Pharmacists and Community Health Workers Improve Medication-Related Process Outcomes among Cambodian Americans with Depression and Risk for Diabetes

Abstract

Background:

Cambodian Americans have high rates of cardiometabolic and psychiatric disorders and disadvantaged social determinants of health (SDOH). These factors can make it challenging to resolve drug therapy problems (DTPs) and improve medication-related outcomes. This manuscript reports planned analyses from a randomized controlled trial in which participants were randomized to one of 3 treatment arms: 1) CHW-delivered lifestyle intervention called Eat, Walk, Sleep (EWS), 2) EWS plus pharmacist/CHW-delivered medication therapy management (EWS+MTM), or, 3) social services (SS: control).

Objectives:

We compared the three arms on changes in self-reported medication adherence, barriers, and beliefs. Within the EWS+MTM arm only, we assessed the impact of EWS+MTM on DTP resolution and examined predictors of DTP resolution.

Methods:

Cambodian Americans aged 35–75 at high risk for developing diabetes and meeting criteria for likely depression (n=188) were randomized (EWS, n=67; EWS+MTM, n=63; SS, n=50; control). For all participants, self-reported surveys were collected at baseline, 12-months and 15-months. DTPs were assessed on the same schedule but only for participants in the EWS+MTM.

Results:

All 3 groups reported a significant decrease in barriers to taking medications. Compared to the other arms, the EWS+MTM arm reported a decrease in forgetting to take medications at 15 months. In the EWS+MTM arm, mean DTPs per patient was 6.57 and 84% of DTPs were resolved. SDOH predictors of DTP resolution included years of education (OR=0.94, p=.016), ability to write English (OR=0.73, p=.015), difficulty communicating with provider (OR=1.39, p<.001), private insurance (OR=1.99, p=.030), disability (OR=0.51, p=.008), and years living under Pol Pot (OR=0.66, p=.045). Medication barriers at baseline significantly predicted DTP resolution (OR=0.79, p=.019) such that each additional barrier was associated with a 21% reduction (1–0.79) in the odds of having a resolution.

Conclusion:

CHWs can reduce medications barriers and help pharmacists reduce DTPs in disadvantaged populations.

Background

Medication-related process outcomes, including barriers to taking medication, beliefs about medications, and forgetting to take medications are important for individuals with depression and cardiometabolic risk.^1–3 Drug therapy problems are also an important medication-related process outcome. A drug therapy problem (DTP) is an unwanted incident related to medication therapy that adversely affects the desired goals of treatment.⁴ DTPs are the 14^th leading cause of morbidity and mortality worldwide⁵, and have been identified as a global public health concern by the World Health Organization.⁶ Therefore, testing innovative strategies to resolve DTPs and improve medication-related process outcomes is a public health priority.

DTPs do not occur equitably across populations.^7–10 Little is known about DTPs in the Cambodian population, but the evidence that does exist suggests that they have high rates¹¹ compared to other populations.^7,9,12–14 Moreover, they often present with clinical complexity including higher rates of cardiometabolic disease, depression, and post-traumatic stress disorder compared to the general U.S. population.^15,16 Cambodian Americans are also disadvantaged by interrelated social structures and economic systems. They have social determinants of health (SDOH) that are characterized by lack of opportunity and resources¹⁷ such as education, economic stability, and health care access and quality.^18,19 A 2014 Southeast Asian American Community Assessment found modal annual income was less than $10,000 and 27% of Cambodians had no formal schooling at all.¹⁹ Cambodian Americans have low access to medical and mental health care due to insurance status, cost, language, and transportation barriers.^20,21 As a refugee group, Cambodian Americans may also have culturally based health beliefs regarding medications as well as mistrust or lack of confidence in providers and pharmacists. Therapeutic alliance is therefore key for developing a good working relationship between a healthcare professional and patient.²² Cambodians who fled the genocidal Pol Pot regime are one of the longest established refugee groups in the U.S.²³ Their long period since resettlement can uniquely inform interventions to resolve DTPs and improve medication-related process outcomes in other complex minority, immigrant, and refugee groups.

