Design of a randomized controlled trial using a community-academic partnership to improve medication adherence: MI-CARE Study protocol and implementation

Abstract

The prevalence of hypertension has been observed to be disproportionately higher among African American and other minoritized populations compared with White Americans. Persistent compounding factors such as sociocultural barriers and medication nonadherence among minoritized populations exacerbate hypertension and cardiovascular disease disparities. The “My Interprofessional Care team for Adherence and Research Engagement (MI-CARE)” intervention focused on collaborative, culturally responsive tailoring of healthcare provision to improve medication adherence. A social-ecological framework was used to design the MI-CARE intervention to address patient needs at multiple levels of influence on medication adherence and delivered by a pharmacist-community health worker team. An intention-to-treat randomized controlled trial (RCT) at a federally qualified health center tested the effectiveness of this team-based, comprehensive care intervention through 6 months post-randomization, compared with usual care, in two minoritized cultural groups: African American and Latino/a. This study assessed pre- to post-intervention changes in medication adherence (proximal outcome), blood pressure (distal outcome), and other comorbid health variables (distal outcomes). To do so, we collected self-reported medication adherence, manual pill count, and blood pressure data at baseline and at 2.5 months and 6 months post-randomization. We also aimed to identify factors associated with intervention effectiveness, including tailored intervention features, intervention dose, medication beliefs, barriers to adherence, and health literacy. Throughout the study, the interprofessional research team’s adaptability and responsiveness supported the implementation of this RCT within a real-world clinical setting. Insights from advances and challenges in clinical operations informed adaptations that strengthened healthcare delivery while maintaining research integrity.

Introduction

Hypertension, widely prevalent in the United States (U.S.), fuels racial/ethnic disparities in cardiovascular disease,¹ with 44% of African American patients diagnosed with hypertension, the highest rate in the world, compared to 28% of White Americans.² Both non-Hispanic Black and Hispanic adults have a higher prevalence of poorly controlled hypertension compared to non-Hispanic White adults.² Additionally, previous research showed that fewer Hispanic patients with hypertension were prescribed medications for it,³ while African American deaths from preventable heart disease and stroke were twice those of White Americans.^4,5 Racial/ethnic disparities in medication adherence amplify the effects of hypertension inequity. Historically, minoritized and immigrant populations have had an increased risk of medication nonadherence due to barriers including systemic racism, medical mistrust, limited healthcare access, and inadequate multilingual healthcare resources.^6-9 Other factors fueling nonadherence included memory-related challenges and culturally variable health and medication beliefs.^10,11

Pharmacist interventions, comprising patient education, provider guidance, and medication management, have demonstrated effectiveness in improving adherence, blood pressure (BP) levels, other health outcomes, and patient satisfaction.^12,13 Systematic reviews and meta-analyses reported that interprofessional teams involving a pharmacist had more favorable effects on patient-centered health outcomes, including medication adherence, than healthcare provided without a pharmacist for adults in the U.S.^14,15 A 2023 meta-analysis assessed the effects of Medication Therapy Management (MTM) on hypertension and found an overall mean reduction in BP across the 27 included studies.¹⁶ Yet limited research has evaluated the effects of pharmacist interventions in federally qualified health centers (FQHCs) serving patients from minoritized groups. A retrospective cohort analysis of Hispanic adults with diabetes treated at an FQHC in California found that after receiving 6-month comprehensive medication management (similar to MTM), 40% of participants reached the hemoglobin A1c (HbA1c) goal of <8%.¹⁷ Overall, however, there is little research reporting the extent to which pharmacists' interventions are delivered and effective among minoritized populations with hypertension.^18,19

FQHCs prioritize hiring from the communities served. Community Health Workers (CHWs), by design and definition, comprise a workforce with longstanding ties and shared lived experience with the community they serve and help to bridge and build trust for patients within their healthcare setting.^20,21 CHWs play a unique role as liaisons among healthcare systems, community resources, and patients while providing health education, access to social resources, and navigation support.^8,22,23 A small body of research has examined pharmacist-CHW partnerships for patients with diabetes or hypertension.^8,24,25 While these studies reveal encouraging improvements in health outcomes among diverse patients, several limitations exist, including less-than-rigorous study designs, small sample sizes, self-report measures, inconsistent follow-up data across patients, and no comparison groups. Moreover, a literature search did not find studies that tested pharmacist-CHW partnerships to improve medication adherence and hypertension management within an FQHC setting, serving the most diverse primary care patient populations disproportionately burdened by health disparities.²⁶

