Multifactorial Interventions to Reduce Duration and Variability in Delays to Identification of Serious Injury After Falls in Hospital Inpatients.

Abstract

Inpatient falls are the most commonly reported safety incidents and are associated with serious injuries. This study aimed to use multifactorial interventions to reduce the delays to the diagnosis of serious injury in a time series analysis after serious incidents relating to falls within a central London Trust. A multiprofessional project team undertook process mapping to identify opportunities for improvement at different stages in the management of a fall. The interventions included an educational teaching session aimed at doctors, a lanyard card designed by doctors using the plan-do-study-act methodology, a falls-specific pager for radiographers, and a new system to refer to portering. Quantitative data were obtained using an serious incident database where serious injury occurred (SI data; n = 65) and routinely collected incident reporting database on falls regardless of injury (IR data; n = 178). Qualitative questionnaire data (n = 70) were also used to evaluate doctors' confidence in falls assessment before and after interventions. Results in the IR data demonstrated a significant reduction in the median (interquartile range) minutes delay in the time to review a patient after a fall from 81 (43-180) to 51 (26-112; P = 0.003) and the time to order imaging from 102 (45-370) to 50 (33-96; P = 0.04). Analysis of the SI database demonstrated a nonstatistically significant reduction in the overall time taken to detect serious injury after a fall from 348 (126-756) to 192 (108-384) minutes (P = 0.070). Furthermore, analysis using statistical process control charts showed evidence of special cause variation and a shift in the process in detecting serious harm after a fall. Junior doctors' confidence in investigations improved from 53% to 76% (P = 0.04) after the intervention. The cumulative application of multiple interventions with small individual effects resulted in a substantial positive effect on delays and variability in diagnosis of serious harm. Given a similar institutional context, the more effective interventions in our study could be adopted elsewhere.