Proven best practices for clinical treatment of these hard-to-reach populations is limited.²⁴ Innovative practice models are needed to address medication-related process outcomes in clinically complex patients with disadvantaged SDOH.²⁵ Medication therapy management (MTM) holds promise. MTM includes a broad range of services provided by pharmacists.^26–28 In MTM, medication-related decisions are coordinated collaboratively among the healthcare team and patient. MTM includes 5 core elements: medication therapy review, a personal medication record, a medication action plan, intervention/referral, and documentation and follow-up.²⁶ During the MTM process, pharmacists can improve health outcomes through identification and resolution of DTPs, by correcting misconceptions about medications, and by promoting adherence. Whereas strong evidence exists that the MTM delivery by pharmacists is effective ^27,29–33, there is limited evidence that MTM can achieve positive outcomes in complex populations.^34,35

Community health workers (CHWs) also hold promise for improving medication-related process outcomes. CHWs play a unique role by engaging medically underserved communities to overcome barriers to care.³⁶ As typically bilingual, bicultural and respected members of the community, CHWs bridge the community and healthcare establishment. Their unique skills can mitigate the deleterious effects of SDOH on health outcomes.

This study engaged CHWs to improve medication-related process outcomes. In one arm of this study, we integrated CHWs into MTM; CHW/pharmacist teams delivered MTM. CHWs optimized MTM by providing language translation, cultural interpretation, education, behavior modification, role modeling, bureaucracy navigation, and advocacy.

Objectives

Data reported here are from the Diabetes Risk Reduction through Eat, Walk, Sleep and Medication Management (DREAM) trial (clinicaltrials.gov identifier NCT02502929). DREAM was a randomized, controlled trial comparing three groups to reduce risk for developing type 2 diabetes among Cambodian Americans with depression. One group received a lifestyle intervention called Eat, Walk, Sleep (EWS). EWS is a cardiometabolic lifestyle curriculum that was created by and for Khmer people and was designed to be delivered by CHWs.^37,38 Another group received EWS plus Medication Therapy Management (EWS+MTM) delivered by pharmacist/CHW teams. The third group received social services (SS) only and was designed as an equitable control group for this underserved population. The primary biological outcomes of the DREAM study were HbA1c and insulin resistance; those outcomes will be reported in forthcoming publications. This report focuses on a priori planned analyses of medication-related process outcomes only, including DTP resolution in the EWS+MTM arm.

In our first set of analyses, we compared the three arms on changes in self-reported medication forgetting, barriers, and beliefs. We expected that EWS+MTM would result in more improvement than the other two arms. In our second set of analyses, we focused on EWS+MTM arm only. Within the EWS+MTM arm we: 1) assessed the impact of EWS+MTM on DTP resolution, 2) examined SDOH predictors of DTP resolution to determine which patients benefited most from the intervention, and, 3) summarized the patient-reported therapeutic alliance with their CHW and pharmacist. We expected that CHW/pharmacist teams would effectively resolve DTPs, that participants with greater disadvantage would have higher DTP resolution, and that participant-reported therapeutic alliance with the CHW and pharmacist and satisfaction with MTM would be high.

Methods

Participants and sampling

Inclusion criteria for the DREAM study were: 1) Cambodian or Cambodian-American; 2) aged 35–75; 3) Khmer speaking; 4) had lived in Cambodia during the Pol Pot regime; 5) currently living in Connecticut, Massachusetts, or Rhode Island (northeastern U.S.); 6) elevated risk for diabetes per the American Diabetes Association Risk Test modified for this population; 7) consumed meals by mouth; 8) ambulatory; and, 9) met criteria for depression by a) elevated depressive symptoms indicative of likely major depressive disorder on the Khmer language Hopkins Symptom Checklist on two occasions that were two weeks apart during a study screening and eligibility period, and/or, b) current antidepressant medication. Exclusion criteria were: pregnancy or planning pregnancy; major medical problems requiring intensive treatment; known type 2 diabetes; has tried to harm self in the past 2 years; has spent 3 or more days in a psychiatric hospital in the past 2 years; serious thinking or memory problems (e.g., from schizophrenia or dementia); seeing or hearing problems that would interfere with group sessions. Data were collected from March 2016 to October 2020.

Pharmacists and community health workers

MTM was conducted by four pharmacists who had experience in providing comprehensive MTM services and were either nationally certified in MTM or board-certified through Board of Pharmacy Specialties.

