Treatment and control group participants completed a baseline survey after consenting to be in the study, and follow-up assessments were conducted at 3-month and 6-month. Surveys were administered in Punjabi by trained interviewers. Additionally, clinical measurements were obtained at these timepoints.

Primary outcomes included weight, BMI, blood pressure, and glucose reduction, access to and utilization of care, and knowledge and practice of physical activity and healthy eating. Weight, blood pressure, and glucose and total cholesterol measurements were obtained by CHWs during baseline and 3- and 6-month follow-up visits. BMI was calculated at each timepoint using current weight and the height recorded at baseline, and Asian BMI categories were used. Waist circumference was obtained at each timepoint. Blood pressure was collected using previously validated methods for community-based settings [55,57]. Three resting blood pressure measurements approximately 2–3 min apart were collected at each timepoint using an OMRON automatic blood pressure monitor with participants in a seated position. The average of the second and third blood pressure readings was recorded as the final value. Two-hour fasting glucose and total cholesterol measurements were obtained via finger stick test.

Demographic variables and access to healthcare measures were adapted were from the Census American Community Survey (ACS) and the Behavioral Risk Factor Surveillance Survey (BRFSS) [58,59]. Self-efficacy questions related to exercise, nutrition, and health-related decisions were adapted from the Bandura Self-Efficacy Scale [60]. Questions on diabetes knowledge were adapted from the Michigan Diabetes Knowledge Scale and risk assessment questions from the American Diabetes Association (ADA) [45,61]. Questions on food behaviors such as portion control, preparation/buying, and planning, as well as intent to engage and motivators of physical activity were adapted from measurement of the behavioral objectives of a weight management intervention [62]. Additionally, treatment group participants were asked a series of questions assessing satisfaction with the intervention that were adapted from other studies. All survey questions were developed in English and translated into Punjabi by bi-lingual translators. Questions were reviewed by project coalition members for accuracy and cultural appropriateness and culturally relevant examples were integrated.

Qualitative data collection was conducted for program evaluation purposes. Qualitative interviews were conducted with the CHW supervisor on a quarterly basis by an independent evaluator during and after intervention completion to assess experiences in implementing the program, including barriers and facilitators to recruitment, retention, and diabetes prevention promotion. The lead investigator and the evaluator developed the interview questions using a review of relevant literature. In addition, a focus group with seven female treatment group participants was facilitated by the CHW supervisor to assess participant satisfaction and to obtain feedback for program improvements after the 6-month intervention period.

Descriptive statistics are summarized, and we compare baseline characteristics of the treatment and control groups for all individuals consented and enrolled into the intervention (n = 126). Changes in outcome variables were reported across baseline, 3-month, and 6-month for all individuals having complete data at all three timepoints. For continuous variables, mean change and standard deviation (SD) were reported; for categorical variables, total n and percent were reported. Within-group differences were assessed using paired-sample t-tests for continuous variables and chi-square tests for categorical variables; between-group significance was reported using t-tests of the total change across time points for continuous variables. Statistical significance was set at p < 0.05. Analyses were conducted using SPSS version 19.0 (IBM Corp., Armonk, NY, USA).

Notes from CHW interviews and focus group transcripts were reviewed and coded by two of the co-authors for themes related to feasibility, acceptability, and changes in outcomes among study participants. Narrative analysis techniques were utilized whereby segments of text relating to themes were identified and core codes and secondary codes were assigned. Relationships between codes within themes were also explored. Discrepancies in coding were resolved by discussion and consensus between the two coders.

Among treatment group participants in the final analyses, 97% (n = 57) completed four of the six intervention sessions, while 92% completed all six intervention sessions. In terms of follow-up phone calls, 85% participated in at least six of 10 follow-up phone calls with the CHW, and 69% participated in all 10 follow-up calls.

Approximately 17% of treatment group participants (n = 17) were lost to follow-up. Among these individuals, n = 9 did not attend any sessions after being enrolled, consented, and completing baseline despite numerous attempts at contact, and n = 8 were lost to follow-up due to lack of time to participate in the intervention, family obligations or work schedules, or leaving the U.S. to travel to the home country. Individuals that were lost to follow-up were not demographically different than the individuals that were retained in the treatment group. Among control group participants, only one individual was lost to follow-up.

Socio-demographic characteristics and baseline clinical information is presented for the 126 individuals with baseline data enrolled into treatment and control groups (Table 3). At baseline, individuals in the treatment group were significantly more likely to be female, to have lived in the U.S. for fewer years, to have higher education levels, to have health insurance, to report their health as good or higher, and to have a lower weight.

