Ranking Institutions by Clustered Hospitalization Records

• Activity Number: 581
• Type: Contributed
• Date/Time: Wednesday, August 1, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Yan Wang*+ and David Zingmond and Honghu Liu
• Companies: University of California at Los Angeles Fielding School of Public Health and University of California at Los Angeles and University of California at Los Angeles School of Dentistry
• Keywords: Survey Logistic Model ; Variable Selection ; Rank ; Clustered observations ; Stroke

Abstract

Ranking institutions according to different measures clustered within certain domains is very common in many fields, e.g. Universities are ranked each year based on majors according to faculties' reputation, students' performance, source of funding, etc.; Facilities are ranked based on diseases according to doctors' specialty, patients' complications, outcome of treatments, etc. Often these measures are not independent and are always clustered within certain domains. This problem can be more complicated because the collection of the data and the validity of each measure. The situation can be simplified as that how the weight for each measure is being assigned statistically and how the final score for ranking is calculated by taking the data complications and the institution characteristics into account. We propose the ranking methods based on statistical models (Survey Logistic model and Proportional Hazard model) to rank institutions. Finally, the methods will be applied to the analytic administrative data collected by California licensed hospitals to rank their performance in stroke acute care.