The Impact of a Mandated Trauma Center Alcohol Intervention on Readmission and Cost per Readmission in Arizona

Abstract

Background: Persons appearing in trauma centers have a higher prevalence of unhealthy alcohol use than the general population. Screening and brief intervention (SBI) is designed to moderate drinking levels and avoid costly future readmissions, but few studies have examined the impact of SBI on hospital readmissions and health care costs in a trauma population.

Research design: This study uses comparative interrupted time-series and the Arizona State Inpatient Database to estimate the effect of the American College of Surgeons Committee on Trauma SBI mandate on the probability of readmission and cost per readmission in Arizona trauma centers. We compare individuals with and without an alcohol diagnosis code before and after the mandate was implemented.

Results: The mandate resulted in a 2.2 percentage point reduction (44%) in the probability of readmission. Total health care and readmission costs were not affected by the mandate.

Conclusions: The estimates are consistent with a differential effect of SBI: SBI reduces readmissions among those who present with a less serious alcohol-related problem. Persons with more serious alcohol problems are less likely to respond to SBI. These higher risk individuals likely have a higher cost, which may explain the lack of change in readmission costs. Our study is a macrolevel intent-to-treat analysis of SBI's impact that corroborates the potential of SBI implied by efficacy studies in trauma centers and other settings. This study provides a framework for future research involving more states and health systems and evaluating other SBI policies.

Appendix Figure A-1.

Average Cost per Admission per Month. The upper panel presents the trends for readmissions, and the lower panel presents trends for the index admissions. The solid lines represent the alcohol-diagnosed group, and dashed lines represent the non-alcohol-diagnosed group. Time ranges from 16 months pre-mandate to 16 months post-mandate.

Appendix Figure A-2.

Distribution of Index Admission Costs. This figure shows the distribution of index admission costs for the six Arizona trauma centers included in the sample, 16 months pre-mandate to 16 months post-mandate.

Appendix Figure A-3.

Distribution of Readmission Costs. This figure shows the distribution of readmission costs for the six Arizona trauma centers included in the sample, 16 months pre-mandate to 16 months post-mandate.

Appendix Figure A-4.

Average Readmission Rate and Alcohol-Diagnosis Rate per Month. The solid line represents the percentage of readmissions in a given month, and dashed lines represent the percentage of admissions with an alcohol diagnosis in a given month. Time ranges from 16 months pre-mandate to 16 months post-mandate.

Figure 1.

CITS Predictions for the Probability of a Readmission. The solid black lines represent the predicted trend from the CITS model in Equation (1) for the alcohol diagnosis code group, and the dashed lines represent the predicted trend for the non-alcohol-diagnosis code group. Solid dots represent the unconditional monthly readmission rate for the alcohol diagnosis code group, smoothed using a 2-month moving average. Hollow dots represent the unconditional monthly readmission rate for the non-alcohol-diagnosed group, smoothed using a 2-month moving average. Time ranges from 12 months pre-mandate to 12 months post-mandate.

Figures 2a and 2b.

CITS Predictions for the Cost per Admission. The left panel (a) contains the log cost per readmission conditional on readmission. The right panel (b) contains the log cost per index admission. The solid black lines represent predicted trend from the CITS model in Equation (1) for the alcohol-diagnosis code group, and the dashed lines represent the predicted trend for the non-alcohol-diagnosis code group. Solid dots represent the unconditional monthly cost per admission for the alcohol-diagnosis code group, smoothed using a 2-month moving average. Hollow dots represent the unconditional monthly cost per admission for the non-alcohol-diagnosis code group, smoothed using a 2-month moving average. Time ranges from 12 months pre-mandate to 12 months post-mandate.

Figures 2a and 2b.

CITS Predictions for the Cost per Admission. The left panel (a) contains the log cost per readmission conditional on readmission. The right panel (b) contains the log cost per index admission. The solid black lines represent predicted trend from the CITS model in Equation (1) for the alcohol-diagnosis code group, and the dashed lines represent the predicted trend for the non-alcohol-diagnosis code group. Solid dots represent the unconditional monthly cost per admission for the alcohol-diagnosis code group, smoothed using a 2-month moving average. Hollow dots represent the unconditional monthly cost per admission for the non-alcohol-diagnosis code group, smoothed using a 2-month moving average. Time ranges from 12 months pre-mandate to 12 months post-mandate.