Intracranial haemorrhage detected by cerebral computed tomography after falls in hospital acute medical wards

Abstract

AbstractBackgroundThere is little published data on brain imaging and intracranial haemorrhage after hospital inpatient falls. Imaging protocols for inpatient falls have been adopted from head injury guidelines developed from data in patients presenting to the Emergency Department. We sought to describe the use of brain computed tomography (CT) following inpatient falls, and determine the incidence and potential risk factors for intracranial haemorrhage.MethodsWe identified inpatient falls in acute medical wards at Monash Health, a large hospital network in the southeast region of Melbourne in Australia, from the incident reporting system during a 32 month period. We examined the post-fall medical assessment form, neurological observation chart and the diagnostic imaging system for details of the fall and brain CT findings. We used survival analysis to evaluate the timeliness of brain imaging and determined potential risk factors for intracranial haemorrhage by logistic regression.ResultsFrom 934 falls in 789 medical inpatients, 191 brain CT scans were performed. The median age of patients was 77 years. Only 55% of falls were from standing height and 24% experienced a head strike. Less than 10% of patients received an urgent scan within one hour, and timeliness of imaging was influenced by anticoagulation status rather than guideline determination of urgency. The overall incidence of intracranial haemorrhage was 0.9%. The factors associated with intracranial haemorrhage were head strike, anticoagulation, loss of consciousness or amnesia, drop in Glasgow Coma Scale and advanced chronic kidney disease.ConclusionsThe incidence of intracranial haemorrhage was low as most inpatient falls were at low risk for head injury. Research is needed to determine if guidelines specific for hospital inpatients may reduce unnecessary scans without compromising case detection, and improve timeliness of urgent scans.