Change in length of hospital stay (LOS) is frequently used to assess the impact and the costs of hospital-acquired complications. In order to compute the attributable change in LOS, it is crucial to account for the timing of events: A complication can only have an effect on LOS, once it has occured. These temporal dynamics can be adequately handled by multistate models; however, there is few software for such models available. We introduce an R-package "changeLOS" for computing change in LOS based on methods described in Schulgen and Schumacher (1996). We will illustrate the program on data from a prospective cohort study on hospital-acquired infections. Main features of the R-package "changeLOS" are R-methods to: (1) describe the multi-state model. (2) compute the Aalen-Johansen estimator for the matrix of transition probabilities P(u-, u) for all observed transition times u.(3) compute the Aalen-Johansen estimator for the matrix of transition probabilities P(s,t); the estimator is a finite matrix product of matrices P(u-,u) for every observed event time in the interval(s,t]. (4) visualize the temporal dynamics of the data, illustrated by transition probabilities. (5) compute and visualize change in LOS. (6) compute bootstrap variances for change in LOS.
| Version: | 2.0.9 |
| Depends: | R (≥ 1.8.1), survival |
| Date: | 2008-10-16 |
| Author: | Matthias Wangler, Jan Beyersmann |
| Maintainer: | Arthur Allignol <arthur.allignol at fdm.uni-freiburg.de> |
| License: | GPL (≥ 2) |
| In views: | Survival |
| CRAN checks: | changeLOS results |
Downloads:
| Package source: | changeLOS_2.0.9.tar.gz |
| MacOS X binary: | changeLOS_2.0.9.tgz |
| Windows binary: | changeLOS_2.0.9.zip |
| Reference manual: | changeLOS.pdf |
| Old sources: | changeLOS archive |