Specific Aims: Bridging Income Generation with GrouP Integrated Care (BIGPIC) Over 80% of cardiovascular disease (CVD) deaths occur in low- and middle-income countries (LMICs). Diabetes, a major risk factor for CVD, is also responsible for substantial morbidity and mortality in LMICs. Elevated blood pressure (BP) increases CVD risk among individuals with diabetes and pre-diabetes; BP control is therefore a powerful way to reduce CVD risk. Cost-effective, culturally appropriate, and context-specific approaches are critical. Two promising strategies to improve health outcomes are group medical visits and microfinance. Both can increase quality of care, clinician-patient trust, self-efficacy, health savings, self-confidence, group cohesion, and social support. While these strategies have been successful in other contexts, their impact on CVD risk reduction among diabetics and pre-diabetics in low-resource settings is not known.

In partnership with the Government of Kenya, the Academic Model Providing Access to Healthcare (AMPATH) Partnership has expanded its clinical scope of work to include diabetes and hypertension. AMPATH has piloted group care and microfinance initiatives among patients with chronic diseases with promising early results. Both strategies are feasible, as is integration of group medical visits into microfinance groups. However, the effectiveness of these strategies individually, and in combination, on improving CVD risk is not known.

Thus, the objective of this proposal is to utilize a transdisciplinary implementation research approach to address the challenge of reducing CVD risk in low-resource settings. The central hypothesis is: group medical visits integrated into microfinance groups will be effective and cost-effective in reducing CVD risk among individuals with diabetes and at increased risk for diabetes in western Kenya, and that the key modifiable CVD risk factor to be addressed is BP. The research team hypothesize that group medical visits and microfinance may each reduce CVD risk, but the integration of group medical visits and microfinance will yield the largest gains. Also further hypothesize is that changes in social network characteristics may mediate the impact of interventions on the primary outcome, and that baseline social network characteristics may moderate the impact of interventions. To test these hypotheses and achieve the overall objectives, the following specific aims will be pursued:

Aim 1: Identify the contextual factors, facilitators, and barriers that may impact integration of group medical visits and microfinance for CVD risk reduction, using a combination of qualitative research methods: 1) baraza (traditional community gathering) form of inquiry; and 2) focus group discussions among individuals with diabetes or at increased risk for diabetes, microfinance group members, and rural health workers.

Subsidiary Aim 1.1: Use identified facilitators and barriers to develop a contextually and culturally appropriate integrated group medical visit-microfinance model to reduce CVD risk among individuals with diabetes or at increased risk of diabetes. This model's acceptability and feasibility will be assessed by conducting focus group discussions with patients, microfinance group members, and health workers.

Aim 2: Evaluate the effectiveness of group medical visits and microfinance groups for CVD risk reduction among individuals with diabetes or at increased risk for diabetes, by conducting a four-arm cluster randomized trial comparing: 1) usual clinical care; 2) usual clinical care plus microfinance groups only; 3) group medical visits only (no microfinance); and 4) group medical visits integrated into microfinance groups. The primary outcome measure will be one-year change in systolic blood pressure (SBP), and a key secondary outcome will be change in QRISK2 CVD risk score, which has been validated for Black Africans.

Subsidiary Aim 2.1: Conduct mediation analysis to evaluate the influence of changes in social network characteristics on intermediate factors and intervention outcomes and moderation analysis to evaluate the influence of baseline social network characteristics on effectiveness of interventions.

Aim 3: Evaluate the incremental cost-effectiveness of each intervention arm of the trial, in terms of costs per unit decrease in SBP, per percent change in CVD risk score, and per disability-adjusted life year saved.

This research project will add to the existing knowledge base on innovative, scalable, and sustainable strategies for reducing CVD risk in diabetes and other chronic diseases in LMICs and other low-resource settings. If proven to be effective, the investigators are poised to expand the approach beyond the trial, thus ensuring that this research will have a significant and positive health impact on a larger population.

