Delirium is a common mental disorder with serious adverse outcomes in hospitalised patients. It is associated with increases in mortality, physical morbidity, length of hospital stay, institutionalisation and costs to healthcare providers. A range of risk factors has been implicated in its aetiology, including aspects of the routine care and environment in hospitals. Prevention of delirium is clearly desirable from patients' and carers' perspectives, and to reduce hospital costs. Yet it is currently unclear whether interventions for prevention of delirium are effective, whether they can be successfully delivered in all environments, and whether different interventions are necessary for different groups of patients.
Our primary objective was to determine the effectiveness of interventions designed to prevent delirium in hospitalised patients. We also aimed to highlight the quality and quantity of research evidence to prevent delirium in these settings.
We searched the Specialized Register of the Cochrane Dementia and Cognitive Improvement Group on 28th September, 2005. As the searches in MEDLINE, EMBASE, CINAHL and PsycINFO for the Specialized Register would not necessarily have picked up all delirium prevention trials, these databases were searched again on 28th October, 2005. We also examined reference lists of retrieved articles, reviews and books. Experts in this field were contacted and the Internet searched for further references and to locate unpublished trials.
Randomised controlled trials evaluating any interventions to prevent delirium in hospitalised patients.
Data collection and quality assessment were performed by three reviewers independently and agreement reached by consensus.
Six studies with a total of 833 participants were identified for inclusion. All were conducted in surgical settings, five in orthopaedic surgery and one in patients undergoing resection for gastric or colon cancer. Only one study of 126 hip fracture patients comparing proactive geriatric consultation with usual care was sufficiently powered to detect a difference in the primary outcome, incident delirium. Total cumulative delirium incidence during admission was reduced in the intervention group (OR 0.48 [95% CI 0.23, 0.98]; RR 0.64 [95% CI 0.37, 0.98]), suggesting a 'number needed to treat' of 5.6 patients to prevent one case. The intervention was particularly effective in preventing severe delirium. In logistic regression analyses adjusting for pre fracture dementia and Activities of Daily Living impairment, there was no reduction in effect size, OR 0.6, but this no longer remained significant [95% CI 0.3,1.3]. There was no effect on the duration of delirium episodes, length of hospital stay, and cognitive status or institutionalisation at discharge. There was also no significant difference in cumulative delirium incidence between treatment and control groups in a sub-group of 50 patients with dementia (RR 0.9 [95% CI 0.59, 1.36]). In another trial of low dose haloperidol prophylaxis, there was no difference in delirium incidence but the severity and duration of a delirium episode, and length of hospital stay were all reduced. We identified no completed studies in hospitalised medical, care of the elderly, general surgery, cancer or intensive care patients. In outcomes, no studies examined for death, use of psychotropic medication, activities of daily living, psychological morbidity, quality of life, carers or staff psychological morbidity, cost of intervention and cost to health care services. Outcomes were only reported up to discharge, with no studies reporting medium or longer-term effects.