Community Health Worker Program Outcomes for Diabetes and Hypertension Control in West Bank Refugee Camps: A Retrospective Matched Cohort Study

Program Model

In early 2018, a home-based CHW program comprised entirely of Palestinian refugees, Health for Palestine (H4P), was initiated as a joint program between 1for3, a Boston-based nongovernmental organization (NGO) serving Palestinian communities globally, and the Lajee Center, a community-based grassroots cultural center in Aida and Beit Jibrin (also commonly known as Al-Azza) camps in Bethlehem, the occupied West Bank. Aida and Beit Jibrin refugee camps, with populations of 5,500 and 2,900, respectively, comprise refugees from more than 27 villages in what is now Israel. The population density of these camps is more than twice that of Manila, the most crowded city in the world.²⁹

The program was created to address the top health concerns of Aida and Beit Jibrin camps’ residents and their most common chronic diseases, including hypertension and T2DM.³⁰ The initial CHWs, all 6 of whom are refugees from the camps aged 19–26 years at the time of the study, were recruited by the Lajee Center using posted announcements in the camps and on social media. An internal medicine physician and a medical student trained the CHWs using a guide modified from published CHW manuals. The training period lasted 6 weeks and included lectures, role playing, hands-on training for blood pressure and blood glucose measurement, and ultimately guided home visits with a physician and/or nurse.

Initial remuneration of the CHWs was 1500 New Israeli Shekels (NIS) (∼US$428) per month in 2018, which, although the minimum wage in the occupied West Bank, is a coveted salary given high rates of unemployment particularly among youth and in refugee camps.³¹ Salaries were increased to NIS1800 (∼US$514) per month in 2019, and 4 additional CHWs were hired that year at the starting salary. CHWs coordinate closely with a pharmacist and physician to adjust patient care plans and to accompany patients to clinics and hospitals as needed, and the CHWs are supervised by a project manager who is also a licensed nurse. A backup medicine supply is maintained to mitigate frequent local stock-outs at both UNRWA and Palestinian Ministry of Health pharmacies. Continuing education is delivered by volunteer physicians and medical students from the United States and Palestine, and the program is financially supported by family foundations and individual donations to 1for3 and the Lajee Center. While no formal costing analysis has yet been published, and despite a more intensive longitudinal focus serving fewer patients per CHW, H4P costs are in the vicinity of an established CHW program in Nepal that includes a noncommunicable disease component.³² H4P’s per capita annual cost of US$6.49 (versus US$3.05), 85% of which is personnel costs (versus 74%), and a CHW-to-population ratio of 1:1400 (versus 1:2017) are comparable, particularly as H4P’s newly trained CHWs had not yet maximized their patient panels (i.e., peak efficiency) during this study’s time window. Furthermore, compared with the 2011 UNRWA cost per visit-minute of US$5.99 in the West Bank,³³ H4P’s cost per visit-minute is more than 16 times lower at US$0.36.

The H4P CHW intervention was designed to mitigate the fragmentation of care resulting from settler colonialism and the complexities of a prolonged refugee crisis produced by the Nakba (“catastrophe”) of 1948, in which 85% of Palestinians were displaced, fled, or were massacred.^30,34 The impacts of a fragmented health system, intergenerational trauma of the Nakba, and ongoing military occupation of the West Bank have resulted in devastating health effects for Palestinians, and particularly refugees within overcrowded refugee camps.³⁵ Poorly connected networks of health care provided by public, private, NGO, charity, and multilateral actors, as well as logistical dilemmas generated by military checkpoints and the colonial architecture of occupation,³⁶ have resulted in a health care system that is difficult to access and navigate. H4P’s work is grounded in the goal of promoting enhanced community control over health, easing patient navigation of these fragmented networks, and simplifying and resisting the resulting complexities in accessing basic, effective, and longitudinal care for refugees living in camps.

