529791 research-article2014
CNRXXX10.1177/1054773814529791Clinical Nursing ResearchTzeng and Yin
Article
Exploring Post-Fall Audit Report Data in an Acute Care Setting
Clinical Nursing Research 2015, Vol. 24(3) 284–298 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1054773814529791 cnr.sagepub.com
Huey-Ming Tzeng, PhD, RN, FAAN1 and Chang-Yi Yin, MA2
Abstract This retrospective, descriptive, chart review study was done to demonstrate one strategy for communicating aggregated and actionable fall data to bedside nurses. It was conducted at a nonprofit acute care hospital in the northwestern United States to analyze the quantitative data captured in post-fall audit reports of patient falls (March 1-December 31, 2012, N = 107 falls). Descriptive and binary statistical analyses were used. The quarterly National Database of Nursing Quality Indicators 2011 and 2012 reports showed that implementation of post-fall audit reports can lead to a lower overall fall rate and a lower fall-injury rate. Increased nursing hours could be a confounding factor of the positive impact of conducting post-fall audits in this study. It is concluded that timely and systematic reporting, analysis, and interpretation of fall data in an electronic format can facilitate prevention of falls and fall injuries. Keywords patient, safety, injury, hospital, fall
1Washington 2Chinese
State University, Spokane, Washington, USA Culture University, Taipei, Taiwan, Republic of China
Corresponding Author: Huey-Ming Tzeng, PhD, RN, FAAN, Dean and Professor, Whitson-Hester School of Nursing, Tennessee Technological University, 10 West 7th Street, Cookeville, Tennessee 38505, USA. Email: [email protected]
Tzeng and Yin
285
Introduction Injurious falls are the most prevalent hospital-acquired adverse events (Schwendimann, Bühler, De Geest, & Milisen, 2006). As the authors have observed, hospitals (e.g., nursing and risk management administrators) commonly review cases of patient falls, and the review processes often vary by hospital (Tzeng, 2010; Tzeng & Yin, 2013; Tzeng, Yin, Tsai, Lin, & Yin, 2007). The review data are rarely communicated back in an aggregated and actionable format. The first author’s recent conversations (unpublished materials) with nursing staff in acute care inpatient units of three nonprofit hospitals in the northwestern United States revealed that registered nurses (RNs) who provided bedside care do not find the current type of documentation of patient falls to be useful. In other words, they want the information, but it needs to be useful and relevant. These RNs wanted timely and regular summative feedback from administration to assist them in effectively preventing falls; such feedback would include the common contributory factors and types of falls. Clinicians are typically provided unit-specific data on two key nursing outcome indicators: the overall number of falls and the number of injury-falls per 1,000 patient-days. These statistics are insufficient for integrating lessons learned into practice. There is an urgent need to help clinicians make sense of post-fall report data so as to prevent falls. The aim of this study was to demonstrate one approach to summarizing the information collected during review of hospital-acquired inpatient falls to communicate back to clinicians in an aggregated and actionable format for bedside use. This article summarized the post-fall audit report data (completed by designated RN fall specialists) using a retrospective, descriptive, chart review study design. It also compared the overall fall rates (including falls with and without injury; that is, total fall rates) and injurious fall rates pre- and post-implementation of the post-fall audit report summaries.
Background As indicated in Preventing Falls in Hospitals: A Toolkit for Improving Quality of Care (Agency for Healthcare Research and Quality [AHRQ], 2013), communications within health care teams regarding patient falls via hospitals’ incident reporting systems and medical records (e.g., updated patients’ fall risk assessment and care plans) are essential. Hill et al. (2010) found that patient falls were less likely to be reported if they occurred subsequent to an earlier fall (the first fall may be reported by the staff, but later falls from the same patient were less likely to be recorded), or when the fall occurred during the morning or afternoon shift. For performance improvement initiatives, the
286
Clinical Nursing Research 24(3)
hospital’s incident reporting system (completed by the primary nurses of fall patients) could be paired with collection of post-fall data by trained fall evaluators (as post-fall audit reports; Hill et al., 2010; Shorr et al., 2008). Hospital-acquired inpatient falls and fall-related injuries have many causes, including but not limited to impaired cognition, mobility, gait, and balance; a history of falling; and dependence in activities of daily living. Inpatient falls can lead to injury, prolonged hospital stays, lack of patient independence, and additional expenses (Schwendimann et al., 2006). Accidental falls (e.g., tripping due to environmental hazards) and anticipated physiological falls (e.g., abnormal gait and balance) are generally preventable, but unanticipated physiological falls (e.g., having a seizure or fainting episodes) and behavioral or intentional falls (e.g., when a patient acts out) are not (Butcher, 2013). Multifactorial interventions for fall prevention in hospital settings have been found to decrease accidental or anticipated physiological injurious fall incidence; however, the evidence is inconclusive and does not identify specific interventions that are key for success (AHRQ, 2013; Boushon et al., 2012; Butcher, 2013; Cameron et al., 2012; Miake-Lye, Hempel, Ganz, & Shekelle, 2013; Raeder, Siegmund, Grittner, Dassen, & Heinze, 2010; Spoelstra, Given, & Given, 2012; Weinberg et al., 2011). The Institute for Healthcare Improvement’s (IHI) recent publication of Transforming Care at the Bedside How-to Guide: Reducing Patient Injuries From Falls, Boushon and associates (2012) emphasized the importance of communicating to all staff the information regarding patients who are at risk of falling. The audit tool provided by IHI enabled nurses to evaluate the reliability of fall risk assessment and interventions in medical-surgical units in hospital settings. As the authors have observed in clinical practice (e.g., hospitals in the Unites States, Taiwan, China) when conducting previous studies (Tzeng, 2010; Tzeng & Yin, 2013; Tzeng et al., 2007), review of hospital-acquired inpatient falls has been incorporated into routine risk management practice to avoid litigation and continually improve care quality in hospital settings. The impact of conducting post-fall reviews and the best practices for conducting such reviews have not been established. Tzeng and Yin (2013) recently identified 81 potential factors for injurious falls in adult inpatients in hospitals in the United States and 75 interventions for prevention. RNs perceived that the five most frequent risk factors were Alzheimer disease, confusion, disorientation, gait problems, and inability to follow safety instructions. Ivziku, Matarese, and Pedone (2011) found that among 179 elderly patients admitted to a geriatric acute care inpatient unit of an Italian hospital, the risk factors most strongly associated with falls were confusion and depression. Another recent study, conducted by Kolla, Lovely,
Tzeng and Yin
287
Mansukhani, and Morgenthaler (2013) in the United States, found that use of zolpidem, a hypnotic agent that may decrease gait balance, was a significant and potentially modifiable risk factor for falls in adult inpatients in acute care settings. In short, bedside prevention of patient falls requires a multifactorial approach and an interdisciplinary health care team, which should include representatives from nursing, medicine, pharmacy, education, physical therapy, and facilities management.
Method Design This retrospective, descriptive, chart review study was done to demonstrate one strategy for communicating aggregated and actionable fall data to bedside nurses. To do so, it summarized the post-fall audit report data using a retrospective descriptive study design. This study analyzed the quantitative data captured in the post-fall audit reports for patient falls. As stated in the document of Fall Policy and Protocol of the study hospital (updated in 2012, p. 1), patient fall is defined as “a sudden, uncontrolled, unintentional, downward displacement of the body to the ground or other object, excluding such resulting from violent blows or other purposeful actions. Incidents where the caregiver intervenes to prevent injury are included in this definition.” This document was reviewed and updated as needed by the fall committee. The post-fall audit tool was also designed by the fall committee of the study hospital as a fall prevention initiative. This interdisciplinary committee included representatives from nursing, pharmacy, education, physical therapy, and facilities management. These falls occurred between March 1 and December 31, 2012, at a nonprofit acute care hospital in the northwestern United States. This project was approved by the institutional review boards of the study hospital and the university affiliated with the first author. The study hospital, a regional medical center, had a Magnet designation. It had 252 certified beds in service and included one adult critical care, medicalsurgical unit (30 beds); one step-down progressive care, medical-surgical unit (20 beds); two adult acute medical units (52 beds); two adult acute surgical units (50 beds); one pediatric unit (11 beds); one birthing center (labor and delivery, postpartum, and neonatal intensive care, 12 beds); one adult acute rehabilitation unit (14 beds); one acute adult psychiatric unit (22 beds); one acute child/adolescent psychiatric unit (26 beds); one acute chemical dependence unit (15 beds); and the emergency room. The prevalence of the hospitalacquired condition of fall and trauma was 1.229 per 1,000 patient discharges (U.S. national rate = 0.527; http://medicare.gov/hospitalcompare).
