Table 3.
Multivariate linear regression results of county characteristics associated with 30-day behavioral health readmission rate for Medicaid enrollees diagnosed with schizophrenia (N=274 counties, 18 states).
| R2=46.3% | Coefficient | Standard Error |
|---|---|---|
| County demographic/comorbidity index (case mix adjuster) | .72*** | .16 |
| Schizophrenia quality of care | ||
| Continuous antipsychotic medication days supply+ | .004 | .02 |
| Visit continuity (no break > 60 days) | .011 | .03 |
| Behavioral health community follow-up within 7 days post hospital discharge++ | -.09*** | .01 |
| Utilization variables that may affect behavioral health readmission | ||
| Behavioral health hospitalization admission rate+ | .27*** | .07 |
| Average behavioral health hospitalization length of stay ++ | .0002 | .0007 |
| Community Treatment Capacity—per 100,000 U.S. Census Population | ||
| Psychiatrists | .0001 | .0002 |
| Free-standing psychiatric care beds | .00005 | .0001 |
| Free-standing alcohol/chemical dependency beds | -.0005 | .0004 |
| Community mental health centers | -.01ˆ | .007 |
| Federally qualified health centers | -.0004 | .002 |
| p. value: * ≤.05 ** ≤.01 *** ≤001 ˆp. value ≤1 | ||
| Predicted Probability of 30 Day Behavioral Health Readmission for Statistically Significant County-Level Rates Measures (adjusted for case mix) for a County | Predicted Probability (%) | 95% CI |
| Behavioral health community follow-up within 7 days post hospital discharge++ | ||
| 85% | 13.3 | 12.9-13.6 |
| 55% | 16.1 | 15.8-16.4 |
| Rate of behavioral health hospitalization admissions in county* | ||
| 30% | 16.7 | 16.4-17.0 |
| 10% | 11.3 | 11.0-11.6 |
Denominator = all schizophrenia diagnosed enrollees.
Denominator = all schizophrenia diagnosed enrollees with at least one MH/SUD hospitalization.
Results weighted based on the number of schizophrenia-diagnosed enrollees with ≥1 behavioral health hospitalization in each county.