National Readmission Rates and Outcomes for Patients Discharged Against Medical Advice
TAN-DOCTOROFMEDICINETHESIS-2018.pdf (392.2Kb)(embargoed until: 2019-05-01)
MetadataShow full item record
CitationTan, Sally Yi-Meng. 2018. National Readmission Rates and Outcomes for Patients Discharged Against Medical Advice. Doctoral dissertation, Harvard Medical School.
AbstractPurpose: Reducing non-elective readmissions is a strategic priority for hospitals. Individuals discharged against medical advice (AMA) are at high risk for readmissions. Previous studies on readmission outcomes after AMA discharge have been limited in their generalizability. This study sought to determine the likelihood of readmissions after AMA discharge, to identify factors associated with readmissions, and to assess in-hospital mortality, inpatient charges and length of stay for these readmissions.
Methods: Using the Agency for Healthcare Research and Quality (AHRQ) all-payer Nationwide Readmissions Database, we conducted a retrospective cohort analysis of 19,882,317 (95% CI: 12,232,775 - 20,535,955) weighted index admissions for patients ≥18 years admitted from January - November 2014. We calculated 30-day non-elective readmission rates, 30-day in-hospital mortality, lengths of stay, and hospital charges by discharge disposition. To assess differences in readmission rates by discharge disposition, we estimated multivariable logistic regression models and adjusted for patient, clinical, and hospital characteristics.
Results: Patients discharged AMA had a 30-day all-cause readmission rate of 21.0% (95% CI: 20.6 - 21.3) versus 10.4% (95% CI: 10.2 - 10.5%) for routine discharge to home or self-care. The difference remained significant (p<0.001) after adjusting for clinical, sociodemographic and hospital characteristics. Younger age, increased number of chronic comorbidities, low household income, and having public insurance were associated with higher readmission rates. Adjusted odds of 30-day in-hospital mortality was estimated to be 10% higher for patients discharged AMA versus routine discharge to home or self-care (aOR 1.10, 95% CI: 1.01 - 1.20). Leaving AMA resulted in the lowest hospital utilization (p<0.001), with a median 30-day total LOS of 5.8 days (IQR: 3.1 - 10.9) and total charges of $48,499 (IQR: $26,494 - $92,488).
Conclusions: The 30-day readmission rate was significantly higher for patients discharged AMA than for routine discharge to home or self-care. While patients discharged AMA had lower overall hospital utilization, in aggregate they had slightly higher adjusted in-hospital mortality. There is potential opportunity to both lower healthcare spending for lower-risk AMA patients and improve outcomes for higher-risk AMA patients. Future research should focus on better risk stratification of this patient population so that such targeted interventions may be implemented.
Citable link to this pagehttp://nrs.harvard.edu/urn-3:HUL.InstRepos:37006457