Healthy Immigrant Community Study Protocol: A Randomized Controlled Trial of a Social Network Intervention for Cardiovascular Risk Reduction among Hispanic and Somali Adults

Abstract

Background:

Immigrants to the United States face structural barriers that contribute to rising cardiovascular risk factors and obesity after immigration. This manuscript describes the development of the Healthy Immigrant Community protocol and baseline measures for a stepped wedge cluster randomized trial to test the effectiveness of a social network intervention for cardiovascular risk reduction among two immigrant populations.

Methods:

We developed a social network-informed, community-based, participatory research-derived health promotion intervention with Hispanic and Somali immigrant communities in Minnesota consisting of mentoring, educational and motivational sessions, group activities, and a community toolkit for healthy weight loss delivered by culturally concordant health promoters (HPs) to their social networks. Using a stepped wedge cluster randomized design, social network-based groups were randomly assigned to receive the intervention either immediately or after a delay of one year. Outcomes, measured at baseline, 6 months, 12 months, and 24 months, were derived from the American Heart Association’s “Life’s Simple 7”: BMI and waist circumference, blood pressure, fasting blood glucose, total cholesterol, physical activity level, and dietary quality.

Results:

A total of 51 HPs were enrolled and randomized (29 Hispanic; 22 Somali). There were 475 participants enrolled in the study, representing a mean social network group size of 8 (range, 5–12). The mean BMI of the sample (32.2) was in the “obese” range.

Conclusion:

Processes and products from this Healthy Immigrant Community protocol are relevant to other communities seeking to reduce cardiovascular risk factors and negative health behaviors among immigrant populations by leveraging the influence of their social networks.

1. Background

By many health measures, immigrants arrive in the United States (US) healthier than the general US population.[1] However, the longer immigrants reside in the US, the more they approximate the high health risk profiles of the general population for cardiovascular disease and related chronic conditions.[2, 3] Approximately 15 years after immigration, the prevalence of obesity among immigrants to the US approximates that of US-born adults.[4] Among immigrant populations to high-income nations, dietary and physical activity behaviors are less healthy than the non-immigrant majority populations.[5, 6]

Evidence-based health promotion programs in group settings have been shown to be effective at improving health behaviors in general populations,[7] but despite calls for interventions to address obesity and health behaviors among immigrant populations,[8] few intervention studies have been reported.[9] One novel approach is to utilize social networks to promote a healthy lifestyle. A range of social network characteristics are highly associated with obesity-related health behaviors in the general populations (nutritional intake and physical activity).[10] These findings apply to obesity; for example, a 32-year longitudinal study showed a person’s chances of becoming obese increased by 57% if their network partners (friends, family, neighbors, etc) became obese, which was mediated by the frequency and intensity of interactions.[11] While a causal link between social network factors and obesity cannot be demonstrated by a single study, a body of evidence has emerged that supports the contribution of social network constructs to the obesity epidemic and cardiovascular risk accumulation in the US.[12–15]

Social network interventions involve purposeful use of existing social networks in the individual’s natural environment to promote positive behavior change and health outcomes.[16] A systematic review of network interventions reported consistent findings of positive behavior change with a social network approach from a variety of settings and across a range of health outcomes.[17] Interventions incorporating social networks for diet and/or weight loss have been identified as a promising and innovative approach.[13] Simulation models suggest traditional weight loss interventions frequently fail because they lack consideration of the participant’s social networks, and that network-driven interventions may be highly effective.[18]

Community based participatory research (CBPR) is a means to collaboratively investigate health topics, whereby community members and academics partner in an equitable relationship through all phases of research.[19] This approach can address the interplay between health behaviors and the social determinants of health because it empowers communities, promotes understanding of culturally pertinent issues, and targets the multi-faceted barriers to health.[20] This study builds on our team’s experience with CBPR through the Rochester Healthy Community Partnership (RHCP), a 19-year community-academic partnership.[21] RHCP has become productive and experienced at deploying data-driven programming with community partners from immigrant populations.[22]

The problem of cardiovascular risk accumulation and weight gain after immigration has been an RHCP priority for the past 15 years. In the Healthy Immigrant Families (HIF) study, RHCP partners implemented a nutrition and physical activity intervention with immigrant families through a CBPR approach.[23] Results showed improvement in dietary quality among adults.[24], but there was no impact on physical activity level. Participants noted in post-intervention focus groups that a more flexible group-based approach with friends and family might promote greater participation, and they recommended incorporating weight management strategies into the intervention. After reviewing those findings in meetings and at an RHCP research summit, community and academic partners embraced the conceptual pivot to an intervention delivered through social networks.

RHCP then conducted an analysis of social networks to identify factors related to health behaviors and obesity among 691 Somali and 610 Hispanic participants in Rochester, MN, through a survey instrument and body mass index (BMI) assessment. Results demonstrated that overweight and obesity clustered by ascertainable social networks for obesity and obesity-related behaviors.[25, 26] Among participants who were overweight or obese, the number of social contacts trying to lose weight was associated with weight loss intentions, and this effect was mediated by social norms for obesity, social support, and social cohesion.[25, 26] These results indicated that social contacts and normative beliefs influence weight status and intentions for weight control and cardiovascular risk reduction, thereby highlighting the importance of targeting social influence in treatment of overweight and obesity and the prevention of cardiovascular disease in these high risk groups.