While FQHC-led research innovation is uncommon, it is gaining more attention recently.^27-29 Community-engaged, community-responsive, and community-based participatory research (CBPR) have been evolving approaches with important theoretical and methodological frameworks to guide successful implementation.^30-32 Established over the past 25 years,³³ the latter approaches are frequently the result of community-academic partnerships with a dedicated commitment to building and sustaining community-engaged research. A randomized controlled trial (RCT) testing a pharmacist-CHW team care model, using FQHC-led research and community-responsive research principles, is needed to minimize bias, produce generalizable results for patients from minoritized groups, and reduce health disparities.

Objectives

Our study sought to fill this gap by designing and implementing a pharmacist-CHW team care model to improve medication adherence and test its effectiveness in an RCT at an FQHC with African American and Latino/a patients. We assessed pre- to post-intervention changes in medication adherence, BP, and comorbid health outcomes. We aimed to identify factors associated with intervention effectiveness, including tailored intervention features, medication beliefs, barriers to adherence, intervention dose, and health literacy, in the two cultural groups.

Overview of the MI-CARE Intervention

The “My Interprofessional Care team for Adherence and Research Engagement (MI-CARE)” intervention focused on collaborative, culturally responsive tailoring to improve medication adherence. The MI-CARE team consisted of an MTM pharmacist and a CHW, linguistically and culturally responsive to the FQHC’s patient population. As shown in Figure 1, the social ecological framework was used to design MI-CARE to address patient needs at multiple levels of influence.³⁴ The team identified and addressed structural, socio-cultural, clinical, and individual barriers to medication adherence. At the structural level, social determinants of health barriers were identified through a standardized social needs screening tool, documented in the electronic health record (EHR), and addressed by linking patients to health center and community resources. At the socio-cultural level, the pre-intervention self-report survey and CHW-patient dialog gathered relevant health and medication beliefs to guide the tailored intervention. At the clinical level, the MTM pharmacist optimized the hypertension medication regimen and provided adherence aids. Patient education was tailored for cultural responsiveness and to the individual's health literacy level.

Figure 1. MI-CARE: A Social Ecological Intervention Model.

This figure shows structural, socio-cultural, clinical, and individual levels of barriers to medication adherence that our intervention targets.

Table 1 presents the MI-CARE intervention visit components, including the responsibilities of the MTM pharmacist and CHW (interventions specifically designed for MI-CARE participants are indicated in red font). We used a wraparound visit format beginning with a 10-minute pre-visit session, in which the CHW built rapport with the participant, asked culturally specific questions, and set expectations for the team visit. During the 20-minute MI-CARE team visits, the pharmacist provided MTM services focused on hypertension and medication adherence management, with the CHW serving as a cultural broker. After the team identified medication adherence barriers and optimized participants’ medication regimens, they offered adherence aids, such as a medication list and blister packs or a pillbox, when appropriate. Following each PharmD-CHW team visit, the CHW met with the participant to reinforce the pharmacist’s education on medication management and adherence aids, address any questions, and provide support for implementing adherence strategies. Between study visits, the CHWs provided ongoing care coordination, social resource navigation, and socio-cultural support.

Table 1.

MI-CARE PharmD-CHW Team Intervention Description

MI-CARE Wraparound Intervention Visit Components and Examples (Red font indicates MI-CARE specific interventions)
CHW Pre-Visit (10mins)	PharmD-CHW Team Visit (20mins) - Use CHW Interview	CHW Post-Visit (10mins) - Use PharmD-CHW Team Intervention
Elicit medication beliefs Concerns Necessity Cultural practices Beliefs/practices for hypertension Elicit any complementary modalities for hypertension and other chronic conditions	Beliefs intervention Address medication concerns Amplify medication necessity Hypertension therapy optimization Review medications and lifestyle Select evidence-based medications and doses Implement lifestyle modifications Monitor clinic/home blood pressure Assess vitals and laboratory results Screen for medication-related problems, including allergies, side effects, and drug interactions Prevent cardiovascular disease onset or progression Adherence intervention Educate on hypertension and medications Tie dosing times to daily routine/activity (e.g., take meds at breakfast) Tie med storage to med taking per daily activity (e.g., keep on breakfast table) Adherence aids Knowledge: medication list with indications and directions for use Organization: blister packs or pillbox Reminder: blister packs, pillbox, alarm, family/social support Monitoring: blister packs or pillbox Access: medication synchronization	Reinforce adherence education Implement goal setting Reinforce info about how to use blister packs or pillbox Gather participant questions for follow-up Cultural factors Implement family/social support Address SDOH barriers to medication adherence (e.g., link to health center or community resources, provide home visits) Assist with structural barriers (e.g., insurance coverage, transportation to health center or pharmacy)