288
Clinical Nursing Research 24(3)
Data Source and Sample The data source for this study was the completed post-fall audit reports collected and maintained by the fall committee of the study hospital. As stated in the study hospital’s fall policy and protocol, post-fall reviews were conducted by one of the two designated RN fall specialists shortly after a patient fall occurred. Both fall specialists were co-chairs of the fall committee. Post-fall review, using a real-time audit tool, was conducted to improve understanding of the history and types of falls. The fall committee approved the post-fall review mechanism as a quality-improvement initiative on February 20, 2012, with the goal of decreasing patient falls beginning March 1, 2012. A RN working at the study hospital recoded data from the post-fall audit reports for individual patients and entered the data into a spreadsheet. The de-identified data were then forwarded to us for analysis. A total of 119 patient falls occurred at the study hospital between March 1 and December 31, 2012. Patient falls were included that occurred in the inpatient care units or emergency room and that the fall specialists had identified as accidental falls (falls related to environmental factors; that is, due to an unsafe environment, such as equipment or water in a patient’s path) or anticipated physiological falls (falls related to intrinsic factors; that is, due to known individual risk factors, such as confusion, weakness, and seizures). Nine unanticipated physiological falls (i.e., unpreventable falls, such as falls due to cardiac arrest syndrome and changes with medications) and three falls that occurred outside of the study units were excluded from the analysis, resulting in a total of 107 fall cases valid for this analysis.
Measurement The one-page post-fall review form was developed by the fall committee of the study hospital. This form included seven sections. Section 1, “Demographic Characteristics,” included the patient’s sex, patient age in years, and admission date. The unit and location where the fall occurred, and the date and time of the fall were also documented. The length of stay in whole days (no decimals), from the admission date to the date of fall, was calculated. In Section 2, “Interventions in Place at Time of Fall as Observed by Fall Specialist,” as observed by the fall specialist, were documented. These included fall/yellow socks, fall magnet outside the patient’s room, the bed in a low position, bed-check on, use of one-to-one sitter, call light within reach of the patient, fall/yellow sticker on patient armband, fall/yellow
Tzeng and Yin
289
sticker on chart, in up-front room (i.e., moving patients to rooms near the nurses’ station), and others. For Section 3, “Interview Questions,” the fall specialist asked the primary nurse of the fall patient two questions: (a) What was the patient’s fall risk prior to this fall? (low, moderate, high, or don’t know), and (b) Do you have any concerns (e.g., did the bed alarm work properly, were there environmental or safety issues or any changes in mental status or medications)? As described in the study hospital’s fall policy and protocol, a staff nurse will assess a patient’s fall risk using the 2006 Johns Hopkins Hospital fall risk assessment tool within the first 8 hr of admission, within each shift assessment, and when there are changes in a patient’s condition. This tool has been incorporated into the study hospital’s computer documentation system. For Section 4, “Chart Review,” the fall specialist conducted a chart review and evaluated the following three items: (a) time since last fall assessment was completed (less than 12 hr, 12-24 hr, more than 24 hr, or none completed), (b) last fall risk assessment (low, moderate, high, or none completed), and (c) care plan for fall risk prior to fall (yes, no, or not a fall risk patient). For Section 5, “Reminders,” the fall specialist reminded the primary nurse to complete the following actions and recorded those discussed with the primary nurse: (a) document fall facts in nursing notes, (b) contact physician if needed, (c) contact family if needed, (d) fill out risk management notification, (e) update care plan, (f) document the fall in the shift report for nursing staff, and (g) update fall risk assessment. In Section 6, “Additional Information,” the fall specialist indicated whether any injury occurred and the type of fall (accidental, anticipated physiological, unanticipated physiological, or intentional). Section 7 included one open-ended question to prompt the fall specialist to recommend any other actions or offer other comments.
Data Analyses Data were analyzed using IBM SPSS Statistics 20 software (SPSS Inc.; Chicago, Illinois, USA). Only the quantitative data collected during the postfall reviews were analyzed. Univariate and bivariate analyses were performed to document patient and hospitalization characteristics related to falls. Univariate analyses (also called descriptive analyses; means, standard deviations [SDs], and frequency tables) were used as appropriate. Bivariate analyses (independent-sample t tests; one-way, between-group ANOVA; and kappa measure of agreement tests) were used, as appropriate to the types of variables; alpha was set at .05.
290
Clinical Nursing Research 24(3)
Results A total of 107 fall cases were included in the analysis. Table 1 summarizes the findings of the descriptive analyses, and the variables are presented in the order they are described in the “Measurement” section. Independent-sample t tests were conducted to compare patient age and length of stay on the date of fall between the accidental fall group (falls related to environmental factors) and the anticipated physiological fall group (falls related to intrinsic factors) and between the fall group with injury and the group without injury. A summary of the findings follows. •• Patients who experienced accidental falls were significantly younger, M = 58.36 years, SD = 21.15 years, n = 50, than those who experienced anticipated physiological falls, M = 67.49 years, SD = 15.90 years, n = 57; t(105) = −2.50, p = .014 (two-tailed). •• There was no significant difference in length of stay at date of fall for the accidental fall group, M = 5.40 days, SD = 5.72 days, n = 50, versus the anticipated physiological fall group, M = 7.02 days, SD = 6.30 days, n = 57; t(105) = −1.38, p = .17 (two-tailed). •• There was no significant difference in patient age for the injury group, M = 63.20 years, SD = 19.09 years, n = 25, versus the no-injury group, M = 64.25 years, SD = 18.52 years, n = 75; t(98) = −0.24, p = .81 (two-tailed). •• There was no significant difference in length of stay at date of fall for the injury group, M = 6.04 days, SD = 6.94 days, n = 25, versus the no-injury group, M = 6.04 days, SD = 4.88 days, n = 75; t(98) = 0.00, p = 1.00 (two-tailed). One-way ANOVA was conducted to explore the impact of the time of fall (in terms of three shifts of 8 hr each [day, evening, and night] and six intervals of 4 hr each) on patient age and length of stay at date of fall. The findings are summarized below. •• There was no significant difference in patient age for falls that occurred during the three 8-hr intervals, F(2, 104) = 1.35, p = .26 (two-tailed): 1. day shift (>7 a.m.-3 p.m.; M = 63.03 years, SD = 17.55 years, n = 34) 2. evening shift (>3 p.m.-11 p.m.; M = 60.19 years, SD = 22.53 years, n = 42) 3. night shift (>11 p.m.-7 a.m.; M = 67.55 years, SD = 14.54 years, n = 31)
291
Tzeng and Yin
Table 1. Analysis of Post-Fall Audit Reports for 50 Accidental Falls Related to Environmental Factors and 57 Anticipated Physiological Falls Related to Intrinsic Factors, March 1 Through December 31, 2012. Variable
n (%)
M (SD)
Section 1: Demographic Characteristics Sex Male 58 (54.2) Female 49 (45.8) Patient age, years 63.22 (19.01)a The unit/department where the fall occurred Critical care unit 3 (2.8) Progressive care unit 7 (6.5) Medical Unit A 36 (33.6) Medical Unit B 14 (13.1) Surgical unit, general 8 (7.5) Surgical unit, orthopedic/neurology 6 (5.6) Birthing center 3 (2.8) Rehabilitation unit 10 (9.3) Psychiatric unit, adult 14 (13.1) Psychiatric unit, child and adolescent 1 (0.9) Emergency room 5 (4.7) Time of fall (a coded variable) >7 a.m.-11 a.m. 15 (14) >11 a.m.-3 p.m. 19 (17.8) >3 p.m.-7 p.m. 25 (23.4) >7 p.m.-11 p.m. 17 (15.9) >11 p.m.-3 a.m. 16 (15) >3 a.m.-7 a.m. 15 (14) Length of stay, days from admission date to date 6.26 (6.07)b of fall (a coded variable) Section 2: Interventions in Place at Time of Fall as Observed by Fall Specialist Fall/yellow socks 48 (44.9) Fall magnet outside patient room 46 (43) Bed in low position 81 (75.7) Bed-check on 24 (22.4) Use of one-to-one sitter 3 (2.8) Call light in reach of patient 79 (73.8) Fall/yellow sticker on patient armband 23 (21.5) Fall/yellow sticker on chart 24 (22.4) In up-front room 50 (46.7) Section 3: Interview Conducted by Fall Specialist With Patient’s Primary Nurse What was the patient fall risk prior to fall? 2.35 (0.85) Low = 1 25 (23.4) Moderate = 2 18 (16.8) (continued)
292
Clinical Nursing Research 24(3)
Table 1. (continued) Variable
n (%)
M (SD)
High = 3 61 (57) Don’t know (coded as a missing value) 2 (1.9) System missing (no information provided) 1 (0.9) Section 4: Chart Review by Fall Specialist Time since last fall assessment completed, hours 1.25 (0.68) 24 = 3 4 (3.7) None completed (coded as a missing value) 3 (2.8) System missing (no information provided) 12 (11.2) Last fall risk 2.38 (0.81) Low = 1 19 (17.8) Moderate = 2 20 (18.7) High = 3 54 (50.5) None completed (coded as a missing value) 4 (3.7) System missing (no information provided) 10 (9.3) Care plan for fall risk prior to fall Yes 65 (60.7) No 7 (6.5) Not a fall risk patient 10 (9.3) System missing (no information provided) 25 (23.4) Section 5: Fall Specialist Reminded Primary Nurse to Complete the Following Actions and Indicated the Ones Addressed Document fall facts in nursing notes 85 (79.4) Contact physician if needed 82 (76.6) Contact family if needed 50 (46.7) Fill out risk management notification 88 (82.2) Update care plan 71 (66.4) Document the fall in the shift report for nursing 82 (76.6) staff Update fall risk assessment 81 (75.7) Section 6: Additional Information Provided by Fall Specialist Injury present Yes 25 (23.4) No 75 (70.1) Unsure (coded as a missing value) 7 (6.5) Type of fall Accidental (related to environmental factors) 50 (46.7) Anticipated physiological (related to intrinsic 57 (53.3) factors) aRange bRange
= 8-93 years. = 1-33 days.
