71 years, with a mean age of 58.3. The typical CHA was married (55.6%) and currently employed (55.6%). Some had high-school education or less (33.3%), others technical school (44.4%) or college education (22.2%). Of the 25 CHAs who responded to the question about annual household income, 44% had household incomes below $25,000, 40% between $25,000 and $49,000, and 16% more than $50,000. The majority of CHAs reported that they had an annual Pap test (90%) or one every other year (7%); most responded that they had a Pap test because they wanted to detect cancer early (74%) or to make sure they did not have cancer (63%) or an infection (44%).

Training of CHAs

When asked about the training, nearly all CHAs agreed that it helped them understand how the ALTS clinical trial was related to the CRIS research study (100%); understand their role as a CHA research partner (100%); understand better the Pap test procedure (93%); provided them with information they used to answer questions about the Pap test procedure (100%); understand the need for an intervention and control group in a study (93%); understand better how research is planned and conducted (85%); understand the need for informed consent when conducting research (96%); understand the importance of their role as a CHA to the ALTS trial (100%); and that role-playing helped them understand and practice what might work best when contacting ALTS participants (100%).

CHAs as Research Partners and Types of Contacts Made

Most CHAs felt that their case load was about right (85%), although several thought it was too low (15%). When CHAs called their ALTS cases, the most common questions they received were about women’s health issues (78%), pelvic exams (74%), and research (74%).

Table 1 presents the CHAs’ self-reported adherence to the CRIS protocol regarding contacts they made with their ALTS cases. For the most part, CHAs did what they were asked. They made the initial call to introduce themselves to their ALTS cases and made reminder calls prior to appointments. Although they were not as compliant with follow-up calls after the appointment, many did call to make sure their cases rescheduled missed appointments. Nearly all the CHAs kept a record of each time they contacted their ALTS cases. Most sent a calendar with stickers as reminders of future appointments, and many sent cards for special occasions. They were less likely to send a card after the initial introductory call or make calls not associated with scheduled visits. When asked what motivated them to call their ALTS participants, most CHAs listed incentives, encouragement and support from other CHAs, and monthly telephone reminder from CRIS staff (Table 2).

Table 1.

Contacts Made by CHAs with ALTS Participants (N=27)

Contacts made	Always %	Most of the time %	Seldom %	Never %
Initial phone call to introduce myself	52	44	4	0
Sent a card after initial introductory call	15	37	37	11
Reminder call 2 days before appointment	31	69	0	0
Follow-up call after appointment	31	23	46	0
Called to make sure they had rescheduled missed appointment	37	44	15	4
Calls not associated with scheduled visits	12	15	35	38
Sent cards for special occasions	20	52	24	4
Sent calendar with stickers as reminders	41	52	4	3
Kept record of each contact I made	41	52	7	0
Forgot to call at appropriate times	4	4	50	42
Called CRIS staff if I had problems with a case	20	36	32	12

Table 2.

Factors Motivating CHAs to Call Their ALTS Cases (N=27)

Motivating Factors	%
Incentives for making calls	77
Encouragement and support of other CHAs	69
Monthly telephone calls from CRIS staff	65
Reminder cards in the mail to call ALTS cases	61
Encouragement and support of CRIS staff	63
A personal desire to help others	48
Monthly CRIS meetings	41
Initial training received as a CHA	33

Incentives for CHAs

CHAs received monetary incentives for the work they performed with the CRIS project. Most were satisfied with the $50 gift card for completing the training (85%), the $10 for attending monthly meetings (96%), the $5 incentive for effort (88.5%), and the $5 incentive for each ALTS patient who successfully completed her visit (88%). Only two CHAs reported that they would have preferred to be employed by the university instead of volunteering on the project.

Retention and Adherence of ALTS Participants

A total of 632 ALTS trial participants consented to be in the CRIS study. On average, ALTS trial participants were 27 years of age. They were predominantly African American, and fewer than half reported that they had completed more than a high-school education. There was no difference in the demographic characteristics of the intervention and the controls groups (Table 3).

Table 3.

Demographic Characteristics of ALTS Participants in the CRIS Study

	Overall N=632		CHA Intervention N=359		Control N=273
	n	%	n	%	n	%
African American	574	90.8	327	91.1	247	90.5
White	56	8.9	30	8.3	26	9.5
Other	2	0.2	2	0.3	0	0.0
High school or less	333	53.0	194	54.0	139	51.0
More than high school	298	47.0	164	46.0	134	49.0
Mean age in years	27.12		27.00		27.28

It was hypothesized that women in the CHA intervention group would report higher adherence rates than women in the control group, who received the standard ALTS trial protocol. Adherence rates were measured based on attendance of the 1,492 remaining ALTS trial follow-up clinic visits (60% of all ALTS visits). Table 4 presents the cumulative adherence rates with ALTS follow-up visits by group. Chi-square analysis indicated that the proportion of follow-up visits was significantly lower in the control group than in the CHA intervention group. Of the 632 CRIS study participants, 251 in the CHA intervention group kept all their expected ALTS clinic appointments (80%), compared to 131 in the control group (65%) (p < .0001).

Table 4.