MI-CARE = My Interprofessional team Care for Adherence and Research Engagement, SDOH = Social Determinants of Health

Follow-up visits were scheduled as clinically indicated, based on the participant’s hypertension control, complexity of medication management, and further education needs. To ensure participants were equipped to self-manage their condition and medications independently, a booster session was offered to the MI-CARE group participants 8 weeks after the initial intervention session. During this visit, the pharmacist-CHW team reinforced patient education offered during the intervention session(s) and addressed participants’ questions.

Usual Care Condition

The participants in the comparison “Usual Care” group, planned as a waitlist control, received medical care as usual at the FQHC. This group received the Usual Care intervention while on a waitlist for the opportunity to receive the MI-CARE intervention after participation in the study, if the study findings revealed positive benefits and no adverse effects of MI-CARE. The FQHC routinely offers comprehensive services beyond general primary care to be culturally and linguistically responsive to the needs of the population served. These services include a fully functional Wellness Center offering a Self-Measured BP program and chronic disease self-management support for conditions such as diabetes and hypertension. Therefore, participants randomized to Usual Care were offered and could have received various services potentially effective for hypertension management.

Research Design and Methods

We conducted an intention-to-treat RCT with African American and Latino/a patients in collaboration with an FQHC research/practice partner serving diverse patients with low income in Massachusetts. The main study hypothesis was that the MI-CARE intervention would improve medication adherence through 6 months post-randomization to a greater extent than Usual Care.

Ethics

This study was reviewed and approved by the Institutional Review Board (IRB) at the University of Arizona (IRB ID: 2106945516). A two-member safety monitoring committee was established to monitor and guide safety considerations for this trial, to which the research team reported monthly study progress and any safety events. The study was also reviewed internally and approved by the FQHC’s consumer-led Board of Directors. The waitlist control design was developed specifically to make the MI-CARE intervention accessible to participants in the Usual Care condition, supporting a more equitable approach between study groups. The enhanced Usual Care condition had continued access to FQHC-wide social care resources that further supported this balance between what study participants and non-participants received. As the study progressed, the FQHC began rolling out blister packs as a standard of care medication adherence resource for all patients and maintained non-study patients' access to CHW and MTM/clinical pharmacist services. This trial was registered with ClinicalTrials.gov (Identifier: NCT05470439) in July 2022 before the study enrollment.

Research Team

The present study resulted from a long-standing partnership between academia and an FQHC with members who have collaborated for up to 20 years. Our team of interdisciplinary investigators has complementary knowledge and expertise in research and practice. We are proficient in MTM, hypertension management, medication adherence, health literacy, public health, social psychology, research design, RCT protocol development and implementation, data analysis, FQHC operations, pharmacy services, and CHW workforce and navigation. We also had the insights and support of FQHC leadership, providers, and staff, as well as the 51% consumer-led board, when developing the MI-CARE intervention.

Setting

Though uncommon in most FQHCs, the partnering FQHC site has developed a fully integrated and sustained research program, and its scholarship includes the development of an FQHC-led research approach that centers on real-world setting research methods.^28,29 Being responsive to and adapting research methods and implementation to align with ever-evolving clinical, operational, and policy realities while maintaining scientific rigor was a key component of conducting this RCT. The partner FQHC site serves a substantial proportion of patients from minoritized backgrounds across three clinical sites. It is the largest refugee health assessment site in the state and a National Committee for Quality Assurance patient-centered medical home that provides 98,000 visits annually to 22,000 patients in over 70 languages. Approximately 36% of the patient population is Latino/a, and 18% is African American. Three-quarters of the patients live at or below the U.S. federal poverty line, making structural barriers to health care common among FQHC patients.²⁶ The health center has an in-house pharmacy and uses OCHIN/EPIC, an EHR system developed specifically for FQHCs.