The CHW intervention takes an accompaniment approach,³⁷ which encompasses transportation facilitation, medication supervision and delivery, adherence support, appointment scheduling and adherence, and monitoring of vital signs and glucose. The CHWs utilize motivational interviewing to address the primary risk factors of diet, exercise, medication adherence, and smoking. Given the known impacts of stress and allostatic load on chronic diseases like diabetes,³⁸ psychosocial counseling is integrated into CHW patient care. For guidance and support, the CHWs meet weekly with a social worker, also a Palestinian refugee, who is supervised by a clinical psychologist. The social suffering from occupation, military incursions, and prolonged displacement influence a range of psychological and physiological outcomes that interact to heighten morbidity as described in syndemic models.^39–41 Understanding the efficacy of H4P’s work among Aida and Beit Jibrin camp residents, a population suffering regular military incursions and the greatest exposure to tear gas in the world,⁴² may offer new insight into the role that CHW programs can play within the Middle East, refugee camps, and zones of chronic violence and social exclusion or oppression.

• Download figure
• Open in new tab
• Download powerpoint

Community health workers in the West Bank discuss symptoms with a patient. © 2018 Nashat Jawabreh/Health for Palestine

As 1 of the first studies to assess an accompaniment-based CHW program in the region, this study aimed to (1) understand the feasibility of a grassroots accompaniment-based CHW program in a high-conflict and limited-resource setting, (2) assess the mean trend in patient blood pressure in hypertensive patients over the course of the first year of the program, and (3) assess the association between being assigned an H4P CHW as a T2DM patient in Aida or Beit Jibrin camp on A1c change. Our hypothesis was that CHW assignment results in greater diabetes progress as indicated by a declining A1c value compared to patients without CHW assignment.

Study Population

At the time of the initiation of the data analysis in late 2019, H4P had cared for 101 patients since its founding in 2018. We recruited patients living within Aida and Beit Jibrin camps into the program using several techniques, including flyers, personal connections to hired CHWs, and most commonly, snowballing. Recruitment targeted no specific centers, geographic areas, or patient types. A simple visual geospatial analysis of a random sample of patients revealed no grouping of patients by location, such as in the camp center, near medical clinics, or around the CHWs’ office at Lajee Center.

To enroll in the CHW program, patients had to have a known diagnosis of hypertension or diabetes, be aged over 18 years, and reside in Aida or Beit Jibrin camp. We confirmed a diagnosis of diabetes or hypertension using a medical history intake form compiled from clinic notes and patient interviews. Study eligibility criteria for diabetes patients included receiving diabetic care solely within the UNRWA Bethlehem clinic, a baseline A1c greater than 6.4, at least 2 A1cs at least 6 months apart during the duration of the program, and 1 A1c no earlier than 5 months before the first visit recorded by the CHWs. For hypertension patients, eligibility criteria for inclusion in the blood pressure analysis were (1) a diagnosis of hypertension, (2) blood pressure values available 9–12 months after enrollment, and (3) baseline blood pressure greater than or equal to 130/80.

Matches were selected from among all diabetic patients receiving diabetic care at the UNRWA Bethlehem clinic, with an A1c test performed between March 2018 and November 2019. These dates correspond with the duration of the H4P program and were selected to ensure similar start and end dates of both groups. We extracted data on matched patients from the UNRWA electronic medical record system.

Patient and Public Involvement

The concept of patient involvement is integral to the CHW intervention and extended to the goals of this analysis. Plans for the retrospective study were born out of patient interest in understanding their own progress and were shared with the 30 patients receiving CHW care. Patient interest in viewing their A1c results extracted from the UNRWA database for the purpose of the analysis resulted in a modified data agreement with UNRWA to allow transfer of patient A1c values into CHW data forms via KoBoToolbox for permanent patient access. All 30 patients receiving CHW care were given updates on both their individual A1c trajectory over time and the average A1c decline among the intervention group. The medical implications of A1c change were shared with the CHWs, who shared this information with patients on an individual level.

