Journal Pre-proof Food insecurity screening in primary care: Patterns during the COVID-19 pandemic by encounter modality Cassandra J. Nguyen PhD , Rachel Gold PhD, MPH , Alaa Mohammed MPH , Molly Krancari MPH , Megan Hoopes MPH , Suzanne Morrissey PhD , Dedra Buchwald MD , Clemma J. Muller PhD PII: S0749-3797(23)00151-4 DOI: https://doi.org/10.1016/j.amepre.2023.03.014 Reference: AMEPRE 7121 To appear in: American Journal of Preventive Medicine Please cite this article as: Cassandra J. Nguyen PhD , Rachel Gold PhD, MPH , Alaa Mohammed MPH , Molly Krancari MPH , Megan Hoopes MPH , Suzanne Morrissey PhD , Dedra Buchwald MD , Clemma J. Muller PhD , Food insecurity screening in primary care: Patterns during the COVID-19 pandemic by encounter modality, American Journal of Preventive Medicine (2023), doi: https://doi.org/10.1016/j.amepre.2023.03.014 This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2023 Published by Elsevier Inc. on behalf of American Journal of Preventive Medicine. 1 TITLE Food insecurity screening in primary care: Patterns during the COVID-19 pandemic by encounter modality AUTHORS Cassandra J. Nguyen, PhD, 1 Rachel Gold, PhD, MPH, 4,5 Alaa Mohammed, MPH 2 Molly Krancari, MPH, 4 Megan Hoopes, MPH, 4 Suzanne Morrissey, PhD, 4 Dedra Buchwald, MD, 2,3 Clemma J. Muller, PhD 2,3 1 Nutrition Department, University of California, Davis, Davis, CA, USA 2 Institute for Research and Education to Advance Community Health, Washington State University, Seattle, WA, USA 3 Elson S. Floyd College of Medicine, Washington State University, Spokane, WA, USA 4 OCHIN Inc., Portland, OR, USA 5 Kaiser Permanente Northwest Center for Health Research, Portland, OR, USA CORRESPONDING AUTHOR Cassandra J. Nguyen 1100 Olive Way Suite 1200 2 Seattle, WA 98101 Phone: 206-708-8621 Fax: N/A Email: Cassandra.nikolaus@yahoo.com ABSTRACT Introduction. Screening for food insecurity in clinical settings is recommended, but implementation varies widely. This study evaluated the prevalence of screening for food insecurity and other social risks in telehealth versus in-person encounters during the COVID-19 pandemic, and changes in screening before versus after widespread COVID vaccine availability. Methods. These cross-sectional analyses used electronic health record (EHR) and ancillary clinic data from a national network of 400+ community health centers with a shared EHR. Food insecurity screening was characterized in 2022 in a sample of 275,465 first encounters for routine primary care at any network clinic during March 11, 2020-December 31, 2021. An adjusted multivariate multi-level probit model estimated screening prevalence based on encounter mode (in-person versus telehealth) and time period (initial pandemic versus after vaccine availability), in a random subsample of 11,000 encounters. Results. Encounter mode was related to food insecurity screening (p<0.0001), with an estimated 9.2% screening rate during in-person encounters, compared to 5.1% at telehealth encounters. There was an interaction between time period and encounter mode (p<0.0001), with higher 3 screening at in-person vs. telehealth encounters after COVID-19 vaccines were available (11.7% v. 4.9%) versus before vaccines were available (7.8% v. 5.2%). Conclusions. Food insecurity screening in first primary care encounters is low overall, with lower rates during telehealth visits and the earlier phase of the COVID-19 pandemic. Future research should explore methods for enhancing social risk screening in telehealth encounters. 4 INTRODUCTION Food insecurity, defined as a lack of sustained access to enough food to lead an active and healthy lifestyle, 1 is associated with diet-related chronic diseases including type 2 diabetes and cardiovascular disease, 2-4 suboptimal disease self-management, 5,6 and premature mortality. 7 Given these outcomes, various initiatives to address food insecurity in the U.S. have been proposed. 8 One such initiative is screening for food insecurity in clinical settings, which is recommended by many national health organizations, including the National Academy of Medicine, the American Academy of Pediatrics, and the American Diabetes Association. 9-11 Such screening can be an essential first step for healthcare providers seeking to refer at-risk patients to food assistance programs, food pantries, and medically tailored meal programs. 