Factors associated with falls in hospitalized adult patients Jill Cox PhD, RN, APN-C, CWOCN, Charlotte Thomas-Hawkins PhD, RN, Edmund Pajarillo PhD, RN, BC, CPHQ, NEA-BC, Susan DeGennaro MS, RN, APN-C, OCN, Edna Cadmus PhD, RN, NEA-BC, FAAN, Miguel Martinez MA PII: DOI: Reference:
S0897-1897(14)00164-5 doi: 10.1016/j.apnr.2014.12.003 YAPNR 50611
To appear in:
Applied Nursing Research
Received date: Revised date: Accepted date:
6 September 2014 5 December 2014 13 December 2014
Please cite this article as: Cox, J., Thomas-Hawkins, C., Pajarillo, E., DeGennaro, S., Cadmus, E. & Martinez, M., Factors associated with falls in hospitalized adult patients, Applied Nursing Research (2014), doi: 10.1016/j.apnr.2014.12.003
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 1
Title: Factors associated with falls in hospitalized adult patients.
T
Authors:
SC
RI P
Jill Cox PhD, RN, APN-C, CWOCN Assistant Professor, Rutgers University School of Nursing, 180 University Ave. Newark NJ Advanced Practice Nurse, Englewood Hospital and Medical Center, Englewood, NJ
[email protected] NU
Charlotte Thomas-Hawkins, PhD, RN Associate Professor, Rutgers University School of Nursing, 180 University Ave. Newark NJ
[email protected] ED
MA
Edmund Pajarillo PhD, RN, BC, CPHQ, NEA-BC Associate Dean, Faculty Services, Rutgers University School of Nursing, 180 University Ave., Newark NJ
[email protected] CE
PT
Susan DeGennaro, MS, RN, APN-C, OCN Advanced Practice Nurse- Oncology, Englewood Hospital and Medical Center, 350 Engle St. Englewood, NJ
[email protected] AC
Edna Cadmus, PhD, RN, NEA-BC, FAAN Clinical Professor, Rutgers University School of Nursing, 180 University Ave, Newark NJ
[email protected] Miguel Martinez, MA Institutional Research Specialist, Rutgers University School of Nursing, 180 University Ave. Newark NJ
[email protected] Corresponding author: Jill Cox Email:
[email protected] Phone: 201-259-5622 (cell) 201-391-3166(home) Fax: 201-391-3589
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 2
Introduction
T
Falls among hospitalized patients occur frequently, and some repeatedly. Despite efforts
RI P
in hospitals to identify patients at risk for falls and to prevent these incidents, falls among hospitalized patients are not a rare event and continue to be a major health care concern. Each
SC
year, between 700,000 and 1,000,000 persons fall in United States hospitals (Currie, 2008). The
NU
Centers for Medicare and Medicaid Services (CMS) considers a fall in a hospitalized patient a never-event. Consequently, healthcare organizations are not reimbursed for these events, further
MA
adding to the financial burden placed on institutions (CMS, 2008). Importantly, the National Patient Safety Goals set forth for hospitals by The Joint Commission (TJC) in 2014 includes a
PT
Background and Significance
ED
specific goal to reduce falls and the associated harm from falls (TJC, 2014).
The National Database of Nursing Quality Indicators (NDNQI) defines a fall as an
CE
unplanned descent to the floor with or without injury to the patient (NDNQI, 2013). The causes
AC
of falls in hospitals are multifactorial including factors intrinsic and extrinsic to the patient, as well as organizational or workforce factors. Intrinsic fall risk factors are specific to the patient’s health status and include conditions such as impaired mental status (confusion), older age (frailty associated with aging), visual disturbances, multiple comorbidities, unsteady gait, use of sedative and hypnotic medications, and history of falling (Currie, 2008; Gray-Miceli & Quigley, 2012). The multitude of intrinsic fall risks for many hospitalized patients makes it difficult to isolate specific risk factors and illustrates the constellation of patient-level risks that can create the “perfect storm” and lead to a patient fall. Extrinsic fall risks include environmental obstacles and the way in which these
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 3
obstructions may influence, or facilitate, accidental falls. For example, tubing from medical equipment, electrical cords, inappropriate footwear, and clutter in the patient’s environment have
T
all been found to increase fall risk in the hospitalized patient (Currie, 2008; Gray-Miceli &
RI P
Quigley, 2012). Fall prevention protocols in acute care hospitals are designed to address both extrinsic and intrinsic fall risk factors commonly encountered in the acute care setting as a means
SC
to minimize fall occurrences. Finally, staffing on inpatient units has also been examined as a
NU
system or unit-level predictor of falls in hospitalized patients. There is empirical evidence that low nurse-to-patient ratios and a higher proportion of registered nurses(RN) to unlicensed
MA
assistive personnel (UAP) on hospital units are important factors to consider in inpatient fall preventions efforts (Lake, Shang, Klaus, & Dunton, 2010; Titler, Shever, Kanak, Picone, & Qin,
ED
2011).
