Principal health systems research question: Can health advocates (HAs) in places of worship (PoWs) be successfully integrated into the primary care pathway to support the prevention and control of NCDs in low-income settings?

Specific objectives: 

a) Describe the organisational, sociocultural and religious contexts of PoWs that influence the confidence, competence and commitment of health advocates (HAs) to promote NCD prevention and management activities.

b) Describe the impact of HAs and their activities on health systems (e.g. service delivery, health workforce, governance), and congregations. 

c) Describe the plausible mechanisms that allow the linkage (or not) between (a) and (b) and how these vary within and between countries.

d) Identify solutions to implementation barriers in health systems and PoWs and construct a concept map to inform up-scaling, if results are positive. 




Permission and ethical approval for the CONTACT Study were received from the Ministry of Public Health and its Ethical Review Committee. In consultation with the Ministry, and based on distribution of religious groups, poverty indices and other census data, we selected Region 3, one of Guyana’s 10 administrative regions, as the primary site for implementation. Three intervention PHCCs were purposefully selected from a list provided by the Ministry, on the basis of population size served, geographic location and staffing complement. Each PHCC was linked to 1 Christian Church, 1 Hindu Mandir (Temple) and 1 Muslim Masjid (Mosque) from within its Ministry-designated catchment area, for a total of 9 intervention PoWs. Eligible PoWs were identified through a ‘groundtruthing’ process. Both qualitative and quantitative methods were used to obtain data, and all project outputs were developed through iterative participatory approaches.


Phase 1: Consultation with Stakeholders
We used concept mapping, a structured participatory methodology, to gather information on factors influencing implementation, prioritise areas for health system improvement, develop conceptual models of implementation, and foster communication and collaboration across stakeholder groups. Concept mapping uses qualitative procedures to generate group reflections on a topic of interest, followed by quantitative methods to synthesize and represent the group’s ideas visually.  We used an adapted version of the 6 concept mapping phases first outlined by William Trochim. 
Purposive sampling was used to recruit multi-disciplinary stakeholders with representation from congregations (Muslims, Christians, Hindus), Primary Health Centre (PHC) practitioners, administrative support staff at PHCs, regional health directors, and Ministry of Public Health staff. Participants completed three focus prompts: ‘Factors influencing the impact of H advocates are likely to be...’; ‘Successful recruitment and training of Health Advocates will be affected by...’; ‘Factors that will affect the ability of health centres to promote this service are.....’ The research team distilled the completed statements, deleted duplicates, clarified ambiguities and combined points. 
Participants were then asked to sort the statements into conceptually similar piles that make sense to them and create a name for the piles. For each statement, participants were asked ‘How important do you think this will be for the prevention and control of NCDs in Guyana?’ and ‘How feasible is it to do this this?’ They were then asked to rate each statement on its importance and changeability/feasibility using a standardised scale. 
Concept Systems MAX software was used to generate concept maps. Analysis of the data generated from the sorting and rating exercise produced visual summaries of the relationships among the strategies. 


Phase 2: Preparedness of PHCCs and PoWs for intervention
(i) We assessed PHCC readiness using a 185-item questionnaire. To our knowledge, there was no existing instrument designed specifically for this purpose, so we created and pre-tested an instrument based primarily on the WHO handbook for Monitoring the Building Blocks of Health Systems. The questionnaire was comprised of sections corresponding to the building blocks, and the guidelines for monitoring health systems were adapted to suit the context of a single PHCC. For example, service delivery was assessed using PHCC-relevant WHO tracer items for general and NCD-specific service. The questionnaire was completed by the most senior staff member at each PHCC (usually the doctor-in-charge), except for an observation section that was completed by trained research staff.
PHCC Questionnaire
 - Part 1 (Interview)

PHCC Questionnaire - Part 2 (Observation)


(ii) Our 66-item POW readiness questionnaire was adapted from the instrument created for the Civil Society NCD Regional Status Report and utilized by the Healthy Caribbean Coalition in 2013. This questionnaire was pretested and updated before administration to the leaders of the 9 POWs in the form of structured interviews. Domains included: governance and financing, existing social and health programmes, religious leaders’ knowledge about NCDs and local PHCCs, and their perspectives on and resources available for the CONTACT intervention.
PoW Questionnaire


Phase 3: Prevalence of NCDs and risk factors among congregants
442 congregants from the 9 PoWs were interviewed by trained survey staff using a 140-item pretested questionnaire that included questions taken from the WHO STEPwise approach to chronic disease risk factor surveillance (STEPS) instrument (Pan-American version), the Major Depression Inventory (MDI-10), and the abbreviated WHO Quality of Life (WHOQOL-BREF) scale. This was a convenience sample, as congregants participated on a voluntary basis further to calls from their religious leaders; it was not feasible to select a random sample, as the PoWs did not maintain registers of members, and attendance varied from one service day to another. Separate surveys were conducted at the PoWs, accompanied by ‘mini’ health fairs to provide incentive for participation and occupy congregants waiting to be interviewed.


Phase 4: Refinement of intervention and development of HA training materials
During this phase the CONTACT Study research team met weekly (more frequently for the local team) to discuss and refine the intervention in light of observations made during Phases 1 to 3. Expert input from Guyana and other Caribbean countries was sought for the development of the health advocate training materials; the local PAHO office provided technical and diplomatic support; and the MoPH provided technical support and funding. Two principal documents served as the foundation for the Health Advocate Training Manual: the Lay Diabetes Facilitator Training Manual for six English-speaking Caribbean countries (PAHO, Bridges) and the PAHO Passport to Healthy Lifestyles.

Training Manual


Phase 5: Health advocate recruitment and training
The study aimed to recruit 2 HAs from each PoW, one male and one female, with at least one person being a senior, respected congregation member. Religious leaders were asked to nominate, in collaboration with their congregation, trusted members to act as potential HAs. Potential HAs had to meet the following minimum criteria: a sound secondary education, age between 25 – 60 years, regular attendance at services, good time management skills and ability to multi-task and prioritise work, engagement in PoW’s social activities, ability to maintain confidentiality, good written and verbal communication skills, strong organisational and planning skills, and willingness and availability to volunteer at the PHCC’s chronic disease clinic. Potential HAs were interviewed and completed a 26-item written test assessing basic comprehension and numeracy skills. 

HA Recruitment poster

Over the course of three months a total of 29 individuals were nominated from the nine PoWs (see Table 1). Twenty-two nominees were females, some of whom were recent high school graduates starting new jobs and unavailable to attend the interview. While selecting the final list of HAs, we also considered their availability to attend the 8-day HA training. Of 19 persons interviewed, 15 HAs were selected and trained in 2 four-day workshops.
Training Manual


Phase 6: Intervention

Intervention Guide and Referral Sheet

Logbook - Activity Log

Logbook - Click Records


Statistical Methods – All Phases
Qualitative Data. CONTACT uses an inductive thematic analysis approach to identify patterns and themes in qualitative data (open-ended questions on readiness questionnaires). All researchers agree on themes to ensure robust analysis. 
Quantitative Data Descriptive statistics are used to assess the characteristics of participants in the congregation surveys. Categorical data are summarized using frequencies and tested using ?2, or Fisher’s exact test if an expected cell value is <5. Continuous data are plotted and, if approximately normally distributed, are summarized using means and standard deviations and tested using the Student’s t-test. If not normally distributed, they are summarized using medians and interquartile ranges and tested with non-parametric tests (equality of medians). All analyses are performed in Stata/IC 13.1 (StataCorp, Texas), and P-values are considered statistically significant if <0.05.