12-14 Despite support for food insecurity screening in clinical settings from healthcare organizations, 9- 11 practitioners, 15 and patients, 16 rates of such screening vary. Between 2017-2018, 30-40% of healthcare administrators surveyed reported that their facilities ever screened for food insecurity, 17 but this estimate does not mean that food insecurity screening occurs at all encounters for all patients at these facilities. Screening rates have been shown to vary by clinic, provider, and patient characteristics, with facility rurality, 17 provider type, 15,18 and several patient characteristics 19,20 associated with lower rates. During the COVID-19 pandemic, food insecurity increased in the U.S. due to the pandemic- driven economic downturn and disrupted food system. 21 Concurrently, social distancing 5 recommendations caused many healthcare services to be delivered via telehealth. 22 Telehealth presents unique barriers to food insecurity screening, including privacy concerns, limited proficiency with telemedicine platforms, 22,23 and disruptive technical difficulties, all of which may impact ability to conduct food insecurity screening. The coinciding rises in telehealth and food insecurity during the pandemic yield a unique opportunity to build knowledge about variation in patterns of food insecurity screening. As high utilization of telehealth is expected to continue, 22 characterizing food insecurity screening in telehealth encounters during the pandemic may provide insights to improve screening practices. Of particular interest is how screening patterns changed during the initial months of the pandemic when social distancing recommendations were the most stringent compared to the period after vaccines became widely available. The relaxed social distancing recommendations which followed vaccine availability may have been associated with changes to in-person healthcare use patterns. Yet, little has been published in this area to date. This study provides observational evidence of food insecurity screening patterns in a national database of community health centers (CHCs) during the COVID-19 pandemic. Analyses characterize variability in patterns of screening for food insecurity and other social risks (e.g., housing stability and transportation access) during telehealth versus in-person encounters, and in different stages of the pandemic which were characterized by changes in social distancing recommendations. The results are informative to clinics seeking to implement or expand food insecurity screening. 6 METHODS Study Sample This analysis used electronic health record (EHR) data from the national OCHIN network of CHCs. OCHIN, Inc. is a nonprofit organization that hosts a shared instance of the Epic EHR for its member CHCs, which primarily serve patients who are publicly insured or uninsured. Social needs screening documentation, including screening for food insecurity, has been available in this EHR since mid-2015. The data requested and received included ambulatory and telehealth encounters at all OCHIN member clinics from January 1, 2019, to December 31, 2021, representing 1,844,921 unique patients served by 492 clinics across 16 states. The procedures for this study were approved by the Kaiser Permanente Northwest Institutional Review Board (Study # 696). The dataset has a hierarchical structure with four distinct levels (healthcare system, clinic, provider, and encounter), each of which is nested within the preceding level. Screening data in the EHR was matched with encounters using date and encounter IDs. Appendix Figure 1 illustrates this structure with a classification and unit diagram. Food insecurity screening was recorded at 2.6% of encounters between January 2019 and December 2021 (Appendix Figure 2). As first visits for routine primary care in this period had the highest rates of food insecurity screening (7.8%), the study sample was restricted to first 7 visits, defined as a patient’s earliest encounter at a study clinic. Routine primary care was identified based on clinic specialty, provider type, encounter descriptions, and level of service codes. Measures Clinic-, provider-, patient-, and encounter-level characteristics of each encounter, described below, were selected based on prior literature and a conceptual model. The model is shown in Appendix Figure 3, with each variable ’ s supportive literature outlined by its hierarchical level of influence. The primary outcome is whether food insecurity screening was documented in the EHR at a given encounter. All clinics had access to identical EHR interfaces where screening results could be documented in structured data entry via a flowsheet. Food insecurity was assessed at the household level with a set of three questions, adapted from the standard U.S. Department of Agriculture’s Food Security Survey Modules. 