PT
Most falls can be categorized as one of four types: accidental, unanticipated physiologic, anticipated physiologic, or intentional, (Currie, 2008; Gray-Miceli & Quigley, 2012; Morse,
CE
2009; Quigley, Palacios & Sephar, 2010). An accidental fall is defined as an unintentional fall
AC
due to extrinsic factors such as patient trips, slips, or falls because of an environmental problem such as spills on the floor or objects strewn in his or her path. An unanticipated physiologic fall occurs due to a patient’s unknown intrinsic risk factors such as a syncopal episode, seizure or pathological hip fractures. These factors would not have been identified during a fall risk assessment and are thus considered unanticipated. Conversely, an anticipated physiologic fall occurs in a patient with a known intrinsic fall risk factor. Examples would include a history of a prior fall, weak or impaired gait, incontinence, or impaired mental status. Lastly, an intentional fall occurs when the patient deliberately falls.
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 4
Inpatient falls can result in injury and death and place an increased economic burden to patients and the health care system (Currie, 2008; Wong, Recktenwald, Jones, Waterman,
T
Bollini, & Dunagan, 2012). In fact, longer hospital stays and additional treatment due to fall-
RI P
related injury contribute to a 61% increase in overall patient care costs (Fitzpatrick, 2011) The Centers for Medicare & Medicaid Services (CMS) limits reimbursement to hospitals for certain
SC
types of traumatic fall injuries such as fractures, dislocations, and intracranial injuries. Fall
NU
related injuries occurs in approximately 30-50% of all falls in hospitals ranging from simple bruising to fractures and even death (Wong et al., 2012). The NDNQI (2012) identifies five
MA
levels of injury based on the severity of the injury sustained. A minor or slight injury is described as those that need a simple intervention, i.e., ice application on the site of the injury or
ED
elevation of the injured limb. A moderate injury would require suturing of a sustained open
PT
wound from the fall or splinting the limb; while a major or severe injury can involve a significant surgical intervention, casting the injured extremity or joint, or requiring additional work-up such
CE
as a neurological consultation, radiologic, hematology, or magnetic resonance imaging. A fall
AC
death is one that is attributed to the injuries sustained during the fall, and not due to the physiologic events that caused the fall. Despite the need for fall prevention approaches that are tailored to each patient’s unique intrinsic or extrinsic risks, most fall prevention strategies are universal and designed to capture fall risk within a diverse patient population. While formalized fall risk assessment using validated tools such as the Hendrich I Falls Risk Assessment Scale (Hendrich, 2003; Hendrich, 2013) or Morse Falls Risk Assessment Scale (Morse, Morse, & Tylko, 1989) yield a composite fall risk score, these scores do not indicate the type of fall for which a patient is at risk, making it difficult to tailor or target fall prevention strategies to complement the patient’s unique
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 5
physiological or environmental risks. Thus, there is a clear need to gain a broader understanding of the factors, both patient-related and environmental, that contribute to falls in hospitalized
T
patients.
RI P
Purpose of the study
The purpose of this study was to examine intrinsic, extrinsic, and workforce factors that
SC
contribute to falls among hospitalized adult patients. The following aims were addressed in this
NU
study: 1) ascertain the number of documented patient falls attributable to one of four causes (accidental, anticipated, unanticipated, intentional); 2) determine the level of injury associated
MA
with each of the four types of falls; and, 3) identify any associations between demographic, patient, environmental, and workforce factors and the odds of fall occurrences.
ED
Design and Methods
PT
This was a descriptive, correlational, retrospective study. The study was conducted in a 500-bed Magnet teaching hospital in northeastern New Jersey.
