Although not without controversy, readmission is entrenched as a hospital quality metric. The Centers for Medicare and Medicaid Services (CMS) in the U.S., for example, currently uses 30-day readmission rates to determine hospital reimbursement rates. To-date, statistical analyses for readmission have hinged on fitting a logistic-Normal generalized linear mixed model. Such an analysis, however, ignores deaths as a competing risk. For clinical conditions with high mortality, such as pancreatic cancer, ignoring death can have profound effects in terms of conclusions regarding readmission; a hospitals seemingly good performance for readmission may be an artifact of it having poor performance for mortality. Furthermore, in such settings, scientific and policy interest may lie in understanding co-variation of performance with respect to readmission and mortality simultaneously. We propose novel multivariate hospital-level performance measures for readmission and mortality that derive from an analysis that of cluster-correlated semi-competing risks data. We also consider a number of profiling-related goals, including the identification of extreme performers and the bivariate classification of hospitals according to whether they have higher-/lower-than-expected readmission and mortality rates. To the best of our knowledge the latter is novel as a profiling goal. Towards achieving these goals this we develop a Bayesian decision-theoretic approach that characterizes hospitals on the basis of minimizing the posterior expected loss for an appropriate loss function. In some settings, particularly if the number of hospitals is large, the computational burden may be prohibitive. To resolve this, we propose a series of analysis strategies that will be useful in practice. Throughout the methods are illustrated with data from CMS on N=17,685 patients diagnosed with pancreatic cancer between 2000-2012 at one of J=264 hospitals in California.
Department of BiostatisticsMeasuring performance for end-of-life care
Sebastien Haneuse, PhD, Associate Professor Department of Biostatistics, Harvard University
February 21, 2019
3:30 pm - 5:00 pm
3755 SPH I
1415 Washington Heights
Ann Arbor, MI 48109-2029
Sponsored by: Department of Biostatistics
Contact Information: Zhenke Wu (zhenkewu@umich.edu) & Peisong Han (peisong@umich.edu)
Although not without controversy, readmission is entrenched as a hospital quality metric. The Centers for Medicare and Medicaid Services (CMS) in the U.S., for example, currently uses 30-day readmission rates to determine hospital reimbursement rates. To-date, statistical analyses for readmission have hinged on fitting a logistic-Normal generalized linear mixed model. Such an analysis, however, ignores deaths as a competing risk. For clinical conditions with high mortality, such as pancreatic cancer, ignoring death can have profound effects in terms of conclusions regarding readmission; a hospitals seemingly good performance for readmission may be an artifact of it having poor performance for mortality. Furthermore, in such settings, scientific and policy interest may lie in understanding co-variation of performance with respect to readmission and mortality simultaneously. We propose novel multivariate hospital-level performance measures for readmission and mortality that derive from an analysis that of cluster-correlated semi-competing risks data. We also consider a number of profiling-related goals, including the identification of extreme performers and the bivariate classification of hospitals according to whether they have higher-/lower-than-expected readmission and mortality rates. To the best of our knowledge the latter is novel as a profiling goal. Towards achieving these goals this we develop a Bayesian decision-theoretic approach that characterizes hospitals on the basis of minimizing the posterior expected loss for an appropriate loss function. In some settings, particularly if the number of hospitals is large, the computational burden may be prohibitive. To resolve this, we propose a series of analysis strategies that will be useful in practice. Throughout the methods are illustrated with data from CMS on N=17,685 patients diagnosed with pancreatic cancer between 2000-2012 at one of J=264 hospitals in California.