RHCP adapted the HIF intervention framework to target weight loss among overweight and obese adults, and to provide the flexibility for delivery by community peer interventionists. The 12-week pilot intervention was tested through a single-arm pilot study with 4 social networks of adults (2 Hispanic, 2 Somali) with 39 participants, who highly rated the intervention on satisfaction, motivation, and confidence to eat a healthy diet, be physically active, and lose weight. On average, participants lost weight, lowered their blood pressure, had more servings of vegetables per day, and increased their physical activity level.[27]

RHCP community and academic partners reflected on the results of the formative research, and collectively derived the Healthy Immigrant Community (HIC) protocol presented here.

2. Materials and Methods

The study design and reporting are aligned with the Consolidated Standards of Reporting Trials (CONSORT) statement[28] and the study is registered with the Clinical Trials Registry (NCT05136339). Protocol reporting follows the Guidelines for Reporting Outcomes in Trial Protocols (SPIRIT-Outcomes 2022 Extension)[29]

2.1. Study team

RHCP community and academic partners have collectively derived the approach represented in this protocol. Upon receipt of funding, RHCP developed 5 working groups, each consisting of community and academic members: Recruitment; intervention materials; measurements; community resources; and communications. Representatives from each group reported back to the RHCP HIC project group every 2 weeks and to the broader RHCP partnership every 2 months.

2.2. Study design

This is a randomized trial to assess the efficacy of a social network-informed CBPR-derived health promotion intervention on obesity and cardiovascular risk factors in two immigrant communities – Hispanic and Somali. Using a stepped wedge cluster randomized design, social network-based groups were randomly assigned to receive the intervention either immediately (step 1) or after a delay of one year (step 2). Outcomes, measured at baseline, 6 months, 12 months, and 24 months, were derived from the American Heart Association’s “Life’s Simple 7”: BMI, waist circumference, blood pressure, fasting blood glucose, total cholesterol, physical activity level, and dietary quality. The intervention consists of community-based mentoring, educational and motivational sessions, group activities, and application of a community toolkit for healthy weight management delivered by trained health promoters to their social networks. During the control period, participants received printed health information developed by the study team during each assessment that is not directly related to cardiovascular risk reduction.

Rationale for addressing obesity.

Minnesota Community Health Needs Assessments conducted in 2016 demonstrated that obesity was a priority for intervention. Survey items included in the RHCP social network analysis (above) indicated that dietary quality, physical activity, and obesity were among the top health priorities for Hispanic and Somali populations. These concerns mirror national trends reflecting increasing obesity and associated co-morbidities among immigrants after arrival to the US. Finally, 80% of adult participants in the HIF project and 66% of participants in the social network analysis were overweight or obese. Therefore, RHCP partners concluded that it is appropriate to focus on health behaviors relevant to obesity in the next phase of work.

Rationale for including Hispanic and Somali populations.

RHCP has a history of collaborative work between many different immigrant groups (and non-immigrant Latinos) towards common goals. The 2017 RHCP partnership evaluation highlighted the positive impact of this unique collaboration on community building, collective advocacy, and health outcomes. For the Healthy Immigrant Families project and for the social network analysis, community leaders and community-based organizations from Hispanic and Somali populations took leadership roles. The work with the two groups has occurred both in parallel and in collaboration, such that best practices and learning opportunities have been shared throughout the process. Furthermore, these represent the two largest immigrant groups to Rochester, MN. If successful, this approach could be used as a model framework for social network level interventions in other populations.

2.3. Setting

Rochester, Minnesota, is a medium-sized metropolitan area in the southeastern part of the state. According to 2022 Census data, 13.6% of residents are foreign-born. The HIC project included community partners from the Hispanic and Somali communities, who recruited participants from Rochester and surrounding communities.

2.4. Participants

Eligibility.

Eligibility criteria included (1) self-identification as Hispanic or Somali immigrant, (2) member of a social network linked to a HP, (3) age ≥ 18, (4) willingness to participate in all aspects of the study; and (5) provision of oral informed consent. Exclusion criteria include (1) pregnancy at the time of enrollment and (2) serious medical conditions or disabilities that, per the participant’s perspective, contraindicate being physically active. To avoid stigmatization, having a healthy BMI (18< BMI < 25) did not exclude individuals from participating in the intervention.

Based on our social network data, we did not restrict the participants to a particular age group, as there was no significant clustering found by age. Likewise, network data, and community feedback helped to determine we did not need to restrict Hispanic groups to gender concordance, but Somali groups were gender-concordant (except for social networks with family members), due to cultural and religious practices. While this study prioritized foreign-born immigrant populations, US-born members of social networks were not excluded.

Sample size.