CE
Sample. The electronic medical records of all patients who were admitted to the hospital
AC
during the year 2012 were delimited to those patients admitted to medical and/or surgical units during this time period. All adult patients admitted to a medical and/or surgical unit in 2012 and who fell during their hospital stay were identified via the hospital’s safety reporting system. These patients were ordered randomly, and every third patient was systematically selected until 50 patients were chosen. In addition, all adult patients admitted to medical and/or surgical units in 2012 who did not fall were ordered randomly and systematically selected until 110 patients were chosen via the hospital’s electronic medical record database. Each of the 110 non-fallers selected were matched two to one with fallers on the unit during the month of admission. Thus,
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 6
the study sample was comprised of 50 fallers and 110 non-fallers admitted to a medical and/or surgical unit in 2012.
T
Inclusion criteria were established for fallers and non-fallers. For fallers, the patient must
RI P
have been 1) 18 years of age or older; 2) admitted to a medical and/or surgical unit in 2012; and 3) have fallen at least once during the hospitalization on the unit. For the non-faller sample,
SC
patients must have been 1) 18 years of age or older; 2) admitted to a medical and/or surgical unit
NU
in 2012; and 3) did not experience a fall during the hospitalization. Data Collection
MA
Data for this study were abstracted from the hospital’s existing electronic medical record (EMR) system and recorded on a data abstraction record developed for this study. All data
ED
abstracted were de-identified to protect subject confidentiality. Staff nurses (RNs) were recruited
PT
to assist with data abstraction for this study and were trained by the Principal Investigator and co-investigators. Co-investigators concurrently audited twenty percent of data abstracted by data
CE
collectors to ascertain compliance and adherence to the data collection process. Institutional
AC
Review Board approval for this study was obtained from both Rutgers University and the participating hospital. Data Analysis
Descriptive statistics were used to summarize and to describe the study sample. Frequencies were recorded for each fall type category and fall injury level. Prior to bivariate and multivariate analyses, analytic weights were developed and applied to the each faller and nonfaller in the study to correct for the overrepresentation of fallers (Magee et al, 2008; Bell et al, 2012). In the target population, fallers represented less than 2 percent of all patients admitted to the medical-surgical units during 2012 but comprised approximately 31 percent of the study
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 7
sample. The application of analytical weights simultaneously restored the proportion of fallers to non-fallers in the target population – from 31 percent to about 2 percent - and expanded it to
T
represent the number of patients in the target population – from 110 to 12,834 patients1. Hence
RI P
forth, we will refer to the non-faller sample to which the analytical weights were applied as the weighted sample (n=12,834) and the other as the non-weighted sample (n=160). For fallers, after
SC
analytical weighting was applied this sample expanded from 50 to 205 patients. Logistic
NU
regression was employed to determine the factors that predicted the odds of a patient fall during the hospital admission.
MA
Results
Descriptive statistics and frequency distributions for the demographic and study variables
ED
for the non-weighted study sample are summarized in Table 1. The mean age for the study
PT
sample was 70 years (SD 18). Caucasian was the dominant race in the sample at 73% (n=117). The top admitting diagnosis was cardiac (31%) and the average length of the hospital admission
CE
was 5.6 days (SD 5). The mean admission Hendrich I fall scale score for the study sample was 7
AC
(SD 5). Most patients (69%) had no prior history of falling in the study sample. Falls were stratified by type (Table 2) to address the first specific aim of this study. The majority of falls (54%, n=27) were classified as anticipated physiologic falls. Accidental falls comprised the second highest (28%) type of falls (n = 14). Only one fall, representing 2% of the fallers, was categorized as an unanticipated physiologic fall. For the second specific aim of this study, frequency distributions were calculated to identify the distribution of fall injury level by fall type (Table 3). The majority of fallers sustained no injury (74%, n = 39). In addition, no patient in this sample sustained a serious or 1
After developing and applying analytical weights the actual number of patients in the target population was verified at 12,727. The difference of less than 1 percent in the number of patients between the target population and the expanded sample is likely due to the rounding of analytical weights.
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 8
fatal injury. Seven patients who experienced an anticipated physiologic fall sustained either slight (n= 6) or moderate injury (n= 1).