The Academic Model Providing Access to Healthcare Partnership (AMPATH) was initiated in Kenya in 2001. In 2009, AMPATH was designated as a Center of Excellence for Cardiovascular and Pulmonary Disease Research by the NHLBI, and the program's research portfolio includes CVD, diabetes, and risk factors. Population-based prevalence of diabetes and hypertension in Kenya is not well known, and estimates vary according to technique, definitions, and geography. Prevalence rates range from 1% to 12% for diabetes, and 5% to 24% for hypertension. The lower ranges of these figures are self-reported disease and likely underestimate true prevalence due to low awareness of and screening for chronic diseases in this region.

In response to this substantial and growing burden of diabetes and hypertension, AMPATH has formed a Chronic Disease Management (CDM) Program , and established productive partnerships with the Kenyan government and local communities, in order to optimize care delivery for chronic diseases, such as diabetes and hypertension. AMPATH has a Memorandum of Understanding with the Kenyan Ministry of Public Health and Sanitation and the Ministry of Medical Services to test and evaluate innovative approaches to chronic disease management. The CDM Program developed a robust diabetes and hypertension management protocol derived from the International Diabetes Federation, World Health Organization (WHO), and Joint National Commission 7 and 8 guidelines for diabetes and hypertension management, using drugs contained in the Kenyan national formulary. The CDM program's interventions are delivered at Ministry of Health facilities with associated personnel. In addition, AMPATH has developed innovative, community-based solutions to ensure a consistent and secure supply of essential medicines. The CDM Program has enrolled over 2,000 patients with diabetes and 3,100 patients with hypertension, who are being cared for at nine rural health centers and 30 rural dispensaries. However, achieving BP control has been challenging; preliminary analysis of CDM Program data indicates that despite an average reduction in systolic BP (SBP) of 9.3 mmHg, only 30-40% of the CDM patient population have controlled BP. Thus, the program is actively seeking out innovative approaches to optimize care delivery and outcomes for patients.

AMPATH is also sensitive to the economic reality of its catchment population, recognizing that the average daily income for a substantial proportion of its clients is less than one US dollar per day. AMPATH has created a Safety Net Program to improve the economic security of clients without encouraging dependency, by promoting income-generating activities. The program has conducted preliminary work organizing microfinance groups, initially starting with HIV patients and pregnant women, and recently expanding to patients with diabetes and hypertension. Based on the model of Village-Level Savings and Loan Associations, AMPATH's microfinance program is a client-driven model that involves the creation of community savings groups. Thus, microfinance group members mobilize and manage their own savings, provide interest-bearing loans to group members, offer a limited form of financial insurance, and contribute to a social fund that is used in cases of emergency or welfare issues of group members. AMPATH also provides group members specific training on agribusiness, entrepreneurship, financial literacy, accounting, marketing, and group dynamics, to support engagement in sustainable income-generating activities. Since the program was launched in November 2010, there have been a total of 9,969 clients (83.8% female, 16.2% male), 508 groups, cumulative deposits of over $250,000 and social fund of nearly $23,000. The microfinance program has increased group membership by 14% with a retention rate of over 98%. An impact evaluation of the microfinance program revealed that group members have increased their monthly income by 10% and increased their weekly expenditure on food per household by 18%.

This research project will be conducted within the AMPATH catchment area, in six Divisions in western Kenya, each of which is geographically and administratively divided into Community Units (CUs) of approximately 6,000 individuals: Burnt Forest (4 CUs), Chulaimbo (8), Endebess (4), Kapsara (6), Matunda (4), and Moi's Bridge (6). Each Division has rural health facilities (rural health centers and rural dispensaries) staffed by clinical officers (mid-level practitioners) and nurses, while CHWs are in the villages. There has been a longstanding positive relationship among AMPATH, healthcare providers, and the communities. The CDM program has an established presence in all these communities.