The five CHWs were bilingual, bicultural, born in Cambodia, and considered trusted in their communities. Training the CHWs in MTM included interactive presentations by our pharmacist investigator and study coordinator totaling 4 hours, shadowing experts complete the medication review form, and role play. Prior to working independently, CHWs were supervised for at least 10 medication review sessions with participants. Training materials included a 34-page toolkit for reading prescription and non-prescription labels, a 2-page chart of commonly prescribed medications, a 14-page manual for conducting medication reviews, and several online videos.

Procedures

Overview.

The UConn Health institutional review board approved all participant procedures on May 28, 2015 (IRB # 15–164S-3). Participants signed written informed consent forms in their preferred language. First, baseline study assessments were conducted. Surveys were administered verbally in the participant’s preferred language, anthropometrics were measured, and blood samples were collected for assay. Following baseline assessments, participants were randomized and began treatment. Later assessments occurred at posttreatment (12-months) and follow-up (15-months). Participants were compensated with $10 pharmacy gift cards for each of the three assessment components. Details of randomization, interventions, assessments, and the CONSORT diagram have been previously published.³⁹ A CONSORT diagram is in Appendix 1, available on JAPhA.org as supplemental content.

EWS and SS Interventions.

Participants assigned to SS were assessed for social service needs such as food and housing assistance and then CHWs provided linkages to community resources for any unmet needs over 15 months as requested; for example, translation of written medical materials or filing taxes. Participants assigned to EWS or to EWS+MTM received EWS (3 individual sessions and 24 group sessions with a CHW) over the course of 12 months and booster sessions between post-treatment and follow-up assessments. The EWS curriculum involved health education and social support to improve health behaviors, as well as a 1-hour group session led by a CHW that covered medication tips and dispelling common misunderstandings that are based on experiences in Cambodia (for example, viewing pharmacists as medication salespeople).

MTM Intervention.

Participants randomized to EWS+MTM, in addition to the EWS sessions, also received MTM. See Table 1 for MTM session flow. First participants met with the CHW in a face-to-face session to create a comprehensive medication record that included pharmacy names and telephone numbers, prescriber names, medication allergies and reaction, tobacco/alcohol/betel nut use, medication name/dose/directions, recency of visit with the prescriber, how the participant is actually taking the medication, and participant satisfaction with medication(s). The CHW collected information regarding all prescription and nonprescription medications, herbal products, traditional Khmer medicines, and dietary supplements.

Table 1:

Delivery of MTM in EWS+MTM arm (n=63)

Type of Activity	Description of Activity
Baseline study assessment: lab work, anthropometries, surveys
Introductory letter and signed release of information sent to provider
CHW-Participant Initial Meeting	CHW and participant meet in person to establish rapport, socialize to MTM, and complete Medication Review Form
Preparation for MTM Visit #1	Pharmacist reviews 1) Medication Review Form 2) demographics 3) baseline biometrics (e.g. HbA1c, body mass index) 4) baseline patient surveys (e.g. Beliefs about Medicine)
MTM Visit #1 (CHW and participant in same location)	Blood pressure and heart rate taken by CHW. Pharmacist remotely connected via secure videoconferencing: review medications, identify DTP number and type, determine DTP resolution status, develop and discuss a Medication Action Plan
Post MTM Visit #1	Pharmacist mails Medication Action Plan to participant and faxes Medication Report to provider
MTM Visit #2 and #3	Same as MTM Visit #1, and assess progress on Medication Action Plan and determine if new actions are indicated
Post MTM Visit #2 and #3	Pharmacist mails Medication Action Plan to participant and if indicated, faxes Medication Report to provider
12-month study assessment: lab work, anthropometries, surveys
Preparation for MTM Visit #4	Pharmacist prepares for final MTM session by reviewing changes to lab work, anthropometrics, surveys
MTM Visit #4	Same as above MTM visits. Pharmacist also recaps MTM sessions and partnership development
15-month study assessment: lab work, anthropometrics, surveys
Follow-up letter with lab results sent to provider

MTM = medication therapy management; EWS = Eat, Walk, Sleep; CHW = community health worker; HbA1c = hemoglobin A1c; DTP = drug therapy problem

Next, the participant, CHW, and pharmacist met together. The participant and CHW were together face-to-face in their locale and used an electronic device (such as tablet, laptop, or computer) to communicate with the pharmacist in his/her office via secure, high-definition videoconferencing.