Statistically significant improvements in clinical variables between baseline and 6-month were observed in glucose, weight, and BMI for the treatment group, but not for the control group (Table 4). Systolic and diastolic blood pressure decreased significantly for treatment and control groups, and in both groups the majority of individuals (96% in both groups) had controlled blood pressure at 6-month. Total cholesterol levels increased significantly for the treatment group between baseline and 6-month (p < 0.01). Additionally, significant between-group differences in the correct direction were observed for glucose, and the change in glucose between baseline and 6-month was significantly greater for the treatment group (p < 0.01). Between-group differences for changes in weight and BMI also approached significance, with greater changes seen in the treatment group.

Among the treatment group, 88.7% reported engaging in any physical activity at 6-month, compared to 3.8% at baseline (p < 0.01). The control group demonstrated a non-significant, smaller increase in physical activity. Both groups showed a significant increase in social interaction related to physical activity (e.g., whether participants were more likely to reach out to friends or family to engage in physical activity). However, this change was greater among the treatment group (3.6 vs. 1.9 at 6-month), and between-group differences in social interaction related to physical activity were also significant between baseline and 6-month (p < 0.01).

In terms of food-related behaviors, the treatment group showed a statistically significant increase in portion control from baseline to 6-month (1.8 to 3.6, p < 0.01), while a non-significant change was seen for the control group (2.9 to 2.7, p = 0.09); the between-group difference in change from baseline to 6-month was also significant (p < 0.01). In addition, 25% of the treatment group reported eating brown rice often or almost always at 6-month compared to 4% at baseline (p < 0.01); no individuals reported eating brown rice among the control, at either time point.

Diabetes knowledge increased significantly for both groups. Results were run using two scales. The treatment group showed a greater increase in knowledge for the ADA scale (6.5 vs. 4.8 at 6-month), while both groups showed similar improvements for the Michigan scale (3.2 vs. 3.1 at 6-month). The between-group differences from baseline to 6-month in both diabetes scales were statistically significant (p < 0.01).

Overall, treatment group participants were satisfied with the intervention and found the CHWs to be useful in helping them to answer questions, facilitating access to a doctor, and changing their behaviors (Table 5). Everyone reported that their CHW treated them with respect and dignity all of the time, while lower percentages were shown for doctors (86%) and other health professionals (62%). Eighty-eight percent of participants were totally satisfied with the intervention, and everyone was at least very satisfied.

Findings from the qualitative data provided contextual insights that expand understanding of the main findings from the quantitative pilot results. Qualitative data further inform the feasibility issues that emerged during the pilot, including recruitment, retention, and the organization and implementation of the pilot. As discussed above, recruitment was facilitated by building on strong relationships with respected Sikh organizations in the community and by the principles of “seva”, which is a call for volunteerism and community service in the Sikh religion, which enabled CHWs to benefit from help provided by volunteers at recruitment events and in getting the word out about the project. The analysis of the qualitative data documented several barriers to recruitment. The main challenges to recruitment that CHW and focus group discussion revealed were: (1) Community members felt that if they were free of symptoms of diabetes, they did not need to take steps to prevent the disease; and (2) community members felt that diabetes was a natural part of aging and could not be prevented. As the CHW supervisor commented: “People say there’s a need for the program but don’t show up. They say, “It’s not a problem for me”. Prevention is not a priority for the community; they have the attitude: “I’ll deal with it when I get the disease”.”. The CHW supervisor further noted that community members seemed more motivated to participate in the intervention if they knew someone with diabetes or if the screening showed them to have an at-risk BMI: “When I tell them their BMI, it is eye-opening and motivates people to join the program”. Indeed, the reasons that focus group participants offered as their motivation to join the program was less about preventing disease than a result of pre-existing issues, such as previous health problems (“I had some (health) problems so I thought I would go and learn something”); the desire to lose weight (“I had some extra weight; I came because of the weight”); and prior gestational diabetes or elevated blood sugar (“I had diabetes when I was pregnant and the doctor told me I could develop diabetes; I also had borderline high blood sugar”). Another recruitment challenge noted by the CHW supervisor was gender-related. She observed that it was difficult to enroll men into the intervention due to their long work hours.

Retention was facilitated by trust in the CHW and consistent follow-up, particularly by CHWs living within the community where the pilot intervention was held. The CHW supervisor’s comments provided insights on challenges that affected program retention. These issues included the extended travel by participants to India during the intervention period; difficulty scheduling education sessions and data collection meetings that would accommodate the participants’ schedule; childcare and family obligations, particularly among woman, that prevented several women from attending sessions on the arranged date and time; and long work hours, particularly among men, which prevented participating in multiple sessions.