24 Each clinic had access to several standard social risk screening tools in their shared EHR; question wording and other details are previously published. 25 Though each clinic had access to the same set of screening tools, they could independently elect whether and how to screen for any social risks, in the form of specific screening tools or individual social risk questions. As study data were restricted to first visits for routine primary care at ambulatory or telehealth encounters, it is unlikely that patients completed such screening prior to study encounters. The secondary outcome is the documentation of screening for any of three social risks (food insecurity, housing instability, and access to 8 transportation) by encounter, to assess if independent variables differ for food insecurity screening and general social risk screening. Independent variables included mode of encounter (in-person or telehealth), identified based on standard procedure codes (e.g., 99421), billing modifiers (i.e., GT, 95), and OCHIN Epic encounter types. The second independent variable, time period, was defined as March 11, 2020- April 30, 2021 (the initial pandemic period) or May 1, 2021-December 31, 2021 (after COVID- 19 vaccines were broadly available) which roughly reflect shifts in public-health related social distancing restrictions in the U.S. Potential clinic-level confounders included: rurality per 2010 Rural-Urban Commuting Area Codes, and whether the clinic was participating in initiatives related to social risks at the time of the encounter. 26 At the provider level, provider type was considered, defined by National Provider Identifier taxonomies, collapsed into seven categories. Patient-level confounders included race, ethnicity, sex, age, household income, documented homelessness, and rurality of patient’s residence. At the encounter level, confounders included whether interpreter services were needed and health insurance status. Statistical Analysis The original plan was to include all qualifying encounters in the analysis, but the complexity of hierarchical linear modeling with over 275,000 observations overloaded the available computing 9 capacity. Therefore, probability sampling was used to select a random subset of 11,000 observations, which was the largest sample size the most powerful computer accessible could accommodate. The sample was constrained to include 250 observations per encounter mode per month, to reduce variance inflation for estimating associations between telehealth and in-person visits compared with the full dataset. Descriptive statistics were calculated to assess how well the random inferential subsample was representative of the larger full sample. Using the 11,000 inferential subsample, all variables were descriptively summarized using frequencies for categorical variables and means, standard deviations (SD), and ranges for continuous variables, as appropriate. Graphs of monthly encounters and food insecurity screening rates by encounter mode (Appendix Figure 2) informed inclusion criteria and model selection, as described in the Data subsection above. Encounters were stratified by both independent variables and descriptively characterized by potential confounders and the outcome variables. Hierarchical linear modeling was used to estimate associations for the primary (food insecurity screening) and secondary (any of the three social risk screenings) outcomes. 27 These models were specified to include random intercepts at the healthcare system, clinic, and provider levels; fixed effects for encounter mode, time, and potential confounders; and a fixed effect for interaction between mode and time. Output for these models has a similar interpretation to ordinary least squares regression and logistic regression, but they appropriately account for the 10 hierarchical nested structure of the dataset. 28 All point estimates were reported with 95% confidence intervals. As a sensitivity analysis t o account for potential bias due to a small number of clinics’ previous experience with two intervention trials to promote social risk screening, 26,29 a second random subsample was generated as described above but excluded clinics participating in one (n=30) or both (n=6) from eligibility. Analyses were repeated and compared with outcomes for the main subsample; these findings were not reported separately unless there were notable differences. RESULTS Figure 1 shows number of first primary care encounters each month in the study period, and the monthly rate of food insecurity screening by mode of encounter (telehealth or in-person). Initially there were intentions to use data between January 2019 and February 2020 for a pre- COVID-19 