Sample

The study participants were African American and Latino/a adult (≥18 years of age) patients at the study FQHC who spoke English or Spanish and were able to provide informed consent. They had a diagnosis of hypertension with at least one daily oral medication taken chronically to control BP, polypharmacy (≥5 chronic medications), and less than adequate hypertension medication adherence (<85% or >120% at study eligibility screening). Polypharmacy included chronic medications that are prescription drugs and over-the-counter agents. Potential participants were excluded if they were already receiving MTM services for hypertension at enrollment. Participants received monetary incentives (totaling $110) for their time and participation in the 6-month study, and bus vouchers for transportation.

Study Procedures

Figure 2 provides an overview of the MI-CARE study design. Participants were primarily recruited by CHWs, using an EHR-based registry of patients with hypertension. Telephone calls were made to those who met the study criteria based on their EHR data. During initial contact with potential participants, CHWs, trained in Human Subjects Protections, stated that they were recruiting for a study to learn how best to help patients take their blood pressure medications; the study involves multiple sessions for data collection, and participants would receive monetary incentives for their participation. With those who expressed interest in participating, the CHWs screened for study eligibility. Other recruitment strategies included posting patient-facing and provider-facing study flyers, requesting provider referrals, and CHWs engaging clinical team members, including medical assistants. After eligibility was determined through self-report data, prescription information, and manual pill counts, the recruiting CHW discussed detailed study information with potential participants. The same CHW reviewed the consent form with potential participants in their preferred language, answered questions, obtained consent, and provided a copy of the consent form. Study participants completed the baseline survey and were randomized to either the MI-CARE or Usual Care group, stratified by cultural group.

Figure 2. MI-CARE Study Design.

This figure illustrates the overview of the trial design from eligibility to assessment.

Power and Randomization

The power analysis was based on a previous research study³⁰ that measured medication adherence of the population source and preliminary data analyses from a pilot project at the same partner FQHC, which indicated a small-sized effect of MTM intervention on BP. We also considered potential effect sizes for improvement in medication adherence and BP based on existing data of community health center clients when determining expected effect size. We conservatively expected an effect size of f2 of .03 to ensure we have adequate analytical power to detect existing effects and avoid Type II errors. The main hypothesized mediational model is a simple model with only one predictor, one mediator, one outcome, and no latent variables, which suggested the need for a relatively small sample size. We conducted a power analysis for the most complicated proposed multiple regression with the lowest power (i.e., the multiple regression including all of the factors of interest in the model as predictors). We used G*Power 3.1.9.4^35,36 to estimate power using a criterion alpha of .05, a conservatively expected small effect size of f² of .03, and 5 predictors in the model. According to G*Power 3.1.9.4, a sample of 209 has power of 0.80 to detect an effect of a predictor on the outcome (e.g., BP) that is of .03 size. Thus, a sample size of 209 provides sufficient power to conduct the proposed analyses, and the current achievements in study enrollment indicate that we can expect to achieve a sample size of at least 209.

The randomization scheme was based on the planned 209 sample size (at least 104 in each cultural group, African American and Latino/a), with around 52 participants from each cultural group assigned to either the MI-CARE or Usual Care group. We used a random number generator to produce a list of randomized group assignments for each subsample with minor adjustments near the end of recruitment to achieve approximately equal intervention group sizes within each cultural group.

Data Collection Overview

Table 2 illustrates the study timeline and data collection. To assess pre- to post-intervention changes in medication adherence (proximal outcome), BP (distal outcome), and other comorbid health distal outcomes, we collected self-report survey data, pill count, and clinical data for both the MI-CARE and Usual Care groups at multiple time points. In total, there were five data collection sessions conducted: 1) study eligibility screening, 2) baseline (2-4 weeks later) that ended with randomization into the study group, 3) 2 weeks post-randomization, 4) 10 weeks (2.5 months) post-randomization, and 5) 24 weeks (6 months) post-randomization.

Table 2.