We expected each HP to recruit approximately 7 participants from their social networks and a loss to follow up of 20% over 12 months. We assumed a conservative intra-cluster correlation in outcomes of $\mathrm{\rho }=0.10$ to calculate the design effect ${\mathrm{D}\mathrm{E}}_{SW}$ based on the formula:[30, 31]

$$\mathrm{D}\mathrm{E}\mathrm{s}\mathrm{w}=\frac{1+\mathrm{p}(\mathrm{k}\mathrm{t}\mathrm{n}+\mathrm{b}\mathrm{n}-1)}{1+\mathrm{p}(1/2\mathrm{k}\mathrm{t}\mathrm{n}+\mathrm{b}\mathrm{n}-1)}\cdot \frac{3(1-\mathrm{p})}{2\mathrm{t}(\mathrm{k}-1/\mathrm{k})}$$

where $\mathrm{k}=2$ is the number of steps (including baseline), $\mathrm{b}=1$ is the number of measurements per participant in the baseline step, $\mathrm{t}=1$ is the number of measurements per participant at each step, and $\mathrm{n}=7$ is the number of participants per cluster. Assuming 80% follow up and equal distribution across clusters, and 80% power, we estimated the detectable standardized effects (SDs) for a range of numbers of enrolled HPs with a maximum probability of a Type II error of $\mathrm{\alpha }=0.05$.

For the primary endpoints, a third standard deviation (0.3SD) difference between groups was considered clinically meaningful.[32] On the basis of the HIF study, we assumed a 0.1 standard deviation change in the control group (“contamination”) so this represents a 0.4SD change in the intervention group. Based these assumptions and methods we planned to accrue 54 clusters (27 per group) with seven participants in each cluster, for a total sample size of approximately 378 network participants.

2.5. Recruitment, enrollment, randomization, and blinding

Recruitment.

Peer interventionists, the Health Promoters (HPs), were recruited by RHCP community partners who approached community members who were known to be opinion leaders about health. RHCP community partners also used the list of opinion leaders identified in the social network analysis for targeted recruitment of HPs.

HPs then invited members of their social networks to participate in the intervention. The network was limited to a maximum of 12 participants per HP. Those recruited by HPs were invited to enrollment events at one of two community-based locations, where eligibility was confirmed, oral informed consent was completed, and baseline measures were obtained.

Randomization.

The units of randomization were social network subgroups. Randomization occurred after all HPs and their social network participants were enrolled and baseline measures completed. Randomization was performed using constrained randomization, with the final assignment selected from a set of allocations balanced on ethnicity (Hispanic or Somali), age, group size, and sex of the HPs.

Blinding.

Study staff performing measurements and named investigators were blinded to treatment condition.

2.6. Participant retention

Study participants received four $75 gift cards, one for each measurement time. HPs received $100 for participating in study orientation (an evening training session), $600 for training (meeting every other week with the health coaches) and delivering the intervention to their network during the first 6 months, and $300 for sustaining the intervention in the second 6 months.

2.7. Theoretical approach

Social Ecological theory postulates how structural and functional characteristics of social groups (e.g., social networks)[33] influence behaviors through several pathways,[13] including shared social activities and social support. Social networks can also indirectly influence health behaviors related to broader conditions for health by addressing modifying and mediating conditions for healthy behaviors.[34] Mediating and modifying factors impacting diet and physical activity including self-efficacy and social support among Hispanic and Somali immigrants to Southeast Minnesota were elucidated in our formative work,.[35, 36] Higher social cohesion, i.e., close relationships among community members with strong mutual trust and reciprocity[37], has been associated with a range of positive health outcomes.[38] For this study, the empiric relationship between cohesion as a mediator of group (e.g., social network) formation, maintenance, and productivity[39] is especially relevant. Group cohesion is an established network-level metric[40] that has been used to assess network intervention implementation progress.[41, 42]

2.8. Interventions

The intervention consists of community-based mentoring and education delivered by trained, culturally and linguistically congruent Hispanic and Somali HPs to their social networks. Over a period of 6 months, HPs facilitated 12 face-to-face or virtual group visits. Each of the twelve sessions targeted a particular behavior for weight loss and/or reduction of cardiovascular risk (Table 3). All sessions included goal setting, review of food and/or activity tracking, and positive reflections. Specific strategies for weight loss included regular weigh-ins, completion of food and physical activity records, reducing portion sizes, planning meals, use of MyPlate[43] for dietary proportions, removing problem foods from the home, increasing physical activity level, and reducing sedentary behavior.

Consistent with the theoretical framework, HPs provided support and monitor progress on goals via six evidence-based strategies: Tracking, Goal Setting, Mindfulness, Social Support, Problem Solving Skills, and Motivational Strategies.[44] Participants were informed that aims of the intervention include weight loss of 3% in overweight or obese individuals, reduction of portion sizes of calorie dense foods, a shift of dietary quality to increase fruit and vegetable consumption, and 150 minutes per week of moderate to vigorous physical activity.