comparison. However, given the very low counts of telehealth encounters during this time period (total n=329; Figure 1), only encounters occurring after March 11, 2020, were analyzed. Data were further restricted to 1) encounters at clinics active throughout the 22-month period (i.e., clinics active on or before March 11, 2020 and through December 31, 2021), 2) primary care encounter types with more than 0% food insecurity screening (e.g., off-site visits, rapid evaluations, and occupational health visits), and 3) observations with non-missing data on patient sex, insurance, and rurality. After application of these inclusion criteria (Figure 2), the analytical sample included 275,465 encounters. 11 Table 1 describes the characteristics of encounters, patients, providers, and clinics in the full sample. Most encounters were in-person (72%), covered by public insurance (53%), and did not require interpreter services (60%). Patients were most commonly non-Hispanic (58%) white (57%), between 18-64 years old (68%), with household incomes ≤ 130% of the federal poverty level (51%) and living in an urbanized area (84%). Most encounters occurred at clinics in urbanized areas (84%) that did not participate in either of the two social risk screening interventions (91%). Food insecurity screening was documented at 7% of included encounters. Rates were higher at in-person (9%) than telehealth (3%) encounters. Patients screened for food insecurity were more likely to be Black, Hispanic, 18 years or older, from lower income households, or reporting homelessness. Food insecurity screening rates were higher in encounters with nurse practitioners or physicians than those with other provider types. Before conducting inferential statistics, the inferential subsample of 11,000 was characterized to assess comparability to the larger analytical sample (Appendix Table 1). The subsample appeared adequately representative of the full sample, and the inferential analysis proceeded. Table 2 shows the association between encounter type or time period and whether food insecurity screening occurred in adjusted multivariable models built with the inferential subsample. Encounters that occurred after COVID-19 vaccines were widely available had higher likelihood of food insecurity screening than those in the first 14 months of the pandemic (estimated probabilities: 8.3; 95% CI 5.0, 11.7 v. 6.5; 95% CI 3.5, 9.5, p<0.0001). Mode of encounter was also significantly associated with likelihood of food insecurity screening, with an 12 estimated 9% screening rate during in-person encounters compared to 5% at telehealth encounters in multivariable models based on the inferential subsample (p<0.0001). An interaction between time period and encounter mode was observed (p<0.0001), as indicated by the higher screening rate after the availability of vaccines evident only for in-person encounters (11.7% vs. 7.8%). In contrast, there was no discernable difference in screening rates between time periods for telehealth encounters (4.9% vs. 5.2%). Results of analyses for the combined indicator of screening for any of the three social risks (food insecurity, housing insecurity, transportation access; Table 2) were nearly identical to results for food insecurity screening alone. In sensitivity analyses, excluding clinics that took part in social risk screening initiatives did not demonstrably affect estimates of the effects of the independent variables (Appendix Table 2). DISCUSSION These analyses characterized variability in screening for food insecurity and other social risks during in-person versus telehealth encounters and during stages of the COVID-19 pandemic. Overall, prevalence of food insecurity screening at first primary care encounters was low. Results suggest lower food insecurity screening at encounters conducted via telehealth vs. in person, and higher screening at encounters after COVID-19 vaccines were available vs. the initial months of the COVID-19 pandemic. Effect of encounter mode interacted with time period. During both periods, the lowest rates of food insecurity screening (5%) occurred during telehealth encounters; the highest rate (12%) was after COVID-19 vaccines were widely available, at encounters conducted in-person. 13 In the same CHC network, prior to the pandemic food insecurity screening occurred at only 1% of encounters during the first two years that its documentation was feasible in structured data in the EHR (2016-2018). 20 Rates of screening for housing security and relationship safety were similar. 