The development of the RISK tool for fall prevention

In-hospital fall prediction using machine learning algorithms and the Morse fall scale in patients with acute stroke: a nested case-control study

BACKGROUND

Falls are one of the most common accidents in medical institutions, which can threaten the safety of inpatients and negatively affect their prognosis. Herein, we developed a machine learning (ML) model for fall prediction in patients with acute stroke and compared its accuracy with that of the existing fall risk prediction tool, the Morse Fall Scale (MFS).

METHODS

This is a retrospective nested case-control study. The initial sample size was 8462 admitted to a single cerebrovascular specialty hospital with acute stroke. A total of 156 fall events occurred, and each fall case was randomly matched with six control cases. Six ML algorithms were used, namely, regularized logistic regression, support vector machine, naïve Bayes (NB), k-nearest neighbors, random forest, and extreme-gradient boosting (XGB).

RESULTS

We included 156 in the fall group and 934 in the non-fall group. The mean ages of the fall and non-fall groups were 68.3 (± 12.2) and 65.3 (± 12.9) years old, respectively. The MFS total score was significantly higher in the fall group (54.3 ± 18.3) than in the non-fall group (37.7 ± 14.7). The area under the receiver operating curve (AUROC) of the MFS in predicting falls was 0.76 (0.73-0.79). XGB had the highest AUROC of 0.85 (0.78-0.92), and XGB and NB had the highest F1 score of 0.44.

CONCLUSIONS

The AUROC values of all of ML algorithms were similar to those of the MFS in predicting fall risk in patients with acute stroke, allowing for accurate and efficient fall screening.

Backward Walking Styles and Impact on Spatiotemporal Gait Characteristics

Forward walking (FW) is a common balance assessment tool. However, its sensitivity is limited by the ceiling effect. Reverse gait, such as backward walking (BW), has been reported to have more advantages than FW for balance assessment. Three factors related to postural instability (i.e., increased speeds, restricted arm swing, and reduced visual feedback) during BW were investigated to determine BW conditions that have the potential to predict falls. Three-dimensional analyses were used to analyze seven walking conditions. FW and BW at self-selected and fast speeds were analyzed to identify the effects of speed. Walking with normal arm swings, crossed arms, and abducted arms during BW was tested to determine the effects of arm position. BW with closed and open eyes was compared to investigate the effects of visual feedback. BW had a significantly shorter step length than FW at high speeds. When the arms were abducted, the stance phase (%) was significantly lower compared to when arms were crossed during BW. Moreover, BW with closed eyes revealed significantly higher mediolateral center of mass (COM) displacements than with open eyes. We observed that BW with fast speeds, a crossed arm position, and closed eyes has the potential to help assess fall risk because it requires higher balance ability through spatiotemporal and COM adjustment.

Impact of Fall Prevention on Nurses and Care of Fall Risk Patients

Purpose of the Study

Falls are common events for hospitalized older adults, resulting in negative outcomes both for patients and hospitals. The Center for Medicare and Medicaid (CMS) has placed pressure on hospital administrators by identifying falls as a "never event", resulting in a zero falls goal for many hospitals. Staff nurses are responsible for providing direct care to patients and for meeting the hospital no falls goal. Little is known about the impact of "zero falls" on nurses, patients and the organization.

Design and Methods

A qualitative study, using Grounded Dimensional Analysis (GDA) was conducted to explore nurses' experiences with fall prevention in hospital settings and the impact of those experiences on how nurses provide care to fall risk patients. Twenty-seven registered nurses and certified nursing assistants participated in in-depth interviews. Open, axial and selective coding was used to analyze data. A conceptual model which illustrates the impact of intense messaging from nursing administration to prevent patient falls on nurses, actions nurses take to address the message and the consequences to nurses, older adult patients and to the organization was developed.

Results

Intense messaging from hospital administration to achieve zero falls resulted in nurses developing a fear of falls, protecting self and unit, and restricting fall risk patients as a way to stop messages and meet the hospital goal.

Implications

Results of this study identify unintended consequences of fall prevention message on nurses and older adult patients. Further research is needed understand how nurse care for fall risk patients.