Over the subsequent 6 months, HPs facilitated an additional 12 virtual or face-to-face weight maintenance sessions to focus on goal setting, social support, motivation for lifestyle changes, review of food and/or activity tracking, and reflections (i.e., no new content).

Intervention guide.

RHCP developed an HP intervention guide that was be printed, and available electronically. The HP Guide consists of 2 sections: Section 1 includes 6 pillars of behavior change.[44] Section 2 includes the 12 session content modules. HPs were also provided with supplementary information that included culturally appropriate recipes with nutrient and calorie information and a physical activity checklist; a Community Resource booklet with a list of low or no cost community resources for physical activity and healthful eating; and HIC Newsletters (one-page document highlighting community success stories).

Training of Health Promoters.

We have previously described the training procedures for interventionists for the HIF study.[45] For this project, language congruent Health Coaches (or study coordinators) met with intervention oversight clinicians (a registered dietitian (KZ), an exercise specialist (TR) and a health psychologist (MMC) with expertise in weight management) every two weeks for one hour to review the content for the next session and to answer questions. In a train the trainer model, the Health Coaches then met with HPs every one to two weeks for up to one hour to provide real time training on the content for upcoming group sessions and to answer questions. HPs were able to ask questions and make suggestions to the Health Coaches, which the Health Coaches communicated to the oversight clinicians.

2.9. Treatment Fidelity

In the design and conduct of this study, we have incorporated recommendations from the Treatment Fidelity Workgroup.[46] We previously described our process for assessing treatment fidelity for community-based behavior change interventions with immigrant groups in a CBPR framework.[45] The HPs utilized a manual containing an outline for each session and a checklist of critical topics to be covered at each session. Health Coaches, who were trained in fidelity assessment for this study, randomly observed at least 10% of sessions with the HPs. Bi-weekly meetings were held with the HPs and health coaches to reinforce treatment fidelity, provide corrective feedback, and to conduct additional training if needed.

2.10. Data collection and outcomes

Data were collected at 0 (baseline), 6, 12 (primary endpoint) and 24 months by trained study staff and volunteers. Participants completed a survey and participated in physical measurements.

Demographic, acculturation, and discrimination measures.

The following demographic data were collected at baseline: age, gender, ethnicity, country of birth, annual household income, education level, and employment status. Participants will report on the following proxies for acculturation: years lived in the US, primary language spoken at home, and level of English language proficiency on a 5-point Likert scale.[47] The everyday discrimination scale[48] was used at baseline to assess perceived discrimination.

Primary outcome measures.

BMI:

Weight was measured to the nearest 0.1 kg using a portable scale (Seca 880 Digital Floor Scale). Height was measured to the nearest 0.1 cm at baseline only using a stadiometer. Participants were asked to remove shoes prior to both measurements. Waist circumference was measured to the nearest 0.1 cm at the narrowest part of the torso between the ribs and the iliac crest. Participants removed excess clothing prior to the measurement and smooth the remaining clothing against the skin.

Secondary outcome measures.

The secondary outcome measure were the “Life’s Simple 7” Composite Score, which is adapted from the American Heart Association standards based on health assessment data[49]: blood pressure, fasting lipid panel and glucose, BMI, self-reported cigarette smoking status, dietary quality, and physical activity patterns. Point values are assigned to each component: 2 points for ideal, 1 point for intermediate, 0 points for poor. The total sum allows for a continuous measure of cardiovascular health ranging from poor to ideal (0–14 points). The final score will be categorized as 0–6 (poor), 7–8 (intermediate), 9–14 (ideal). In previous studies, adherence to the ideal Life’s Simple 7 metrics was found to be associated with decreased risks of cardiovascular disease[50] and all-cause mortality[51].

Seated blood pressure (systolic and diastolic) measurements were made on the right arm using an automated device after sitting quietly for five minutes.[52] Blood pressure was measured three times and the average of the second and third readings was used in analyses. Fasting glucose and cholesterol was collected by a single finger prick. The portable Whole Blood Lipid Screen Cholestech LDX Analyzer was used to analyze specimens.

Dietary Quality.

A 24-hour dietary recall was collected at each measurement using the Automated Self-administered 24-hour Recall (ASA24) system, a National Cancer Institute web-based tool that enables multiple automated 24-hour recalls. The recalls were performed using a computer under the supervision of study staff and with an interpreter, as needed. Through these intake estimates, the effect of the intervention on diet quality (Healthy Eating Index plus intake of fruits, vegetables, sweetened beverages, percent energy from fat and total calorie consumption) may be examined.

Physical activity.

The International Physical Activity Questionnaire (IPAQ) [53] was used to assess the number of minutes of mild, moderate, and vigorous physical activity over the previous 7 days.

Health-related Quality of Life was measured by single-item linear analog scale assessments of physical, emotional, and overall health developed by members of our research team[57, 58].

Theory-based measures.