Cumulative Adherence Rate to ALTS Follow-up Visits by CRIS Study Group

Attendance	Control		CHA Intervention
	n	%	n	%
Visit shows	449	65	642	80
Visit no shows	238	35	163	20
Total appointments	687		805

P < .0001

Further analysis of adherence suggested that the intervention was most effective with repeated contact over the course of the ALTS trial. Of the 129 women with only one remaining visit, 99% of the intervention group (n=97) versus 97% of the control group (n=32) completed the visit (p = .44). Of the 258 women who had two visits remaining, 73% of the intervention group (n=124) compared to 53% of the control group (n=134) made both visits (p =.001). Of the 133 women with three visits remaining, 54% of the intervention group (n=92) compared to 56% of the control group (n=41) completed all three visits (p = .85). Of the 112 women with four visits remaining, 30% of the intervention group (n=46) compared to 8% of the control group (n=66) completed all four visits (p = .002).

Discussion

Results of the CRIS study strongly indicate that the recruitment and training of CHAs to serve as research partners, albeit labor-intensive, can yield significant results in the retention and adherence of participants in clinical trials. Overall, the study produced two significant findings. First, CHAs aided in the retention and of clinical trial participants and their adherence with scheduled visits, and supported the hypothesis that CRIS study intervention participants would comply with scheduled follow-up visits at significantly higher rates than their counterparts in the control group. It is important to note that all ALTS participants, including the control group, received strategies to increase adherence to protocols, such reminder calls and cards from clinic nurses, transportation assistance, as well as incentives to keep the four scheduled appointments. This result is very credible because the data are based on clinic appointment records, not on participant self-report.

The second finding demonstrates that volunteer CHAs can be trained and can serve as research partners. Previous research indicates that African American women seek social support and advice regarding health from women they deem trustworthy and knowledgeable, a basic premise of the Community Health Advisor (CHA) model. Unlike media or written solicitation, the CHA model relies on the inherent strengths of the community, which include informal social networks and grassroots health education and promotion. The model invites respected members of the community to participate as partners in research.^8,10 Furthermore, the model is operationally feasible and can be included in the outreach phase of any clinical trial.

Limitations

Because the ALTS trial was already underway (with 40% of all visits already made) when the CRIS ancillary study was approved for implementation, the number of visits that could be impacted by the CHAs was reduced. This was not considered to be problematic because the purpose of the CRIS project was to determine the effectiveness of utilizing natural helpers to improve the retention of African American and low-income women and their adherence to scheduled appointments.

CHAs followed only those women who consented to be in the CRIS project (N=632). The CRIS staff had no way of determining if the CRIS participants were different from the other ALTS participants, which includes a residual potential for selection bias.

Finally, although ALTS participants were recruited from health department clinics, which typically provide care to low-income people, the trial did not collect information on their income. Thus, in terms of income distribution, direct comparison between CHAs and study participants is not possible.

Conclusion

Study outcomes indicated that CHAs can take on roles of research staff and perform such tasks with increased effectiveness. The CHAs were able to market, mentor, motivate, and monitor study participants more effectively than the standard practices. Because they were part of the community, CHAs were connected in the community and were trusted by those “hard to reach.” The CHAs successfully marketed health education information to the community as long as they had a simple standard message to relay and were supported by the research staff. CHAs encouraged healthy behaviors by modeling those behaviors themselves. They received regular Pap smears and thus were able to discuss the importance of the screenings with the younger ALTS women. CHAs motivated the ALTS women by telephone calls and cards. Although the CHAs were motivated by incentives, encouragement, and support from the research staff, they had a personal desire to help other women. The CHAs monitored the ALTS participants by recording their activities. They generally followed the protocol and were not reticent to communicate to the research team concerns about contacts they felt were not important or necessary. The CHAs were more likely to make the initial phone call, the reminder calls, and calls to reschedule appointments. CHAs complied with sending out calendars and reminders and keeping a record of each contact made. In this team effort, CHAs worked to adhere to their roles, and the ALTS participants worked to adhere to the ALTS study protocols. Our findings provide an insight into the collaborative research effort necessary to improve clinical trial retention and adherence among underrepresented populations.

Implications for Practice

The findings have implications for the retention of minorities and underrepresented women in clinical trials and their adherence to trial protocols. Three lessons were learned in the evaluation of the effectiveness of using CHAs as research partners. First, although the CHAs were able to monitor adherence to protocols, they preferred making calls associated with scheduled visits and did not like to verify kept appointments. It is important to acknowledge that the CHAs have insights into the timing and appropriateness of calls; they preferred making reminder calls than monitoring of protocol adherence. Acknowledging these preferences can help the project team streamline and simplify the process associated with monitoring adherence.

Second, the CHAs helped the research team understand the importance of being a volunteer rather than an employee of the research institution. As volunteers, CHAs perceived themselves primarily as helping the participants and only secondarily as helping the investigative team. This was important to the CHAs because it supported them in their role as natural helpers. However, CHAs considered monetary incentives important for following the CRIS protocols.

Last, the CHAs role as mentors and role models seemed to affect the adherence of ALTS participants. The CHAs were older than the ALTS participants but were from the same community and understood their problems and concerns. It is therefore important to recruit, train, and support CHAs who have similar values and beliefs as the research participants. Reminder calls and even monetary incentives do not replace the role of CHAs who can understand and empathize with the research participants.