Activity Monitoring and Heart Rate Variability as Indicators of Fall Risk: Proof-of-Concept for Application of Wearable Sensors in the Acute Care Setting

Growing concern for falls in acute care settings could be addressed with objective evaluation of fall risk. The current proof-of-concept study evaluated the feasibility of using a chest-worn sensor during hospitalization to determine fall risk. Physical activity and heart rate variability (HRV) of 31 volunteers admitted to a 29-bed adult inpatient unit were recorded using a single chest-worn sensor. Sensor data during the first 24-hour recording were analyzed. Participants were stratified using the Hendrich II fall risk assessment into high and low fall risk groups. Univariate analysis revealed age, daytime activity, nighttime side lying posture, and HRV were significantly different between groups. Results suggest feasibility of wearable technology to consciously monitor physical activity, sleep postures, and HRV as potential markers of fall risk in the acute care setting. Further study is warranted to confirm the results and examine the efficacy of the proposed wearable technology to manage falls in hospitals. [Journal of Gerontological Nursing, 43(7), 53-62.].

Risk assessment of fall-related occupational accidents in the workplace

Objectives:

This study aimed to examine effective assessment methods of falls in the workplace.

Methods:

There were 436 employees (305 males and 131 females) of electrical appliance manufacturers included in this study. In 2014, a baseline survey was conducted using the fall scores questionnaire and the self-check risk assessment of falls and other accidents in the workplace (physical function measurement and questionnaire). In 2015, the occurrence of falls in the past year was investigated. Multivariate logistic regression analyses were performed to examine factors relevant to falls.

Results:

In total, 62 subjects (14.2%) fell during the year, including those who fell during off-hours. The occurrence of falls during that one year was only associated with having experienced falls during the past year in the baseline survey (odds ratio [OR] 5.0; 95% confidence interval [CI] 2.5-9.7). Falls during that year were also related to the inability to walk 1 km continuously (OR 0.1; 95% CI 0.1-0.6), tripping sometimes (OR 4.0; 95% CI 1.6-9.9), step height differences at home (OR 3.0; 95% CI 1.3-6.8), and working in the production section (OR 0.2; 95% CI 0.1-0.5). Measurements of physical functions, such as muscle strength, balance, and agility, were not different between subjects who fell and those who did not.

Conclusions:

Our results showed that the questionnaire assessing falls during the past year could be useful to assess the risk of falls in the workplace. Annual checks for falls may contribute to fall prevention programs in the workplace.

Predicting falls: considerations for screening tool selection vs. screening tool development

AIMS

This paper discusses considerations for falls risk screening tool selection vs. the need to develop new tools.

BACKGROUND

Inpatient falls are a complex patient safety issue that represent a significant burden for the healthcare system. In the inpatient context, falls risk screening tools are most often used for predicting falls, but in some populations assessment tools are more suited, however in others, a clinician's clinical judgment may be just as effective. Limited external validity is a central issue with falls risk screening tools when used in different populations than the original study. There is clinical need for guidance regarding screening tool selection vs. the need to development new tools and how to effect change in relation to the prediction of falls.

DESIGN

Discussion paper.

DATA SOURCES

This discussion paper is based on our own experiences and research and is supported by literature.

IMPLICATIONS FOR NURSING

This paper provides clinicians with a better understanding of considerations for falls risk screening tool selection vs. the need to develop new tools. In doing so, it provides clinicians guidance on how to critique the efficacy and utility of their falls risk screening tool. This paper equips clinicians for effecting change in relation to the prediction of falls.

CONCLUSION

Falls risk prediction is a particularly complex patient safety issue. Clinicians need to be aware of the limitations of their tool used to predict falls.

Instruments for assessing the risk of falls in acute hospitalized patients: a systematic review and meta-analysis

Background

Falls are a serious problem for hospitalized patients, reducing the duration and quality of life. It is estimated that over 84% of all adverse events in hospitalized patients are related to falls. Some fall risk assessment tools have been developed and tested in environments other than those for which they were developed with serious validity discrepancies. The aim of this review is to determine the accuracy of instruments for detecting fall risk and predicting falls in acute hospitalized patients.

