170 - Food Insecurity Screening: Does Context Affect Disclosure Rates?

Poster Number: 170
Publication Number: 170.412

Kristen E. Samaddar, Phoenix Children's Hospital, Phoenix, AZ, United States; Jennifer Farabaugh, Phoenix Children's Hospital, Phoenix, AZ, United States; Caitlin Parke, University of Arizona College of Medicine - Phoenix, Phoenix, AZ, United States

Background: Food insecurity (FI) affects 14.8% of households with children. FI has negative impacts including impaired cognition, decreased self esteem, foregone medical care, and increased emergency room and admission rates. The American Academy of Pediatrics introduced the 2 question Hunger Vital Sign (HVS) in 2017. Researchers at Phoenix Children's Hospital wanted to determine the effect on disclosure rates with different methods of FI screening.

Objective: To compare disclosure of FI using a verbal screen and 2 written formats using the Hunger Vital Sign

Design/Methods: Between June 2018 and December 2020, 3 screening formats were utilized. Using retrospective chart review, we compared responses for each group: 6/18-12/18 verbal screening; 6/19-12/19 written HVS in Healthy Habits Tool; 6/20-12/20 written HVS in Adverse Childhood Experiences/SDOH Screen (Figure 1). Results were analyzed to determine percent of positive response by screening format. Sub-groups were insurance, primary language, age, BMI and medical complexity. Chi-squared and Fisher’s Exact tests were used to compare categorical variables. Univariate and multivariate logistic regression was used to ascertain the association between screening format and the odds of FI.

Results: A total of 5,741 families completed FI screenings. Of those, 5.07% screened positive. Written screeners had higher disclosure rates (verbal 2.59%, written Healthy Habits 7.1%, written ACE/SDOH 5.1%, p= < 0.001%). The same FI screen included on the Healthy Habits Tool had statistically significantly more positive responses than the ACE/SDOH instrument (p=0.004). Subgroups were generally not associated with significant changes in positive screening apart from insurance and language.Conclusion(s): Our study assessed if the format of FI screening influences disclosure and the variability in different populations. Similar to prior literature, we found an increase in disclosure rates using Likert-style written tools vs. dichotomous verbal screening. Novel to our study, higher positive disclosure rates were seen in screens imbedded in a lifestyle questionnaire compared to an ACE screener. Disclosure rates were increased in some expected subgroups (insurance and language) but not BMI, age or medical complexity. Study limitations include possible repeated sampling of patients over time and the potential impact of broader social and economic events during the study period.

Our study demonstrates that not only do the questions and delivery method matter but also the context of screening. Based on our results, separating FI screening from ACE screening may confer higher disclosure rates.
CParke_CV.pdf
FI Disclosure Rates with HVS in 2 Different Screeners