A claims-based algorithm to reduce relapse and cost in schizophrenia.

Area of Science:

• Health Services Research
• Psychiatric Epidemiology
• Health Informatics

Background:

• Schizophrenia relapse poses significant healthcare burdens.
• Accurate identification of high-risk patients is crucial for effective intervention.
• Managed care organizations seek algorithms to predict and mitigate patient relapse.

Purpose of the Study:

• To refine a predictive algorithm for identifying schizophrenia patients at high risk of relapse.
• To evaluate the effectiveness of a case management (CM) program utilizing this algorithm within a Medicaid population.

Main Methods:

• A predictive algorithm was developed using Medicaid claims data (medical and pharmacy) from August 2009 to July 2014 for 12,353 schizophrenia patients.
• The algorithm's performance was assessed using metrics like positive predictive power, negative predictive power, sensitivity, and specificity.
• A case management program targeted the highest-risk patients identified by the algorithm, comparing outcomes (hospitalizations, ED visits) against a control group.

Main Results:

• The refined algorithm demonstrated strong predictive capabilities: 64.0% positive predictive power, 94.7% negative predictive power, 40.2% sensitivity, and 97.9% specificity.
• Following case management intervention, the CM group saw a decrease in inpatient admissions (23.3% to 13.3%) and a ~15% reduction in monthly ED visits.
• Conversely, the control group experienced an increase in these adverse events.

Conclusions:

• The developed algorithm shows promise as an effective case-finding tool for managed care plans.
• The findings suggest the algorithm can aid in mitigating hospitalizations among high-risk schizophrenia patients.
• Further research may be warranted to confirm statistical significance and optimize intervention strategies.