T
All variables found to be significant in bivariate analysis in the weighted sample were
RI P
entered into logistic regression analysis to address the third specific aim of the study. Age (Odds Ratio [OR] = 1.17; p = 0.027; 95% confidence interval [CI], 1.01 - 1.35), narcotic/sedative use
SC
(OR = 16.64; p = 0.001; 95% CI, 2.96 - 93), and overnight shift (OR = 3.12; p = 0.00; 95% CI,
NU
1.69 - 5.75) significantly and independently predicted the likelihood of a fall during the hospitalization. Conversely, cardiovascular comorbidities (OR = 0.10; p = 0.001; 95% CI, 1.01 -
MA
1.35), neuro/musculoskeletal disease (OR = 0.233, p = 0.000; 95% CI, 0.145 - 0.373), evening shift (OR = 0.011; p = 0.035; 95% CI, 0.00 - 0.729), implementation of fall prevention strategies
ED
(OR = 0.12; p = 0.00; 95% CI, 0.06 - 0.26), and a higher RN-to-unlicensed assistive personnel
PT
(UAP) staffing ratio (RN/UAP) (OR = 0.19; p = 0.001; 95% CI, 0.06 - 0.58) were significantly
(Table 4).
CE
and independently associated with a decreased likelihood of a fall during the hospitalization
AC
Marginal probabilities of experiencing a fall based on risks were calculated revealing that patients identified at the time of admission to be at high risk for a fall experienced a 17% higher probability of falling during the hospitalization episode, as compared to patients admitted at low or moderate fall risk levels. Discussion and Clinical Application Falls in hospitalized patients can occur for many reasons. Factors such as variations in patients’ age, health status, medication history, comorbid conditions, functional abilities as well as environmental issues must all be considered when determining fall risks for persons who are hospitalized. Moreover, determining the type of fall experienced by a patient is important. This
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 9
is critical in identifying the most appropriate fall prevention strategies based on the patient’s unique needs, in assisting caregivers to understand the overall nature of falls and in supporting
T
the staff in their efforts to prevent falls in future patients. Findings from this study support the
RI P
multi-etiologic nature of falls in hospitalized patients. Some factors identified as fall risks in this study are consistent with the empirical literature. Other fall risks, while divergent, provide
SC
insight into the root cause of patient falls and are worthy of future investigation.
NU
The first aim of our study was to stratify each fall according to type. While the literature supports stratification of falls by type, there is a gap in the empirical literature regarding use of
MA
this classification system. Morse (2009) stated that the use of this classification system in clinical practice might likely result in the stratification of approximately 78% of falls as anticipated
ED
physiologic events. Individualized and timely application of safety measures will potentially
PT
avert a particular type of fall when nursing staff are aware and recognize the various stratifications of falls. In this particular study, anticipated physiologic falls were the most
CE
common type of fall identified. The goal of universal fall- and injury-prevention efforts in
AC
hospitals is to decrease or prevent adverse outcomes for patients who are deemed at risk for falling. However, for a patient who has a unique identified physiological risk, i.e., visual impairment, a more tailored approach to fall prevention is appropriate as compared to a person with no known physiological risks (Oliver, Healy, & Haines, 2010). More than 25% of the falls in this study were extrinsically driven and due to environmental hazards. Recognition of environmental hazards and orienting patients to the environment are key strategies for patients at risk for accidental falls and are often part of a comprehensive fall prevention plan. It is widely recognized, however, that the prevention of falls is difficult even with the implementation of multimodal intervention strategies aimed at both
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 10
intrinsic and extrinsic fall risk factors in the complex and evolving clinical environment of hospitals (Colen-Emeric, Schenck, Gorospe et al, 2006). Thus, ongoing evaluations of fall- and
T
injury-prevention interventions should be conducted to determine the effectiveness of these
RI P
efforts as best practices.
The second aim of our study was to determine the fall injury level by fall type. In our
SC
sample, 74% of the fallers experienced no injury as a result of a fall, which may be attributed in
NU
part to successful recognition of fall risk and the implementation of universal fall prevention interventions. It may also be attributed to the limitation in the sampling methodology. Moderate
MA
level injuries, which represented 6% (n=3) of the faller sample, occurred in falls categorized as either anticipated physiologic (n=1), unanticipated physiologic (n=1), or falls due to risk factors
ED
inherent in a fall categorized as a combination of an accidental or anticipated physiologic fall
PT
(n=1). Future research that examines events surrounding falls that result in moderate or severe injuries may elucidate fall risk factors that are potentially modifiable (i.e., environmental
CE
factors), from factors that could be considered non-modifiable (i.e., patient comorbidities).
