Survival Analysis of Black Adolescent Mental Health Services Engagement

Background/Purpose: The no-show rate is high in ambulatory psychiatric care, especially community mental health centers, which usually result in higher amounts of early termination from mental health services. Patients who miss appointments and experience early termination from services also are more unwell and more functionally impaired, contributing to a higher risk and rate of hospitalization. Few studies have explored this critical treatment concern. More attention is needed on this phenomenon, especially focused at what demographic and clinical factors may correlate with or directly cause these clinical care adherence outcomes.

Conceptual Framework: The Andersen Model of Heath Care Utilization guided this study. Population level characteristics including gender, age, perceived service need, mental health diagnoses, and global assessment of functioning were examined for direct and indirect impact on mental health service use behavior as it relates to the outcome of early termination from services. Additionally, environmental factors were explored via Diagnostic and Statistical Manual of Mental Disorders Axis IV social conditions (economic, education, housing, and legal system interaction). These were tested for effect on the study sample experiencing psychiatric care and mental health services early termination.

Methods: Participants were recruited from a community mental health agency in a metropolitan, Midwestern city. The sample included 89 Black adolescent (12-17 years old) and parent dyads. Survival analysis was conducted to observe the service termination patterns of Black adolescents using an 18-month study window. Kaplan-Meier estimator was used to describe the survival function and discern group differences on service termination. A Cox proportional hazards model was employed to seek significant predictors of service termination. A data-driven process was followed to derive the final set of predictors retained in the model, including statistically significant interaction terms. Diagnostics including tests of the proportionality assumption and detection of influential cases were conducted to ensure that findings of the final model were robust.

Results: Within an 18-month study window, a total of 60 participants (67.4%) terminated services, with the number of months using services ranging from one day to just under 17.5 months. Statistically significant covariates predicting early service termination are stigma (p<.001), depression (p<.006), behavior disorder (p<.033), housing issues (p<.001), legal issues (p<.001), GAF (p<.007), and the interaction of economic issues with housing issues (p<.001). Educational issues showed a statistical trend (p<.053).

Conclusions and Implications: Nearly two-thirds of mental health service participants experienced an unacceptable early termination from services. The factors contributing to this worrisome finding are multiple and complex. Housing issues, stigma, and legal system interaction are the most powerful predictors of faster speed to early termination and likely priorities to address in service delivery, respectively. Additionally, those in need of housing services and suffering from economic issues may be more inclined to continue participating in treatment. Individuals experiencing depressive disorder (indicated by scale and not clinical diagnosis), behavioral disorders, and lower GAF scores are at reduced risk.