Survey items were used to assess self-efficacy and social support for diet and physical activity that were adapted for our HIF study from instruments developed for the Patient-centered Assessment and Counseling for Exercise plus Nutrition (PACE+) program for low-income, ethnically diverse adults[59] that included nutrition self-efficacy[60] and social support,[61] physical activity self-efficacy[60] and social support.[61] Test-retest reliabilities and internal consistency reliabilities were acceptable for all of these constructs, and association with behaviors suggested criterion-related validity[60, 61]. A 6-item measure of group social cohesion with excellent reliability (Cronbach alpha 0.95)[62] used in previous network intervention studies and in our pilot studies[41] was used. Number of social contacts trying to lose weight and social norms for weight loss was assessed using a 4-item measure with good to excellent reliability and item loadings.[63]

For survey instruments available only in English, we edited the English-language version of each item with community partners from both groups for cultural relevance, followed by forward-translation, panel discussion, backward translation, a pre-test, a cognitive briefing and consensus on the final version by a core group of community members from each group [64]. We have previously described this process of adapting the World Health Organization translation procedure for use with survey instruments in a CBPR framework.[65]

2.11. Data analysis

All analyses are appropriate for cluster randomized stepped wedge trials. In the direct effects model, the co-primary endpoints are changes in BMI and waist circumference from baseline to month 12. We will use mixed effects regression models to adjust for correlations within social network clusters and over time.[66] Models will include indicators for intervention status and (to account for secular effects) study step. Covariates indicated above including demographics and social covariates will be included in profile analysis models to assess the sensitivity of the primary analysis conditional on these potential concomitant (mediating) influences. All secondary endpoints will be analyzed using the same methods as used for the primary endpoints.

To account for missing data, simple imputation methods such as minimum, average, or maximum values carried forward will be followed by multiple imputation methods to assess the robustness of the study results in the presence of the missing data. All models will be assessed for goodness of fit, with effects reported with confidence intervals.

2.12. Ethics and monitoring

Prior to study initiation, the Mayo Clinic IRB approved the protocol. All participants provided oral informed consent and a signed HIPAA document prior to study participation. The project manager monitors quality control by reviewing data and conducting annual audits. Confidentiality is ensured by assigning participants a de-identified study number. The association between the study number and the participant identity is kept by study staff in a locked file cabinet, and this information will be destroyed after completing the study. All participant data will be securely stored in a locked file cabinet. All data elements and transfer activities are strictly compliant with the Health Insurance Portability and Accountability Act.

2.13. Dissemination

Consistent with CBPR principles, dissemination of research findings will be a collaborative, multi-pronged effort by academic and community partners at RHCP meetings, community forums, and academic conferences as well as in peer-reviewed publications.

2.14. Positionality of researchers

The research team includes Black, Indigenous, People of Color, and White faculty, staff, and community partners. All team members have experience and knowledge about the Hispanic and/or Somali community and health inequities through lived experience and/or scholarship.

3. Results

A total of 51 HPs were enrolled and randomized (29 Hispanic; 22 Somali). There were 475 participants enrolled in the study linked to these 51 HPs, representing a mean social network group size of 8 (range, 5–12). Table 2 reports the characteristics of the participants by intervention group (early versus delayed), along with the standardized mean differences (SMDs) for each. The only SMD > 0.2 was for self-reported consumption of fruits and vegetables as snacks, which was somewhat more common in the early intervention group. The mean BMI of the sample (32.2) was in the “obese” range.

Table 2.