In preparation for the first MTM session, pharmacists reviewed the current medication list and medication use history form. Relevant data from the research record were available to the pharmacist, including blood assays (i.e., HbA1c and lipid profile), physical assessment findings (i.e., body mass index and blood pressure), medication-related patient surveys (i.e., beliefs about medicine, barriers to taking medicines, medication forgetting), and patient lifestyle habits (i.e., self-reported sleep and physical activity).

During the first MTM session, the CHW facilitated introductions and provided the participant’s blood pressure and heart rate readings to the pharmacist in real time. The pharmacist interviewed the participant, with translation by the CHW if needed, to reconcile the medication list. Together, they collaborated to identify and categorize DTPs. If the DTP could be resolved, the pharmacist also documented DTP resolution status and person(s) involved in resolving the DTP, and action taken. Together, the pharmacist, CHW, and participant developed and discussed a Medication Action Plan detailing action steps.

After the session, the pharmacist mailed the Medication Action Plan to the participant. The pharmacist also faxed a medication report to the provider outlining the findings of the first MTM session, the reconciled medication list, vitals taken during first MTM session, baseline biometrics, and recommendations.

The second and third MTM sessions followed the same procedure as the first MTM session, with the exception that the medication report was only faxed to the provider if warranted. The first, second, and third MTM sessions were completed prior to the 12-month study assessment which included repeated lab work, anthropometrics, and surveys. The pharmacist then reviewed the 12-month results in preparation for the fourth and final MTM session, which occurred prior to the 15 month study assessment. During the final MTM session, the pharmacist followed the same steps as above and also recapped the MTM sessions and partnership development. Every attempt was made to keep the same pharmacist assigned to a given participant, but this was not always possible and some participants worked with more than one pharmacist.

Measures

Demographic Characteristics and Social Determinants of Health. Participants self-reported demographics including age, sex, and marital status. SDOH included income, years of education, English language proficiency, Khmer language proficiency, difficulty communicating with healthcare providers, disability status, insurance status (Medicare, Medicaid, private, none), and access to a car and driving. Participants also reported historical disadvantage, i.e., the number of years they lived under the Pol Pot regime in Cambodia and in a refugee camp.

Clinical Characteristics. Symptoms of depression and anxiety were assessed with the depression subscale (15-item and 10-item respectively) of the Khmer language Hopkins Symptom Checklist using the published mean cutoffs of 1.75 to determine likely major depressive disorder and anxiety disorder.⁴⁰ Cronbach’s coefficient alpha in this study was 0.93 and 0.94 respectively. Symptoms of post-traumatic stress disorder were assessed with the Khmer language version of the Harvard Trauma Questionnaire⁴¹ using the published cutoff of mean = 2.5 to determine likely post-traumatic stress disorder (PTSD). Coefficient alpha in this study was 0.93.

Medications. Participants across all study arms self-reported number of medications. They also indicated whether they took medications for psychiatric conditions including depression, anxiety, PTSD, or ‘broken courage’ or ‘thinking too much’ (the last two are Cambodian cultural idioms of psychological distress).

Medication-Related Process Outcomes Across Treatment Arms

Medication Forgetting. One item asked about frequency of forgetting to take medications, with response options on a Likert scale from 0 = ‘never’ to 4 = “always”.

Barriers to Taking Medicines. Based on previous research in the target population^11,19, we created an 8-item scale to assess the frequency with which structural barriers interfere with medication-taking such as language, transportation, cost, and medication cost offsets such as food and housing. Response options are 1 = yes, often, 2 = yes, sometimes, or 3 = no, never. Higher scores indicate higher barriers. Coefficient alpha was 0.80.

Beliefs About Medicines - Specific. The Specific scale of the Beliefs about Medicines Questionnaire (BMQ⁴² - Specific) assesses the respondent’s perception of his or her own medicine(s). It comprises two 5-item subscales: the ‘Specific Necessity’ subscale (i.e. beliefs about the necessity of taking one’s medications to remain healthy) and the ‘Specific Concerns’ subscale (i.e. concerns about the negative effects of one’s medications). Response options are on a 5-point Likert scale from 1 = strongly agree to 5 = strongly disagree. Higher scores reflect higher perceived need and concerns, respectively. Coefficient alpha in this sample was 0.84 for ‘Specific Necessity’ and 0.74 for ‘Specific Concern’.

