Cross-sectional study of the prevalence, causes and management of hospital-onset diarrhoea

Abstract

The National Health Service in England advises hospitals collect data on hospital-onset diarrhoea (HOD). Contemporaneous data on HOD are lacking. To investigate prevalence, aetiology and management of HOD on medical, surgical and elderly-care wards. A cross-sectional study in a volunteer sample of UK hospitals, which collected data on one winter and one summer day in 2016. Patients admitted ≥72 h were screened for HOD (definition: ≥2 episodes of Bristol Stool Type 5-7 the day before the study, with diarrhoea onset >48 h after admission). Data on HOD aetiology and management were collected prospectively. Data were collected on 141 wards in 32 hospitals (16 acute, 16 teaching). Point-prevalence of HOD was 4.5% (230/5142 patients; 95% confidence interval (CI) 3.9-5.0%). Teaching hospital HOD prevalence (5.9%, 95% CI 5.1-6.9%) was twice that of acute hospitals (2.8%, 95% CI 2.1-3.5%; odds ratio 2.2, 95% CI 1.7-3.0). At least one potential cause was identified in 222/230 patients (97%): 107 (47%) had a relevant underlying condition, 125 (54%) were taking antimicrobials, and 195 (85%) other medication known to cause diarrhoea. Nine of 75 tested patients were Clostridium difficile toxin positive (4%). Eighty (35%) patients had a documented medical assessment of diarrhoea. Documentation of HOD in medical notes correlated with testing for C.difficile (78% of those tested vs 38% not tested, P<0.001). One-hundred and forty-four (63%) patients were not isolated following diarrhoea onset. HOD is a prevalent symptom affecting thousands of patients across the UK health system each day. Most patients had multiple potential causes of HOD, mainly iatrogenic, but only a third had medical assessment. Most were not tested for C.difficile and were not isolated.