Cost-effectiveness analysis of Health care associated infection
Source:    Publish Time: 2012-08-11 07:19   3193 Views   Size:  16px  14px  12px
Author: Xuanqian XieHealth care associated (or called hospital-associated) infection (HAI) is one of important safety is

Author: Xuanqian Xie

Health care associated (or called hospital-associated) infection (HAI) is one of important safety issues in the health care system. The common HAIs include bloodstream infections, organism specific infection (such as C diff), pneumonia, urinary infection and surgical site infection (Foley et al., 2011). It was estimated that there are 2 million HAI per year in US, leading approximately 100,000 patients direct or indirect death (Foley et al., 2011). It is even worse in the developing countries. Quite a few articles provided estimates of HAI induced mortality (or called cause-specific mortality), cost and length of stay. But, the precision of these estimates can be doubted for many reasons. Primarily, no well established methods can be used to estimate the attributable death, attributable length of stay(LoS) and cost due to HAI. As far as I know, although some researchers proposed some methodology papers on this issue, few of them were applied in practice. Most present estimates of HAI related costs and LoS were based on less robust methods. Secondly, HAI related costs and LoS vary greatly across countries, health care systems, outbreak period or not, levels of hospitals and so on. Therefore, it is almost impossible to combine various sources of data to get “generic” estimates of HAI related costs and LoS. 

The ideal model of patient path can be constructed as follows.


Although this model is clear for most readers, the real calculation is not simple. I present some simplified calculation procedure here.

1.       Hazard ratio of infection versus without infection: Firstly, we can define a time-dependent covariate, infection or not, to all inpatients. Secondly, we define death as the event and censor patients when they are discharged. Then, we use time-dependent Cox proportional hazard model to estimate the hazard ratio of infection versus without infection for death. (For more advanced models for survival analysis, please have a look of R package. See link: )

2.       Attributable length of stay and cost: Both LoS and cost data are positive and left skewed. Therefore, in the methodology perspective, we can use same methods to estimate length of stay and cost. In the analysis, one of principles is only the LoS and cost after infection should be counted. The most common method used in practice is matching. The Multistate Models were applied in Beyersmann et al. 2006. And they wrote a methodology article of matching time-varying variable later (Wolkewitz, 2009). Forster et al. published an article to estimate the LoS for Clostridium difficile recently (Forster et al., 2012). Authors used the Kaplan–Meier curves with the time-varying covariate of C Diff. status. “A patient who acquires C. difficile would have two lines of data; the first corresponding to the time between admission and acquisition of C. difficile and the second corresponding to the time to discharge or death.” Authors used this way to estimate the median LoS of patients with C Diff. or not. But, the group of patients with C diff. actually are nested to those no infection (not independent samples for two groups). This simply comparison is not robust. Also, KM curve does not adjust other risk factors. Anyway, due to the absence of good statistical models, it is acceptable to do so.

An Example of Cost Effectiveness Analysis of Technology for Infection Control


Ventilator-associated pneumonia (VAP) is a common nosocomial infection in mechanically ventilated (MV) patients. Intubation with a modified endotracheal tube (ETT) that allows subglottic secretion drainage (SSD) may reduce incidence of VAP (Xie et al, 2011). Our primary interest was to estimate the cost-effectiveness of use of SSD ETT compared to standard ETT. The simple decision tree can be constructed.


Our analysis considered two costs: the cost of the SSD ETT device and the cost of VAP treatment. We assumed that resource consumption was balanced in both arms with the exception of these two costs. For simplicity, we ignored clinical outcomes (e.g. VAP mortality) whose cost is difficult to estimate. (Note: A meta-analysis (Muscedere et al. 2011) showed that SSD ETT did not significantly reduce ICU mortality, hospital mortality, or risk of adverse events.) The estimate of the baseline risk of VAP was obtained from data provided by the Infection Control Department, MUHC. The estimated efficacy of SSD ETT was based on the predicted risk ratio from our meta-analysis. Meta-analyses and cost-effectiveness analysis were performed using Winbugs 1.4. Monte Carlo simulations were used to obtain the credible intervals. More details of data inputs and results can be found in the original paper (Xie et al, 2011).

This model is simple, straightforward and widely used, even in the comprehensive report, such as cost-effectiveness analysis of coated catheters in preventing bloodstream infections (Hockenhull et al.,  2008).



Beyersmann J, Gastmeier P, Grundmann H, Barwolff S, Geffers C, Behnke M et al. Use of multistate models to assess prolongation of intensive care unit stay due to nosocomial infection. Infect Control Hosp Epidemiol 2006; 27(5):493-499.

Wolkewitz M, Beyersmann J, Gastmeier P, Schumacher M. Efficient risk set sampling when a time-dependent exposure is present: matching for time to exposure versus exposure density sampling. Methods Inf Med 2009; 48(5):438-443.

Forster AJ, Taljaard M, Oake N, Wilson K, Roth V, van WC. The effect of hospital-acquired infection with Clostridium difficile on length of stay in hospital. CMAJ 2012; 184(1):37-42.

Muscedere J, Rewa O, McKechnie K, Jiang X, Laporta D, Heyland DK. Subglottic secretion drainage for the prevention of ventilator-associated pneumonia: A systematic review and meta-analysis. Critical Care Medicine 2011; 39(8):1985-1991.

Steven L. Foley (Editor), Anne Y. Chen (Editor), Shabbir Simjee (Editor), Marcus J. Zervos (Editor). Molecular Techniques for the Study of Hospital-Acquired Infection, 2011, Wiley-Blackwell

Xie X, Nicolau I, McGregor M, Dendukuri N. Subglottic secretion drainage endotracheal tubes for prevention of ventilator-associated pneumonia. Montreal (Canada): Technology Assessment Unit (TAU) of the McGill University Health Centre (MUHC); 2011 Dec 21. Report no. 56. 38 p. Available from :

Hockenhull JC, Dwan K, Boland A, Smith G, Bagust A, Dündar Y, et al. The clinical effectiveness and cost-effectiveness of central venous catheters treated with anti-infective agents in preventing bloodstream infections: a systematic review and economic evaluation. Health Technol Assess 2008;12(12).