BACKGROUND

Healthcare-associated infections (HAIs) are a major threat to patient safety, and are associated with mortality rates varying from 5% to 35%. Important risk factors associated with HAIs are the use of invasive medical devices (e.g. central lines, urinary catheters and mechanical ventilators), and poor staff adherence to infection prevention practices during insertion and care for the devices when in place. There are specific risk profiles for each device, but in general, the breakdown of aseptic technique during insertion and care for the device, as well as the duration of device use, are important factors for the development of these serious and costly infections.

OBJECTIVES

To assess the effectiveness of different interventions, alone or in combination, which target healthcare professionals or healthcare organisations to improve professional adherence to infection control guidelines on device-related infection rates and measures of adherence.

SEARCH METHODS

We searched the following electronic databases for primary studies up to June 2012: the Cochrane Effective Paractice and Organisation of Care (EPOC) Group Specialised Register, the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, and CINAHL. We searched reference lists and contacted authors of included studies. We also searched the Cochrane Database of Systematic Reviews and Database of Abstracts of Reviews of Effectiveness (DARE) for related reviews.

SELECTION CRITERIA

We included randomised controlled trials (RCTs), non-randomised controlled trials (NRCTs), controlled before-after (CBA) studies and interrupted time series (ITS) studies that complied with the Cochrane EPOC Group methodological criteria, and that evaluated interventions to improve professional adherence to guidelines for the prevention of device-related infections.

DATA COLLECTION AND ANALYSIS

Two review authors independently extracted data and assessed the risk of bias of each included study using the Cochrane EPOC 'Risk of bias' tool. We contacted authors of original papers to obtain missing information.

MAIN RESULTS

We included 13 studies: one cluster randomised controlled trial (CRCT) and 12 ITS studies, involving 40 hospitals, 51 intensive care units (ICUs), 27 wards, and more than 3504 patients and 1406 healthcare professionals. Six of the included studies targeted adherence to guidelines to prevent central line-associated blood stream infections (CLABSIs); another six studies targeted adherence to guidelines to prevent ventilator-associated pneumonia (VAP), and one study focused on interventions to improve urinary catheter practices. We judged all included studies to be at moderate or high risk of bias.The largest median effect on rates of VAP was found at nine months follow-up with a decrease of 7.36 (-10.82 to 3.14) cases per 1000 ventilator days (five studies and 15 sites). The one included cluster randomised controlled trial (CRCT) observed, improved urinary catheter practices five weeks after the intervention (absolute difference 12.2 percentage points), however, the statistical significance of this is unknown given a unit of analysis error. It is worth noting that N = 6 interventions that did result in significantly decreased infection rates involved more than one active intervention, which in some cases, was repeatedly administered over time, and further, that one intervention involving specialised oral care personnel showed the largest step change (-22.9 cases per 1000 ventilator days (standard error (SE) 4.0), and also the largest slope change (-6.45 cases per 1000 ventilator days (SE 1.42, P = 0.002)) among the included studies. We attempted to combine the results for studies targeting the same indwelling medical device (central line catheters or mechanical ventilators) and reporting the same outcomes (CLABSI and VAP rate) in two separate meta-analyses, but due to very high statistical heterogeneity among included studies (I(2) up to 97%), we did not retain these analyses. Six of the included studies reported post-intervention adherence scores ranging from 14% to 98%. The effect on rates of infection were mixed and the effect sizes were small, with the largest median effect for the change in level (interquartile range (IQR)) for the six CLABSI studies being observed at three months follow-up was a decrease of 0.6 (-2.74 to 0.28) cases per 1000 central line days (six studies and 36 sites). This change was not sustained over longer follow-up times.

AUTHORS' CONCLUSIONS

The low to very low quality of the evidence of studies included in this review provides insufficient evidence to determine with certainty which interventions are most effective in changing professional behaviour and in what contexts. However, interventions that may be worth further study are educational interventions involving more than one active element and that are repeatedly administered over time, and interventions employing specialised personnel, who are focused on an aspect of care that is supported by evidence e.g. dentists/dental auxiliaries performing oral care for VAP prevention.