Tzeng and Yin
293
•• There was no significant difference in patient age for falls occurring during the six 4-hr intervals, F(5, 101) = 0.66, p = .66 (two-tailed): 1. 2. 3. 4. 5. 6.
>7 a.m. to 11 a.m. (M = 63.47 years, SD = 18.29 years, n = 15) >11 a.m. to 3 p.m. (M = 62.68 years, SD = 17.45 years, n = 19) >3 p.m. to 7 p.m. (M = 61.84 years, SD = 22.13 years, n = 25) >7 p.m. to 11 p.m. (M = 57.76 years, SD = 23.59 years, n = 17) >11 p.m. to 3 a.m. (M = 68.94 years, SD = 15.34 years, n = 16) >3 a.m. to 7 a.m. (M = 66.07 years, SD = 14.00 years, n = 15)
•• There was a significant difference in the length of stay on the date of fall for falls that occurred during the three 8-hour intervals, F(2, 104) = 3.63, p = .03 (two-tailed); the post hoc tests showed that mean length of stay on the date of fall was significantly shorter for the falls that occurred during the evening shift compared with those that occurred during the night shift (M difference = −3.39 days, p = .046): 1. day shift (>7 a.m.-3 p.m.; M = 7.26 days, SD = 5.59 days, n = 34) 2. evening shift (>3 p.m.-11 p.m.; M = 4.36 days, SD = 4.34 days, n = 42) 3. night shift (>11 p.m.-7 a.m.; M = 7.74 days, SD = 7.84 days, n = 31) •• There was no significant difference in the length of stay on the date of fall for the six 4-hr intervals, F(5, 101) = 1.79, p = .12 (two-tailed): 1. 2. 3. 4. 5. 6.
>7 a.m. to 11 a.m. (M = 8.07 days, SD = 5.64 days, n = 15) >11 a.m. to 3 p.m. (M = 6.63 days, SD = 5.62 days, n = 19) >3 p.m. to 7 p.m. (M = 4.36 days, SD = 4.54 days, n = 25) >7 p.m. to 11 p.m. (M = 4.35 days, SD = 4.17 days, n = 17) >11 p.m. to 3 a.m. (M = 6.56 days, SD = 4.79 days, n = 16) >3 a.m. to 7 a.m. (M = 9.00 days, SD = 10.18 days, n = 15)
A kappa measure of agreement test was used to examine the degree of agreement on fall risk (low, moderate, or high) prior to fall as assessed by the primary nurse (in an interview) versus the fall specialist (based on a chart review). The kappa measure of agreement value was 0.85 (p < .001), representing a very good agreement. In other words, the primary nurses’ understanding about patients’ fall risk prior to falls was consistent with the documented fall risk. In addition, the authors used the study hospital’s archived data to investigate the impact of the post-fall review on the overall fall rates and injurious
294
Clinical Nursing Research 24(3)
fall rates. The quarterly report of the National Database of Nursing Quality Indicators (NDNQI), prepared by the American Nurses Association, was used to help interpret any changes that occurred after the tool was implemented. The pre-implementation NDNQI data included the second, third, and fourth quarters of 2011. The post-implementation NDNQI data were composed of the data from the second, third, and fourth quarters of 2012. Based on the acuity-adjusted information from all reporting adult acute inpatient care units combined, the average number of falls per 1,000 patient-days was 4.84 in 2011 and 4.43 in 2012. The average number of injury-falls per 1,000 patient-days was 0.86 in 2011 and 0.67 in 2012. In other words, the average overall fall rate was 8.47% lower in 2012 compared with 2011. The average injury-fall rate was 22.09% lower in 2012, with improved patient outcomes, compared with 2011. Following the NDNQI guidelines, the total number of hours per patient-day was 11.6 in 2011 and 15.08 in 2012. The average RN hours per patient-day were 10.94 in 2011 and 15.15 in 2012, and RNs supplied an average of 65.64% of all nursing hours in 2011 and 65.9% in 2012. Therefore, the number of nursing hours per patient-day was 30% greater at the study hospital in 2012, and the number of RN hours per patient-day was 38.48% greater. It is possible that increases in the number of staff from 2011 to 2012 are a confounding factor in the present study of the post-fall review and overall fall rate and injury-fall rate.
Discussion The objective to this study was to demonstrate one approach for summarizing the information collected during review of hospital-acquired inpatient falls to communicate back to clinicians in an aggregated and actionable format for bedside use. A retrospective descriptive study was used to achieve this objective. About half (n = 50, 47%) of the falls occurred in the adult medical units. The next-most common fall sites were the adult psychiatric unit (n = 14, 13.1%) and the surgical units (n = 14, 13.1%). Mean length of stay on the date of fall was shorter for falls that occurred during the evening shift rather than the night or day shifts. Analysis by time of fall (by shift) could provide additional insights into possible modifications to structure and process indicators (e.g., staffing levels and intervention choices to prevent falls among new patients during the evening shift). As for the interventions in place at the time of the fall, the two used most frequently were keeping patient beds in a low position and keeping the call light within reach of the patients. Only about 75% of the patients had one or both of these interventions in place. More than 20% (n = 25, 23.4%) of the patients had a low fall risk prior to fall, and nurses had completed the last fall assessment for 75.7% (n = 81) of them within
Tzeng and Yin
295
12 hr before the fall. In addition, the patients who experienced accidental falls were significantly younger (M = 58.36 years, SD = 21.15 years, n = 50) than those who experienced anticipated physiological falls (M = 67.49 years, SD = 15.90 years, n = 57). This study’s finding on the fall rates pre- and post-implementation of the post-fall audit report suggested that the post-fall audit report was a useful quality-improvement initiative at the study hospital. Because of limited data points for trending and statistical analyses, the authors could not conclude that implementing the post-fall audit report significantly decreased the overall fall rate and the injury-fall rate. The downward trends for both rates are encouraging. As the authors are not employed by the study hospital, the first author gave a presentation of the summary of the findings to the fall committee at the study hospital in mid-June 2013. The fall committee members found the data summary to be helpful and to verify their impressions about characteristics of falls at their hospital. This presentation was the first time that the study hospital had the post-fall data in an electronic format and subject to research analysis. Discussion focused on the data in Section 2 (Table 1). The fall committee members thought that additional analyses by unit might help with developing unit-specific fall prevention initiatives. As a courtesy, a week prior to the presentation, the first author shared the findings with an executive inpatient services. After the presentation to the fall committee, the presentation materials were also shared with the Research Committee of the study hospital.
