Patient falls : a key issue in patient safety in hospitals

Patient safety issues in hospital settings gained worldwide attention within the adverse events discourse launched by the landmark report “to err is human” by the Institute of Medicine in 2000. In this report it was estimated that health care errors and adverse events (AE’s) may ac-count for up to 98,000 patient deaths per year in the USA. Research in AE’s revealed that be-tween 2.9% and 16.6% of hospitalized patients experience at least one AE during a hospital epi-sode. Permanent disability or death due to AE’s has been experienced by up to 15.9% of the pa-tients. Although AE’s have primarily focused on adverse events associated with surgical proce-dures and adverse drug reactions, in-patient falls and associated injuries deserve increasing atten-tion as they have shown to be most frequent AE’s in hospital settings.
Patient falls in the hospital care setting are recognized as a serious health problem since they are common and may result in injuries and complications which prolongs hospitalization, decreases patients’ functional capacities and leads to increased health care costs. The impact a fall can have on a patient’s perception of safety and well-being may inhibit the patient’s ability and willing-ness to participate in activities of daily living and rehabilitation due to fear of falling again. Many aspects of in-patient falls in hospitals such as circumstances, patient characteristics and fall risk factors as well as interventions to prevent patient falls during hospitalization have been widely researched. Yet, there remain gaps in the evidence which guided this research program. More specifically, 1) little information was available regarding fall characteristics among clinical departments of single acute care hospitals, 2) there was a need for further validation of screening instruments to identify in-patients at risk for falling during hospitalization and 3) findings on the effectiveness of multifactorial falls prevention programs in acute care settings and their sustain-ability in daily clinical practice was conflicting.
This research program consisted of a series of retro- and prospective studies addressed the cited gaps. Using clinical and demographic patient data of more than 34,000 hospitalized patients from the years 1999 to 2003 of the “Stadtspital Waid”, an urban public hospital in Zurich, Switzer-land, and findings in relation to the following six research areas are summarized.
First, in a 5 year population-based retrospective study we examined characteristics associated with hospital in-patient falls across clinical departments using incident reporting data and admin-istrative patient data. In a population of 34,972 hospitalized patients (mean age: 67.3 years; fe-male 53.6%, mean length of stay: 11.9 days), 7.2% of the in-patients experienced at least one fall during their hospitalization (surgical department: 1.9%, medical department: 8.8% and geriatric
department 24.8%). Comparison of fallers and non-fallers revealed that fallers were on average 13.5 years older, consisted of 3.8% more females and stayed on average 13.1 days longer in the hospital. Two third (64.8%) of the patients who fell were not injured, 30.1% experienced minor injuries and 5.1% sustained major injuries. Three out of four patients (75.7%) fell in their bed-rooms. Patients fell most often while ambulating (43%) and transferring (35%). Fall risk factors in patients who fell included: impaired mobility (83.1%), impaired cognition (55.3%), use of narcotics (38.6%), and use of psychotropics (25.4%). Half of the patients (50.1%) who fell while hospitalized had a pre-hospital history of falls. These findings are in line with international find-ings indicating that in-patient falls in hospitals are common especially in departments of geriat-rics and internal medicine. Characteristics of falls identified in this study in relation to the time, location, and consequences are similar to findings of previous studies. It appears that in-patient falls should be regarded as an important safety issue especially since one in three falls resulted in at least a minor injury. We recommend giving attention to identifying patients at risk for falling and implementing effective interventions to prevent patient falls and to minimize fall related injuries.
Second, we investigated the association between hospital in-patient fall rates and days of the week, months and lunar cycles. Previous reports indicated that health care professionals hold perceptions that in-patient falls may increase during times of full moon. We therefore compared adjusted fall rates per 1,000 patient days with days of the week and months within 62 complete lunar cycles. The fall rates fluctuated slightly over the entire observation time, ranging from 8.4 to 9.7 falls per month (p=0.757), and from 8.3 falls on Mondays to 9.3 falls on Saturdays (p=0.587). The fall rates within the lunar days ranged from 7.2 falls on lunar day 17 to 10.6 falls on lunar day 20 (p=0.575). Our study revealed that inpatient fall rates were not associated with days of the week, months, or seasons or with lunar cycles such as a full moon or new moon. Therefore, existing perceptions that falls are associated with full moon were not confirmed. We suggest that preventive strategies focus on patients’ modifiable fall risk factors (e.g. gait instabil-ity) and the provision of a safe hospital environment.
Third, we contributed to the further validation of fall risk instruments with a prospective cohort study in which we evaluated the diagnostic value of the Morse Fall Scale (MFS). The goal was to identify risk for falling in hospitalized patients analyzing different MFS cut-offs to determine which score was most useful in identifying in-hospital patients at risk for falls. A consecutive sample of 386 hospitalized patients of the department of internal medicine was studied. The pri-mary nurses completed the MFS (fall risk items: history of falling, secondary diagnosis, ambula-tory aids, intravenous therapy, type of gait, and mental status) for each newly hospitalized patient
within 24 hours of admission. ROC analysis showed that a cut off of 55 points on the MFS had the highest diagnostic value (AUC: 0.701) with a sensitivity of 74.5%, a specificity of 65.8%, and positive and negative predictive values of 23.3%, and 94.9% respectively. While the high negative predictive values (e.g. 95% of the non falling patients were identified as not at risk for falling) may give appropriate reassurance for patients with low risk for falling, the scale seems to be of limited operational value since positive predictive values were only between 12% and 24%. While screening patients for risk for falling may lead to more targeted assessment and subse-quent modification of risk factors using multifactorial interventions, we recommended that the MFS undergo local validation to determine the best cut off score for a given setting before its clinical use.
