Comparison of Two Interquartile Ranges Utilizing the Bootstrap Method

• Activity Number: 259
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Barbara Napolitano and Martin L. Lesser*+ and Renee Pekmezaris and Dana Lustbader
• Companies: Feinstein Institute for Medical Research and Feinstein Institute for Medical Research and North Shore LIJ Health System and North Shore University Hospital
• Keywords: interquartile range ; bootstrap ; hospital length of stay

Abstract

Objective: To test the hypothesis that the "spread" of a hospital length of stay (LOS) distribution for an intervention was different from that of a control group. Method: The ratio (R) of the two interquartile ranges (IQRs) of the two groups was computed and, utilizing the bootstrap method with 1000 replications, the hypothesis that the mean ratio was 1 (equality) versus not equal to 1 (different) was tested with a z-test. Results: Median hospital LOS for the two treatment groups were similar, however, the IQR for the control group was 22 days, while the IQR of the intervention was 16 days, demonstrating a "tightening" of the hospital LOS distribution in the intervention group. The mean value of R (ratio of IQRs, 0.74 +/-0.07 SD) was significantly less than 1 (bootstrap z-test, P< 0.0003), demonstrating that the LOS variability for the intervention group was smaller than for control.