Study Timeline

Intervention and Data Collection Timeline in Weeks Elapsed since Enrollment (Plain text items = both MI-CARE and Usual Care groups; Bold items = MI-CARE group only)
→ Time elapsed in weeks	Week 0 (Eligibility)	Week 2 (Baseline)	Week 4	Weeks 5-8a	Week 12 (2.5 mo. post- randomization)	Week 26 (6 mo. post- randomization)
Activity	Initial Screening, Enrollment	Randomization	Initial MI-CARE Needs assessment	On-going MI-CARE Intervention	MI-CARE Booster
Pill count	Pill count 1A	Pill count 1B			Pill count 2A	Pill count 2B
Other data collection		Baseline Survey,b BP	BP		BP	Follow-up Survey,b BP

^aIntervention follow-up was scheduled as needed during this time period.

^bBaseline and Follow-up Surveys contain all interview items, including the subjective adherence scale.

We collected pill count data at four timepoints. Pre-intervention adherence was assessed using a validated self-report scale at baseline and pill count data collected during screening and baseline sessions, which reflect adherence during the time between these two sessions. Post-intervention medication adherence was assessed using the validated self-report scale at 6 months post-randomization and pill count data collected at 2.5 months and 6 months post-randomization sessions, reflecting adherence between these two sessions.

At baseline and 6 months post-randomization, we collected self-report survey data and BP measures. We collected additional BP measures at 2 weeks and 2.5 months post-randomization sessions. Self-report surveys assessed participants’ medication adherence, medication beliefs, medication management, food insecurity, social support, barriers to adherence, and health literacy and numeracy. Surveys were administered orally by bilingual/bicultural CHW data collectors in the participant’s preferred language, with responses recorded directly into an online survey platform. For other health outcomes (e.g., HbA1c), we leveraged existing data collected by providers during non-study-related clinical care visits with study participants. After all participants have completed their 6-month post-randomization data collection session, these study-related health outcome data will be used for this study, according to the “release of information” permission participants provided.

The study CHWs served dual roles as members of the intervention team and as data collectors. To establish rapport between the CHW and participants, baseline data was collected by the CHW who would deliver the intervention to that participant. To maintain blinding post-randomization, follow-up data collection was performed by a different study CHW who was blinded to randomized group assignment and was not involved in that participant’s intervention delivery. Structured protocols and training guided CHWs in their intervention and data collection activities across all aspects of the study.

Fidelity Monitoring

To support ongoing implementation fidelity of the RCT, the investigator team met biweekly, and the full study team met biweekly to review protocol checks, to discuss emerging issues from the field, and determine any needed adjustments. Feedback from CHWs and MTM pharmacists, gathered through these regular meetings, was incorporated to help refine study processes. The investigator team retained responsibility for determining, documenting, and implementing any resulting changes, with decisions based on whether proposed changes upheld scientific rigor and study fidelity.

Even though the national healthcare staff shortages following the COVID-19 pandemic affected our project and initial hiring, we experienced minimal staff turnover and the pharmacist-CHW team remained intact during the MI-CARE study intervention period. A process evaluation procedure was used throughout the RCT to monitor and document any changes to the study protocol and decisions made by the teams. Drawing on the seminal work by Fixsen and colleagues (2005),³⁷ our study used the core implementation components of their framework, which includes practitioner selection, pre-service and in-service training, ongoing coaching and supervision, practitioner performance evaluation, decision support data systems, facilitative administrative supports, and system interventions. The development of “SmartPhrases” (placed within the visit note template in the EHR) for MTM pharmacists and CHWs is an example of many support systems implemented to support ongoing fidelity monitoring.

Study Measures and Outcomes

Table 3 presents study measures. Primary outcome measures included medication adherence and BP. We assessed medication adherence pre- and post-intervention, objectively using manual pill counts and subjectively via MARS-5.³⁸ The MARS-5 assesses the frequency of medication nonadherence with a 5-point Likert scale. For the objective pill counts, the first count for each period established “the number of pills possessed by each participant” at that time. The second count established “the number of remaining pills” at that time to calculate the number of pills taken. During each pill count, the CHW counted all chronic daily-use oral medications. Based on these data, medication refills obtained during the measurement period, and prescribed medication dose and dosing schedule, we computed medication adherence during the measurement period: (the number of pills taken / the number of pills that should have been taken) x 100. We calculated the adherence rate for each eligible medication and the mean adherence rate across all eligible medications.^30,39,40 We obtained BP values twice in the seated position, 2 minutes apart, according to the 2017 ACC/AHA High Blood Pressure Clinical Practice Guideline procedures.⁵ We measured BP using a validated, automated upper-arm sphygmomanometer. The BP measure used for each relevant data collection session was the average of the two BP values.