AC
Prevention efforts specifically tailored to intrinsic patient and environmental patient risks can be implemented and evaluated. The third aim of our study was to determine risk factors significantly associated with the likelihood of fall occurrences in hospitalized medical and /or surgical patients. In multivariate analysis, three intrinsic factors - age, narcotic/sedative use, and high fall risk assessment score were all found to significantly predict the likelihood of a fall. Patient factors, such as age and the use of narcotic/sedative use have been supported in varying degrees in the empirical literature. Advancing age as a fall risk factor is strongly supported among individuals living in the community setting, while among those who are hospitalized, this variable has not emerged as a
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 11
strong predictor of falls (Currie, 2008; Hendrich, Bender, & Nyhuis, 2003). The mean age for fallers in this study was 69 years, representing the lower end of the older adult spectrum. Our
T
findings regarding narcotic and sedative use supports the empirical evidence linking these
RI P
medications to patient falls (Costa-Dias, Oliveira, Martins, Araujo, Santos, Moreira, & Jose, 2014; Titler et al., 2011). Additionally, patients prescribed narcotics or sedatives were 16 times
SC
more likely to fall, similar to the result in a retrospective analysis of 193 hospitalized patients by
NU
Costa-Dias and colleagues (2014) who found that these medications increased the likelihood of a fall nine times.
MA
Temporal variation in fall occurrence was also demonstrated in this study. Results indicated that the likelihood of a fall increased during the night shift, while the evening shift was
ED
associated with a lower likelihood of a fall. Findings from another study of 711 fall events set in
PT
the acute care setting (Kerzman, Chetrit, Brin, & Toren, 2004) showed a contrasting outcome that the daytime shift was significantly associated with falls. Currently, there is a lack of
CE
evidence to conclusively link shift or time of day to fall occurrence.
AC
Patients in our study who were determined to be at high risk for falls using the Hendrich I fall scale were nearly 17% more likely to fall during the hospitalization. The predictive nature of a fall indicated by a high score on the Hendrich I scale is significant for two reasons. First, this supports the predictive validity of the Hendrich I risk assessment tool that has been previously described in the literature to demonstrate robust specificity and sensitivity (Hendrich, Bender & Nyluis, 2003.). Secondly, it underscores the value of fall risk assessment as an important nursing intervention aimed at fall prevention that can be simply and efficiently incorporated into clinical practice.
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 12
Factors found to decrease the likelihood of a fall in our sample included cardiovascular disease, neurological/musculoskeletal disease, higher RN/UAP staffing ratios, and the evening
T
shift. While the literature supports comorbid disorders such as arrhythmias, syncope, multiple
RI P
sclerosis and Parkinson’s disease as risk factors for falls (Currie, 2008), our results did not align with these previous findings. This might be an aberrant outcome among our study sample but it
SC
is also equally plausible that patients with these disorders were recognized as high fall risk on
NU
admission and fall prevention strategies were appropriately initiated, thereby averting potential falls among these patients.
MA
The evening shift hours were also found to significantly decrease the likelihood of a fall in our study sample. As previously stated, there is no consensus in the literature regarding shift
ED
variations and fall likelihood (Currie, 2008). It is plausible that less falls on the evening shift
PT
could be the result of increased patient visitation in the evening. With patients having company during this time period, it may be less likely for them to attempt independent ambulation either
CE
to go to the bathroom or to get up and be out of bed on their own. Further research on the time of
AC
fall occurrence is warranted to understand the diurnal impact on fall occurrence. The implementation of fall prevention strategies was also found to significantly decrease the likelihood of a fall in our study. This underscores the positive impact regarding the implementation of evidence-based fall prevention interventions on overall fall occurrence. In the hospital where this study was conducted, a standardized fall risk assessment protocol was implemented based on the strategies outlined by the Agency for Health Care Quality and Research (Degelau, Belz, Bungum, Flavin, Harper, Leys, Lundquist, & Webb, 2012). In our study sample, 88% (n=140) of patients had a fall prevention protocol in use. Results from this
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 13
study lend support to the assertion that evidence-based fall prevention interventions can be effective modalities that can decrease fall occurrence in hospitalized patients.