20 This may reflect the slow initial implementation processes of relatively new screening recommendations. Greater societal awareness of food insecurity brought about by the COVID-19 pandemic may explain why food insecurity screening occurred in almost 3% of all encounters between 2019-2021 in this network (Appendix Figure 2). The highest rate of screening observed (12%) was at in-person encounters after vaccines were widely available, possibly as a result of continued implementation efforts and / or providers’ improving familiarity with screening processes and workflows. Findings complement and expand on other studies of food insecurity screening. Prior to the pandemic, in one Southeastern Michigan community health center, food insecurity screening occurred at almost 25% of primary care encounters, 30 a prevalence far higher than this study’s estimate of 7% of encounters. This difference may be driven by differences in patient populations, incentive models, or individual clinic characteristics, such as having a clinic champion for screening. 31 It also aligns with other studies showing that individual community health centers ’ screening rates vary considerably. 20 Regardless, these relatively low rates are likely caused in large part by known barriers to healthcare providers conducting social risk screening, which include concerns about stigmatizing patients, 14,32 limited time, resources, or available staff, 33,34 and ambiguous or labor-intensive processes for responding to a positive 14 screen. 14,35 As a growing number of community clinics are being required to screen for social risks, identifying ways to address such barriers to screening will be increasingly important. The results presented here also suggest that food insecurity and other social risk screening is less likely at first primary care encounters conducted via telehealth in comparison to in-person. This is likely driven – at least in part – by the known challenges of providing any care via telehealth, such as concerns about patient privacy, 22,23 limited proficiency with telemedicine platforms and technical disruptions. 36 Furthermore, it is possible that first primary care encounters conducted via telehealth addressed more urgent needs in comparison to in-person encounters; additional research on the types of care scheduled for telehealth encounters would be needed to confirm this. However, additional barriers unique to screening for social risks may further drive these differences. It is possible that food insecurity screening is more likely to happen during in-person encounters as a result of greater rapport between patient and provider 37 or because additional observations made during the visit serve as a stimulus to social risk screening. Further research is needed to explore how, why, and under what conditions food insecurity screening occurs. This research should also assess whether facilitators and barriers to screening at in-person encounters differ when compared to audio-only vs. video-based telehealth encounters. The observational findings presented here can serve as a foundation for this needed research by identifying opportunities to improve rates of food insecurity screening across encounter modes. Limitations 15 These results should be interpreted with attention to analysis limitations. Analyses were limited to CHCs in the OCHIN network and limited by the available variables. For example, encounter length likely differs between mode (in-person or telehealth) and may be related to social risk screening, but encounter length was not available. Another limitation relates to the provider characteristics assessed. One provider is associated with each encounter, but it is unclear whether this provider was the individual responsible for asking social risk questions or entering the associated data. Alternatively, providers may have elected to note social risk information outside of the structured flowsheet in progress notes which would not have been captured here. Future studies might consider using natural language processing or similar methods to identify food insecurity documentation outside of structured fields. Finally, the current analysis was restricted to first primary care encounters, which were among those with the highest screening prevalence in the dataset. Though this homogeneity was prioritized in the present study, future work may want to consider other specialty encounters, such as urgent care, where social risk screening occurs. CONCLUSIONS In this study, observed food insecurity screening rates in first primary care encounters were greater than estimates pre-pandemic, although they remained low overall. Findings suggest a lower likelihood of social risk screening during telehealth encounters