RESUMEN
This paper examines risk factors influencing food insecurity during the first year of the COVID-19 pandemic in a state in the U.S. heavily impacted by it and offers recommendations for multi-sector intervention. The U.S. Census Bureau Household Pulse Survey was analyzed to evaluate the impacts of COVID-19 on food security in Massachusetts from April 2020 through March 2021 using a study sample of 57,678 participants. Food security was defined as a categorical variable (food security, marginal food security, low food security, very low food security) and binary variable (food security and food insecurity). Known or suspected factors that contribute to it, such as childcare, education, employment, housing, and transportation were examined in multivariate logistic regression models. Data imputation methods accounted for missing data. Sociodemographic characteristics, including lower education level and living in a household with children, were determinants of food insecurity. Another factor that influenced food insecurity was economic hardships, such as unemployment, being laid off due to COVID-19, not working due to concerns about contracting or spreading COVID-19, or not having enough money to buy food. A third factor influencing food insecurity was food environment, such as lack of geographic access to healthy foods. Some of these factors have been exacerbated by the pandemic and will continue to impact food security. These should be addressed through a comprehensive approach with public health efforts considering all levels of the social ecological model and the context created by the pandemic.
RESUMEN
INTRODUCTION: Substance use disorders (SUD) are chronic conditions that often warrant coordinated medical care throughout a relapsing and remitting course. However, SUD treatment is frequently measured as a binary outcome, where individuals either receive or do not receive care following the immediate treatment seeking event. This study aimed to describe longitudinal treatment seeking behaviors by assessing service use patterns among individuals with diagnosed SUDs in a safety net hospital that offers a "no wrong door" care model. This study also examined whether certain patient subgroups were more likely to transition to service use patterns that support recovery or treatment disengagement over time. METHODS: The team conducted a retrospective cohort study using electronic health record data from adult patients diagnosed with SUDs (n = 1157) who regularly accessed services at a safety net hospital over a five-year period. The study used latent class analysis (LCA) and latent profile analysis to empirically identify distinct treatment utilization profiles of individuals with SUDs. We used multinomial logistic regression to evaluate predictors of class membership and transitions over a five-year period. RESULTS: The research team identified five distinct service use classes, including patients who disengaged from services (42.4%), or those who predominantly used outpatient substance use services (7.0%), mental health services (13.0%), primary care services (24.7%), or other specialty care services (13.1%). Being female and an older adult were statistically significant predictors for membership in any service use-driven status. Black patients had increased odds for "substance use service" and "primary care" service statuses over time. CONCLUSION: LCA and latent transition analysis (LTA) methods are novel approaches for identifying profiles of patients with higher risk for health services disengagement. SUD treatment engagement strategies are needed to reach males, young adults, and individuals with non-opioid SUDs.