Characteristics of Healthy Immigrant Community study participants

Characteristic	Early Intervention (N=246)	Delayed Intervention (N=229)	All (N=475)	SMD
Age, mean (SD)	44.2 (15.1)	44.7 (15.0)	44.5 (15.0)	−0.037
Gender, N (%)				0.035
Male	93 (37.8%)	87 (38.0%)	180 (37.9%)
Female	139 (56.5%)	134 (58.5%)	273 (57.5%)
Other	11 (4.5%)	6 (2.6%)	17 (3.6%)
Ethnicity, N (%)				−0.039
Hispanic	146 (59.3%)	129 (56.3%)	275 (57.9%)
Somali	96 (39.0%)	97 (42.4%)	193 (40.6%)
Other	3 (1.2%)	1 (0.4%)	4 (0.8%)
Education, N (%)				0.139
8th grade or less	60 (24.4%)	61 (26.6%)	121 (25.5%)
Some High School	35 (14.2%)	36 (15.7%)	71 (14.9%)
High School graduate or GED	62 (25.2%)	63 (27.5%)	125 (26.3%)
Some College	46 (18.7%)	41 (17.9%)	87 (18.3%)
College or Graduate Degree	41 (16.7%)	26 (11.3%)	67 (14.1%)
Household Income, N (%)				−0.041
Less than $10,000	55 (22.4%)	41 (17.9%)	96 (20.2%)
$10,000 to $19,999	23 (9.3%)	32 (14.0%)	55 (11.6%)
$20,000 to $29,999	43 (17.5%)	46 (20.1%)	89 (18.7%)
$30,000 to $39,999	40 (16.3%)	30 (13.1%)	70 (14.7%)
$40,000 to $49,999	27 (11.0%)	19 (8.3%)	46 (9.7%)
$50,000 to $74,999	30 (12.2%)	29 (12.7%)	59 (12.4%)
$75,000 or higher	15 (6.1%)	19 (8.3%)	34 (7.2%)
Employment Status, N (%)				−0.019
Full time	109 (44.3%)	101 (44.1%)	210 (44.2%)
Part time	61 (24.8%)	54 (23.6%)	115 (24.2%)
Unemployed, retired, or disabled	75 (30.5%)	73 (31.9%)	148 (31.2%)
Health Insurance, N (%)				0.138
Yes	147 (59.8%)	150 (65.5%)	297 (62.5%)
No	98 (39.8%)	75 (32.8%)	173 (36.4%)
Born in United States, N (%)	22 (8.9%)	29 (12.7%)	51 (10.7%)	0.149
Father born in U.S., N (%)	6 (2.4%)	4 (1.7%)	10 (2.1%)	−0.045
Mother born in U.S., N (%)	4 (1.6%)	3 (1.3%)	7 (1.5%)	−0.026
Year arrived to U.S., N (%)				0.188
1970–1995	25 (10.2%)	31 (13.5%)	56 (11.8%)
1996–2000	33 (13.4%)	34 (14.8%)	67 (14.1%)
2001–2005	40 (16.3%)	36 (15.7%)	76 (16.0%)
2006–2015	59 (24.0%)	42 (18.3%)	101 (21.3%)
2016–2022	36 (14.6%)	27 (11.8%)	63 (13.3%)
Language spoken at home, N (%)				0.190
English	20 (8.1%)	21 (9.2%)	41 (8.6%)
Somali	61 (24.8%)	71 (31.0%)	132 (27.8%)
Spanish	113 (45.9%)	91 (39.7%)	204 (42.9%)
Other	6 (2.4%)	1 (0.4%)	7 (1.5%)
Self-rated English language speaking ability, N (%)				−0.051
Not at all	51 (20.7%)	42 (18.3%)	93 (19.6%)
Not well	86 (35.0%)	82 (35.8%)	168 (35.4%)
Well	52 (21.1%)	50 (21.8%)	102 (21.5%)
Very well	53 (21.5%)	52 (22.7%)	105 (22.1%)
Everyday discrimination scale score, mean (SD)a	9.6 (4.3)	9.8 (4.9)	9.7 (4.6)	−0.057
Perceived discrimination categoriesb, N (%)
Ancestry/Origin	56 (22.8%)	51 (22.3%)	107 (22.5%)	0.012
Language	84 (34.1%)	69 (30.1%)	153 (32.2%)	0.086
Gender	14 (5.7%)	14 (6.1%)	28 (5.9%)	−0.018
Race	57 (23.2%)	56 (24.5%)	113 (23.8%)	−0.030
Religion	29 (11.8%)	28 (12.2%)	57 (12.0%)	−0.013
Weight	43 (17.5%)	25 (10.9%)	68 (14.3%)	0.188
Appearance	18 (7.3%)	18 (7.9%)	36 (7.6%)	−0.020
Sexual Orientation	7 (2.8%)	0 (0.0%)	7 (1.5%)	0.242
Education/Income	25 (10.2%)	20 (8.7%)	45 (9.5%)	0.049
BMI, mean (SD)	31.8 (7.1)	32.6 (6.0)	32.2 (6.6)	0.117
Total cholesterol, mg/dL, mean (SD)	178.7 (40.7)	178.6 (40.7)	178.7 (40.7)	0.001
HDL cholesterol, mg/dL, mean (SD)	43.4 (14.8)	42.8 (13.5)	43.1 (14.2)	0.048
Triglycerides, mg/dL, mean (SD)	162.2 (116.9)	158.8 (114.6)	160.5 (115.7)	0.030
LDL cholesterol, mg/dL, mean (SD)	107.1 (35.0)	108.9 (32.8)	108.0 (34.0)	0.054
Glucose, mg/dL, mean (SD)	108.1 (41.8)	107.6 (43.3)	107.9 (42.5)	0.011
Waist circumference, cm, mean (SD)	101.7 (14.7)	101.8 (12.9)	101.7 (13.8)	0.008
Systolic blood pressure, mm Hg, mean (SD)	125.5 (18.1)	125.4 (17.6)	125.5 (17.9)	0.008
Diastolic blood pressure, mm Hg, mean (SD)	81.6 (10.4)	81.4 (11.2)	81.5 (10.8)	0.015
Eat fruits and vegetables as snacks, N (%)				0.238
