Comparing rules for truncating hospital length of stay

Details

Serval ID
serval:BIB_C7D551F51BC9
Type
Article: article from journal or magazin.
Collection
Publications
Institution
Title
Comparing rules for truncating hospital length of stay
Journal
Casemix Quarterly
Author(s)
Ruffieux Christiane, Paccaud Fred, Marazzi Alfio
Publication state
Published
Issued date
2000
Volume
2
Number
1
Pages
3-11
Notes
SAPHIRID:47974
Abstract
Most distributions of hospital length of stay are asymmetric, with a long right tail and some very large observations (outliers). These features vitiate the reliability of many statistical summaries, such as the arithmetic mean, and comparisons based on them. A common remedy is to truncate (i.e., remove) values outside some limits and take the arithmetic mean of the remaining values. In general, the limits are based on a position measure (e.g., mean, median, quartiles) and a scale measure (e.g., standard deviation, median absolute deviation, interquartile range). In addition, a scale transformation (usually the logarithm) is frequently used.Using a data base with almost five millions hospital stays from five European countries, this paper explores the performance of five common truncation rules combining various options on transformation, position and scale. These rules are compared with a new one called " approximated quartile based truncated mean " or AQTM. The AQTM is based on a parametric model which takes into account the shape of the data distribution. This paper shows that the usual truncation rules produce very different estimates, and that most of the usual rules are biased with respect to the AQTM. There is one exception based on symmetric truncation beyond the interquartile range on the logarithmic scale. Since the AQTM is computationally as simple as the usual rules, but has a better foundation, it is preferred. [Authors]
Keywords
Length of Stay , Bias (Epidemiology) , Statistical Distributions , Outliers, DRG , Statistics as Topic
Create date
14/03/2008 10:20
Last modification date
20/08/2019 15:43
Usage data