Study Limitations The main limitation of this study is that its generalizability is restricted to a single source of data, a nonteaching hospital. In addition, the post-fall audit report is an in-house, custom-made tool. Because the tool used in this study did not capture whether each patient had had a previous fall during the study period, each fall was treated as an independent incident. Another limitation of the study is the fact that patient fall data were based on voluntary reporting by staff; it is possible that some falls were not reported. In addition, the sample size was small. It is also noted that there was a significant increase in the number of nursing hours on the units between 2011 and 2012. Nursing hours could be a confounding factor (also called the confounder) making it difficult to interpret the positive impact of conducting post-fall audits on reducing falls and fall injuries. This study was not designed to analyze both changes in nursing hours and post-fall audits on the single hospital’s rates of falls and fall injuries.
296
Clinical Nursing Research 24(3)
Conclusion This descriptive study summarizes the information collected in post-fall audit reports as one approach to communicate back to clinicians in an aggregated and actionable format for bedside use, particularly for RNs who submit the patient incident reports to the hospital’s incident reporting system and assist in completing the post-fall audit reports. A reduction in the fall rates and injury-fall rates after implementing the post-fall audit review as a qualityimprovement initiative was observed in this study. While it is premature to conclude that the reduction was a result of this quality-improvement initiative, these preliminary findings are promising.
Relevance to Clinical Practice To promote patient safety during hospital stays (e.g., preventing hospitalacquired injurious falls), it is important to inform practice in a timely and systematic manner. Clinicians, risk management officers, quality-improvement specialists, and administrators should consider storing collected fallrelated data (e.g., fall risk assessment and post-fall audit data) in an electronic format and developing queries or computer programs to facilitate statistical analyses, reporting, and interpretation of data. It is also essential to develop a hospital-based post-fall audit report form that accounts for the hospital’s safety culture, care environment (e.g., design, layout, and infrastructure for human resources), and practice (e.g., fall policies and protocol). The post-fall report may be used as a stand-alone audit tool for performance improvement or in combination with other data sources for validation (e.g., the hospital’s incident reporting system). As risk factors for accidental falls, anticipated physiological falls, and unanticipated physiological falls are very different, it would be helpful to summarize the data by type of fall. In addition, it would be useful to produce fall profiles individualized for different units to inform bedside nurse practice. The fall profile of an acute rehabilitation unit could differ considerably from that of a critical care unit or a birthing unit. This effort could facilitate development of unit-specific fall prevention initiatives. Additional research is needed to refine the post-fall audit report template to capture the most relevant faller information. Conducting interviews or focus group meetings with clinicians, risk management officers, quality-improvement specialists, and administrators is one way to refine the tool to facilitate data collection and report generation to inform bedside fall prevention care. For example, it would be helpful to add more specific contributory factors to the post-fall audit (e.g., medication related, floor surface, confusion).
Tzeng and Yin
297
Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.
References Agency for Healthcare Research and Quality. (2013, January). Preventing falls in hospitals: A toolkit for improving quality of care (AHRQ Publication No. 13-0015EF, Prepared by RAND Corporation, Boston University School of Public Health, and ECRI Institute). Rockville, MD: Author. Boushon, B., Nielsen, G., Quigley, P., Rutherford, P., Taylor, J., Shannon, D., & Rita, S. (2012). Transforming care at the bedside how-to guide: Reducing patient injuries from falls. Cambridge, MA: Institute for Healthcare Improvement. Retrieved from http://www.ihi.org/knowledge/Pages/Tools/ TCABHowToGuideReducingPatientInjuriesfromFalls.aspx Butcher, L. (2013). The no-fall zone. Hospitals & Health Networks, 87(6), 26-30. Cameron, I. D., Gillespie, L. D., Robertson, M. C., Murray, G. R., Hill, K. D., Cumming, R. G., & Kerse, N. (2012). Interventions for preventing falls in older people in care facilities and hospitals. The Cochrane Database of Systematic Reviews, 12, Article CD005465. doi:10.1002/14651858.CD005465 Hill, A. M., Hoffmann, T., Hill, K., Oliver, D., Beer, C., McPhail, S., . . . Haines, T. P. (2010). Measuring falls events in acute hospitals—A comparison of three reporting methods to identify missing data in the hospital reporting system. Journal of the American Geriatrics Society, 58, 1347-1352. Ivziku, D., Matarese, M., & Pedone, C. (2011). Predictive validity of the Hendrich fall risk model II in an acute geriatric unit. International Journal of Nursing Studies, 48, 468-474. Kolla, B. P., Lovely, J. K., Mansukhani, M. P., & Morgenthaler, T. I. (2013). Zolpidem is independently associated with increased risk of inpatient falls. Journal of Hospital Medicine, 8, 1-6. Miake-Lye, I., Hempel, S., Ganz, D. A., & Shekelle, P. G. (2013). Inpatient fall prevention programs as a patient safety strategy. Annals of Internal Medicine, 158, 390-396. Raeder, K., Siegmund, U., Grittner, U., Dassen, T., & Heinze, C. (2010). The use of fall prevention guidelines in German hospitals—A multilevel analysis. Journal of Evaluation in Clinical Practice, 16, 464-469. Schwendimann, R., Bühler, H., De Geest, S., & Milisen, K. (2006). Falls and consequent injuries in hospitalized patients: Effects of an interdisciplinary falls prevention program. BMC Health Services Research, 6, Article 69. Retrieved from http:// www.ncbi.nlm.nih.gov/pmc/articles/PMC1534028/pdf/1472-6963-6-69.pdf
298
Clinical Nursing Research 24(3)
Shorr, R. I., Mion, L. C., Chandler, A. M., Rosenblatt, L. C., Lynch, D., & Kessler, L. A. (2008). Improving the capture of fall events in hospitals: Combining a service for evaluating inpatient falls with an incident report system. Journal of the American Geriatrics Society, 56, 701-704. Spoelstra, S. L., Given, B. A., & Given, C. W. (2012). Fall prevention in hospitals: An integrative review. Clinical Nursing Research, 21, 92-112. Tzeng, H. M. (2010). Understanding the prevalence of inpatient falls associated with toileting in adult acute care settings. Journal of Nursing Care Quality, 25, 386-394. Tzeng, H. M., & Yin, C. Y. (2013). Frequently observed risk factors for fall-related injuries and effective preventive interventions: A multihospital survey of nurses’ perceptions. Journal of Nursing Care Quality, 28, 130-138. Tzeng, H. M., Yin, C. Y., Tsai, S. L., Lin, S., & Yin, T. J. (2007). Patient falls and open visiting hours: A case study in a Taiwanese medical center. Journal of Nursing Care Quality, 22, 145-151. Weinberg, J., Proske, D., Szerszen, A., Lefkovic, K., Cline, C., El-Sayegh, S., . . . Weiserbs, K. F. (2011). An inpatient fall prevention initiative in a tertiary care hospital. Joint Commission Journal on Quality and Patient Safety, 37, 317-325.
Author Biographies Huey-Ming Tzeng earned her PhD in Nursing from the University of Michigan, School of Nursing, Ann Arbor, Michigan, in 1997. She takes both macro and micro approaches to address patient safety and health care quality issues. She is currently researching the promotion of safe hospital stays, with a focus on fall prevention for adult inpatients. Chang-Yi Yin earned his Master’s degree in History from National Taiwan University, Taipei, Taiwan. He is an internationally well-known historian in Chinese and world history and area studies. He started collaborating with Dr. Tzeng since 2003. Together they have addressed patient safety issues, such as, infection control and environment of care.