Methods

Systematic review and meta-analysis. Main databases, related websites and grey literature were searched. Two blinded reviewers evaluated title and abstracts of the selected articles and, if they met inclusion criteria, methodological quality was assessed in a new blinded process. Meta-analyses of diagnostic ORs (DOR) and likelihood (LH) coefficients were performed with the random effects method. Forest plots were calculated for sensitivity and specificity, DOR and LH. Additionally, summary ROC (SROC) curves were calculated for every analysis.

Results

Fourteen studies were selected for the review. The meta-analysis was performed with the Morse (MFS), STRATIFY and Hendrich II Fall Risk Model scales. The STRATIFY tool provided greater diagnostic validity, with a DOR value of 7.64 (4.86 - 12.00). A meta-regression was performed to assess the effect of average patient age over 65 years and the performance or otherwise of risk reassessments during the patient’s stay. The reassessment showed a significant reduction in the DOR on the MFS (rDOR 0.75, 95% CI: 0.64 - 0.89, p = 0.017).

Conclusions

The STRATIFY scale was found to be the best tool for assessing the risk of falls by hospitalized acutely-ill adults. However, the behaviour of these instruments varies considerably depending on the population and the environment, and so their operation should be tested prior to implementation. Further studies are needed to investigate the effect of the reassessment of these instruments with respect to hospitalized adult patients, and to consider the real compliance by healthcare personnel with procedures related to patient safety, and in particular concerning the prevention of falls.

Person-environment interactions contributing to nursing home resident falls

Although approximately 50% of nursing home residents fall annually, the surrounding circumstances remain inadequately understood. This study explored nursing staff perspectives of person, environment, and interactive circumstances surrounding nursing home falls. Focus groups were conducted at two nursing homes in the mid-Atlantic region with the highest and lowest fall rates among corporate facilities. Two focus groups were conducted per facility: one with licensed nurses and one with geriatric nursing assistants. Thematic and content analysis revealed three themes and 11 categories. Three categories under the Person theme were Change in Residents' Health Status, Decline in Residents' Abilities, and Residents' Behaviors and Personality Characteristics. There were five Nursing Home Environment categories: Design Safety, Limited Space, Obstacles, Equipment Misuse and Malfunction, and Staff and Organization of Care. Three Interactions Leading to Falls categories were identified: Reasons for Falls, Time of Falls, and High-Risk Activities. Findings highlight interactions between person and environment factors as significant contributors to resident falls.

Instruments for measuring fall risk in older adults living in long-term care facilities: an integrative review

Lack of consistency in the literature regarding the use of fall risk assessment in long-term care settings and the uniqueness of the environment demand a critical analysis of fall risk instruments specific to older adults living in long-term care facilities. This integrative review of the existing literature on instruments used to measure fall risk in residents living in long-term care facilities revealed a total of 16 fall risk assessment tools from 13 studies. Of the 13 studies, only 8 reported sensitivity, specificity, and interrater reliability values of the tools. Only the Morse Fall Scale and Mobility Fall Chart demonstrated high predictive values in their initial developmental studies. This review can help clinicians make informed choices about tools to predict falls in their settings and establish appropriate preventive interventions.

Fall incidence and fall prevention practices at acute care hospitals in Singapore: a retrospective audit

OBJECTIVE

To investigate the incidence of falls and explore fall prevention practices at acute care hospitals in Singapore.

DESIGN

A retrospective audit to collect baseline data on (1) incidence of falls (patient fall rates and fall injury rates) and (2) fall prevention practices, was conducted in five acute care hospitals in Singapore from December 2004 to March 2005.

STUDY PARTICIPANTS

Medical record data (n = 6000) of patients admitted into the medical, surgical and geriatric units in the five hospitals.

OUTCOME MEASURES

Fall incidence was obtained from the hospital's fall databases and incident reports for the period of June 2003 to May 2004. In total, 6000 medical records from five hospitals were randomly selected, retrieved and reviewed to determine whether falls, fall assessments and interventions were being initiated and documented.