No	15 (6.1%)	20 (8.7%)	35 (7.4%)
Sometimes	132 (53.7%)	140 (61.1%)	272 (57.3%)
Often	48 (19.5%)	38 (16.6%)	86 (18.1%)
Everyday	46 (18.7%)	27 (11.8%)	73 (15.4%)
Fruit drinks, punch, or sports drinks, N (%)				−0.118
No	101 (41.1%)	77 (33.6%)	178 (37.5%)
Sometimes	114 (46.3%)	121 (52.8%)	235 (49.5%)
Often	19 (7.7%)	21 (9.2%)	40 (8.4%)
Everyday	8 (3.3%)	7 (3.1%)	15 (3.2%)
Regular soda drinks, N (%)				−0.024
No	96 (39.0%)	87 (38.0%)	183 (38.5%)
Sometimes	111 (45.1%)	105 (45.9%)	216 (45.5%)
Often	29 (11.8%)	24 (10.5%)	53 (11.2%)
Everyday	9 (3.7%)	11 (4.8%)	20 (4.2%)
Servings of fruits per day, mean (SD)	1.9 (1.2)	1.7 (1.4)	1.8 (1.3)	0.155
Servings of vegetables per day, mean (SD)	1.8 (1.4)	1.7 (1.4)	1.8 (1.4)	0.030
Healthy Eating Indexc, mean (SD)	53.5 (14.4)	52.0 (14.5)	52.8 (14.5)	0.099
Perceived access to healthy foods, N (%)				0.062
Never	60 (24.4%)	57 (24.9%)	117 (24.6%)
Once in a while	62 (25.2%)	58 (25.3%)	120 (25.3%)
Sometimes	57 (23.2%)	63 (27.5%)	120 (25.3%)
Often	41 (16.7%)	31 (13.5%)	72 (15.2%)
Always	26 (10.6%)	20 (8.7%)	46 (9.7%)
Perceived access to safe places for physical activity, N (%)				0.177
Never	34 (13.8%)	44 (19.2%)	78 (16.4%)
Once in a while	50 (20.3%)	62 (27.1%)	112 (23.6%)
Sometimes	92 (37.4%)	59 (25.8%)	151 (31.8%)
Often	38 (15.4%)	38 (16.6%)	76 (16.0%)
Always	32 (13.0%)	24 (10.5%)	56 (11.8%)
IPAQ Score, mean (SD)d	2799.0 (3879.1)	2471.2 (2865.0)	2651.6 (3458.4)	0.096
IPAQ Category, N (%)e				−0.054
Inactive	37 (15.0%)	25 (10.9%)	62 (13.1%)
Minimally Active	60 (24.4%)	54 (23.6%)	114 (24.0%)
Active	67 (27.2%)	55 (24.0%)	122 (25.7%)
Active Smoked 100+ cigarettes in lifetime, N (%)	33 (13.4%)	47 (20.5%)	80 (16.8%)	0.188
Current cigarette smoker, N (%)	13 (5.3%)	20 (8.7%)	33 (6.9%)	0.172
Social cohesionf	5.7 (1.3)	5.8 (1.3)	5.8 (1.3)	−0.052
Social support for healthy eating, mean (SD)g	3.4 (1.1)	3.3 (1.1)	3.4 (1.1)	0.108
Social support for physical activity, mean (SD)g	3.2 (1.1)	3.1 (1.1)	3.2 (1.1)	0.085
Number of close friends/family who encourage healthy eating, N (%)				−0.075
None	8 (3.3%)	11 (4.8%)	19 (4.0%)
1–2	94 (38.2%)	80 (34.9%)	174 (36.6%)
3–4	77 (31.3%)	68 (29.7%)	145 (30.5%)
5–6	42 (17.1%)	38 (16.6%)	80 (16.8%)
7+	25 (10.2%)	31 (13.5%)	56 (11.8%)
Number of close friends/family who encourage physical activity, N (%)				−0.011
None	10 (4.1%)	17 (7.4%)	27 (5.7%) 178
1–2	95 (38.6%)	83 (36.2%)	(37.5%) 147
3–4	78 (31.7%)	69 (30.1%)	(30.9%)
5–6	44 (17.9%)	33 (14.4%)	77 (16.2%)
7+	18 (7.3%)	27 (11.8%)	45 (9.5%)
Social norms for obesity, mean (SD)h	5.5 (2.7)	5.8 (2.8)	5.6 (2.8)	−0127
Social norms for weight loss, mean (SD)h	2.9 (1.2)	2.9 (1.2)	2.9 (1.2)	0.009
Number of social contacts trying to lose weight, N (%)				−0.070
Nobody	21 (8.5%)	19 (8.3%)	40 (8.4%)
A few of them	90 (36.6%)	74 (32.3%)	164 (34.5%)
Half of them	47 (19.1%)	53 (23.1%)	100(21.1%)
Most of them	59 (24.0%)	57 (24.9%)	116 (24.4%)
All	25 (10.2%)	26 (11.4%)	51 (10.7%)
How likely are you to try to lose weight in the next 3 months?, N (%)				−0.018
Very unlikely	9 (3.7%)	10 (4.4%)	19 (4.0%)
Unlikely	11 (4.5%)	10 (4.4%)	21 (4.4%)
Neutral	25 (10.2%)	24 (10.5%)	49 (10.3%)
Likely	104 (42.3%)	88 (38.4%)	192 (40.4%)
Very likely	94 (38.2%)	97 (42.4%)	191 (40.2%)
Quality of life: physical, N (%)				0.158
Poor	11 (4.5%)	23 (10.0%)	34 (7.2%)
Fair	54 (22.0%)	52 (22.7%)	106(22.3%)
Good	90 (36.6%)	81 (35.4%)	171 (36.0%)
Very Good	48 (19.5%)	36 (15.7%)	84 (17.7%)
Excellent	43 (17.5%)	37 (16.2%)	80 (16.8%)
Quality of life: emotional, N (%)				0.105
Poor	9 (3.7%)	9 (3.9%)	18 (3.8%)
Fair	50 (20.3%)	41 (17.9%)	91 (19.2%)
Good	65 (26.4%)	86 (37.6%)	151 (31.8%)
Very Good	62 (25.2%)	48 (21.0%)	110 (23.2%)
Excellent	59 (24.0%)	44 (19.2%)	103 (21.7%)