Medication-Related Process Outcomes in the EWS+MTM Arm Only

Drug Therapy Problems (DTPs). Due to resource limitations, DTP data were only collected from participants assigned to the EWS+MTM arm. Determination of DTPs followed guidelines of the Pharmacists’ Patient Care Process set forth by the Joint Commission of Pharmacy Practitioners.⁴³ The classification system has four major categories: appropriateness (is the medication indicated for the condition in this particular patient?), effectiveness (is the medication meeting clinical target?), safety (does the patient have, or is the patient at risk for, adverse drug reactions?), and adherence (is the patient taking the medication as prescribed?). DTPs are assessed and classified in that sequence so that problems associated with the medication itself are not attributed to patient non-adherence. The pharmacist identified DTPs by reviewing the participant’s medication review form that had been filled out by the CHW and discussing it with the participant. If applicable, the drug and/or condition was documented. The pharmacist coded resolution status using the following: R = Resolved, P = Partial Resolution, N = Not resolved. Partial vs full resolution was defined based on whether the DTP required an additional intervention by the health team member or participant.⁴⁴

Therapeutic Alliance.⁴⁵ Therapeutic alliance was assessed at post assessment with a modified version of the Therapeutic Alliance scale. Items tapped into trusting, having confidence in, liking, and being listened to by the CHW and the pharmacist, respectively. All participants reported alliance with their CHW; only those participants assigned to EWS+MTM reported alliance with their pharmacist. Response options were from 0 = not at all to 3 = a lot, with higher scores indicating higher alliance. Cronbach’s alpha for alliance was 0.82 with CHW, and alliance 0.94 with pharmacist. Participants also reported satisfaction with the MTM intervention from 0 = not at all to 3 = a lot, with higher scores indicating higher satisfaction.

Data Analysis.

First, descriptive statistics were used to illustrate the demographic and SDOH characteristics of participants. Second, to compare the EWS+MTM arm to the SS and EWS arms on medication forgetting, barriers to medication, and beliefs about medications, a series of linear mixed models were used with contrast statements to test if the EWS+MTM arm differed from the other two arms at each of the three time points of the study and change in outcomes between time points. Third, within the EWS+MTM arm, descriptive statistics were used to describe the number of DTPs and percentage of resolutions. Fourth, within the EWS+MTM arm, we examined which patients benefited most from the EWS+MTM intervention. A generalized linear model with binomial distribution and logit link was used to test whether demographic characteristics, SDOH, and forgetting, barriers, and beliefs predicted the percentage of DTP resolutions. Finally, descriptive statistics were used to report therapeutic alliance with CHWs and pharmacists, respectively. Analyses were conducted using SPSS v27.

Results

Sample

Table 2 provides characteristics of participants assigned to the EWS+MTM arm. Participants were mostly female, average age of 55, low household income, little formal education, most received public health insurance, most were clinically depressed and anxious, spent an average of 4 years in refugee camp, and lived in the US for 30 years. We have previously reported that the three treatment arms were similar at baseline along key characteristics.³⁹ Three-quarters (77%) of participants in the EWS+MTM arm completed all 4 MTM sessions.

Table 2:

Demographic characteristics for participants assigned to the EWS+MTM arm (n=63)

