The study drew on data from 750 adult Medicaid recipients enrolled at five community health centers, a population that represents some of the most economically vulnerable people in the United States. Medicaid recipients live on low incomes by definition of the program’s eligibility rules, and many work in service-sector jobs that were disproportionately disrupted by lockdowns, reduced hours, and illness during the pandemic. Participants in the cohort had a mean age of 43.2 years, with a standard deviation of 11.2 years, and the sample was racially and ethnically diverse: 15.3 percent of participants identified as Black and 54.9 percent as Hispanic. This demographic composition makes the cohort particularly informative, because communities of color in the United States experienced both higher rates of pandemic-related economic disruption and elevated burdens of COVID-19 illness itself.

Methodologically, the study is a secondary analysis of a longitudinal, observational cohort originally designed to examine other health outcomes. All key variables were self-reported at two time points: baseline and a one-year follow-up. Financial stress was measured with validated self-report items capturing how strained participants felt by their economic circumstances. Depressive symptoms were assessed using the PHQ-8, an eight-item version of the Patient Health Questionnaire that covers the core diagnostic criteria for major depressive episodes, and anxiety symptoms were measured with the GAD-7, the seven-item Generalized Anxiety Disorder scale widely used in both research and primary care. Social support was measured as perceived social support, a subjective rating of how much emotional and practical help participants felt was available to them from the people around them.

The analytical approach relied on multivariable linear regression models, the standard statistical tool for estimating how a continuous outcome changes as a function of one or more predictors while adjusting for covariates. The central question was whether social support moderated the relationship between changes in financial stress and changes in mental health symptoms. In statistical terms, moderation means that the slope of the association between the predictor and the outcome differs depending on the level of a third variable. The researchers tested this by including interaction terms between social support and financial stress change in their regression models, with separate models for depressive symptoms and anxiety symptoms as outcomes.

The results were striking in their clarity. Worsening financial stress over the one-year period was significantly associated with increases in depressive symptoms, with a p-value below 0.01, and with increases in anxiety symptoms, also with a p-value below 0.01. In other words, participants whose financial situation deteriorated during the pandemic reliably reported more depression and more anxiety a year later. This dose-response-like pattern between economic deterioration and psychological decline is consistent with a large body of research linking financial strain to common mental disorders, but the longitudinal design strengthens the causal inference considerably compared with cross-sectional surveys that measure stress and symptoms at a single moment in time.

The surprise came when the researchers examined the role of social support. Perceived social support on its own was not related to changes in depressive symptoms at all, with a p-value of 0.83, nor to changes in anxiety symptoms, with a p-value of 0.28. More importantly for the stress-buffering hypothesis, social support did not significantly moderate the effect of financial stress on depressive symptoms, with an interaction p-value of 0.51. For anxiety symptoms, the interaction approached but did not reach conventional statistical significance, with a p-value of 0.07, leaving open a faint possibility of a weak buffering effect that the study was not powered to confirm. Taken at face value, however, the findings indicate that having people to lean on did not protect these low-income adults from the psychological consequences of losing financial ground.

Why might social support fail as a buffer against financial stress when it has shown protective effects against other kinds of adversity? The authors and related literatures offer several plausible mechanisms. Financial stress is fundamentally a material problem: it stems from inability to pay rent, buy food, or cover medical bills. Emotional encouragement from friends and family, however sincere, cannot substitute for money. In a pandemic context, entire social networks were simultaneously strained by the same economic shock, which may have degraded the quality and availability of support precisely when it was most needed. There is also the possibility of contagion effects, in which the financial distress of close contacts amplifies rather than relieves one’s own worry. Additionally, perceived social support is a subjective measure, and people under severe economic pressure may perceive less support even when networks remain nominally intact, a phenomenon sometimes called support erosion.

The study’s limitations are worth keeping in view. As a secondary analysis of an observational cohort, it cannot definitively establish causation, although the temporal ordering of measurements helps. All measures were self-reported, introducing the possibility of reporting bias, and the sample was drawn from community health centers in one state, which may limit generalizability to other regions or to uninsured populations. The one-year follow-up window captures a specific and turbulent phase of the pandemic, and trajectories of financial and psychological recovery may differ over longer horizons. The near-significant interaction for anxiety symptoms suggests that future studies with larger samples should continue probing whether buffering effects exist for some outcomes or subgroups.

Nevertheless, the practical implications are significant for public health policy and clinical practice. If social support cannot be relied upon to protect low-income adults from the mental health effects of financial shocks, then interventions must target the financial stressor itself. Direct economic supports, such as expanded income assistance, unemployment insurance, food assistance, and eviction moratoria, may function as mental health interventions in their own right. Clinicians serving Medicaid populations should screen for financial strain and recognize that patients with strong family networks remain at risk of depression and anxiety when their economic circumstances worsen. The findings also challenge the broader cultural narrative that resilience is primarily a matter of personal relationships and coping skills, pointing instead toward structural determinants of mental health.

The research, funded by the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health under grant R01 DK124145-01, with additional support to Dr. Thorndike from NIH grant K24 HL16307 and to Dr. Cheng from the National Heart, Lung, and Blood Institute, was approved by the Partners Healthcare System Institutional Review Board and conducted in accordance with the Declaration of Helsinki. As the United States continues to grapple with the long economic shadow of the pandemic, the message of this study is sobering but useful: for the millions of Americans living on the economic margin, money problems are mental health problems, and no amount of social warmth fully insulates the mind from the weight of a shrinking wallet. Policies that stabilize household finances may do as much for psychological well-being as any therapy or support group, particularly for the low-income communities where the pandemic’s economic damage ran deepest.

Subject of Research: The moderating role of social support in the relationship between financial stress and mental health symptoms among low-income Medicaid recipients during the COVID-19 pandemic

Article Title: Moderating effect of social support on the association between changes in financial stress and depressive and anxiety symptoms during the COVID-19 pandemic in Medicaid recipients: a longitudinal study

Article References: Williams, P., Cheng, J., & Thorndike, A. N. (2026). Moderating effect of social support on the association between changes in financial stress and depressive and anxiety symptoms during the COVID-19 pandemic in Medicaid recipients: a longitudinal study. BMC Public Health. https://doi.org/10.1186/s12889-026-28862-4

Image Credits: AI Generated

DOI: 10.1186/s12889-026-28862-4

Keywords: social support, financial stress, depression, anxiety, Medicaid, COVID-19, mental health, low-income population, longitudinal study, public health, PHQ-8, GAD-7