T
Finally, a higher RN/UAP ratio significantly predicted a decreased likelihood of falls.
RI P
Prior studies support this outcome (Lake et al., 2010; Titler et al., 2011) that the RN is a critical component of a successful fall prevention program. Clearly, effective communication of fall
SC
risks to ancillary caregivers or the incorporation of effective critical thinking skills in patient care
NU
that is aimed at decreasing patient falls can be essential components of fall prevention strategies employed by RNs. Qualitative research is needed to explore and disentangle the specific and
MA
likely implicit strategies used by RNs to prevent falls in hospitalized patients. Importantly, our research supports the evidence that the presence of the RN is essential to fall prevention in the
ED
hospital setting.
PT
Limitations
The retrospective nature of this study is recognized as a limitation. However, most of the
CE
data abstracted from the EMR and patient reporting systems represent objective data that would
AC
not vary regardless of the study approach employed. One benefit of a prospective study would be the advantage of participation in a post-falls analysis in real time which could provide additional clarity to the nature and circumstance surrounding a fall event. Our sampling methodology of sampling non-fallers to fallers at a 2:1 ratio may also be considered a limitation of this study. The use of the statistical methodology of analytical weighting was used to overcome this limitation in order to provide greater meaning to the data obtained. Use of a single research site also diminishes the generalizability of the study findings.
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 14
Conclusions Patient falls cannot be completely avoided and will continue to occur in hospitalized
T
patients, sometimes resulting in serious injury that may have lasting effects on the patient’s
RI P
future quality of life, long after hospital discharge. In order to successfully prevent a fall, risk factors must be accurately assessed and prevention strategies must be effectively implemented.
SC
Results of our study support these two key elements (assessment and prevention) as significant
NU
considerations that are fundamental to a comprehensive fall reduction program. A patient with a high fall risk score determined through the use of a validated tool was shown in our research to
MA
be a significant predictor of a fall during the hospitalization. Additionally, the initiation of fall
likelihood of fall occurrence.
ED
prevention strategies following a prescribed fall prevention protocol was found to decrease the
PT
Exploring the types of falls can lead to prevention interventions that can be specifically tailored to meet the fall risks identified for each patient. While fall prevention is the role of all
CE
caregivers, the RN plays a pivotal role in both risk determination and prevention; the value of
AC
this role cannot be underestimated in a comprehensive fall prevention program. Without accurate risk determination, the implementation of fall prevention approaches cannot appropriately take place. While the prediction of a fall with injury remains somewhat elusive at this time, there is a need for the development and implementation of a risk assessment tool that has the potential to accurately predict the potential for a fall with injury. The ultimate outcome is a safer, cost effective hospital admission that will eventually improve the patient’s quality of care. Acknowledgements We would like to acknowledge the following staff RNs who so diligently assisted us with the data collection for this study: Nelio Abdon, Nicole Chvasta, Lauren Dotson, Carole Eastman, Aurora Garcia, Norma Maine, Janet Mantel, Jung Sang, Diana Torres, Jamie Valdez.
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 15
References
RI P
and Medicaid by addressing never events. Available at
T
Center for Medicare and Medicaid Services (2008). CMS improves patient safety for Medicare
http://www.cms.gov/Newsroom/MediaReleaseDatabase/Fact-Sheets/2008-Fact-Sheets-
SC
Items/2008-08-042.html.
NU
Colon-Emeric, C., Schenck A., Gorospe, R. McArdle, J., Dobson, L.,DePorter, C., & McConnell, E. (2006). Translating evidence-based falls prevention into clinical practice
MA
in nursing facilities: Results and lesson from a quality improvement collaborative. Journal of the American Geriatric Society, 54(9), 1414-1418.
ED
Costa-Dias, M. J., Oliveira, A. S., Martins, T., Araujo, F., Santos, A. S., Moreira, C. N., & Jose,
PT
H. (2014). Medication fall risk in old hospitalized patients: a retrospective study. Nurse Education Today 34, 171 - 176.
