Which prognostic factors predict both in-hospital delirium and falls? An umbrella review of meta-analyses

Introduction

In-hospital delirium and in-hospital falls are leading causes of harm to older people in hospital and are inter-related. Interventions to prevent in-hospital delirium have been shown also to prevent in-hospital falls. To better target interventions to older in-patients at risk both of delirium and falls, a prognostic model that predicts both delirium and falls risk would be clinically useful. Prognostic models intended for routine clinical care are most effective when based on information routinely captured in electronic health records (EHRs).

Objective

To identify overlapping prognostic factors, routinely captured in EHRs, that predict either in-hospital delirium or falls in older adults and grade the strength of association.

Method

Systematic reviews with meta-analyses of prognostic factors, routinely captured in EHRs, for either in-hospital delirium or falls were included. Pooled effect estimates from multi-variable analyses were prioritised.  A MeaSurement Tool to Assess Systematic Reviews 2 (AMSTAR-2) was used to rate methodological quality of systematic reviews; reviews rated critically low were excluded from the final analysis. Strength of association was graded using established criteria for umbrella reviews.

Results

Twenty-six non-overlapping systematic reviews with 189 meta-analyses (164 delirium, 26 falls) were included in the final analysis. Seventy meta-analyses (36.8%) were from multi-variable analyses only. In pooled analyses, the following prognostic factors were associated with an increased risk of both in-hospital delirium and falls: age (years), male sex, cognitive impairment, hyponatraemia, history of falls, cerebrovascular disease, Parkinson’s disease and peripheral arterial disease. Frailty was significantly associated with delirium, but there was no high-quality meta-analysis examining its prognostic effect for falls.

Conclusions
This review identifies EHR-available prognostic factors for inclusion in prognostic models to predict both in-hospital delirium and falls in older adults. Future work should prioritise high-quality primary research of prognostic factors for in-hospital falls.