Risk Factors and Injury Associated With Falls in Elderly Hospitalized Patients in a Community Hospital Jacques E. Chelly, MD, PhD, MBA,*Þ Linda Conroy, RN, MSN, MPM,þ Gregory Miller, AS,§ Marc N. Elliott, PhD,|| Jennifer L. Horne, BA,* and Mark E. Hudson, MD* Objective: Hospitalized patient falls are a major cause of disability, functional impairment, and even death. The objective of this pros- pective study was to assess the frequency and contributing factors of falls in hospitalized patients. Methods: Between December 2004 and November 2005, data related to falls in hospitalized patients were recorded: patient demographics, the patient’s functionality, mental status, surrounding circumstances, and the level of injury. Outcome measures were overall rates of pa- tient falls as a function of ward, shift, month, age, and incidence rate ratios (IRRs). Results: Falls were recorded in 611 patients. The overall patient fall rate was 4.36 (95% confidence interval, 4.02 to 4.72) per 1000 patient-days. The fall rate was significantly above the mean for the neurosurgical (IRR = 2.32; P G 0.001) and short-stay wards (IRR = 1.69; P = 0.013). Patients aged 56 to 70 years or older than 70 years fell 1.45 and 1.78 times more frequently, respectively, versus patients aged 55 years or younger (P G 0.001). The odds that a fall resulted in injury were multiplied by 1.19 for each additional decade of age (P = 0.018), and the age-adjusted injury rate for the oncology ward (46.4%) was significantly higher than the overall average (P = 0.001). Conclusions: Age and the patient condition before and during hos- pitalization resulted as the most important determinants of falls in hospitalized patients. Key Words: patient safety, elderly, inpatient falls, hospitalized patient falls (J Patient Saf 2008;4:178Y183) F alls in hospitalized patients are a major cause of disability, functional impairment, and even death, especially among the elderly. 1Y3 This represents a worldwide concern as indicated by reports published from Australia, 3,4 Denmark, 5 Hong Kong, 6 Israel, 7,8 Italy, 9,10 the Netherlands, 11 Sweden, 10 and Taiwan. 12 Despite recent improvements in health care and early attempts to reduce patient falls, 13Y15 data from the Pennsylvania Health Care Cost Containment Council indi- cated that the estimated annual fall rates per 1000 patients increased from 5 to 7.2 in Pennsylvania hospitals from 1994 to 1997. 16 A cross-sectional analysis of hospital inpatient discharge records for patients aged 65 years or older from the Wisconsin Bureau of Health Information reported an unintentional fall rate of 6.9%; similar data have also been reported in Massachusetts. 17 Recently, the Joint Council on Accreditation of Healthcare Organizations stressed the importance of reducing the frequency of falls and the risk of injury resulting from falls in hospitalized patients 18 and requested to implement a Bfall reduction program.[ 19 Several studies have investigated the risk fac- tors 8,17,20Y24 and drugs 24 associated with falls among hospi- talized patients. However, most fall-related articles have been based on retrospective analyses and/or analyses of a small sample of patients and have provided a limited understanding of this important issue. The few observational prospective studies that have been conducted were not carried out in the United States, where health care system resources are not comparable to most other countries, and the length of hospi- tal stays are generally longer. 25 Establishing the risk factors for falls among U.S. patients is critical for determining the resources and/or appropriate prevention programs necessary for reducing their prevalence and identifying the patients on whom intervention efforts should be concentrated. This prospective study was designed to assess the frequency of patient falls and to identify the contributing factors to patient falls and injuries in an American community hospital. METHODS Between December 1, 2004, and November 30, 2005, each hospitalized patient fall at the University of Pittsburgh Medical Center Presbyterian YShadyside Hospital in Pittsburgh, PA, was recorded using the RiskMaster software program (Northamptonshire, United Kingdom). Falls took place in one of the following 15 hospital wards: thoracic-cardiac surgery (thor-card surg), surgical oncology (surg oncology), medicine, pulmonary medicine (pulmonary), oncologyYstem cell, orthopedic ward (orthope- dics), neurosurgical (neurosurg), oncology, cardiologyYpost catheterization (cardiologyYpost cath), short stay, fam- ily practice (fam pract), medical cardiology telemetric (cardiology-telemetry), cardiothoracic intensive care unit (ICU), surgical/neurosurgical ICU (S/NS ICU), and medical and cardiac ICU (MICU/CCU). For each patient, an independent observer recorded patient demographics, the hospital ward in which the fall took place, the date and time of the fall, the type of surgery performed (when applicable), the patient’s functionality at home and in the hospital (i.e., use of a cane, walker, or crutches), any alteration of mental status before the fall, and contributing medications according to the Pennsylvania Department of Health BPatient Safety Reporting ORIGINAL ARTICLE 178 J Patient Saf & Volume 4, Number 3, September 2008 From the *Department of Anesthesiology, University of Pittsburgh Medical Center, and †Department of Orthopedic Surgery, ‡Shadyside Hospital, §UPMC Center for Quality Improvement & Innovation, ||RAND Corporation, Santa Monica, California. Correspondence: Jacques E. Chelly, MD, PhD, MBA, Shadyside Hospital, Department of Anesthesiology, Suite M104 (Posner Pain Center), 5230 Centre Ave, Pittsburgh, PA 15232 (e-mail: chelje@anes.upmc.edu). Copyright * 2008 by Lippincott Williams & Wilkins Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.