**Background:** Chronic conditions such as diabetes, obesity, heart disease, and depression are highly prevalent and frequently co-occur with food insecurity in communities served by US community health centers. Racial/ethnic minority communities and those with lower socioeconomic status are disproportionately affected — for example, 12% of Black adults and 11% of Latinx adults have diabetes, 1.7 and 1.6 times higher than non-Hispanic white adults. 'Food as Medicine' approaches, particularly produce prescriptions, have shown promise for decreasing food insecurity and increasing fruit/vegetable intake, but evidence on combining produce prescriptions with group medical visits is limited. One prior study combining produce prescriptions with group medical visits found daily fruit/vegetable consumption increased from 5.2 to 6.4 servings at 4 months, systolic blood pressure decreased from 146.1 mm Hg to 129.9 mm Hg among those with hypertension, and depressive symptoms decreased from 14.5 to 7.7 among those with depression.
**Methods:** This quasi-experimental study evaluates Recipe4Health, a multi-sectoral collaboration between Alameda County, Community Health Center Network, Open Source Wellness, and Dig Deep Farms. The intervention has two components: (1) Food Farmacy — 16 weekly doorstep deliveries of organic produce (~16 servings/week) from Dig Deep Farms; (2) Behavioral Pharmacy — a 4-month weekly group medical visit series on Zoom for up to 24 patients, led by trained health coaches with primary care provider participation, targeting physical activity, healthy eating, social connection, and stress reduction. Adult patients (18+) with food insecurity and/or chronic conditions (obesity, pre-diabetes, type 2 diabetes, hypertension, depression, anxiety) are referred by clinic staff who receive 2 hours of food insecurity screening training plus 8 hours of clinical nutrition training. Pregnant women are excluded. The study plans to recruit 250 in Food Farmacy only and 140 in Food Farmacy plus Behavioral Pharmacy. Data collection runs from August 2021 to December 2024. The primary outcome is daily fruit/vegetable consumption measured by the 10-item Dietary Screener Questionnaire (DSQ-10). Secondary outcomes include physical activity (Exercise Vital Sign), health-related quality of life (CDC HRQOL-4), social isolation (UCLA Loneliness 3-item), food insecurity (6-item Household Food Security Short Form), depressive symptoms (PHQ-9), anxiety (GAD-7), and EHR-derived outcomes (HbA1c, blood glucose, lipids, BMI, blood pressure, medication use, ED visits, hospitalizations). Surveys are collected at baseline and 4 months. EHR data are collected at baseline (4 months prior to referral and 1 month after) and at 6-month and 12-month follow-up (with 3-month allowable windows). Propensity score matching uses facility-level pair-matching plus individual-level matching on sociodemographics (age, race/ethnicity, sex), clinical characteristics (ICD-9/ICD-10 codes, medication classes), and health outcomes (HbA1c, LDL cholesterol) from the past 18 months, with up to four controls per participant. Mixed models will compare within-group and between-group changes, adjusting for baseline values and clinic-level clustering. With 140 in Food Farmacy + Behavioral Pharmacy and 1:1 matching, the study has 80% power to detect an effect size of 0.4 or greater (α=0.025, two-sided). With 250 in Food Farmacy only, 80% power to detect an effect size of 0.3 or greater. For EHR exploratory analyses, the study anticipates at least 2000 matched pairs, sufficient to detect an effect size of 0.10.
**Clinical Implications:** This study is designed to provide evidence to inform policies addressing food insecurity and nutrition-sensitive chronic conditions in healthcare settings. National, state, and local policies are increasingly supporting addressing social determinants of health — some states are obtaining Medicaid waivers to address food insecurity, and California is considering pilot projects similar to Recipe4Health for Medicaid (Medi-Cal) beneficiaries. The quasi-experimental design has limitations: lack of randomization, inability to collect patient-reported outcomes from the control group, and potential sparseness of EHR data due to increased telehealth use during COVID-19. However, stakeholder engagement through a Community Advisory Board and partnership with clinics, farms, and policy makers aims to maximize direct policy impact. Results will be disseminated through scientific literature, updates to the Alameda County Board of Supervisors, clinic dashboards for real-time screening/referral data, and the Dig Deep Farms newsletter.