Use of Predictive Analytics to Identify Unhealthy Opioid Use and Guide Intervention

Objective:

The authors aimed to use the newly developed Opioid Risk Stratification Tool to identify individuals who may be at risk for unhealthy opioid use and to examine the impact of applying a mailing and engagement intervention to this population and their prescribers, with the goal of reducing high-risk prescribing behaviors, opioid medication use, and mortality rates.

Methods:

A nonrandomized controlled study was conducted with members from two Medicaid managed care organizations. In both the intervention (N=131) and control (N=187) groups, an algorithm identified members at moderate to high risk for hazardous opioid use. Members at increased risk in the intervention group and their prescribers received a letter from the managed care organization, and members still at risk 3 months after the mailing were contacted by a care coordinator. Individuals in the control group were not contacted. Medicaid claims data were used to compare opioid use and prescribing practices between groups before and after the intervention.

Results:

Individuals in the intervention group were less likely to have any opioid prescription postintervention compared with those in the control group (OR=0.55, p<0.001), and the intervention group had a greater reduction in the number of individuals with concurrent opioid and benzodiazepine prescriptions (OR=0.49, p=0.042). Practices such as multiple opioid prescriptions and multiple prescribers of opioids were not affected by the intervention.

Conclusions:

An intervention targeting individuals at risk for hazardous opioid use was associated with a reduction in some high-risk prescribing practices. Future research should determine the ideal mix of interventions to reduce as many risk factors as possible.