Table 3.

Study Measures

Measure	Assessment	Available language
Primary outcome: Pill counts	Objective medication adherence measuring % of medications taken	N/A
Primary outcome: Medication Adherence Report Scale-5 (MARS-5)33	5-item self-reported medication adherence using a 5-point Likert scale	English and Spanish
Primary outcome: Blood pressure	Mean of 2 measures, using a validated, automatic upper-arm cuff device	N/A
Secondary outcomes: Weight, body mass index, hemoglobin A1c, lipid profile, and statin use	Health outcomes from electronic health records	N/A
Beliefs about Medicines Questionnaire (BMQ)37	18-item scale used to assess patients' beliefs about medications in general (BMQ-General, 8 items) and specific medications prescribed to them (BMQ-Specific, 10 items)	English and Spanish
Behavioral Risk Factor Surveillance System (BRFSS)38 - adapted (2-item)	Cost and transportation barriers to adherence in the last 30 days	English and Spanish
Medication Management Practice Questions	8 categorical questions developed by the investigator team to assess complex regimen, medication reminder tools, medication management, refill strategies, and provider/pharmacist communication	Developed in English and translated into Spanish
Short Assessment of Health Literacy-Spanish & English (SAHL-S&E)39	18-item scale measuring Health Literacy	English and Spanish
Short Numeracy Understanding in Medicine instrument (sNUMi)40	8-item scale measuring health numeracy	English and Spanish
Medical Outcomes Study (MOS)-Social Support Survey,41 abbreviated	Perceived emotional and pragmatic social support	English and Spanish
Patient Health Questionnaire (PHQ-9)42	9-item questionnaire assessing depression	English and Spanish
Social Risk Factors, adapted from National Association of Community Health Centers’ Protocol for Responding to and Assessing Patients' Assets, Risks, and Experiences (NACHC PRAPARE)	A study site-specific tool in electronic health records, assessing social determinants of health, housing, transportation, and income	English and Spanish
Medical Mistrust Scale43	12-item scale used to assess experiences of discrimination in medical settings and trust in health care providers	English and Spanish
United States Department of Agriculture (USDA) Food Insecurity Scale44	6-item survey to assess the severity of food insecurity	English and Spanish
Patient Satisfaction Survey for Comprehensive Medication Management, adapted from Moon et al.45	10-item questionnaire to assess patient satisfaction with the pharmacy care received	Available in English and translated into Spanish
Electronic Health Record (EHR) Template for MI-CARE Protocol	Participant-level report tailored to measure intervention dose	N/A

Secondary outcome measures, to-be-obtained from the EHR, included participants’ weight, body mass index, HbA1c for participants having diabetes, and lipid panel, and use of statin medication for those having dyslipidemia. Self-reported surveys also collected demographic data (e.g., age, sex, marital status, income, education), English proficiency, Beliefs about Medicines,^41,42 cost and transportation barriers,⁴³ medication management practices, health literacy and numeracy,^44,45 social support,⁴⁶ depression screening,⁴⁷ medical mistrust,⁴⁸ food insecurity,⁴⁹ and patient satisfaction.⁵⁰

Results

We are 4.5 years into this 5-year study and have completed enrollment activities with 212 participants enrolled. Figure 3 displays participant flow from attempted contact of potential participants to study enrollment. Of 1,339 potential participants contacted, only 303 (16.5%) were screened; the remaining participants declined to participate in healthcare research, were determined ineligible, or were lost to follow-up before screening could be completed. However, of 303 screened, 235 (77.6%) were eligible, and 212 (70.0%) were enrolled in the study (see Figure 3). The total enrollment slightly exceeded the target sample of 209, which was determined by power analysis to achieve a power of .80. Thirty-three African American and 74 Latino/a participants were randomized to the MI-CARE group, and 34 African American and 71 Latino/a participants to the Usual Care group. We closed study enrollment in October 2025, with follow-up data collection planned for completion in March 2026.

Figure 3. MI-CARE Participant Flow to Enrollment.

This figure depicts the participant flow of recruitment, screening, and enrollment.

