Are Rates of Bed Occupancy and Nurse Absenteeism Significant Risk Factors for Hospital Acquired Clostridium difficile-Associated Disease?

Background: Since the early 2000s, the NAP1/BI/O27 strain of Clostridium difficile has been implicated in large outbreaks of C difficile-associated disease (CDAD) worldwide. In England this occurred concurrently with the introduction of policies aimed at improving financial efficiency within the National Health Service. These included reductions in the numbers of hospital in-patient beds and decreases in the length of stay. Mismatches in these changes have contributed to a health service that can be typically described as high throughput with elevated rates of bed occupancy and nurse absenteeism.

Objective: To establish if bed occupancy rates and nurse absenteeism are significant risk factors for CDAD in adult medical and surgical in-patients. 

Methods: University Hospitals of Leicester (UHL) NHS Trust is an 1800 bed, three sites, mixed secondary/tertiary medical school NHS hospital. A retrospective cohort study was carried out using a dataset constructed from demographic and pathology records of adult patients admitted to medical or surgical wards in UHL between April 2006 and April 2008 and contemporaneous staff records. Univariate and multivariate Cox regression modelling was used to examine the strength and significance of the associations of rates of bed occupancy and nurse absenteeism with rates of CDAD.  Variables collected included patient characteristics, antibiotic exposure, case-mix, and monthly ward absenteeism and bed occupancy rates.  CDAD was diagnosed using enzyme immunoassay.

Results: Over 24 months, 178, 673 patients were admitted (915,919 patient days). Median ward bed occupancy was 91% (range 0%-100%).  Median nurse absenteeism was 4.8% (range 0%-100%). 1627 patients developed CDAD giving an overall infection rate of 1.71 per 1000 patient days. Multivariate Cox regression analysis found a strong, positive and highly statistically significant association between bed occupancy and CDAD risk.  In the adjusted model, an exposure-response was observed across the first 4 levels of exposure.  The sharpest increase in risk occurred at occupancy levels above 81%.   Above this level the risk of CDAD increased from 14% to 65%.  The risk of CDAD peaked at occupancy levels of between 88.1% and 95%.  Patients on wards with between 88.1% and 95% bed occupancy were twice as likely to have CDAD (HR 1.99, 95% CI 1.43 – 2.75) when compared to those on wards with the lowest bed occupancy.  After adjusting for age, antibiotic exposure and case mix, the measure of effect was attenuated but remained highly statistically significant (p <0.001).  No meaningful association was found between nurse absenteeism and CDAD.

Conclusions: There was strong evidence of an association between high bed occupancy and CDAD rates. Target maximum bed occupancy rates should be an important part of both infection control and bed management policies.