|
Activity Number:
|
90
- Novel Statistical Methods for COVID Pandemic and Other Current Health Policy Issues
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 9, 2021 : 10:00 AM to 11:50 AM
|
|
Sponsor:
|
Health Policy Statistics Section
|
|
Abstract #318471
|
|
|
Title:
|
Assessing Classification Accuracy of Hospitals' Performance on Discrete Outcome
|
|
Author(s):
|
Zhou Lan* and Zhenqiu Lin and Haiqun Lin and Shu-Xia Li and Chengan Du
|
|
Companies:
|
Yale School of Medicine and Yale New Haven Hospital and Rutgers University and Yale New Haven Hospital and Yale School of Medicine
|
|
Keywords:
|
hospital quality;
health policymaking;
classificaiton accuracy;
binary outcome measure;
Monte Carlo method;
case-mixing
|
|
Abstract:
|
The assessment and report of hospital quality is a critical component in health policymaking. Agencies have developed outcome measures to help policymakers categorize hospitals based on their performance measures. Adams et al. (2010) and He et al. (2014) calculate the rate by giving the probabilistic discrepancy that classification results based on the true values v.s. estimated values. However, these attempts were based on linear mixed models. It prohibits their applications to other outcome measures. To address this issue, we extend the previous work to an outcome measure under more generalized settings (Normand and Shahian 2007), where the mixed logistic model is fitted, and the risk-standardized mortality rate and interquartile range categorization are used to classify the hospitals. Our proposed classification accuracy method extends the method of Adams et al. (2010) and He et al. (2014) to accommodate the situation with binary outcomes and patient case-mix by adopting a Monte Carlo method to calculate classification. Methodological details are given to show our proposed scheme is proper and valid. Factors driving the accuracy rates are revealed through numerical studies.
|
Authors who are presenting talks have a * after their name.
