Food insecurity affects approximately 10.5% of households in the US, 13.1% in Texas, and 16.3% in the Greater Houston area (1,2) and is strongly correlated with the prevalence of chronic disease (3). Food prescription programs are gaining popularity as an evidence-based strategy to reduce food insecurity (4). Emerging partnerships between health care clinics and food banks leverage established community points of contact (eg, food banks, food pantries) to provide healthy food to those in need; these partnerships have demonstrated health improvement outcomes (5). Success of such partnership programs is largely dependent on staff referral and care coordination; however, few data exist on how clinic staff perceive such partnerships (6). Since 2018, the Houston Food Bank (HFB) has operated a food prescription program, Food Rx, across 21 health care clinics in the Greater Houston area. Our study examined the perceptions of clinic staff about the importance of the program and challenges they encountered in program implementation. Our objective was to determine what clinicians and other clinic staff members perceived as the benefits of the HFB food prescription program and the barriers encountered in implementing it.

We conducted a retrospective secondary analysis of the data HFB collected from May 2018 through March 2021. Electronic surveys were sent by HFB to all clinic staff in 21 participating clinics (federally qualified health centers, charity clinics, health systems, a city health department, and an academic health center) involved in implementing HFB’s food prescription program, HFB Food Rx; 17 responded (81% response rate at the clinic level) (Figure). A convenience sample of 252 health care staff members in the 17 organizations responded to the survey. One open-ended question was asked in the survey: “Briefly (in 250 words or less) provide qualitative feedback related to progress of the Food Rx program and partnership with the Food Rx program, including feedback from participants about Food Rx and ways to improve Food Rx for participants.” Clinic partners gave consent for data use in our research as part of HFB partnership agreements. The primary investigator (J.W.M.) used de-identified data to review the open-ended question and response frequency and to create a preliminary codebook. Response text was then assigned to existing codes with modification as needed to capture emerging themes by using constant comparative methodology (7). Subsequently, 2 additional investigators (J.A., J.T.) blind-coded data by using the codebook as a reliability check. Any code discrepancies were resolved by consensus among the 3 investigators (J.W.M., J.A., J.T.). Data were analyzed in June 2021 by using Dedoose version 9.0.15 (Dedoose: SocioCultural Research Consultants, LLC).

This study was part of a parent study to assess the clinical impact of the HFB food prescription program. The study was approved by The University of Texas Health Committee for Protection of Human Subjects.

We identified 4 broad areas of experience and perception among survey respondents (Table). The first was the importance of having a value-based care strategy and the role of food prescription programs in that strategy to address food and health needs in populations vulnerable to food insecurity. The second area was patient and food pantry barriers to success in implementation of Food Rx. The third area was the importance of coordinated patient care within and between organizations to implement an effective and standardized partnership-based food prescription program. Coordination included recognition of the value of establishing consistent communication channels between clinic and patients. The fourth area was the need for improved within-clinic workflow integration and quality improvement metrics for clinicians to understand their current performance in screening and responding to patient needs associated with social determinants of health. Respondents also noted the particular importance of food prescription programs during the COVID-19 pandemic.

The perspectives of clinicians and clinic staff in collaborative food prescription programs are critical to understanding program shortcomings and guiding implementation and dissemination design to improve participation and drive overall program success. A consistent theme that emerged from our survey was the perception that a food prescription program is an effective value-based care strategy to address social determinants of health, indicating a positive shift of clinicians’ thinking toward offering such programs as part of the standard of care for patients who are food insecure and struggling with diet-related diseases. A recent systematic review and meta-analysis (5) demonstrated a positive effect of food prescription programs on health outcomes and the desire of clinic staff to continue these programs. As seen in our study, future sustainability efforts need to consider reimbursement for these programs by health insurance companies (8).

Challenges identified by participating clinic staff for program success included the variability of fresh food availability at the food pantry. Although food quality is important to food banks, a recent qualitative study of food bank leadership found that maintaining sufficient stocks of a variety of fresh produce and refrigerated items poses a challenge for food pantries that historically provided only shelf-stable foods (9). This challenge may be exacerbated by other barriers such as patient lack of transportation (10,11). Program success may depend on ensuring a sufficient inventory of fresh produce, maintaining timely restocking practices at pantries, and minimizing patients’ personal barriers.

Additionally, respondents emphasized the importance of interorganizational care coordination, which was reported in prior qualitative research on clinician perspectives on food bank–food prescription partnerships (12). Currently, there is no warm referral (having clinic staff contact another agency or clinician on the client’s behalf) or standardized reminder across clinicians and food pantries to facilitate prescription redemption among patients. In the absence of referrals and reminders, a substantial proportion of patients may fail to redeem their prescriptions, use prescriptions within the 6-month period when the prescriptions are valid, or ever return to the clinic for follow-up.

Finally, we identified clinic workflow integration and use of quality metrics around food prescription programs as important predictors of implementation success. These 2 processes involve developing and training clinic staff in detailed clinic workflow on who will be screened for food insecurity and how, what the clinic response will be to a positive screen (eg, food prescription program referral), and potentially integrating food prescriptions into regular patient care for food-insecure patients. Other studies have reported barriers to program implementation resulting from poor workflow integration, including lack of consistent screening for food insecurity resulting from time constraints and inadequate training of clinic staff (10,13). Furthermore, developing and implementing quality improvement metrics in health care for social determinants of health need to be considered. Measures of the Healthcare Effectiveness Data and Information Set (HEDIS) help clinicians understand their current performance in key areas as well as plan quality improvement efforts. HEDIS measures encompass a number of different health care processes and quality areas, and most recently the National Committee for Quality Assurance (NCQA), the body governing HEDIS measure development, proposed similar quality measures for social determinants of health (14).

Clinicians noted the importance of maintaining communication with patients during the COVID-19 pandemic. In the initial phase of the pandemic, issues related to the food supply chain and unemployment increased rates of food insecurity (15), and lockdowns caused widespread disruption. These conditions reduced access to health clinics and altered food pantry distribution; however, Food Rx pivoted and remained continuously active, which was important to maintaining communication about the program among clinic staff, food pantries, and patients. As one health care organization noted, “HFB has been great at communicating with us and letting us know what their plans [are] and have offered their assistance to us as well. We are very thankful for their partnership, and hope to continue this at our other sites.”

Our study had limitations related to the evaluation of clinician feedback. We used a convenience sample rather than a random sample, which may have resulted in selection bias. The time point of survey administration varied across different clinics, and our evaluation approach was not guided by a validated framework, which could strengthen future studies. However, our scaled evaluation across 21 health care organizations of various types, and a high response rate at the clinic level with a wide range of clinic respondents, provided important feedback related to produce prescription implementation with strong external validity.

In conclusion, our study reports perspectives of clinicians and clinic staff in diverse health care settings on the benefits of food prescription programs for patients, as well as challenges and facilitators to program implementation. Food prescription programs are gaining popularity nationwide, and our study provides insight into successful implementation of such programs. Future studies should incorporate a mixed-methods approach guided by a validated framework to obtain regular input from implementation partners.