Review of perioperative falls.

Falls are a known public health problem, and there is increasing recognition of the importance of perioperative falls for risk prediction and quality assessment. Our objective was to review existing literature regarding the occurrence, injuries, and risk factors of preoperative and postoperative falls. A systematized search of PubMed entries between 1947 and November 2015 produced 24 articles that met inclusion criteria. Most studied orthopaedic surgery patients older than 65 yr. Four were rated 'good' quality. Interrater reliability for the quality assessment was moderate (κ = 0.77). In the 3-12 months before surgery, the proportion of preoperative patients who fell ranged from 24 to 48%. Injuries were common (70%). The rate of postoperative falls ranged from 0.8 to 16.3 per 1000 person-days, with a gradual decline in the months after surgery. Injuries from postoperative falls occurred in 10-70% of fallers, and 5-20% experienced a severe injury. Risk factors were not well studied. Prospective studies reported a higher percentage of falls and fall-related injuries than retrospective studies, suggesting that there may be underdetection of falls and injuries with retrospective studies. Perioperative falls were more common than falls reported in the general community, even up to 12 months after surgery. Surgery-related falls may therefore occur beyond the hospitalization period. Future studies should use a prospective design, validated definitions, and broader populations to study perioperative falls. In particular, investigations of risk factors and follow-up after hospitalization are needed. REGISTRY NUMBER: PROSPERO registration number CRD42015029971.