CNRXXX10.1177/1054773814529791Clinical Nursing ResearchTzeng and Yin
Article
Exploring Post-Fall Audit Report Data in an Acute Care Setting
Clinical Nursing Research 2015, Vol. 24(3) 284–298 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1054773814529791 cnr.sagepub.com
Huey-Ming Tzeng, PhD, RN, FAAN1 and Chang-Yi Yin, MA2
Abstract This retrospective, descriptive, chart review study was done to demonstrate one strategy for communicating aggregated and actionable fall data to bedside nurses. It was conducted at a nonprofit acute care hospital in the northwestern United States to analyze the quantitative data captured in post-fall audit reports of patient falls (March 1-December 31, 2012, N = 107 falls). Descriptive and binary statistical analyses were used. The quarterly National Database of Nursing Quality Indicators 2011 and 2012 reports showed that implementation of post-fall audit reports can lead to a lower overall fall rate and a lower fall-injury rate. Increased nursing hours could be a confounding factor of the positive impact of conducting post-fall audits in this study. It is concluded that timely and systematic reporting, analysis, and interpretation of fall data in an electronic format can facilitate prevention of falls and fall injuries. Keywords patient, safety, injury, hospital, fall
1Washington 2Chinese
State University, Spokane, Washington, USA Culture University, Taipei, Taiwan, Republic of China
Corresponding Author: Huey-Ming Tzeng, PhD, RN, FAAN, Dean and Professor, Whitson-Hester School of Nursing, Tennessee Technological University, 10 West 7th Street, Cookeville, Tennessee 38505, USA. Email: [email protected]
Tzeng and Yin
285
Introduction Injurious falls are the most prevalent hospital-acquired adverse events (Schwendimann, Bühler, De Geest, & Milisen, 2006). As the authors have observed, hospitals (e.g., nursing and risk management administrators) commonly review cases of patient falls, and the review processes often vary by hospital (Tzeng, 2010; Tzeng & Yin, 2013; Tzeng, Yin, Tsai, Lin, & Yin, 2007). The review data are rarely communicated back in an aggregated and actionable format. The first author’s recent conversations (unpublished materials) with nursing staff in acute care inpatient units of three nonprofit hospitals in the northwestern United States revealed that registered nurses (RNs) who provided bedside care do not find the current type of documentation of patient falls to be useful. In other words, they want the information, but it needs to be useful and relevant. These RNs wanted timely and regular summative feedback from administration to assist them in effectively preventing falls; such feedback would include the common contributory factors and types of falls. Clinicians are typically provided unit-specific data on two key nursing outcome indicators: the overall number of falls and the number of injury-falls per 1,000 patient-days. These statistics are insufficient for integrating lessons learned into practice. There is an urgent need to help clinicians make sense of post-fall report data so as to prevent falls. The aim of this study was to demonstrate one approach to summarizing the information collected during review of hospital-acquired inpatient falls to communicate back to clinicians in an aggregated and actionable format for bedside use. This article summarized the post-fall audit report data (completed by designated RN fall specialists) using a retrospective, descriptive, chart review study design. It also compared the overall fall rates (including falls with and without injury; that is, total fall rates) and injurious fall rates pre- and post-implementation of the post-fall audit report summaries.
Background As indicated in Preventing Falls in Hospitals: A Toolkit for Improving Quality of Care (Agency for Healthcare Research and Quality [AHRQ], 2013), communications within health care teams regarding patient falls via hospitals’ incident reporting systems and medical records (e.g., updated patients’ fall risk assessment and care plans) are essential. Hill et al. (2010) found that patient falls were less likely to be reported if they occurred subsequent to an earlier fall (the first fall may be reported by the staff, but later falls from the same patient were less likely to be recorded), or when the fall occurred during the morning or afternoon shift. For performance improvement initiatives, the
286
Clinical Nursing Research 24(3)
hospital’s incident reporting system (completed by the primary nurses of fall patients) could be paired with collection of post-fall data by trained fall evaluators (as post-fall audit reports; Hill et al., 2010; Shorr et al., 2008). Hospital-acquired inpatient falls and fall-related injuries have many causes, including but not limited to impaired cognition, mobility, gait, and balance; a history of falling; and dependence in activities of daily living. Inpatient falls can lead to injury, prolonged hospital stays, lack of patient independence, and additional expenses (Schwendimann et al., 2006). Accidental falls (e.g., tripping due to environmental hazards) and anticipated physiological falls (e.g., abnormal gait and balance) are generally preventable, but unanticipated physiological falls (e.g., having a seizure or fainting episodes) and behavioral or intentional falls (e.g., when a patient acts out) are not (Butcher, 2013). Multifactorial interventions for fall prevention in hospital settings have been found to decrease accidental or anticipated physiological injurious fall incidence; however, the evidence is inconclusive and does not identify specific interventions that are key for success (AHRQ, 2013; Boushon et al., 2012; Butcher, 2013; Cameron et al., 2012; Miake-Lye, Hempel, Ganz, & Shekelle, 2013; Raeder, Siegmund, Grittner, Dassen, & Heinze, 2010; Spoelstra, Given, & Given, 2012; Weinberg et al., 2011). The Institute for Healthcare Improvement’s (IHI) recent publication of Transforming Care at the Bedside How-to Guide: Reducing Patient Injuries From Falls, Boushon and associates (2012) emphasized the importance of communicating to all staff the information regarding patients who are at risk of falling. The audit tool provided by IHI enabled nurses to evaluate the reliability of fall risk assessment and interventions in medical-surgical units in hospital settings. As the authors have observed in clinical practice (e.g., hospitals in the Unites States, Taiwan, China) when conducting previous studies (Tzeng, 2010; Tzeng & Yin, 2013; Tzeng et al., 2007), review of hospital-acquired inpatient falls has been incorporated into routine risk management practice to avoid litigation and continually improve care quality in hospital settings. The impact of conducting post-fall reviews and the best practices for conducting such reviews have not been established. Tzeng and Yin (2013) recently identified 81 potential factors for injurious falls in adult inpatients in hospitals in the United States and 75 interventions for prevention. RNs perceived that the five most frequent risk factors were Alzheimer disease, confusion, disorientation, gait problems, and inability to follow safety instructions. Ivziku, Matarese, and Pedone (2011) found that among 179 elderly patients admitted to a geriatric acute care inpatient unit of an Italian hospital, the risk factors most strongly associated with falls were confusion and depression. Another recent study, conducted by Kolla, Lovely,
Tzeng and Yin
287
Mansukhani, and Morgenthaler (2013) in the United States, found that use of zolpidem, a hypnotic agent that may decrease gait balance, was a significant and potentially modifiable risk factor for falls in adult inpatients in acute care settings. In short, bedside prevention of patient falls requires a multifactorial approach and an interdisciplinary health care team, which should include representatives from nursing, medicine, pharmacy, education, physical therapy, and facilities management.
Method Design This retrospective, descriptive, chart review study was done to demonstrate one strategy for communicating aggregated and actionable fall data to bedside nurses. To do so, it summarized the post-fall audit report data using a retrospective descriptive study design. This study analyzed the quantitative data captured in the post-fall audit reports for patient falls. As stated in the document of Fall Policy and Protocol of the study hospital (updated in 2012, p. 1), patient fall is defined as “a sudden, uncontrolled, unintentional, downward displacement of the body to the ground or other object, excluding such resulting from violent blows or other purposeful actions. Incidents where the caregiver intervenes to prevent injury are included in this definition.” This document was reviewed and updated as needed by the fall committee. The post-fall audit tool was also designed by the fall committee of the study hospital as a fall prevention initiative. This interdisciplinary committee included representatives from nursing, pharmacy, education, physical therapy, and facilities management. These falls occurred between March 1 and December 31, 2012, at a nonprofit acute care hospital in the northwestern United States. This project was approved by the institutional review boards of the study hospital and the university affiliated with the first author. The study hospital, a regional medical center, had a Magnet designation. It had 252 certified beds in service and included one adult critical care, medicalsurgical unit (30 beds); one step-down progressive care, medical-surgical unit (20 beds); two adult acute medical units (52 beds); two adult acute surgical units (50 beds); one pediatric unit (11 beds); one birthing center (labor and delivery, postpartum, and neonatal intensive care, 12 beds); one adult acute rehabilitation unit (14 beds); one acute adult psychiatric unit (22 beds); one acute child/adolescent psychiatric unit (26 beds); one acute chemical dependence unit (15 beds); and the emergency room. The prevalence of the hospitalacquired condition of fall and trauma was 1.229 per 1,000 patient discharges (U.S. national rate = 0.527; http://medicare.gov/hospitalcompare).