CE
Currie, L. (2008). Fall and injury prevention. In R. Hughes (Ed.). Patient Safety and Quality.
AC
An Evidence Handbook for Nurses. Available at http://www.ahrq.gov/legacy/qual/nurseshdbk/ Degelau, J., Belz, M., Bungum, L., Flavin, P. L., Harper, C., Leys, K., Lundquist, L., & Webb, B. (2012). Prevention of falls (acute care). Health care protocol. Bloomington (MN): Institute for Clinical Systems Improvement (ICSI), 2012 April. Fitzpatrick, M.A. (2011). Meeting the challenge of falls reduction. American Nurse Today, 6(2), 1 - 20. Gray-Miceli, D. & Quigley, P.A. (2012). Fall prevention: Assessment, diagnosis, and
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 16
intervention strategies. New York University Hartford Institute for Geriatric Nursing. Available at http://consultgerirn.org/topics/falls/want_to_know_more.
T
Hendrich, A. (2013). Fall risk assessment for older adults: The Hendrich II Fall Risk ModelTM .
RI P
In Try This: Best Practices to Nursing Care of Older Adults. Greenberg, S. (Ed.). New York University, The Hartford Institute for Geriatric Nursing, (8). Available at
SC
http://consultgerirn.org/uploads/File/trythis/try_this_8.pdf.
NU
Hendrich, A. L., Bender, P. S., & Nyhuis, A. (2003).Validation of the Hendrich II fall risk model: A large concurrent case/control study of hospitalized patients. Applied Nursing
MA
Research, Feb 2003, 16(1), 9 - 21.
Kerzman, H., Chetrit, A. Brin, L. Toren, O. (2004). Characteristics of falls in hospitalized
ED
patients. Nursing and Health Care Management and Policy, 47(2), 223 - 229.
PT
Lake, E., Shang, J. Klaus, S, Dunton, N. (2010). Patients falls: Association with hospital Magnet status and nursing unit staff. Research in Nursing and Health, 33, 413 - 425.
CE
Magee, T., Lee, S. M., Giuliano, K. K., & Munro, B. Generating new knowledge from
S56
AC
existing data: The use of large data sets for nursing research. Nurs Res, 2006; 55(2), S50-
Morse, J. M., Morse, R. M., & Tylko, S. J. (1989). Development of a scale to identify the fallprone patient. Canadian Journal on Aging, 8, 366 – 367. Morse, J.M.(2009) Preventing patient falls.(2nd ed), Springer Publishing, New York. National Database of Nursing Quality Indicators (2013). Changes to NDNQI fall indicator coming for 2Q 2013. NDNQI Nursing Quality News 14(1), 2. National Database of Nursing Quality Indictors (2012) Guidelines for the American Nurses Association National Quality Forum endorsed measures. Accessed at:
http://www.nursingquality.org/Content/Documents/NQF-Data-Collection-Guidelines.pdf
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 17
Oliver, D., Healy, F., & Haines, T. P. (2010). Preventing falls and fall-related injuries in hospitals. Clinics in Geriatric Medicine, 26, 645 - 692. doi: 10.1016/j.cger.2010.06.005.
T
Quigley, P.A., Palacios, P., & Spehar, A. M. (2006). Veterans’ fall risk profile: A
RI P
prevalence study. Clinical Interventions in Aging, 1(2), 169 - 173.
The Joint Commission. (2014). National patient safety goals. Available at
SC
http://www.jointcommission.org/standards_information/npsgs.aspx.
NU
Titler, M. Shever, L., Kanak, M. Picone, D, & Qin, R. (2011). Factors associated with falls
Practice, 25(2), 127 - 151
MA
during hospitalization in an older adult population. Research and Theory for Nursing
Wong, C. A., Recktenwald, A. J., Jones, M. L., Waterman, B. M., Bollini, M. L., & Dunagan, W.
ED
C. (2012). The cost of serious fall-related injuries at three Midwestern hospitals. Journal of
AC
CE
PT
Quality & Patient Safety, 37(2), 81 - 87.