compared to in-person encounters, which was particularly pronounced after COVID-19 vaccines were widely available. As telehealth encounters have a continued presence after the COVID-19 pandemic, 16 demonstrating how social risk screening varies in these encounters provides foundational evidence to inform future research. Future research should explore why screening varies between encounter types, and healthcare systems interested in increasing their social risk screening should consider evaluating their screening patterns by encounter type and modality. These research and programmatic steps could be valuable to initiatives and policies designed to increase rates of food insecurity screening at healthcare encounters. Acknowledgements Not applicable Funding Sources This publication was supported by a grant from the National Center For Advancing Translational Sciences of the National Institutes of Health (KL2TR000421) and a grant from the National Institute of Diabetes and Digestive and Kidney Diseases (1R18DK114701). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The funders had no role in study design; collection, analysis, and interpretation of data; writing the report; and the decision to submit the report for publication. Conflicts of Interest No conflicts of interest were reported by the authors of this paper. Financial Disclosure No financial disclosures were reported by the authors of this paper. REFERENCE 17 1. Coleman-Jensen A, Rabbitt MP, Gregory CA, Singh A. Household Food Security in the United States in 2020. 2021. https://www.ers.usda.gov/publications/pub- details/?pubid=102075. Accessed 9/7/2022. 2. Nikolaus CJ, Zamora-Kapoor A, Hebert LE, Sinclair K. Association of food security with cardiometabolic health during young adulthood: cross-sectional comparison of American Indian adults with other racial/ethnic groups. BMJ Open. 2022;12(6):e054162. doi:10.1136/bmjopen-2021-054162. 3. Nikolaus CJ, Hebert LE, Zamora-Kapoor A, Sinclair KI. Risk of Food Insecurity in Young Adulthood and Longitudinal Changes in Cardiometabolic Health: Evidence from the National Longitudinal Study of Adolescent to Adult Health. J Nutr. 2022;152(8):1944-1952. doi:10.1093/jn/nxac055. 4. Laraia BA. Food insecurity and chronic disease. Adv Nutr. 2013;4(2):203-212. doi:10.3945/an.112.003277. 5. Berkowitz SA, Seligman HK, Choudhry NK. Treat or eat: food insecurity, cost-related medication underuse, and unmet needs. Am J Med. 2014;127(4):303-310 e303. doi:10.1016/j.amjmed.2014.01.002. 6. Berkowitz SA, Karter AJ, Corbie-Smith G, et al. Food Insecurity, Food "Deserts," and Glycemic Control in Patients With Diabetes: A Longitudinal Analysis. Diabetes Care. 2018;41(6):1188-1195. doi:10.2337/dc17-1981. 7. Men F, Gundersen C, Urquia ML, Tarasuk V. Association between household food insecurity and mortality in Canada: a population-based retrospective cohort study. CMAJ. 2020;192(3):E53-E60. doi:10.1503/cmaj.190385. 8. Seligman HK, Berkowitz SA. Aligning Programs and Policies to Support Food Security and Public Health Goals in the United States. Annu Rev Public Health. 2019;40:319-337. doi:10.1146/annurev-publhealth-040218-044132. 9. Council On Community P, Committee On N. Promoting Food Security for All Children. Pediatrics. 2015;136(5):e1431-1438. doi:10.1542/peds.2015-3301. 10. American Diabetes A. 1. Strategies for Improving Care. Diabetes Care. 2016;39 Suppl 1:S6-12. doi:10.2337/dc16-S004. 11. IOM (Institute of Medicine). Capturing social and behavioral domains and measures in electronic health records: phase 2. Washington D.C.2014. 12. Mirsky JB, Zack RM, Berkowitz SA, Fiechtner L. Massachusetts General Hospital Revere Food Pantry: Addressing hunger and health at an academic medical center community clinic. Healthc (Amst). 2021;9(4):100589. doi:10.1016/j.hjdsi.2021.100589. 13. Berkowitz SA, Delahanty LM, Terranova J, et al. Medically Tailored Meal Delivery for Diabetes Patients with Food Insecurity: a Randomized Cross-over Trial. J Gen Intern Med. 2019;34(3):396-404. doi:10.1007/s11606-018-4716-z. 14. Stenmark SH, Steiner JF, Marpadga S, Debor M, Underhill K, Seligman H. Lessons Learned from Implementation of the Food Insecurity Screening and Referral Program at Kaiser Permanente Colorado. Perm J. 2018;22:18-093. doi:10.7812/TPP/18-093. 15. Schickedanz A, Hamity C, Rogers A, Sharp AL, Jackson A. Clinician Experiences and Attitudes Regarding Screening for Social Determinants of Health in a Large Integrated Health System. Med Care. 2019;57 Suppl 6 Suppl 2:S197-S201. doi:10.1097/MLR.0000000000001051. 16. Kopparapu A, Sketas G, Swindle T. Food Insecurity in Primary Care: Patient Perception and Preferences. Fam Med. 2020;52(3):202-205. doi:10.22454/FamMed.2020.964431. 