288
Clinical Nursing Research 24(3)
Data Source and Sample The data source for this study was the completed post-fall audit reports collected and maintained by the fall committee of the study hospital. As stated in the study hospital’s fall policy and protocol, post-fall reviews were conducted by one of the two designated RN fall specialists shortly after a patient fall occurred. Both fall specialists were co-chairs of the fall committee. Post-fall review, using a real-time audit tool, was conducted to improve understanding of the history and types of falls. The fall committee approved the post-fall review mechanism as a quality-improvement initiative on February 20, 2012, with the goal of decreasing patient falls beginning March 1, 2012. A RN working at the study hospital recoded data from the post-fall audit reports for individual patients and entered the data into a spreadsheet. The de-identified data were then forwarded to us for analysis. A total of 119 patient falls occurred at the study hospital between March 1 and December 31, 2012. Patient falls were included that occurred in the inpatient care units or emergency room and that the fall specialists had identified as accidental falls (falls related to environmental factors; that is, due to an unsafe environment, such as equipment or water in a patient’s path) or anticipated physiological falls (falls related to intrinsic factors; that is, due to known individual risk factors, such as confusion, weakness, and seizures). Nine unanticipated physiological falls (i.e., unpreventable falls, such as falls due to cardiac arrest syndrome and changes with medications) and three falls that occurred outside of the study units were excluded from the analysis, resulting in a total of 107 fall cases valid for this analysis.
Measurement The one-page post-fall review form was developed by the fall committee of the study hospital. This form included seven sections. Section 1, “Demographic Characteristics,” included the patient’s sex, patient age in years, and admission date. The unit and location where the fall occurred, and the date and time of the fall were also documented. The length of stay in whole days (no decimals), from the admission date to the date of fall, was calculated. In Section 2, “Interventions in Place at Time of Fall as Observed by Fall Specialist,” as observed by the fall specialist, were documented. These included fall/yellow socks, fall magnet outside the patient’s room, the bed in a low position, bed-check on, use of one-to-one sitter, call light within reach of the patient, fall/yellow sticker on patient armband, fall/yellow
Tzeng and Yin
289
sticker on chart, in up-front room (i.e., moving patients to rooms near the nurses’ station), and others. For Section 3, “Interview Questions,” the fall specialist asked the primary nurse of the fall patient two questions: (a) What was the patient’s fall risk prior to this fall? (low, moderate, high, or don’t know), and (b) Do you have any concerns (e.g., did the bed alarm work properly, were there environmental or safety issues or any changes in mental status or medications)? As described in the study hospital’s fall policy and protocol, a staff nurse will assess a patient’s fall risk using the 2006 Johns Hopkins Hospital fall risk assessment tool within the first 8 hr of admission, within each shift assessment, and when there are changes in a patient’s condition. This tool has been incorporated into the study hospital’s computer documentation system. For Section 4, “Chart Review,” the fall specialist conducted a chart review and evaluated the following three items: (a) time since last fall assessment was completed (less than 12 hr, 12-24 hr, more than 24 hr, or none completed), (b) last fall risk assessment (low, moderate, high, or none completed), and (c) care plan for fall risk prior to fall (yes, no, or not a fall risk patient). For Section 5, “Reminders,” the fall specialist reminded the primary nurse to complete the following actions and recorded those discussed with the primary nurse: (a) document fall facts in nursing notes, (b) contact physician if needed, (c) contact family if needed, (d) fill out risk management notification, (e) update care plan, (f) document the fall in the shift report for nursing staff, and (g) update fall risk assessment. In Section 6, “Additional Information,” the fall specialist indicated whether any injury occurred and the type of fall (accidental, anticipated physiological, unanticipated physiological, or intentional). Section 7 included one open-ended question to prompt the fall specialist to recommend any other actions or offer other comments.
Data Analyses Data were analyzed using IBM SPSS Statistics 20 software (SPSS Inc.; Chicago, Illinois, USA). Only the quantitative data collected during the postfall reviews were analyzed. Univariate and bivariate analyses were performed to document patient and hospitalization characteristics related to falls. Univariate analyses (also called descriptive analyses; means, standard deviations [SDs], and frequency tables) were used as appropriate. Bivariate analyses (independent-sample t tests; one-way, between-group ANOVA; and kappa measure of agreement tests) were used, as appropriate to the types of variables; alpha was set at .05.
290
Clinical Nursing Research 24(3)
Results A total of 107 fall cases were included in the analysis. Table 1 summarizes the findings of the descriptive analyses, and the variables are presented in the order they are described in the “Measurement” section. Independent-sample t tests were conducted to compare patient age and length of stay on the date of fall between the accidental fall group (falls related to environmental factors) and the anticipated physiological fall group (falls related to intrinsic factors) and between the fall group with injury and the group without injury. A summary of the findings follows. •• Patients who experienced accidental falls were significantly younger, M = 58.36 years, SD = 21.15 years, n = 50, than those who experienced anticipated physiological falls, M = 67.49 years, SD = 15.90 years, n = 57; t(105) = −2.50, p = .014 (two-tailed). •• There was no significant difference in length of stay at date of fall for the accidental fall group, M = 5.40 days, SD = 5.72 days, n = 50, versus the anticipated physiological fall group, M = 7.02 days, SD = 6.30 days, n = 57; t(105) = −1.38, p = .17 (two-tailed). •• There was no significant difference in patient age for the injury group, M = 63.20 years, SD = 19.09 years, n = 25, versus the no-injury group, M = 64.25 years, SD = 18.52 years, n = 75; t(98) = −0.24, p = .81 (two-tailed). •• There was no significant difference in length of stay at date of fall for the injury group, M = 6.04 days, SD = 6.94 days, n = 25, versus the no-injury group, M = 6.04 days, SD = 4.88 days, n = 75; t(98) = 0.00, p = 1.00 (two-tailed). One-way ANOVA was conducted to explore the impact of the time of fall (in terms of three shifts of 8 hr each [day, evening, and night] and six intervals of 4 hr each) on patient age and length of stay at date of fall. The findings are summarized below. •• There was no significant difference in patient age for falls that occurred during the three 8-hr intervals, F(2, 104) = 1.35, p = .26 (two-tailed): 1. day shift (>7 a.m.-3 p.m.; M = 63.03 years, SD = 17.55 years, n = 34) 2. evening shift (>3 p.m.-11 p.m.; M = 60.19 years, SD = 22.53 years, n = 42) 3. night shift (>11 p.m.-7 a.m.; M = 67.55 years, SD = 14.54 years, n = 31)
291
Tzeng and Yin
Table 1. Analysis of Post-Fall Audit Reports for 50 Accidental Falls Related to Environmental Factors and 57 Anticipated Physiological Falls Related to Intrinsic Factors, March 1 Through December 31, 2012. Variable
n (%)
M (SD)
Section 1: Demographic Characteristics Sex Male 58 (54.2) Female 49 (45.8) Patient age, years 63.22 (19.01)a The unit/department where the fall occurred Critical care unit 3 (2.8) Progressive care unit 7 (6.5) Medical Unit A 36 (33.6) Medical Unit B 14 (13.1) Surgical unit, general 8 (7.5) Surgical unit, orthopedic/neurology 6 (5.6) Birthing center 3 (2.8) Rehabilitation unit 10 (9.3) Psychiatric unit, adult 14 (13.1) Psychiatric unit, child and adolescent 1 (0.9) Emergency room 5 (4.7) Time of fall (a coded variable) >7 a.m.-11 a.m. 15 (14) >11 a.m.-3 p.m. 19 (17.8) >3 p.m.-7 p.m. 25 (23.4) >7 p.m.-11 p.m. 17 (15.9) >11 p.m.-3 a.m. 16 (15) >3 a.m.-7 a.m. 15 (14) Length of stay, days from admission date to date 6.26 (6.07)b of fall (a coded variable) Section 2: Interventions in Place at Time of Fall as Observed by Fall Specialist Fall/yellow socks 48 (44.9) Fall magnet outside patient room 46 (43) Bed in low position 81 (75.7) Bed-check on 24 (22.4) Use of one-to-one sitter 3 (2.8) Call light in reach of patient 79 (73.8) Fall/yellow sticker on patient armband 23 (21.5) Fall/yellow sticker on chart 24 (22.4) In up-front room 50 (46.7) Section 3: Interview Conducted by Fall Specialist With Patient’s Primary Nurse What was the patient fall risk prior to fall? 2.35 (0.85) Low = 1 25 (23.4) Moderate = 2 18 (16.8) (continued)
292
Clinical Nursing Research 24(3)
Table 1. (continued) Variable
n (%)
M (SD)
High = 3 61 (57) Don’t know (coded as a missing value) 2 (1.9) System missing (no information provided) 1 (0.9) Section 4: Chart Review by Fall Specialist Time since last fall assessment completed, hours 1.25 (0.68) 24 = 3 4 (3.7) None completed (coded as a missing value) 3 (2.8) System missing (no information provided) 12 (11.2) Last fall risk 2.38 (0.81) Low = 1 19 (17.8) Moderate = 2 20 (18.7) High = 3 54 (50.5) None completed (coded as a missing value) 4 (3.7) System missing (no information provided) 10 (9.3) Care plan for fall risk prior to fall Yes 65 (60.7) No 7 (6.5) Not a fall risk patient 10 (9.3) System missing (no information provided) 25 (23.4) Section 5: Fall Specialist Reminded Primary Nurse to Complete the Following Actions and Indicated the Ones Addressed Document fall facts in nursing notes 85 (79.4) Contact physician if needed 82 (76.6) Contact family if needed 50 (46.7) Fill out risk management notification 88 (82.2) Update care plan 71 (66.4) Document the fall in the shift report for nursing 82 (76.6) staff Update fall risk assessment 81 (75.7) Section 6: Additional Information Provided by Fall Specialist Injury present Yes 25 (23.4) No 75 (70.1) Unsure (coded as a missing value) 7 (6.5) Type of fall Accidental (related to environmental factors) 50 (46.7) Anticipated physiological (related to intrinsic 57 (53.3) factors) aRange bRange
= 8-93 years. = 1-33 days.
