Initial remarks: The following is a brief post, written by Dr. Muscarella, that reviews two well-known published studies that discuss the effectiveness of a bundle of best practices for reducing the rate of central line-associated blood stream infections (or, CLABSIs) in intensive care units, or ICUs.

In 2006, a prospective cohort study evaluated the effectiveness of a bundle of practices on the rate of central line-associated bloodstream infections (CLABSIs) in intensive care units (ICUs) in Michigan.21 This bundle included a popular checklist of five evidence-based infection-control practices.

Click here to read Dr. Muscarella’s review, entitled “Dear CDC,” that raises questions about a report by the Centers for Disease Control and Prevention (CDC), published in its renown journal Morbidity and Mortality Weekly Report, that concludes that federal efforts, in part, by the CDC are reducing the rates of CLABSIs in ICUs.

Published in The New England Journal of Medicine, this study found:

  1. that the rate of CLABSIs was significantly reduced within 3 months (compared to the baseline infection rate); and
  2. that this reduction was sustained for 15 more months after the intervention’s implementation.

Click here to read the peer-reviewed article Dr. Muscarella wrote – “Assessment of the Reported Effectiveness of Five Different Quality-Improvement Initiatives for the Prevention of Central Line-Associated Bloodstream Infections in Intensive Care Units” – that complements this post.

Published in the British Medical Journal in early 2010, a second study evaluating this same bundle’s effectiveness found that this reduced CLABSI rate (published in 2006) was sustained for an additional 18 months.22

Like an open-label drug study, many prospective cohort studies evaluating the effectiveness of an intervention on the CLABSI rate are not sufficiently rigorous to avoid or control for biases.

These two studies, however, like several others evaluating the effectiveness of an intervention,18-24,25 were not randomized, controlled, or blinded. Like an open-label drug study (which does not “blind” the researchers or the participants from the study’s intent or the administered drug), many prospective cohort studies evaluating the effectiveness of an intervention on the CLABSI rate are not sufficiently rigorous to either avoid bias or to control for the impact that one or more unrecognized confounding factors might have on the measured outcome.

The reader is directed to Table 1 of Dr. Muscarella’s more complete article (in PDF format) about CLABSIs to which this post refers by clicking here to download it.

Nor did either study evaluate or confirm the staff’s adherence to the bundle’s practices.21,22 As a consequence of these (and other) considerations (the reader is directed to Table 2 of Dr. Muscarella’s more complete article, in PDF format, about CLABSIs to which this post refers by clicking here to download it) these prospective cohort studies (as some of their authors acknowledge21,22) cannot exclude the null hypothesis, namely, that one or more confounding factors—not the (staff’s adherence to the) bundle of studied practices—caused the measured reduction in the rates of CLABSI, notwithstanding the common intimation by these studies that the bundle’s implementation was responsible for the observed reduction in the CLABSI rates (see: Table 2 of this article).

Click here to read a letter Dr. Muscarella wrote – “Dear Pediatrics: An Assessment of the Effectiveness of Bundles and Checklists for the Prevention of Central line-associated Bloodstream Infections in the Neonatal Intensive Care Units (NICUs) of New York State Hospitals” – that complements this post’s discussion.

Examples of factors that can cause the measured rate of CLABSI to under-report the true incidence of infection are listed in Table 1 (click here to download this table) and include:

  • “feedback”21,22 provided to clinicians about the study’s progress and their efforts to reduce the rate of CLABSI in ICUs;
  • rewarding ICUs reporting a reduced infection rate with “incentive payments”;22 and
  • the use of surveillance methods that do not remain fixed and, not only are not standardized among participating ICUs, but also may become less sensitive after the bundle’s implementation in ICUs and fail to detect and count every CLABSI10,11 (click here to download Table 1).

Click here to read Dr. Muscarella’s complementary postThree Facts and Myths about Central Line-Associated Bloodstream Infections.”

Indeed, none of the data collected by either study was independently validated for accuracy. Therefore, that the conclusions of these and other similar prospective cohort studies might have unwittingly over-exaggerated the percentage by which the studied intervention reduced the CLABSI rate, and assigned to the intervention an effect caused instead by a confounding factor (a “false-positive” result) cannot be ruled out (see: Table 1 and Table 2 – click here to download both).

References to this post are available by clicking here.

Post/Blog by: Lawrence F Muscarella PhD posted on 1-10-2013; updated 5-21-2013

Note: Click here to read the peer-reviewed article Dr. Muscarella wrote – Medical Errors, Infection-Control Breaches and the Use of Adulterated and Misbranded Medical Devices – that complements this post’s discussion and provides insights into both assessments of risk associated with identified infection-control lapses and the FDA’s regulation of devices in
tended to prevent healthcare-associated infections, or HAIs.

Leave a Reply

Your email address will not be published. Required fields are marked *