Conceptual Framework Both microfinance and group medical visits can improve health outcomes through many mechanisms. The research team hypothesize that both interventions can cause changes in social network characteristics, leading to improved intermediate factors (diet, physical activity, medication adherence, and retention in care), ultimately yielding CVD risk reduction. It is also possible that both interventions can impact the intermediate factors and health outcomes independent of social network characteristics. Likewise, changes in social network characteristics may mediate the relationship between the interventions and CVD risk reduction independent of the intermediate factors. In addition, the interventions may be related to the health outcomes by other independent pathways. Finally, baseline differences in social network characteristics may moderate the impact of microfinance and group medical visits on CVD risk reduction.

Microfinance can increase savings (specifically health savings) and income, leading to increased ability to pay for healthier foods, medicines, diagnostic services, specialty consultation, and health insurance, thereby improving diet, medication adherence and retention in care. Microfinance can also lead to improved financial literacy, self-confidence, and decision-making agency, which can positively impact all of the intermediate factors, and subsequently CVD risk reduction. Likewise, group medical visits can increase the efficiency of care delivery, improve quality of care, encourage self-efficacy, and increase levels of clinician-patient trust, all of which can improve intermediate factors and thereby achieve CVD risk reduction. Both interventions can change social network characteristics within two to three months, and can create group cohesion, social support, and shared experiences, leading to a social environment more conducive to healthy behavior choices, medication adherence, and retention in care, maximizing their effectiveness. Social networks influence behavior through several theoretical mechanisms: information for behavior change, perceptions of social norms, how-to knowledge, and social comparisons. Through all of these mechanisms, individuals with diabetes or at increased risk of diabetes can achieve CVD risk reduction. It is likely that integrating group medical visits into microfinance groups may yield even better outcomes, through a combination of the mechanisms described above. To assess the independent effects of group medical visits, microfinance, and integrated group medical visits-microfinance groups, a four-arm trial is proposed. In addition, the investigators will assess whether changes in social network characteristics mediate the relationship between each intervention, the intermediate factors, and the primary outcome. Finally, the research team will assess if baseline differences in social network characteristics moderate the relationship between each intervention and the primary outcome. In sum, the Bridging Income Generation with GrouP Integrated Care (BIGPIC) study will be able to assess comprehensively the relationship between microfinance, group medical visits, social networks, intermediate factors, and CVD risk reduction among individuals with diabetes and at increased risk of diabetes.

Subsidiary Aim 1.1: Develop a contextually and culturally appropriate integrated group medical visit-microfinance model for CVD risk reduction AMPATH's Community Strategy Initiative routinely engages with existing community-based governance structures to gather input and feedback on any community-based initiative. In addition, the AMPATH Safety Net Program routinely gathers feedback from community members, and the microfinance intervention has evolved as a result of that input. The research team will adhere to these same principles in this proposed project. A participatory, iterative design process will be pursed and create a "design team" consisting of one facilitator, two research team members, two clinicians, two CHWs, two patients, and the research assistant. The design team will meet for several design sessions over a period of four weeks. The design process will involve three phases: brainstorming, conceptualization, and creation. During brainstorming, the facilitator will first acquaint the team with the findings of the qualitative study and specify the criteria for the integrated group medical visit-microfinance intervention (i.e. prototype as described in Aim 2 below, while incorporating content from the focus groups and mabaraza). In the conceptualization phase, the team will evaluate advantages and disadvantages of ideas resulting from brainstorming, and will develop a more concrete model. The creation phase will involve the actual creation of the integrated group medical visit-microfinance model for CVD risk reduction among individuals with diabetes and at increased risk for diabetes. Educational materials will be developed in English, Kiswahili, and other local languages using standard approaches to translation and back-translation. Once the integrated group medical visit-microfinance model is developed, two FGDs will be conducted each with patients, microfinance group members, and health workers, in order to assess acceptability and feasibility. The investigators have previously used this type of approach with positive community feedback.

