Relationship of Clinical Factors, Self-reported Quality Metrics, Social Determinants of Health, and Treatment Frequency with Long-term Suicide Severity Outcomes
Abstract
Background: Suicide among adolescents and young adults is a major public health concern, highlighting the need to understand risk and protective factors. Yet, little is known about the clinical, psychosocial, and treatment-related factors influencing long-term suicide outcomes. This study aims to identify such factors among youth receiving community behavioral health (BH) care.
Methods: The study included 100 adolescents who received outpatient (BH)treatment in a safety-net health system. In 2023, participants completed a follow-up survey, including the (Computerized Adaptive Test – Mental Health; CAT-MH) for current suicide risk and two retrospective self-report measures: treatment experiences (Consumer Assessment of Healthcare Providers and Systems; CAHPS) and social determinants of health (CONNECT-S). Covariates associated with suicide severity were identified using a LASSO model including EHR-based sociocontextual, diagnostic, and service use variables, time to follow-up, CONNECT-S and CAHPS scores. Ordinary Least Squared (OLS) regression provided estimates of the associations between LASSO-selected covariates and treatment frequency (≥4 vs <4 visits per 3-month quarter) and interaction between CAHPS and treatment frequency.
Results: Ten covariates were selected by the LASSO model. OLS found that any emergency department use during the initial treatment period was associated with an increase in suicide severity of 17.1 [95% CI 1.9, 32.2; p = 0.03]. Preliminary evidence indicated that higher treatment frequency was associated with lower suicide risk severity approximately five years later among youth reporting more positive care experiences; in this group, a 1-point increase in care rating corresponded to a -0.4 reduction in suicide severity [95% CI: -0.8, 0.02; p = 0.06].
Conclusions: Preliminary findings suggest higher intensity care is only likely to be of value if it is of high patient-reported quality. Larger sample confirmatory studies are needed.
