Objectives To clarify the risk factors for falls in hospital settings and to propose the use of such factors to identify high-risk persons at admission.

Design Prospective cohort study.

Setting Fukushima Medical University Hospital, Japan, from August 2008 and September 2009.

Participants 9957 adult consecutive inpatients admitted to our hospital.

Methods Information was collected at admission from clinical records obtained from a structured questionnaire conducted in face-to-face interviews with subjects by nurses and doctors and fall events were collected from clinical records.

Results The proportion of patients who fell during follow-up was 2.5% and the incidence of falls was 3.28 per 100 person-days. There were significant differences in age, history of falling, cognitive dysfunction, planned surgery, wheelchair use, need for help to move, use of a remote caring system, rehabilitation, use of laxative, hypnotic or psychotropic medications and need for help with activities of daily living (ADL) between patients who did and did not fall. Multivariable adjusted ORs for falls showed that age, history of falls and need for help with ADL were common risk factors in both men and women. Using psychotropic medication also increased the risk of falling in men while cognitive dysfunction and use of hypnotic medication increased the risk of falling in women. Planned surgery was associated with a low risk of falls in women.

Conclusions To prevent falls in inpatients it is important to identify high-risk persons. Age, history of falling and the need for help with ADL are the most important pieces of information to be obtained at admission. Care plans for patients including fall prevention should be clear and considered.