Aim 2: Evaluate the effectiveness of group medical visits and microfinance for CVD risk reduction. The effectiveness of group medical visits and microfinance groups for CVD risk reduction will be evaluated among individuals with diabetes or at increased risk for diabetes, by conducting a four-arm cluster randomized trial . The primary outcome measure will be one-year change in SBP, and a key secondary outcome measure will be change in CVD risk score. Randomization will occur at the level of the CU (cluster), and will be stratified by Division. Each of the 32 CUs will be randomly allocated (eight CUs per arm).

Preliminary Data As above, AMPATH's CDM Program has an extensive network of rural health facilities delivering care for diabetes and hypertension. Although there have been notable improvements in clinical measures, nearly 60% do not link to care and nearly 50% do not remain in care once enrolled. Women are twice as likely to engage in care as men. The CDM Program has piloted various forms of group care for diabetes and CVD risk factors and has determined that group care is feasible and acceptable in this setting. The program implemented diabetes self-management support groups providing community-based peer support as a supplement to health facility-based clinical care, education, and counseling. These groups included a peer leader with specific training, and they had regularly scheduled meetings during which members learned about behavioral modification, self-management strategies, and problem-solving approaches. Analysis of 148 subjects at three months revealed that 76% of members attended group meetings at least once per month. Median hemoglobin A1C improved from 9.6% to 9.1%, but there were no significant changes in BP, body-mass index, or waist circumference.

The CDM Program and Safety Net Program have also collaborated to integrate microfinance groups and group medical visits, and have demonstrated the feasibility of this integrated approach. One-year results of this preliminary work (BIGPIC Pilot) are now available and appear promising. Six integrated groups were newly formed comprising 160 members, of whom 90 (over 50%) were patients with diabetes and hypertension. More women attended the initial screening (72%), and were therefore over-represented. After invitation to create integrated group medical visit- microfinance groups, women were slightly more likely to participate than men (83% vs. 74%).

Linkage, retention, and change in SBP comparing usual care vs. BIGPIC Pilot. Model of Care Linkage to care Retention in care

∆ SBP after 1 year Usual Care 42% 50%

• 9.3 mmHg BIGPIC Pilot 66% 66%
• 21.0 mmHg At one year, members had accumulated deposits totaling $7,500, with a social fund of $1150. The loan repayment rate was 97.8%. Progress out of Poverty Index mean score increased from 36 to 44, in addition to other economic indicators. Table 1 provides a summary of linkage, retention, and SBP change, comparing usual care and the BIGPIC Pilot. In year 2, membership has increased by 86.9%, indicative of a very favorable response by the community.

Participants Within each CU, AMPATH has initiated home-based testing and community-level screening of glucose and BP, with a plan to cover one-third of each CU's population every year. Adult individuals with elevated BP (SBP ≥ 140 or diastolic BP (DBP) ≥ 90) or elevated random blood glucose (≥ 8.1 mmol/L) will be assigned a unique AMPATH Medical Record System identification number and referred to the local rural health facility for further evaluation and confirmation of disease status. At the health facility, each individual will have a repeat BP and fasting glucose measured, and those who meet the designated cutoffs (repeat SBP ≥ 140 or DBP ≥ 90; fasting glucose ≥ 7 mmol/L) will be entered into the CDM Program.

Inclusion criteria:

Adults in the CDM Program who:

• Have diabetes (fasting glucose ≥ 7 mmol/L);
• Are at increased risk for developing diabetes (impaired fasting glucose (fasting glucose 5.6 - 6.9 mmol/L);
• Have an elevated Leicester Risk Assessment score (≥ 7)

Exclusion criteria:

• Acute illness requiring immediate medical attention;
• Terminal illness;
• Refusal to provide informed consent;
• Individuals who are pregnant or have HIV will be excluded, (automatically referred to a higher level of care within the AMPATH care system).

Participation in the trial will not be contingent upon membership in a group. An intention-to-treat analysis will be conducted and the denominator in each cluster will include all eligible individuals who consent, irrespective of whether they are in a group, thus minimizing selection bias.