Characteristic	Frequency	Percentage
Gender
Female	51	81.0%
Male	12	19.0
Age
Mean ± SD	55.1 ± 8.5
Marital Status
Married	33	51.6%
Household Income
Under $20,000	25	39.7%
$20,000 – 30,000	16	25.4
$31,000 – 40,000	7	11.1
Over $40,000	6	9.5
Don’t know/Refuse	9	14.3
Able to drive
Yes	48	76.2%
Access to a car
Yes	47	74.6%
Years Education
Mean ± SD	6.3 ± 5.1
% English language proficient
Speak	20	31.7%
Read	18	28.6%
Write	15	23.8%
% Khmer language proficient
Speak	61	96.9%
Read	34	54.0%
Write	27	42.8%
Past-year difficulty communicating with healthcare provider due to language
Never	24	38.7%
Rarely	9	14.5
Sometimes	9	14.5
Often	14	22.6
Always	6	9.7
Health Insurance
Medicare	17	27.0%
Medicaid	18	28.6
Private	22	34.9
None/Don’t know	6	9.6
% Disabled
Yes	17	27.0%
% Meeting cutoff for clinical depression (>=1.75 Hopkins depression)	36	57.1%
% Meeting cutoff for clinical anxiety (mean >=1.75 Hopkins anxiety)	33	52.4%
% Meeting cutoff for post-traumatic stress disorder (mean >=2.5 Harvard)	14	22.2%
Years living under Pol Pot
Mean ± SD	3.3 ± 1.1	N/A
Years in refugee camp
Mean ± SD	3.9 ± 8.0	N/A
Years in the U.S.
Mean ± SD	30.0 ± 11.0	N/A
Total number of medications
Mean ± SD	5.4 ± 4.1	N/A

EWS = Eat, Walk, Sleep; MTM = medication therapy management; SD = standard deviation

Changes in Forgetting, Barriers, and Beliefs Across Arms

Figure 1 panels A through D show the scores for medication forgetting, medication barriers, and BMQ beliefs (necessity and concerns) for each treatment group at three data collection time points. As seen in Panel A, forgetting decreased from baseline to 12 months in the EWS+MTM group while it increased in the SS group. At 12 months the difference between EWS+MTM and other groups was marginally significant (p=.094) and at 15 months there was a significant difference (p=.030) such that the average score in the EWS+MTM arm was 0.26 units lower than the average of the other two arms.

Figure 1:

Changes in Self-reported Medication Surveys Over Time by Treatment Arm

EWS = Eat, Walk, Sleep; MTM = medication therapy management

Panel B shows the trajectory for barriers to medication. For all groups there was a noticeable and significant decrease from baseline to 12 month (p<.001) that was maintained at 15 months but there was no difference between the EWS+MTM group and other groups at 12 months (p=.668) or 15 months (p=.338). Panels C and D show the BMQ necessity and concern subscale scores and no discernable pattern is evident. The EWS+MTM group did not differ from the other groups at 12 months (necessity: p=.872, concern: p=.899) or 15 months (necessity: p=.533, concern: p=.873).

DTP Resolution in the EWS+MTM Arm

Table 3 shows the number of DTPs by type detected over the course of the study and the percentage resolved. Using a conservative approach, participants who were not able to complete all four MTM sessions were still included in the analysis. The average number of DTPs per participant was 6.6 (SD=4.3) and the most frequent type was for safety (M=2.6, SD=2.8) and least frequent type was for effectiveness (M=0.7, SD=0.9). Nearly 84% of the DTPs were resolved and varied from 74% for effectiveness to 94% for adherence. The great majority (88%) of DTP resolutions were managed by the CHW-pharmacist team and did not require intervention from another healthcare provider.

Table 3:

Resolution^* of drug therapy problems over 15 months within the group receiving EWS+MTM (n=63)

DTP Type	Number DTPs	Number Resolutions*	Percentage Resolved
Indication   Mean ± SD	1.90 ± 2.09	1.43 ± 1.74	76.3%
Effectiveness  Mean ± SD	0.68 ± 0.93	0.52 ± 0.84	73.5%
Safety  Mean ± SD	2.57 ± 2.82	2.05 ± 2.56	81.2%
Adherence   Mean ± SD	1.41 ± 2.02	1.33 ± 1.95	94.3%
Total  Mean ± SD	6.57 ± 4.34	5.33 ± 3.94	83.8%

^*Resolution includes partial and full resolution

MTM = medication therapy management; DTP = drug therapy problem

Predictors of DTP Resolution

The regression results for predictors of DTP resolution are shown in Table 4. Significant SDOH predictors were years of education (OR=0.94, p=.016), English writing ability (OR=0.73, p=.015), difficulty communicating with provider (OR=1.39, p<.001), having private insurance (OR=1.99, p=.030), being disabled (OR=0.51, p=.008), and years living under Pol Pot (OR=0.66, p=.045). Among the medication surveys at baseline, medication barriers was a significant predictor (OR=0.79, p=.019) such that each additional barrier was associated with a 21% reduction (1–0.79) in the odds of a having a resolution, even after adjusting for significant demographic characteristics (OR=0.80, 95%CI:0.66–0.98, p=.028). Neither forgetting to take medications nor beliefs (necessity or concerns) predicted DTP resolution.