18 17. Fraze TK, Brewster AL, Lewis VA, Beidler LB, Murray GF, Colla CH. Prevalence of Screening for Food Insecurity, Housing Instability, Utility Needs, Transportation Needs, and Interpersonal Violence by US Physician Practices and Hospitals. JAMA Netw Open. 2019;2(9):e1911514. doi:10.1001/jamanetworkopen.2019.11514. 18. Acquah OO, Honsvall Hoefler AM, Zoller I, Manning LP, Pine DJ, Mitchell RF. Improving Identification of Food-Insecure Patients in an Outpatient Clinic Setting. PRiMER. 2020;4:3. doi:10.22454/PRiMER.2020.115304. 19. Fiori KP, Heller CG, Flattau A, et al. Scaling-up social needs screening in practice: a retrospective, cross-sectional analysis of data from electronic health records from Bronx county, New York, USA. BMJ Open. 2021;11(9):e053633. doi:10.1136/bmjopen-2021- 053633. 20. Cottrell EK, Dambrun K, Cowburn S, et al. Variation in Electronic Health Record Documentation of Social Determinants of Health Across a National Network of Community Health Centers. Am J Prev Med. 2019;57(6 Suppl 1):S65-S73. doi:10.1016/j.amepre.2019.07.014. 21. Leddy AM, Weiser SD, Palar K, Seligman H. A conceptual model for understanding the rapid COVID-19-related increase in food insecurity and its impact on health and healthcare. Am J Clin Nutr. 2020;112(5):1162-1169. doi:10.1093/ajcn/nqaa226. 22. Kichloo A, Albosta M, Dettloff K, et al. Telemedicine, the current COVID-19 pandemic and the future: a narrative review and perspectives moving forward in the USA. Fam Med Community Health. 2020;8(3). doi:10.1136/fmch-2020-000530. 23. Center for Health and Social Care Integration. Tele-Social Care: Implications and Strategies. 2021. https://www.nationalcomplex.care/wp-content/uploads/2021/02/Tele- Social-Care_Implications-and-Strategies_Final.pdf. Accessed 3/13/2023. 24. Bickel G, Nord M, Price C, Hamilton W, Cook J. Guide to Measuring Household Food Security. 2000. 25. Gold R, Bunce A, Cowburn S, et al. Adoption of Social Determinants of Health EHR Tools by Community Health Centers. Ann Fam Med. 2018;16(5):399-407. doi:10.1370/afm.2275. 26. Gold R, Bunce A, Cottrell E, et al. Study protocol: a pragmatic, stepped-wedge trial of tailored support for implementing social determinants of health documentation/action in community health centers, with realist evaluation. Implement Sci. 2019;14(1):9. doi:10.1186/s13012-019-0855-9. 27. Raudenbush SW, Bryk AS. Hierarchical Linear Models: Applications and Data Analysis Methods. 2 ed. Thousand Oaks, CA: Sage Publications, Inc.; 2002. 28. Dickinson LM, Basu A. Multilevel modeling and practice-based research. Ann Fam Med. 2005;3 Suppl 1(Suppl 1):S52-60. doi:10.1370/afm.340. 29. Centers for Medicare & Medicaid Services. Accountable Health Communities Model. 2022. https://innovation.cms.gov/innovation-models/ahcm. Accessed 11/29/2022. 30. McLeod MR, Vasudevan A, Warnick S, Jr., Wolfson JA. Screening for Food Insecurity in the Primary Care Setting: Type of Visit Matters. J Gen Intern Med. 2021;36(12):3907- 3909. doi:10.1007/s11606-020-06474-x. 31. Gruss I, Bunce A, Davis J, Dambrun K, Cottrell E, Gold R. Initiating and Implementing Social Determinants of Health Data Collection in Community Health Centers. Popul Health Manag. 2021;24(1):52-58. doi:10.1089/pop.2019.0205. 19 32. Palakshappa D, Vasan A, Khan S, Seifu L, Feudtner C, Fiks AG. Clinicians' Perceptions of Screening for Food Insecurity in Suburban Pediatric Practice. Pediatrics. 2017;140(1). doi:10.1542/peds.2017-0319. 33. Okafor M, Chiu S, Feinn R. Quantitative and qualitative results from implementation of a two-item food insecurity screening tool in healthcare settings in Connecticut. Prev Med Rep. 2020;20:101191. doi:10.1016/j.pmedr.2020.101191. 34. Greenwood-Ericksen M, DeJonckheere M, Syed F, Choudhury N, Cohen AJ, Tipirneni R. Implementation of Health-Related Social Needs Screening at Michigan Health Centers: A Qualitative Study. Ann Fam Med. 2021;19(4):310-317. doi:10.1370/afm.2690. 35. Barnidge E, LaBarge G, Krupsky K, Arthur J. Screening for Food Insecurity in Pediatric Clinical Settings: Opportunities and Barriers. J Community Health. 2017;42(1):51-57. doi:10.1007/s10900-016-0229-z. 36. Crotty BH, Hyun N, Polovneff A, et al. Analysis of Clinician and Patient Factors and Completion of Telemedicine Appointments Using Video. JAMA Netw Open. 2021;4(11):e2132917. doi:10.1001/jamanetworkopen.2021.32917. 37. Nederveld AL, Holtrop J, Duarte KF, Skalecki M, Broaddus-Shea ET. Multistakeholder Perspectives on Data Sharing to Address Patient Food Insecurity. J Am Board Fam Med. 2022;35(1):85-95. doi:10.3122/jabfm.2022.01.210093.