Rethinking social needs and health care utilization-treating the symptom, not the cause

Sidebottom and colleagues examined the association between health-related social needs (HRSNs) and health care utilization within the Accountable Health Communities (AHC) model, a national demonstration designed to test whether screening for and addressing social needs could reduce avoidable health care use. Their analysis explored 2 key questions: (1) whether baseline HRSNs were associated with health care utilization, and (2) whether the resolution of these needs led to reductions in inpatient and emergency department use. They found that 5 HRSNs—housing instability, food insecurity, transportation challenges, affordability of utilities, and interpersonal safety—were significantly associated with higher emergency department utilization at baseline. In addition, housing instability and transportation challenges were associated with increased inpatient admissions at baseline, even after adjusting for confounders. Despite these associations, the resolution of HRSNs was not associated with lower utilization in a fully adjusted model controlling for demographics, comorbidities, and additional HRSNs. 

Is this truly the case, or do these findings instead reflect a limitation in how researchers define resolution and measure it in this line of research? In the AHC model, resolution was operationalized as a participant no longer self-reporting a need during follow-up screening. This binary measure fails to capture the gradations and fluctuations in need severity. For example, the AHC food insecurity question asks, “Within the past 12 months, you worried that your food would run out before you got money to buy more.” Respondents who answer often true or sometimes true are both coded as food insecure. If someone moves from often true to sometimes true, that improvement, although meaningful, is not reflected in the results. Such oversimplification risks obscuring genuine progress and underestimating the nuanced ways in which social interventions affect people’s lives.

Moreover, the instability of many social conditions complicates measurement. Social needs are dynamic, and short follow-up periods (eg, 6 months) may not capture the cyclical nature of hardship. A person may be food secure one month and insecure the next; a family may find temporary housing only to lose it again when rent increases. Additionally, a patient may have one need resolved, but a new need may arise months later that can impact utilization. This reality highlights the need for a continued, adaptive challenged to address these challenges. Understanding how these needs evolve requires more longitudinal data and refined metrics that go beyond binary indicators.