Table 4:

Binomial Regression Model Predicting Resolution of Total DTPs

Predictor	OR (95% CI)	P-Value
Demographics and Social Determinants of Health
Male	1.16 (0.56 – 2.41)	.688
Age	1.02 (0.99 – 1.06)	.145
Married	1.11 (0.67 – 1.83)	.685
Years of education	0.94 (0.90 – 0.99)	.016
Household income	1.09 (0.84 – 1.41)	.510
Can drive	1.07 (0.62 – 1.86)	.799
Access to a car	0.96 (0.56 – 1.65)	.880
Speak English	0.76 (0.57 – 1.01)	.060
Read English	0.84 (0.67 – 1.07)	.161
Write English	0.73 (0.56 – 0.94)	.015
Difficulty communicate provider	1.39 (1.15 – 1.67)	<.001
Private insurance	1.99 (1.07 – 3.72)	.030
Disabled	0.51 (0.31 – 0.84)	.008
Elevated depression symptoms	0.72 (0.43 – 1.23)	.228
Elevated anxiety symptoms	0.63 (0.37 – 1.06)	.082
Elevated post-traumatic stress symptoms	1.24 (0.69 – 2.21)	.470
Years lived under Pol Pot	0.66 (0.44 – 0.99)	.045
Year lived in refugee camp	1.03 (0.94 – 1.14)	.486
Years in US	0.98 (0.96 – 1.01)	.255
Number of medications	0.96 (0.90 – 1.01)	.108
Self-reported surveys at baseline
Forgetting to take medications	0.90 (0.68 – 1.19)	.461
Medication barriers	0.79 (0.65 – 0.96)	.019
BMQ Beliefs: Necessity	0.96 (0.89 – 1.03)	.270
BMQ Beliefs: Concerns	1.02 (0.96 – 1.10)	.495

DTP = drug therapy problem; BMQ = Beliefs about Medicines Questionnaire

Therapeutic Alliance and Satisfaction

On a 4 point scale (0 = not at all, 3 = a lot) the average rating for therapeutic alliance with the CHW (mean = 2.94), pharmacist (mean = 2.92), and satisfaction with the MTM intervention (mean = 2.90) were near the maximum value. The time spent to obtain and record this information on the Medication Review Form ranged from 50 to 80 minutes per patient, which averaged to 10–15 minutes per medication. The time spent on the initial MTM visit ranged from 20 to 90 minutes depending on patient complexity and number of medications, and the follow up MTM visits ranged from 15 to 60 minutes.

Discussion

The main finding from this study is that CHWs and pharmacists successfully improved important medication-related process outcomes in this clinically complex and hard-to-reach population with disadvantaged SDOH. Participant attendance at sessions, alliance with the pharmacist and CHW, and satisfaction with MTM were all high.

First, all treatment arms reported reduced barriers to taking medications. This suggests that CHWs alone were highly effective in reducing some structural barriers to medication such as transportation and language. We speculate that their most important skills in this regard include language translation, bureaucracy navigation, and advocacy.

Second, EWS+MTM with CHW/pharmacist teams was superior to the other treatments in decreasing participants’ forgetting to take their medication. This improvement occurred even in the absence of changing patient beliefs about medication necessity or concerns. Concerns about treatments exist among racial and ethnic minorities and may be tied to historical roots. For people from Cambodia, this may include mistrust of pharmacists or worry about counterfeit medications or the safety of how medications are manufactured. Such concerns may impact beliefs about disease, self-management approaches, and medication adherence.^46,47 Yet, the CHW/pharmacist team decreased forgetting despite no change in beliefs.

Third, within the EWS+MTM arm, a very high rate of DTPs (84%) were resolved by the CHW/pharmacist teams. The overall high resolution rate would likely be viewed favorably in any population, but it is particularly compelling given the sample’s clinical complexity and disadvantaged SDOH. The rate of DTP resolution far exceeded that of other published studies in populations with mental health issues, though it should be noted that those were much larger studies than the one reported here.^48,49 Although we were not able to compare DTP resolution across all three arms, it is encouraging to see that inclusion of CHWs in the MTM process can have such a positive effect.