RESULTS

The number of fallers for all hospitals was 825. Analysis showed that patient fall rates ranged from 0.68 to 1.44 per 1000 patient days, and the proportion of falls associated with injury ranged from 27.4% to 71.7%. The use of a fall risk assessment tool by nurses was recorded in 77% of all the nursing records.

CONCLUSION

This study has laid the foundation for further research for fall prevention in Singapore by describing current fall rates, fall-associated injury rates and the status of fall prevention practices in acute care settings. The results will be used to inform the development of a tailored multifaceted strategy to facilitate the implementation of Fall Prevention Clinical Practice Guidelines to reduce the burden of falls and fall injuries in hospitals in Singapore.

Fall prediction in inpatients by bedside nurses using the St. Thomas's Risk Assessment Tool in Falling Elderly Inpatients (STRATIFY) instrument: a multicenter study

OBJECTIVES

To assess the predictive value of the St. Thomas's Risk Assessment Tool in Falling Elderly Inpatients (STRATIFY) instrument, a simple fall-risk assessment tool, when administered at a patient's hospital bedside by nurses.

DESIGN

Prospective multicenter study.

SETTING

Six Belgian hospitals.

PARTICIPANTS

A total of 2,568 patients (mean age+/-standard deviation 67.2+/-18.4; 55.3% female) on four surgical (n=875, 34.1%), eight geriatric (n=687, 26.8%), and four general medical wards (n=1,006, 39.2%) were included in this study upon hospital admission. All patients were hospitalized for at least 48 hours.

MEASUREMENTS

Nurses completed the STRATIFY within 24 hours after admission of the patient. Falls were documented on a standardized incident report form.

RESULTS

The number of fallers was 136 (5.3%), accounting for 190 falls and an overall rate of 7.3 falls per 1,000 patient days for all hospitals. The STRATIFY showed good sensitivity (> or = 84%) and high negative predictive value (> or = 99%) for the total sample, for patients admitted to general medical and surgical wards, and for patients younger than 75, although it showed moderate (69%) to low (52%) sensitivity and high false-negative rates (31-48%) for patients admitted to geriatric wards and for patients aged 75 and older.

CONCLUSION

Although the STRATIFY satisfactorily predicted the fall risk of patients admitted to general medical and surgical wards and patients younger than 75, it failed to predict the fall risk of patients admitted to geriatric wards and patients aged 75 and older (particularly those aged 75-84).

An Evidence-based Approach to Fall Risk Assessment, Prevention, and Management

Nurses at an academic medical institution undertook a fall safety initiative. Using an evidence-based approach, they created a risk stratification tool, developed a comprehensive protocol, investigated fall-prevention products and technologies, and piloted the protocol and products/technologies before the full implementation. This article describes their journey and lessons learned along the way, the most compelling of which is the need for a simple, guided, and time-efficient approach to implementing the best practices into clinical care.

Hospital fall risk assessment tools: a critique of the literature

There are numerous fall risk assessment tools developed for use by nurses in hospitals; however, few of them have established validity and reliability. This article will examine the current state of knowledge in regard to fall risk assessment tools through review and critique of the literature on the topic. Recommendations for future research on fall risk assessment tools will be made and a conceptual framework detailing the relationship between the variables involved in assessing the accuracy of fall risk assessment tools will be presented.

Prospective study of fall risk assessment among institutionalized elderly in Japan

The purpose of the present study was to identify risk factors for falls among institutionalized elderly, using the standardized risk assessment tool developed by Izumi. We examined 746 patients from three types of facilities: rehabilitation wards in four general hospitals, three long-term care facilities, and three nursing homes, for up to three months. The incidence of falls within all facilities was 12.5%. Patterns of relative risks of falling differed among types of facilities. The highest relative risk of fall in long-term care facilities and nursing homes was nurses' prediction, followed by history of fall and altered mentation. In contrast, that in general hospitals was mobility. In long-term care facilities, history of falls (odds 3.68, 95CI: 1.47-9.23) and interaction (history of falls and assistance with toileting) (odds 3.13, 95CI: 1.48-6.64) showed significance on adjusted-odds ratios for fall. History of falls, altered mentation, and assistance with toileting may be used to screen patients at a high risk for fall at admission.

