Predictive validity of two process-of-care quality measures for residential substance use disorder treatment

Alex H. S. Harris, Shalini Gupta, Thomas Bowe, Laura S. Ellerbe, Tyler E. Phelps, Anna D. Rubinsky, John W. Finney, Steven M. Asch, Keith Humphreys, Jodie Trafton

In order to monitor and ultimately improve the quality of addiction treatment, professional societies, health care systems, and addiction treatment programs must establish clinical practice standards and then operationalize these standards into reliable, valid, and feasible quality measures. Before being implemented, quality measures should undergo tests of validity, including predictive validity. Predictive validity refers to the association between process-of-care quality measures and subsequent patient outcomes. This study evaluated the predictive validity of two process quality measures of residential substance use disorder (SUD) treatment. Washington Circle (WC) Continuity of Care quality measure is the proportion of patients having an outpatient SUD treatment encounter within 14 days after discharge from residential SUD treatment. The Early Discharge measure is the proportion of patients admitted to residential SUD treatment who discharged within 1 week of admission. The predictive validity of these process measures was evaluated in US Veterans Health Administration patients for whom utilization-based outcome and 2-year mortality data were available. Propensity score-weighted, mixed effects regression adjusted for pre-index imbalances between patients who did and did not meet the measures' criteria and clustering of patients within facilities. For the WC Continuity of Care measure, 76 % of 10,064 patients had a follow-up visit within 14 days of discharge. In propensity score-weighted models, patients who had a follow-up visit had a lower 2-year mortality rate [odds ratio (OR) = 0.77, p = 0.008], but no difference in subsequent detoxification episodes relative to patients without a follow-up visit. For the Early Discharge measure, 9.6 % of 10,176 discharged early and had significantly higher 2-year mortality (OR = 1.49, p < 0.001) and more subsequent detoxification episodes. These two measures of residential SUD treatment quality have strong associations with 2-year mortality and the Early Discharge measure is also associated with more subsequent detoxification episodes. These results provide initial support for the predictive validity of residential SUD treatment quality measures and represent the first time that any SUD quality measure has been shown to predict subsequent mortality.