The fourth main finding from the study was that, within the EWS+MTM arm, those with the greatest disadvantage had the greatest reduction in DTPs. We propose that these are precisely the types of patients who can maximally benefit from the services offered by bicultural, bilingual CHWs paired with pharmacists. Likelihood of DTP resolution was predicted not by number of medications, but by SDOH, i.e., lower education, less ability to write English, and more difficulty communicating with healthcare providers. Whereas we would not expect these relatively stable characteristics to change, we hypothesize that there is more opportunity for CHWs to optimize MTM services for these patients. Having lived fewer years under the Pol Pot regime and not being disabled were also associated with higher likelihood of DTP resolution. We reason that participants with relatively less exposure to the brainwashing genocidal regime, and those without a disability, may have more psychological and social resources that can potentiate benefiting from health interventions.^50,51 Higher DTP resolution was also associated with having private insurance and lower barriers to medication taking. Lack of private insurance may impede provider visits during which medication changes can be addressed. Financial barriers can greatly impede obtaining medications. Barriers related to insurance and cost were largely outside of CHW control.

We assert that the CHWs were crucial to the success of this innovative approach to MTM. The CHWs helped increase participant trust and confidence regarding the pharmacist recommendations. They helped calm anxious participants and motivate those with depression. They decoded for the pharmacist cultural patterns of symptom reporting, deference to authority, and nonverbal communication. These indispensable ‘soft’ skills were in addition to conventional CHW activities such as language translation, appointment scheduling, assistance with insurance paperwork, and picking up prescriptions. They also provided considerable time-saving because the CHW, rather than the pharmacist, conducted the medication history-taking and interpretation. The participants’ high therapeutic alliance with the both CHWs and pharmacists and satisfaction with the CHW/pharmacist intervention documents high acceptability among patients.

Cambodian Americans are a relatively small ethnic group in the U.S., yet lessons learned from this population can be applied to other minority, immigrant, and refugee populations.^10,52 Racial disparities in medication use and DTPs exist.⁵³ For example, older Black adults have been found to have fewer prescription medications but more medication-related problems such as nonadherence, compared to white people.^10,54 We believe that our approaches are applicable to other groups with collective trauma histories and associated mental illnesses who face similar SDOH, i.e., both immigrant groups (e.g., Somali, Iraqi, Afghani) and American groups (e.g., Native American). Interventions such as ours should be modified and tested in other populations vulnerable to DTPs.

Limitations

The EWS+MTM arm was a small and unique sample that may not be generalizable to other groups. DTPs were only assessed in the EWS+MTM arm as grant funding was not sufficient for data collection across all arms. Therefore, DTP number, type, and resolution rate cannot be compared to the other study arms. Notwithstanding this limitation, randomization should have resulted in groups that did not differ in number of DTPs at baseline. Because the same CHWs that worked in the EWS+MTM arm also worked the EWS arm, there is some possibility of contamination. For example, the patient education provided by the pharmacist during MTM sessions could have been relayed by the CHW to participants in the EWS or SS arms. There may have been variability between pharmacists in coding DTPs, and a bias toward coding them as resolved. Lastly, not all participants in the EWS+MTM arm received all four MTM sessions due to scheduling issues and COVID-19 occurring during this study.

Conclusions

Policymakers should recognize the overall value of the CHW in the MTM process with the pharmacist. Community or ambulatory-based pharmacists are ideally positioned to work closely with CHWs who act as an effective liaison between the community and medical establishment. However, integrating pharmacists and CHWs requires new payment models/incentives.²⁵ If findings can be replicated on a larger sample, reimbursement should be provided for CHW MTM services.

Key Points.

What was already known:

1. Cambodian Americans are hard to reach and have disadvantaged social determinants of health.

2. Proven best practices for clinical treatment in such populations is limited.

3. Community health workers (CHWs) can engage medically underserved communities to overcome barriers to care.

What this study adds:

1. Medication therapy management (MTM) delivered by a cross-cultural team with a CHW and pharmacist resolved 84% of drug therapy problems.

2. CHWs substantially contributed to the MTM process by conducting the medication history-taking and language translation.

3. Patients who maximally benefited from the intervention were those with greater disadvantage.