Lessons learned

Conducting an RCT in an FQHC environment brought unique opportunities and challenges. The research team had to be responsive to evolving FQHC clinical practices with logistical staffing and workflows that required a flexible, real-time approach to adapt to changes without compromising research integrity. Balancing intervention fidelity with adaptability was instrumental throughout the RCT duration. The research team employed positive adaptations in study implementation to facilitate real-time problem-solving by leaning into the question, “How and what is going to change?”^51,52 rather than “Should we permit reinvention?”⁵³ The iterative process evaluation guided our adaptations throughout the study, yielding several key lessons. First, the overlap between the FQHC partner’s standard MTM services and the MI-CARE intervention required careful differentiation. We adapted the intervention delivery to emphasize the culturally tailored components and the structured pharmacist-CHW team care model, distinguishing it from routine pharmacist-only MTM service with potential referral to CHWs. The integration of CHWs into the care team to provide tailored support was critical, as CHWs are often better positioned to build trust and rapport with patients from minoritized groups compared to pharmacists alone. CHW recommendations, developed by soliciting participant feedback, also contributed to an effective study adaptation.

Second, staffing shortages presented challenges to recruitment and data collection, particularly the need for culturally and linguistically responsive CHW interventionists and data collectors. The CHWs played a critical role in patient engagement to combat reluctance to participate in healthcare research that required multiple visits to the health center over 6 months and involved pill counts and EHR data collection, which may be perceived as invasive. We developed a CHW dual-role model in which CHWs served as both intervention providers and data collectors, addressing staffing constraints while creating a CHW workforce development opportunity. A focused publication regarding this dual role is in progress to inform future studies and FQHC practices.

Third, the dynamic FQHC environment necessitated study team members with implementation science skills to lead, including translating the proposed study protocol into an actionable set of procedures. This pragmatic approach, informed by ongoing process evaluation, optimized implementation by adapting to the real-world clinical setting and workflows, and by continuous quality improvement, while preserving scientific rigor and research integrity. Detailed methodological and implementation considerations for conducting community-responsive, adaptive, and feasible RCTs in FQHC settings will be disseminated through a forthcoming publication.

Planned analyses

We will perform all analyses using the intention-to-treat principle, including all randomized participants in analyses. For lost to follow-up, we will duplicate their baseline or latest data at the post-randomization timepoints to indicate no change over time. We utilized this method instead of others in order to facilitate appropriate translation to practice by avoiding overstating the effectiveness of the intervention and unnecessary complication of statistical procedures. In addition to adjusting for baseline adherence and BP as appropriate, we will include covariates such as regimen complexity and depression in our statistical models to address other potential alternative explanations of differences in outcomes by intervention group. To test our hypothetical mediational model (Figure 4) that MI-CARE would improve medication adherence (proximal outcome), which in turn will improve BP and other secondary health outcomes (distal outcomes) to a greater extent than Usual Care, we will use the Baron and Kenny approach⁵⁴ that includes Sobel’s test⁵⁵ of the indirect effect. We will first assess mediating effects by examining 1) the overall association between intervention and the outcome variable (e.g., BP), 2) the overall association between intervention and the mediator (medication adherence), 3) the association between adherence and the outcome variable, and 4) the association between intervention and the outcome variable while controlling for adherence. Second, we will use Sobel’s test⁵⁵ to determine if the mediation is statistically reliable. We will also conduct analyses to identify factors associated with intervention effectiveness, including a) tailored features of intervention, b) intervention dose, c) barriers to adherence, d) medication beliefs, e) health literacy, and f) cultural group.

Figure 4. Hypothetical Mediational Model.

This figure shows the mediational model used in the planned analysis for this study.

Discussion

To our knowledge, this is the first RCT testing a pharmacist-CHW team intervention to improve medication adherence for African American and Latino/a patients with low income and a hypertension diagnosis conducted in an FQHC. MI-CARE, a novel interprofessional community-responsive intervention, has the potential to address structural, socio-cultural, clinical, and individual barriers to medication adherence for study participants. The pharmacist-CHW team collaboratively addressed barriers to adherence, such as health insurance co-pays, transportation, culturally variable medication beliefs and concerns, suboptimal medication regimen for hypertension, and low health literacy. Health-related outcomes, including medication adherence and BP, have been measured.

