Artikel
Change in Length of Hospital Stay due to Hospital Infections: a counterfactual approach to deal with time-dependent confounders
Suche in Medline nach
Autoren
Veröffentlicht: | 6. September 2007 |
---|
Gliederung
Text
Introduction / Background: The length of hospital stay (LOS) is presumably affected by hospital-acquired infections (HI). Quantifying the extra stay due to HI is complicated because infection status changes over time. Furthermore, LOS and the risk of HI depend on health status and other patient's conditions. If those factors vary over time, they might act as time-dependent confounders. For example, the use of artificial ventilation affects LOS and is also associated with the occurrence of HI. Additionally, HI might increase the need for artificial ventilation. This means, after HI has occurred, the causal relation between ventilation and infection status is reversed.
Materials and methods: We analyze data from a prospective cohort study of hospital infections in intensive care units (ICU). Baseline characteristics and daily measurements of clinical variables are available. We address pneumonia as one of the most frequent and severe hospital-acquired infections.
To quantify the impact of HI on LOS, we use a method proposed by Robins which includes time-dependent confounders in an appropriate way. There, modeling is done within the counterfactual framework.
We give an introduction to counterfactual variables and review Robins' approach with focus on handling time-dependent confounders. To assess the extra stay, we take advantage of counterfactual variables and define the change in LOS as comparison to never having acquired HI.
Results: Of 1876 admitted patients, 151 obtained hospital-acquired pneumonia which lead to an estimated mean extra stay of 5.3 days (95% confidence interval [1.3;9.1]).
Discussion / Conclusions: The counterfactual approach avoids overestimation, is interpretationally appealing and allows adjustment for confounders. An extension to address the competing endpoints discharge and death is discussed.
References
- 1.
- Schulgen G and Schumacher M. Estimation of prolongation of hospital stay attributable to nosocomial infections: New approaches based on multistate models. Lifetime Data Analysis 1996; 2:219-240
- 2.
- Robins JM, Blevins M, Ritter G et al. G-Estimation of the effect of prophylaxis therapy for Pneumocystis carnii pneumonia on the survival of AIDS patients. Epidemiology 1992; 3:319-36
- 3.
- Robins JM. Structural nested failure time models. In: Survival Analysis. In: Armitage P, Colton T, eds. Encyclopedia of Biostatistics. Chichester, UK: John Wiley and Sons, 1998
- 4.
- Keiding N. Event history analysis and inference from observational epidemiology. Statistics in Medicine 1999; 18: 2353-63