elderly, epidemiology, falls, hospital, nursing care, nursing work shift, time of falls, time series



  1. Lopez-Soto, Pablo J.
  2. Smolensky, Michael H.
  3. Sackett-Lundeen, Linda L.
  4. De Giorgi, Alfredo
  5. Rodriguez-Borrego, Maria A.
  6. Manfredini, Roberto
  7. Pelati, Cristiano
  8. Fabbian, Fabio


Background: A potentially important factor yet to receive adequate study is the time when hospital falls occur. A prior study conducted before the system-wide introduction of preventive measures revealed a biphasic 24-hour pattern of hospital falls with major peak in the morning.


Objectives: The purpose was to identify the temporal patterning of falls among elderly patients in hospitals with comprehensive fall prevention programs in place.


Methods: A 4-year observational study was conducted by the local health authority in the five nonteaching public hospitals located in the province of Ferrara, Italy. Fall records involving patients of ages >=65 years hospitalized in the general medical departments were used. Single- and multiple-component cosinor (time series) analyses were used to explore 24-hour, weekly, and annual patterns of falls.


Results: A total of 763 falls were experienced by 709 different elderly hospitalized patients. Falls typically took place in the patient's hospital room (72%) and bathroom (23%). Major causes were patient instability (32%) and accident (13%), and most occurred when not wearing footwear (45%) or wearing inappropriate sling-back open-toe shoes (39%). Falls happened while standing (39%), while seated (21%), and while getting into, out of, or laying in bed (32%)-either with the bed rails raised or lowered. Fall outcome usually involved no injury (58%) or slight injury (35%), but some (7%) were disabling. Fall occurrence was higher during the night (46%) compared to either the morning (30%) or afternoon (24%) shift. Patterns across 24 hours were characterized by a single major and one or more minor peaks that seemed to be associated with a variety of scheduled patient, hospital, and nursing activities. Multiple-component cosinor analysis identified significant (p < .05) prominent day-night patterns according to fall location, patient position, cause, injury severity, and type of footwear. Falls were more frequent, but not significantly so, on Fridays, Sundays, and Mondays compared with Tuesdays, and were more frequent in winter and spring (p = .003).


Discussion: Documentation by cause and circumstance of these moderate- to high-amplitude temporal patterns in hospital falls of elderly patients advances the knowledge of fall epidemiology by identifying the times of day, week, and year and nursing shifts of elevated risk that is of critical importance to improving hospital patient safety programs.