^aEach response is given a value according to the Likert scale (‘never’=1 to ‘almost everyday’=6). Responses are summed across items to produce a score ranging from 6 to 30

^bAmong participants who perceived any discrimination, they were asked to state whether they experienced discrimination according to these domains (yes/no)

^cPossible scores range from 0–100, with 100 being the healthiest dietary quality

^dInternational Physical Activity Questionnaire, MET minutes per week

^eBased on MET minutes per week

^f7-point Likert scale, with 1 being the lowest response option and 7 being the highest

^g5-point Likert scale, with 1 being the lowest response option and 7 being the highest

^h6-point Likert scale, with 1 being the lowest response option and 7 being the highest acceptance of obesity and support for weight loss

4. Discussion

Healthy Immigrant Community is a social network intervention for cardiovascular risk reduction among Hispanic and Somali immigrant populations, derived through a CBPR approach. This is among the first social network interventions to prioritize health promotion in immigrant populations. Social network interventions have been shown to successfully engage difficult to reach populations by activating network and social environment mechanisms for health promotion.[67] This is important because immigrant populations may be less likely to participate in evidence-based health promotion interventions within traditional healthcare settings.

Network interventions with “difficult to reach” populations is synergistic with the CBPR approach, where community and academic partners co-create the intervention and evaluation, while centering community health priorities.[19] Community participation in the development, implementation, and evaluation of health promotion interventions has been shown to yield positive health outcomes.[68]

One potential limitation of the intervention is that the programmatic intervention dose is lower than what is needed to impact health behaviors and achieve cardiovascular risk reduction. The rationale for the intensity of the intervention is that 1) the intervention will be more impactful when delivered by a member of participant’s social network,[69] and 2) non-programmatic health promoting activities and interactions are likely to occur within and between social networks (indirect intervention effect) throughout the study interval.[70] This approach highlights the potential advantages of social network interventions for scalability and sustainability.[16, 71]

Another limitation of the study is that full implementation of interventions in community-based studies is typically not achieved. The importance of process evaluation in these settings has been highlighted,[72] and we intend to prospectively assess our research process at every phase. Furthermore, there is a risk of indirect intervention dissemination to the delayed-intervention control group, which is common in social network interventions.[70] We have attempted to address this in our sample size calculation, and the study exceeded enrollment targets. Finally, while investigators are blinded to treatment condition, it was not feasible to achieve blinding of study team members with direct participant contact.

5. Conclusions

This manuscript describes the development of the HIC protocol for a stepped wedge cluster randomized trial to test the effectiveness of a social network intervention for cardiovascular risk reduction among two immigrant populations. At baseline, participants (on average) were in the obese range with elevated fasting glucose levels, moderate levels of dietary quality and physical activity, and normal blood pressure and cholesterol, which suggests an opportunity to re-shape cardiovascular risk among this relatively young cohort (mean age=44 years) with a tailored intervention developed through CBPR. Details from this protocol may be relevant to other researchers, practitioners, or community-based organizations who are developing health promotion interventions with similar goals.

Figure 1.

Study Design and Intervention Schedule

Figure 2.

Logic Model of Intervention Strategies and Outcomes

Figure 3.

Intervention outline and session topics

Table 1.

Life’s Simple 7 domains and derivation of its composite score[54–56]

“Life’s Simple 7” Domain	Optimal (2 points)	Intermediate (1 point)	Poor (0 points)
Weight	BM<25	BMI 25–29.9	BMI ≥30
Blood pressure	SBP<120 and DBP<80	SBP 120–139 and DBP 80–89	SBP≥140 or DBP≥90
Total cholesterol	<200 mg/dL	200–239 mg/dL	≥240 mg/dL
Fasting glucose	<100 mg/dL	100–125 mg/dL	≥126 mg/dL
Smoking status	Never or quit >12 months ago	Former or quit ≤12 months ago	Current
Diet	HEI≥69.3	HEI 56.9–69.2	HEI<56.9
Physical activity	≥150 min/wk moderate or ≥75 min/wk vigorous or ≥150 min/wk moderate+vigorous	1–149 min/wk moderate or 1–74 min/wk vigorous or 1–149 min/wk moderate+vigorous	None
Composite Score	9–14	7–8	0–6

Data availability

Data will be made available on request.