Tzeng and Yin
293
•• There was no significant difference in patient age for falls occurring during the six 4-hr intervals, F(5, 101) = 0.66, p = .66 (two-tailed): 1. 2. 3. 4. 5. 6.
>7 a.m. to 11 a.m. (M = 63.47 years, SD = 18.29 years, n = 15) >11 a.m. to 3 p.m. (M = 62.68 years, SD = 17.45 years, n = 19) >3 p.m. to 7 p.m. (M = 61.84 years, SD = 22.13 years, n = 25) >7 p.m. to 11 p.m. (M = 57.76 years, SD = 23.59 years, n = 17) >11 p.m. to 3 a.m. (M = 68.94 years, SD = 15.34 years, n = 16) >3 a.m. to 7 a.m. (M = 66.07 years, SD = 14.00 years, n = 15)
•• There was a significant difference in the length of stay on the date of fall for falls that occurred during the three 8-hour intervals, F(2, 104) = 3.63, p = .03 (two-tailed); the post hoc tests showed that mean length of stay on the date of fall was significantly shorter for the falls that occurred during the evening shift compared with those that occurred during the night shift (M difference = −3.39 days, p = .046): 1. day shift (>7 a.m.-3 p.m.; M = 7.26 days, SD = 5.59 days, n = 34) 2. evening shift (>3 p.m.-11 p.m.; M = 4.36 days, SD = 4.34 days, n = 42) 3. night shift (>11 p.m.-7 a.m.; M = 7.74 days, SD = 7.84 days, n = 31) •• There was no significant difference in the length of stay on the date of fall for the six 4-hr intervals, F(5, 101) = 1.79, p = .12 (two-tailed): 1. 2. 3. 4. 5. 6.
>7 a.m. to 11 a.m. (M = 8.07 days, SD = 5.64 days, n = 15) >11 a.m. to 3 p.m. (M = 6.63 days, SD = 5.62 days, n = 19) >3 p.m. to 7 p.m. (M = 4.36 days, SD = 4.54 days, n = 25) >7 p.m. to 11 p.m. (M = 4.35 days, SD = 4.17 days, n = 17) >11 p.m. to 3 a.m. (M = 6.56 days, SD = 4.79 days, n = 16) >3 a.m. to 7 a.m. (M = 9.00 days, SD = 10.18 days, n = 15)
A kappa measure of agreement test was used to examine the degree of agreement on fall risk (low, moderate, or high) prior to fall as assessed by the primary nurse (in an interview) versus the fall specialist (based on a chart review). The kappa measure of agreement value was 0.85 (p < .001), representing a very good agreement. In other words, the primary nurses’ understanding about patients’ fall risk prior to falls was consistent with the documented fall risk. In addition, the authors used the study hospital’s archived data to investigate the impact of the post-fall review on the overall fall rates and injurious
294
Clinical Nursing Research 24(3)
fall rates. The quarterly report of the National Database of Nursing Quality Indicators (NDNQI), prepared by the American Nurses Association, was used to help interpret any changes that occurred after the tool was implemented. The pre-implementation NDNQI data included the second, third, and fourth quarters of 2011. The post-implementation NDNQI data were composed of the data from the second, third, and fourth quarters of 2012. Based on the acuity-adjusted information from all reporting adult acute inpatient care units combined, the average number of falls per 1,000 patient-days was 4.84 in 2011 and 4.43 in 2012. The average number of injury-falls per 1,000 patient-days was 0.86 in 2011 and 0.67 in 2012. In other words, the average overall fall rate was 8.47% lower in 2012 compared with 2011. The average injury-fall rate was 22.09% lower in 2012, with improved patient outcomes, compared with 2011. Following the NDNQI guidelines, the total number of hours per patient-day was 11.6 in 2011 and 15.08 in 2012. The average RN hours per patient-day were 10.94 in 2011 and 15.15 in 2012, and RNs supplied an average of 65.64% of all nursing hours in 2011 and 65.9% in 2012. Therefore, the number of nursing hours per patient-day was 30% greater at the study hospital in 2012, and the number of RN hours per patient-day was 38.48% greater. It is possible that increases in the number of staff from 2011 to 2012 are a confounding factor in the present study of the post-fall review and overall fall rate and injury-fall rate.
Discussion The objective to this study was to demonstrate one approach for summarizing the information collected during review of hospital-acquired inpatient falls to communicate back to clinicians in an aggregated and actionable format for bedside use. A retrospective descriptive study was used to achieve this objective. About half (n = 50, 47%) of the falls occurred in the adult medical units. The next-most common fall sites were the adult psychiatric unit (n = 14, 13.1%) and the surgical units (n = 14, 13.1%). Mean length of stay on the date of fall was shorter for falls that occurred during the evening shift rather than the night or day shifts. Analysis by time of fall (by shift) could provide additional insights into possible modifications to structure and process indicators (e.g., staffing levels and intervention choices to prevent falls among new patients during the evening shift). As for the interventions in place at the time of the fall, the two used most frequently were keeping patient beds in a low position and keeping the call light within reach of the patients. Only about 75% of the patients had one or both of these interventions in place. More than 20% (n = 25, 23.4%) of the patients had a low fall risk prior to fall, and nurses had completed the last fall assessment for 75.7% (n = 81) of them within
Tzeng and Yin
295
12 hr before the fall. In addition, the patients who experienced accidental falls were significantly younger (M = 58.36 years, SD = 21.15 years, n = 50) than those who experienced anticipated physiological falls (M = 67.49 years, SD = 15.90 years, n = 57). This study’s finding on the fall rates pre- and post-implementation of the post-fall audit report suggested that the post-fall audit report was a useful quality-improvement initiative at the study hospital. Because of limited data points for trending and statistical analyses, the authors could not conclude that implementing the post-fall audit report significantly decreased the overall fall rate and the injury-fall rate. The downward trends for both rates are encouraging. As the authors are not employed by the study hospital, the first author gave a presentation of the summary of the findings to the fall committee at the study hospital in mid-June 2013. The fall committee members found the data summary to be helpful and to verify their impressions about characteristics of falls at their hospital. This presentation was the first time that the study hospital had the post-fall data in an electronic format and subject to research analysis. Discussion focused on the data in Section 2 (Table 1). The fall committee members thought that additional analyses by unit might help with developing unit-specific fall prevention initiatives. As a courtesy, a week prior to the presentation, the first author shared the findings with an executive inpatient services. After the presentation to the fall committee, the presentation materials were also shared with the Research Committee of the study hospital.