Intervention

The intervention analyzed is the aggregate chronic disease program conducted by H4P’s CHWs, whose core tasks are implemented as home-based care. Home visits are prioritized according to 3 levels of disease control, with the patients with the most controlled parameters being visited at least twice a month and those with the most uncontrolled parameters up to multiple times per week. Patients are graduated from the CHW program once they have spent 6 consecutive months at the highest level of control. At its foundation, each visit includes an assessment of the patient’s (1) blood pressure, (2) random blood glucose level, and (3) medication supply and adherence. CHWs measure glucose with locally available fingerstick testing glucose monitors according to manufacturer instructions, and measure blood pressure using automatic instruments according to standard protocol.¹⁰ CHWs engage in active listening and trauma-informed discussion of health challenges, with motivational interviewing and counseling on diet, exercise, smoking, and medication adherence. All patients receive a pillbox from the CHWs with instructions on its proper use. A local physician and a CHW manager closely supervise the CHWs to ensure regular visits and high-quality reporting. International advisors assist in strategic planning and assessments of data quality. The nonintervention group received the standard of care at the Bethlehem UNRWA clinic, which typically consisted of annual follow-up visits with a physician at the clinic.

• Download figure
• Open in new tab
• Download powerpoint

Community health workers in the West Bank check a patient’s blood glucose during one of their first home visits. © 2018 Nashat Jawabreh/Health for Palestine

Study Design

Feasibility measures include visit metrics, specifically number of visits accomplished and patient acceptance of medical assistance from a CHW, including unscheduled home visits. CHWs record each visit, including its length and the metrics collected, within a customized data form on KoBoToolbox for later access and analysis. Baseline blood pressures for hypertensive patients were measured by CHWs at entry into the CHW program and at each subsequent visit using an automatic blood pressure cuff.

Assessment of early outcomes for diabetic patients included analysis of patient A1c values over the duration of their enrollment in the program, from March 2018 to November 2019. As part of the CHW program, all enrolled diabetic patients had undergone A1c testing (ranging from 1 to 4 tests) over the duration of the program’s intervention, the results of which were extracted from the UNRWA database for the purpose of this study. Timing of intervention, as well as timing of repeat measurements, varied from patient to patient. Each patient who received CHW care was retrospectively matched to 3 T2DM patients receiving diabetic care primarily at the UNRWA Bethlehem clinic, by prespecified characteristics and ranges, specifically: baseline A1c (+/− 0.2), age (+/− 3 years), gender (female/male), and follow-up time (+/− 4 months). Logical statements were used for the matching (with replacement) process. Baseline age, gender, history of hypertension, stroke, myocardial infarction, and A1c levels were collected from UNRWA records for all patients.

Statistical Analysis

We accessed average blood pressure readings between March 2018 and November 2019 for all patients at baseline and 9–12 months post-enrollment, using the last blood pressure reading closest to 12 months, and analyzed the data using Excel. Standard deviations are reported and statistical significance for comparisons between baseline and 1 year were computed using 2-sided paired t-tests. Blood pressure data for controls was not uniformly available and so no direct comparison was made.

We created a multivariable regression model to test the hypothesis that CHW assignment results in greater diabetes progress as indicated by a declining A1c value compared to patients without CHW assignment. Unadjusted differences in means for patients in the treatment and control groups are also reported. Given the vast literature on the subject of A1c change, we used clinical reasoning to select covariates for the multivariable model. Despite the unsystematic entry of participants into the CHW program, 5 prespecified covariates were included given their role as potential confounders: history of myocardial infarction, history of stroke, physical activity status, smoking status, and hypertension status. We also performed an assessment without outliers since only 1 follow-up A1c value for each patient was available for the analysis. To account for this second analysis, we used a Bonferroni correction, with 2-tailed P-values of .025 considered statistically significant (.05/2). However, since the analysis was conducted for presentation of results, we did not use a Bonferroni correction for the inclusion of unadjusted results. All confidence intervals (CIs) were calculated with the use of heteroskedasticity-robust standard errors. Because our data set includes 120 subjects and we considered 6 variables in our model, overfitting was not a concern.

Due to the small number of patients within the CHW group, Fisher's exact analysis was used to compare proportions of patients under diabetic control (defined as A1c<8%). Data were analyzed using SAS version 9.4.

Ethical Approval

Institutional review board approval was not required for this retrospective matched cohort study per the Institutional Review Board of Harvard Medical School (Protocol 19-0081), which was locally reviewed for ethical risk and approved by H4P’s partnering organization, Lajee Center, in Aida camp. All participants gave their informed verbal consent or had already done so as part of the care received by the governing organizations.