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 18
Table 1. Frequency Distribution of the Demographic and Study Variables Valuea
T
Variables
RI P
Demographic Factors Age, Mean (SD), range
70,(17.6),20-102
SC
Sex Male
85(53%) 75(47%)
NU
Female Race
MA
Caucasian Black/African American
ED
Asian/Pacific Islander Hispanic
PT
Other/Missing
117(73%) 19 (12%) 9(6%) 9(6%) 6 (3%)
GI Infection Neurological
AC
Cardiac
CE
Hospital Admitting Diagnosis (Top 4)
Length of Hospital Admission (days)
50(31%) 26(16%) 24(15%) 24(15%)
5.6,(5), 1-30
Patient Admitted from: Home
137(86%)
Long Term Care Facility
7(4%)
Transfer from another facility
9(5%)
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 19 Missing data= 7 Patient Related Factors
85(53)
Needs escort
32(20)
Bedrest
19(12)
SC
RI P
Independent Ambulation
T
Functional mobility on admission
Cane
5(3) 2(1)
NU
Orthopedic/Prosthetic device Walker
20(13)
MA
Medication Profile on Admission Cardiac
ED
Narcotic/sedative Diuretic/Laxative
PT
Anticoagulant
CE
Insulin
119(74) 51(32) 45(28) 68(43) 24(15)
Comorbidities on Admission
AC
Cardiovascular
131(82)
Musculoskeletal/Neurologic
116(73)
Metabolic Disorders
82(51)
Psychiatric Disorders
18(11)
Fecal/urinary incontinence
21(13)
Hearing/Visual Impairment
59(37)
Environmental Factors Unit Type Medical
104(65)
Surgical
18(11)
ACCEPTED MANUSCRIPT Falls In Hospitalized Patients 20 Medical/Surgical
32(20)
Workforce Factors 6.5 (SD1.5;2-11.5)
Number of UAPs/shift
3.1 (SD1.2; .93-7.2)
RI P
T
Number of RNs/shift
Fall Related Factors
Low(0-2)
7, (5.4), 0-22
SC
Admission Hendrich I Fall Risk Score (Mean, (SD), range)
NU
3(2%)
Moderate (3-6)
78(49%)
MA
High (7 or greater)
Use of fall prevention protocol a
ED
Previous history of falls
75(47%)
50(31%) 140(88)
AC
CE
PT
Values are number(%) unless otherwise indicated. Due to rounding, not all percentages total 100%.
ACCEPTED MANUSCRIPT
RI P
T
Falls In Hospitalized Patients 21
Table 2. Falls by fall type
N (%) of fallers
Accidental
14(28%)
Anticipated Physiological
27 (54%)
NU
SC
Fall Type
1(2%)
Intentional
MA
Unanticipated Physiological
0
AC
CE
PT
ED
Accidental/Anticipated Physiologic
8(16%)
ACCEPTED MANUSCRIPT
RI P
T
Falls In Hospitalized Patients 22
Table 3. Fall Injury Severity by Fall Type
Accidental
12(24%)
2(4%)
Anticipated Physiologic
18(36%)
6(12%)
Unanticipated Physiologic
0
AC
MA
CE
PT
*fall injury level not indicated
Severe
Total
0
0
14(28%)
1(2%)
0
25(40%) 2*(4%)
0
1(2%)
0
1(2%)
0
1(2%)
0
8(16%)
ED
Accidental/anticipated 7(14%) physiologic
Moderate
SC
Slight
NU
None
ACCEPTED MANUSCRIPT
RI P
T
Falls In Hospitalized Patients 23
Fall Variable
Β(Odds Ratio)
Std Err.
Age
1.17
0.084798
C-V Disease
.010
0.069091
Neuro/Musculoskeletal Disease
.233
Narcotics/sedatives
16.64
Evening Shift
0.0112603
Night Shift
3.120289
P>z
95% CI
2.21
0.027
1.01-1.35
-3.35
0.001
0.026-0.38
0.056026
-6.06
0.000
0.145-0.373
14.65894
3.19
0.001
2.96-93
0.0239634
-2.11
0.035
0.0001740.7294847
0.9758852
3.64
0
1.6903545.759863
MA
NU
Z
ED
PT
CE
SC
Table 4. Significant Predictors of Falls in Hospitalized Patients
0.1982041
0.109837
-2.92
0.003
0.06689750.5872397
Fall Prevention Protocol
0.1281108
0.0476625
-5.52
0.000
0.0617880.265624
AC
Staffing ratio (RN/UAP)