Study Limitations The main limitation of this study is that its generalizability is restricted to a single source of data, a nonteaching hospital. In addition, the post-fall audit report is an in-house, custom-made tool. Because the tool used in this study did not capture whether each patient had had a previous fall during the study period, each fall was treated as an independent incident. Another limitation of the study is the fact that patient fall data were based on voluntary reporting by staff; it is possible that some falls were not reported. In addition, the sample size was small. It is also noted that there was a significant increase in the number of nursing hours on the units between 2011 and 2012. Nursing hours could be a confounding factor (also called the confounder) making it difficult to interpret the positive impact of conducting post-fall audits on reducing falls and fall injuries. This study was not designed to analyze both changes in nursing hours and post-fall audits on the single hospital’s rates of falls and fall injuries.
296
Clinical Nursing Research 24(3)
Conclusion This descriptive study summarizes the information collected in post-fall audit reports as one approach to communicate back to clinicians in an aggregated and actionable format for bedside use, particularly for RNs who submit the patient incident reports to the hospital’s incident reporting system and assist in completing the post-fall audit reports. A reduction in the fall rates and injury-fall rates after implementing the post-fall audit review as a qualityimprovement initiative was observed in this study. While it is premature to conclude that the reduction was a result of this quality-improvement initiative, these preliminary findings are promising.
Relevance to Clinical Practice To promote patient safety during hospital stays (e.g., preventing hospitalacquired injurious falls), it is important to inform practice in a timely and systematic manner. Clinicians, risk management officers, quality-improvement specialists, and administrators should consider storing collected fallrelated data (e.g., fall risk assessment and post-fall audit data) in an electronic format and developing queries or computer programs to facilitate statistical analyses, reporting, and interpretation of data. It is also essential to develop a hospital-based post-fall audit report form that accounts for the hospital’s safety culture, care environment (e.g., design, layout, and infrastructure for human resources), and practice (e.g., fall policies and protocol). The post-fall report may be used as a stand-alone audit tool for performance improvement or in combination with other data sources for validation (e.g., the hospital’s incident reporting system). As risk factors for accidental falls, anticipated physiological falls, and unanticipated physiological falls are very different, it would be helpful to summarize the data by type of fall. In addition, it would be useful to produce fall profiles individualized for different units to inform bedside nurse practice. The fall profile of an acute rehabilitation unit could differ considerably from that of a critical care unit or a birthing unit. This effort could facilitate development of unit-specific fall prevention initiatives. Additional research is needed to refine the post-fall audit report template to capture the most relevant faller information. Conducting interviews or focus group meetings with clinicians, risk management officers, quality-improvement specialists, and administrators is one way to refine the tool to facilitate data collection and report generation to inform bedside fall prevention care. For example, it would be helpful to add more specific contributory factors to the post-fall audit (e.g., medication related, floor surface, confusion).
Tzeng and Yin
297
Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.
References Agency for Healthcare Research and Quality. (2013, January). Preventing falls in hospitals: A toolkit for improving quality of care (AHRQ Publication No. 13-0015EF, Prepared by RAND Corporation, Boston University School of Public Health, and ECRI Institute). Rockville, MD: Author. Boushon, B., Nielsen, G., Quigley, P., Rutherford, P., Taylor, J., Shannon, D., & Rita, S. (2012). Transforming care at the bedside how-to guide: Reducing patient injuries from falls. Cambridge, MA: Institute for Healthcare Improvement. Retrieved from http://www.ihi.org/knowledge/Pages/Tools/ TCABHowToGuideReducingPatientInjuriesfromFalls.aspx Butcher, L. (2013). The no-fall zone. Hospitals & Health Networks, 87(6), 26-30. Cameron, I. D., Gillespie, L. D., Robertson, M. C., Murray, G. R., Hill, K. D., Cumming, R. G., & Kerse, N. (2012). Interventions for preventing falls in older people in care facilities and hospitals. The Cochrane Database of Systematic Reviews, 12, Article CD005465. doi:10.1002/14651858.CD005465 Hill, A. M., Hoffmann, T., Hill, K., Oliver, D., Beer, C., McPhail, S., . . . Haines, T. P. (2010). Measuring falls events in acute hospitals—A comparison of three reporting methods to identify missing data in the hospital reporting system. Journal of the American Geriatrics Society, 58, 1347-1352. Ivziku, D., Matarese, M., & Pedone, C. (2011). Predictive validity of the Hendrich fall risk model II in an acute geriatric unit. International Journal of Nursing Studies, 48, 468-474. Kolla, B. P., Lovely, J. K., Mansukhani, M. P., & Morgenthaler, T. I. (2013). Zolpidem is independently associated with increased risk of inpatient falls. Journal of Hospital Medicine, 8, 1-6. Miake-Lye, I., Hempel, S., Ganz, D. A., & Shekelle, P. G. (2013). Inpatient fall prevention programs as a patient safety strategy. Annals of Internal Medicine, 158, 390-396. Raeder, K., Siegmund, U., Grittner, U., Dassen, T., & Heinze, C. (2010). The use of fall prevention guidelines in German hospitals—A multilevel analysis. Journal of Evaluation in Clinical Practice, 16, 464-469. Schwendimann, R., Bühler, H., De Geest, S., & Milisen, K. (2006). Falls and consequent injuries in hospitalized patients: Effects of an interdisciplinary falls prevention program. BMC Health Services Research, 6, Article 69. Retrieved from http:// www.ncbi.nlm.nih.gov/pmc/articles/PMC1534028/pdf/1472-6963-6-69.pdf
298
Clinical Nursing Research 24(3)
Shorr, R. I., Mion, L. C., Chandler, A. M., Rosenblatt, L. C., Lynch, D., & Kessler, L. A. (2008). Improving the capture of fall events in hospitals: Combining a service for evaluating inpatient falls with an incident report system. Journal of the American Geriatrics Society, 56, 701-704. Spoelstra, S. L., Given, B. A., & Given, C. W. (2012). Fall prevention in hospitals: An integrative review. Clinical Nursing Research, 21, 92-112. Tzeng, H. M. (2010). Understanding the prevalence of inpatient falls associated with toileting in adult acute care settings. Journal of Nursing Care Quality, 25, 386-394. Tzeng, H. M., & Yin, C. Y. (2013). Frequently observed risk factors for fall-related injuries and effective preventive interventions: A multihospital survey of nurses’ perceptions. Journal of Nursing Care Quality, 28, 130-138. Tzeng, H. M., Yin, C. Y., Tsai, S. L., Lin, S., & Yin, T. J. (2007). Patient falls and open visiting hours: A case study in a Taiwanese medical center. Journal of Nursing Care Quality, 22, 145-151. Weinberg, J., Proske, D., Szerszen, A., Lefkovic, K., Cline, C., El-Sayegh, S., . . . Weiserbs, K. F. (2011). An inpatient fall prevention initiative in a tertiary care hospital. Joint Commission Journal on Quality and Patient Safety, 37, 317-325.
Author Biographies Huey-Ming Tzeng earned her PhD in Nursing from the University of Michigan, School of Nursing, Ann Arbor, Michigan, in 1997. She takes both macro and micro approaches to address patient safety and health care quality issues. She is currently researching the promotion of safe hospital stays, with a focus on fall prevention for adult inpatients. Chang-Yi Yin earned his Master’s degree in History from National Taiwan University, Taipei, Taiwan. He is an internationally well-known historian in Chinese and world history and area studies. He started collaborating with Dr. Tzeng since 2003. Together they have addressed patient safety issues, such as, infection control and environment of care.
![[Implementation of focused ultrasonography in an acute care setting].](/assets/img/hold.png)