The fourth study focused on better predicting a patient’s risk of falling. We assessed the predic-tive value of the STRATIFY instrument, a simple fall-risk assessment tool, administered by nurses. Our prospective multi-center study was carried out in six Belgian hospitals during a 3-month period. A total of 2,568 patients expected to be hospitalized for at least 48 hours (mean age: 67.2 years; female: 55.3%) and who were admitted to four surgical (n=875; 34.1%), eight geriatric (n=687; 26.8%), and four general medical wards (n=1,006; 39.2%) were included in this study at the time of their hospital admission. Nurses completed the STRATIFY within 24 hours after admission of the patient. Subsequent falls were documented on a standardized inci-dent report form. The number of fallers was 136 (5.3%), accounting for 190 falls. The STRAT-IFY showed good sensitivity (≥85%) and high negative predictive value (≥99%) for the total sample, for patients admitted to general medical and surgical wards, and for patients younger than 65 years. The STRATIFY, however, showed moderate (67%) to low (57%) sensitivity and high false negative rates (33% and 43%) for patients admitted to geriatric wards and for patients 65 years or older. Thus, although the STRATIFY satisfactorily predicted the fall risk of patients admitted to general medical and surgical wards and patients younger than 65 years, it failed to predict the fall risk of patients admitted to geriatrics wards and patients 65 years and older.
The fifth study was an intervention study, using a quasi-experimental design. More specifically, we evaluated the effectiveness of a nurse-led fall prevention program in a hospital. In a four month study period, 409 patients from an internal medicine department were included in an in-tervention group (n=198) or usual care group (n=211). The program consisted of training nurses in the use of the Morse Fall Scale and the implementation of 15 preventive interventions such as orienting patients to hospital environment and schedules, assisting patients with transfers and ambulation, and providing safe footwear and clothing. Patient falls were registered using the standardized falls incident report form. In the intervention group the proportion of patients at risk
for falls was higher (p=0.048), and fewer patients with multiple falls were observed (p=0.009). The intervention program was effective in preventing multiple falls but not first falls. A pro-longed time to a first fall in a subgroup of fallers in the intervention group may indicate that there was increased nurse awareness of patients at risk for falling and the appropriateness of the interventions utilized. The findings indicate that the intervention program was not successful in preventing falls during the first four days of hospitalization, while some effect can be seen there-after. Based on the experiences with this intervention protocol, an interdisciplinary hospital falls prevention program has been implemented.
In the final study, we examined in-patient fall rates and consequent injuries before and after the implementation of this interdisciplinary falls prevention program (IFP) using a serial survey de-sign. While the fifth study tested the efficacy of the intervention program, this study assessed effectiveness in daily life. The population under study included 34,972 patients (mean age: 67.3 years; female 53.6%, mean length of stay: 11.9 days, mean nursing care time per day: 3.5 hours), hospitalized in the departments of internal medicine, geriatrics, and surgery from 1999 to 2003. Overall, a total of 3,842 falls affected 2,512 (7.2%) of the hospitalized patients. From these falls, 2,552 (66.4%) were without injuries, while 1,142 (29.7%) falls resulted in minor injuries, and 148 (3.9%) falls resulted in major injuries. The fall rates per 1,000 patient days fluctuated slightly from 9.1 falls in 1999 to 8.6 falls in 2003 (p=0.086). After the implementation of the IFP, in 2001 a slight decrease to 7.8 falls per 1,000 patient days was observed until the end of the same year. The annual proportion of minor and major injuries did not decrease after the imple-mentation of the IFP. From 1999 to 2003, patient characteristics changed in terms of slight in-creases (female gender, age, nursing care time) or decreases (length of hospital stay), as did the prevalence of fall risk factors (up to 46.8%) in those patients who fell. In conclusion, following the implementation of the interdisciplinary falls prevention program, neither the frequencies of falls nor consequent injuries decreased substantially. We have hypothesized that lack of adher-ence to the fall prevention program lead to this ineffectiveness. Future studies need to incorpo-rate strategies to maximize and evaluate ongoing adherence to interventions in hospital falls pre-vention programs.
The results of our research program contributed to the evidence based on hospital falls. First, it added detailed knowledge on characteristics of in-patient falls in departments of medicine, geri-atrics and surgery within a single hospital. Second, it established for the first time evidence that in-patient falls and lunar cycles are not associated. Third, it showed that identifying in-patients at risk for falling using specific tools does at best offer an addition to clinical judgement and as-sessment within falls prevention programs. Fourth, it showed that a multifactorial nurse led intervention program has the potential to reduce multiple falls but not first falls in hospitalized medical patients, and fifth, it revealed that the implemented interdisciplinary hospital falls pre-vention program was not able to substantially decrease, either the frequency of falls or conse-quent injuries despite the use of a state of the art intervention protocol.
Future research on in-patient falls should focus on modifying hospital falls prevention strategies. The awareness of health care professionals of the problem of falls in hospitalized patients needs to be addressed in order to support the clinicians’ adherence to evidence based intervention pro-tocols. Furthermore, commitment to changing practice must be improved and professional skills such as assessment and treatment of in-patients at risk for falling need to be further developed to strengthen interdisciplinary health care teams.