This RCT was designed to uphold scientific rigor. Its implementation within a dynamic FQHC environment underscored the critical need for and the essential role of adaptability and real-world responsiveness in community-based FQHC-led research. FQHCs possess inherent strengths for conducting research, including established community trust, diverse multilingual staff, and a deep understanding of the patient populations they serve. Furthermore, FQHCs are leaders in quality improvement-based evidence for best practices. The dynamic setting, responsive to evolving policy requirements, community needs, and rapid integration of practice innovations, requires flexible adaptive approaches to conducting an RCT. Real-world responsiveness is not a compromise to research rigor but rather an essential function that enables feasible implementation in settings where clinical workflows, staffing realities (e.g., nationwide healthcare workforce shortages), and organizational priorities continuously evolve. This study demonstrates that rigorous study designs and real-world responsiveness can coalesce when guided by CBPR underpinnings, iterative process evaluation, and recognition of FQHC organizational assets and challenges as key components to study execution.

This innovative RCT advances the literature focused on medication adherence in people with hypertension who are from minoritized groups. While prior studies have examined barriers to medication adherence among minoritized populations with hypertension and diabetes, medication adherence as a primary outcome for hypertension control remains understudied in these communities.^8,56 An existing RCT examined the comparative impact of three interventions by MTM pharmacists and/or CHWs on medication-related outcomes for Cambodian patients living in Connecticut, Rhode Island, or Massachusetts.²⁴ The findings indicated that participants assigned to a CHW education curriculum group and the combined CHW curriculum and MTM pharmacist visits group had improved diabetes knowledge, health status, nutrition, and social determinants of health.²⁵ The MI-CARE intervention also combined culturally responsive patient education by CHWs and MTM pharmacists, but with a focus on medication adherence, identifying and addressing adherence barriers through culturally responsive education and adherence aids. In contrast to existing studies with less rigorous designs or focused on other chronic conditions, the MI-CARE model was designed for African American and Latino/a patients at an FQHC with hypertension, polypharmacy, and nonadherence at baseline. The innovative, interprofessional team-based, FQHC-led RCT builds on a community-responsive approach and extensive prior research on medication beliefs and practices among minoritized patients. The interprofessional intervention by MTM pharmacists and CHWs advanced the conventional MTM session with collaborative team-based care, simultaneously addressing culturally variable medication beliefs and structural and clinical barriers to adherence.

Limitations

This RCT could not include all culturally minoritized populations served by the study FQHC, so there was a potential for selection bias. There was also potential for non-response and contact bias as FQHC patient populations frequently experience changes to their addresses and telephone numbers, and broader digital access barriers (e.g., limited or inconsistent phone/internet connectivity) that can further affect contact ability and follow-up. Additionally, measurement awareness from participation in pill counts may have influenced adherence behavior. We had to limit study enrollment to African American and Latino/a patients, though we had originally aimed to include Vietnamese immigrant patients also. Although the included cultural groups do not represent all diverse populations with hypertension and nonadherence, our focus was on two groups of minoritized patients who are at high risk for nonadherence and poor health outcomes. If effective, the MI-CARE model with cultural tailoring developed here can be implemented by culturally responsive care teams for patients from diverse communities. Additionally, because bilingual CHWs served dual roles as interventionists and data collectors, complete blinding of data assessment was not possible. However, we achieved blinding of the intervention condition of post-randomization data collection activities.

Next steps

When data analyses for this RCT is completed, we plan to publish the main study outcomes paper. As an FQHC focused on translating research findings into practice, the FQHC intends to use the study findings to integrate Pharmacist-CHW team care into standard practice going forward. Subsequently, we plan to disseminate additional findings and implications from this interdisciplinary public health research project. These findings and implications may help to inform or guide FQHC-led real-world, practice-based research and contribute to knowledge generation and scientific evidence. Moreover, the academic community may learn strategies for community-based practice research co-created, led, and implemented by mutually respectful collaboration with FQHC-academic research and practice partners, specifically in translating proposed interventions into real-world clinical practice settings.

Highlights:

• “My Interprofessional Care team for Adherence and Research Engagement (MI-CARE)” was developed

• MI-CARE used a pharmacist-community health worker (CHW) team

• MI-CARE was designed for low-income African American and Latino/a patients

• MI-CARE involved culturally responsive care to improve medication adherence

• Adaptations were made to conduct the MI-CARE RCT in a real-world clinical setting