In regards to intervention organization and implementation, attempts to facilitate program participation in order to address the scheduling barriers noted above included: hosting sessions in community locations convenient for participants, during weekends and mid-day when women have a break from household responsibilities and childcare, and offering session make-ups. The majority of the focus group’s comments indicated that the participants appreciated the CHWs’ consistent follow-up and willingness to offer make-up sessions: “(The CHW) would call, remind about classes, and make sure we attended. Even when you couldn’t make it she would make it up. She’s very nice. She would make sure we attended all the sessions”. Some of the comments offered by the participants even suggested a preference for classes more often, saying: “It’s better to have them every 2 weeks. Whenever we come we learn something. If sessions are every 4 weeks we would forget everything” and “That’s right, the sooner you call us the quicker we will learn”. The CHWs’ reports revealed that scheduling challenges were a frequent issue, as face-to-face meetings were required to collect the survey questionnaires and clinical measurements in addition to asking the program participants to attend six group sessions.

The qualitative data also provided insight into facilitators and barriers to improving health outcomes and behaviors. A number of the comments made by the focus group participants reinforced quantitative findings that indicated the intervention improved knowledge on diabetes prevention and health behaviors, as illustrated in the following statement offered by one participant, “I got to learn a lot and practice what I learned. I watched my diet and took care to keep diabetes away. I had borderline high sugar but I am ok now”. Another benefit that several participants felt they received from the program was the program’s focus on prevention was helpful in increasing knowledge and motivating healthy behavior change. As one participant explained: “A person who has high blood pressure generally knows what to eat and what not to eat. Not in too much detail though. That I found out about and experienced here. Like I used to eat a lot of sweets, I reduced that quite a bit. You get scared of developing diabetes”. Similarly, a few participants also highlighted the value of the curriculum’s focus on monitoring health outcomes in helping to motivate them to further behavior change. This sentiment is reflected in the following observation made by one participant: “The knowledge that we received…we were able to implement. For example, we would come to class and learn. Then we would go home and moderate our diet. Otherwise we wouldn’t pay attention and eat what’s made. Additionally there were the (blood) tests, we would find out what’s going on with us (our numbers)…”. Qualitative data also explicated how the participants experienced the social support derived from their program participation was both an important motivator for engaging in the program and also played a role in helping some of these participants make lifestyle changes. For example, one participant shared that she initially joined “for fun” before becoming more serious about participating: “When (the CHW) told us to attend we thought “let’s go”. We would just sit outside chatting in the evenings anyway. We just came for fun. But when we attended, we got serious”. Similarly, another participant explained further that: “sometimes we call each other and discuss what to eat or what not to eat. Even if we are eating it but we will tell each other not to”.

The qualitative data also provided further illumination on the challenges that participants experienced in terms of changing health behaviors. Although many of the participants reported that the knowledge they received from the intervention was helpful, some participants also noted that their social contexts sometimes prevented them from implementing the changes learned through the intervention. As one participant observed, “We come from big families so that's the catch”, explaining that dietary changes were harder to make at the family level. The CHWs’ statements also contained the observed that in general goal-setting exercises were challenging for participants. As the CHW supervisor noted: “It is hard for people to set goals. People say, “You give me a goal and I’ll do it”. But they are unwilling to set their own goals. They ask for an example, and if I give them one, they will just use the example”.

The qualitative data expanded upon the importance of cultural relevance and use of the CHW model in the acceptability of the intervention. The project CHWs shared participants’ cultural backgrounds and language, and were able to leverage their knowledge of community resources and religious networks (e.g., gurdwaras and prayer groups) to increase their outreach efforts and ensure the relevance and acceptability of the study to the community. There were a number of features of the CHW model approach, which promotes experiential- and co- learning, that participants found satisfactory. Some of the participants expressed satisfaction with the group format: “It was good. We were sitting in a group, asking and talking with each other, we learned a lot”. Another positive aspect of the program that a few participants commented upon was their relationship with the CHWs, some of these comments described this relationship in warm, familial terms: “She was very, very good. She explained things in a way that made me feel like a family member”. As is custom in the Punjabi language, all participants and CHWs referred to each other in their statements as “bhenji”, meaning “sister”, reflecting congruency with the principle of peer-learning inherent in the CHW model approach. Another further feature that impressed most of the participants was that the program was culturally-tailored for the Sikh community. As one participant shared: “(The CHW) told us (the program) was made specially (sic) for our community”. In addition, a number of participants noted that they were appreciative that this Centers for Disease Control and Prevention-funded program demonstrated an interest by the government in learning more about the Sikh community and in promoting health, as the following statements reflect: “It was good to know (the government is) interested in getting the knowledge” and “The government is concerned in our health, is taking care of us”. The project’s connection to a reputable Sikh organization in the community as well as to an academic medical school facilitated trust of the community; and securing the support of the gurdwara leadership was critical to acceptance by the community. The CHW supervisor noted: “I’ve learned the importance of working closely with the gurdwaras, which are at the heart of the community, and where Sikhs share many meals together. We have been fortunate to have the support of the Sikh Cultural Society, Inc., a well-established gurdwara in the community”.