Screening for Food Insecurity: A Curriculum for Medical Students

In 2021, 13.5 million US households (10.2%) were food insecure.¹ Food insecurity (FI) is defined as a lack of access to enough food for an active, healthy life.² The social and structural determinants of health (SDOH), including FI, are major contributors to chronic illnesses in the United States.³ FI is also associated with higher health care utilization, higher health care costs, and adverse health outcomes, especially for children.^4–6

There is a growing interest in training health professionals to screen for the SDOH, including FI.⁷ The American Academy of Pediatrics recommends universal screening for FI⁸; other organizations plan to consider modifiable SDOH⁹ in their recommended screenings. Yet, most medical schools do not have an FI screening curriculum; existing programs are primarily for residents and attending physicians.^10–12 Given the high prevalence of FI in Washington, DC¹³ where our medical students train, we designed a longitudinal curriculum to address this gap.

The curriculum is a novel attempt to explicitly teach FI screening to medical students. Our study objective was to determine whether participation in this FI screening curriculum improves students’ knowledge, self-efficacy, and behavior in screening for and intervening in cases of FI.

This was a prospective, single-institution study at a midsized, private medical school in the US mid-Atlantic region. Participants included students from the classes of 2022 (intervention cohort) and 2021 (control cohort).

The longitudinal FI curriculum (“curriculum”) was introduced during the first year of medical school. It included three components completed over 3 years (Figure 1).

Year 1. We presented information on FI screening during a 2-hour workshop and distributed the Food Insecurity Screening Informational Pocket Card. Students were then paired to practice screening, using scenarios based on real clinical encounters. The 10-item Qualtrics survey and the Pocket Card can be found in the STFM Resource Library.¹⁴

Year 2. The curriculum was reinforced with a similar 20-minute FI screening role-playing exercise.

Year 3. Students participated in an existing objective structured clinical exam (OSCE) with a patient with poorly-controlled diabetes. The patient had recently struggled with FI but did not present with specific complaints of FI.

During Year 1, students were emailed an informed consent script to participate in the study. We then administered a 10-item survey via Qualtrics. In Year 3, students completed the same survey, postintervention.

Five items assessed students’ FI knowledge using multiple-choice questions and were scored as correct or incorrect. Five items also assessed students’ attitudes using a 5-point Likert scale (1=strongly disagree to 5=strongly agree); we calculated the percentage of respondents who agreed or strongly agreed.

Furthermore, intervention students submitted written reflections on the following prompts after the workshop:

1. What went well?
2. What was challenging?
3. What questions or insights did this exercise raise for you?

Both the intervention cohort and a historical control cohort completed the OSCE, which remained unchanged between years. The historical control group had not participated in the FI curriculum. Standardized patients scored “yes” or “no” on one OSCE checklist item, “Did the student screen for FI?”

We calculated descriptive statistics for intervention students’ pre- and postintervention survey responses. Categorical variables were reported as count and frequency; continuous data were reported as mean and standard deviation. We used Wilcoxon Signed-Rank and McNemar tests to compare the paired pre- and postdata. We did not perform multiple testing corrections. We performed comparison of successful OSCE screening between intervention and control groups using the χ² test. Incomplete data in pre/post knowledge questions with at least one answer were considered incorrect. If none were answered, it was considered missing data and was excluded from analysis. Significance was defined as P<.05. SAS v9.4 was used for all analyses.

We qualitatively analyzed postworkshop reflections to determine predominant themes. Two researchers not involved in the intervention independently completed the data coding and analysis using the software data management program, Ethnograph v6 (Qualis Research, Colorado Springs, CO). They developed a coding dictionary based on the reflection questions asked as part of the assessment in the FI workshop in year 1 and identified emerging themes through review of the data. After independently coding the data, the coded text was extracted to populate an analysis grid to support and finalize themes.  

This study (#20181123) was determined to be exempt from review by our institutional review board.

Of the 204 students who received the intervention, 61 (30.0%) completed both pre- and postintervention surveys, and 137 (67.5%) submitted reflections. Four observations were missing for the postsurvey knowledge questions and seven were missing for postsurvey attitude questions. The OSCE was completed by 156 and 140 students in the intervention, and control cohorts, respectively.

When considering preintervention attitudes, only 24 (39.34%) students agreed/strongly agreed that they were comfortable discussing FI with patients; even fewer (12, 19.67%) agreed/strongly agreed they were confident referring to community resources. However, 57 (93.44%) students agreed/strongly agreed that screening for FI was important for physicians to do with their patients. Some students (23, 37.70%) disagreed/strongly disagreed that their patients would be harmed by screening due to embarrassment and 39 (63.93%) students disagreed/strongly disagreed their screening will divert visit time from more important medical issues (Table 1).

In terms of outcomes, students were significantly more likely to feel extremely or quite confident in referring patients to community resources for nutrition assistance at postintervention compared to preintervention (19.67% vs 42.59%, respectively). They were also significantly more likely to strongly disagree or disagree to “patients may be harmed by screening for food insecurity due to embarrassment” (37.70% vs 81.48%) and “screening for food insecurity will divert visit time from more important medical issues” (63.93% vs 79.63%; Table 1).

Qualitative results were determined based on students' reflections after the workshop. Five salient themes supported students’ preintervention attitudes:

1. Learning New Knowledge and Skills
2. New Perspectives Regarding FI
3. Resource Barriers
4. Physician-Patient Communication
5. Using FI Screening in Clinical Practice.

Explanation of the themes and representative quotes are shown in Table 2.

Regarding knowledge questions, students showed a statistically significant increase in mean percent scores pre- and postintervention (45.24% vs 74.74%, P<.001; Table 3). In their behaviors, students in the intervention cohort were significantly more likely than the control cohort to screen for FI (28.2% vs 10.7%, P<0.001; Table 4).

Our curriculum added less than 3 hours of curricular time and improved students’ knowledge. Moreover, students who participated in the curriculum were more likely than students who had not participated to both recognize the need and screen for FI. This is particularly significant as the standardized patient did not present with specific complaints of FI, so students were not explicitly prompted.

Our study was limited to one intervention cohort at one institution. The number of students who completed both the pre-and postintervention surveys was low. Student behavior was assessed using simulated patients in one case. The results may therefore not be generalizable across different student populations within or outside our institution, in other health profession educational settings, or in practice.

A longitudinal curriculum can improve students’ knowledge, self-efficacy, and behavior in screening for FI. Reinforcing knowledge and skills for behavior change may require minimal additional curricular time if integrated into the existing curriculum.