Breaking the fall factor

Successful fall prevention entails identifying at-risk patients and implementing measures that impede or minimize patient injury. Here, learn practical start-up steps to devise a program in your department.

Fall risk assessment measures: an analytic review

BACKGROUND

Clinicians are often unaware of the many existing scales for identifying fall risk and are uncertain about how to select an appropriate one. Our purpose was to summarize existing fall risk assessment scales to enable more informed choices regarding their use.

METHODS

After a systematic literature search, 21 articles published from 1984 through 2000 describing 20 fall risk assessments were reviewed independently for content and validation by a panel of five reviewers using a standardized review form. Fourteen were institution-focused nursing assessment scales, and six were functional assessment scales.

RESULTS

The majority of the scales were developed for elderly populations, mainly in hospital or nursing home settings. The patient characteristics assessed were quite similar across the nursing assessment forms. The time to complete the form varied from less than 1 minute to 80 minutes. For those scales with reported diagnostic accuracy, sensitivity varied from 43% to 100% (median = 80%), and specificity varied from 38% to 96% (median = 75%). Several scales with superior diagnostic characteristics were identified.

CONCLUSIONS

A substantial number of fall risk assessment tools are readily available and assess similar patient characteristics. Although their diagnostic accuracy and overall usefulness showed wide variability, there are several scales that can be used with confidence as part of an effective falls prevention program. Consequently, there should be little need for facilities to develop their own scales. To continue to develop fall risk assessments unique to individual facilities may be counterproductive because scores will not be comparable across facilities.

Mechanical restraints, rehabilitation therapies, and staffing adequacy as risk factors for falls in an elderly hospitalized population

This case-control study of fall-related risk factors for elderly hospitalized patients hypothesized that the use of mechanical restraints, participation in a rehabilitation program, and staffing inadequacy increase the risk of falls. The study included 252 patients, also known as "cases," aged 60 to 85 years, who fell during the period between March 1 and December 31, 1993, in a large metropolitan hospital; and 250 "controls," randomly selected patients aged 60 to 85 years, who were matched with the cases in terms of length of stay to the day of the fall. Analysis using conditional logistic regression revealed that individuals who had been placed in a mechanical restraint at any point during their hospital stay prior to the fall (for the cases) or the selected day (for the controls) had approximately twice the risk of falling as did patients who had not been placed in restraints. Cases and controls showed no significant differences in terms of their participation in occupational therapy, physical therapy, or cardiac rehabilitation, and staffing adequacy was similar for both groups. While the finding of 2 to 4.7 times the risk for falls for those placed in a mechanical restraint only approached statistical significance, the results indicate that the risk of falling is highest soon after a patient has had to be placed in a mechanical restraint.

Clinical nursing judgment related to reducing the incidence of falls by elderly patients

The incidence of falls among elderly patients has been and continues to be a major challenge for nurses. Falls add physical injury and mental stress to patients' existing health problems, are a deterrent to rehabilitation, and increase healthcare costs. This study describes the variables that nurses identify as influencing their clinical decision making and the nursing behaviors associated with preventing patient falls. The study was grounded in the theory that discretionary nursing behaviors are related to nursing expertise, and the study was guided by the assumption that such behaviors are proactive and anticipatory. An analysis of interviews of registered nurses (n = 14) working on a geriatric rehabilitation unit in a medical center in Ohio focused on the zones of association and the contextual meanings of language used by the nurses when discussing patient falls. Four themes emerged: the reasons for patient falls, identifying patients who are likely to fall, preventing falls, and nurses' feelings when patients fall.

Reducing falls in a rehabilitation setting: a safer environment through team effort

A 2-year falls prevention project has included a concurrent analysis of all falls and a program to prevent falls in a 160-bed rehabilitation hospital. The program has been designed on the basis of the data collected. A consistent follow-up program involves patients and staff in the continuous monitoring of patient safety.