Yearlong Evaluation of Fall Risk Determinants Among 40 Older Adults in Two Residential Assisted Living Facilities

Introduction

Falls and their consequences on the elderly are a huge social and economic problem [1,2]. Falls are the second most common cause of death from unintentional injury worldwide [3]. Each year there are 37.3 million falls that are serious enough to require medical attention [3]. In addition, an estimated 684 000 people worldwide die each year from falls, with more than 80% of them in low- and middle-income countries [4,5]. According to studies, about 30% of people aged over 65 years and 40% to 50% of people over 80 years will fall at least once a year [4–6]. An even higher risk of falling, as high as 50% to 67%, applies to residents living in long-term nursing facilities [7]. The incidence of falls in nursing homes is twice as high as that in community-dwelling older adults [8]. The probability of falling increases with age and the number of risk factors [9].

Researchers list many risk factors for falls, including a history of falls, low physical activity levels, low gait speed, low muscle strength, balance and gait problems, cognitive impairment and dementia, depressive disorders, post-fall syndrome, comorbidities, and polypragmasy [9,10].

Such a large number of factors associated with falls raises the need for multiple tests to assess not only physical but also cognitive-emotional functioning in older people [11].

The most commonly used tests to assess physical performance in older adults are the Short Physical Performance Battery test, 10-meter walk test (10MWT), 2-minute step test (2MST), lower extremity strength, and hand grip strength [12]. The most commonly used tests to assess cognitive status are the Mini Mental State Examination (MMSE) and Montreal Cognitive Assessment (MoCA); however, the MoCA is better than the MMSE in detecting mild cognitive impairment in patients over 60 years of age [13]. Also, the Hospital Anxiety and Depression Scale (HADS), which measures symptoms of anxiety and depression, is recommended to assess psychological distress among a general population 65 to 80 years old [14].

Fall risk factors can be categorized as modifiable and non-modifiable, encompassing external and internal, biological, behavioral, environmental, and socioeconomic aspects. When a single risk factor is present, the likelihood of falling is 18%, whereas the presence of 4 or more factors increases this risk to 78% [10,15].

Just as there are many risk factors for falls, there are also many clinical screening tests that assess the risk of falls [16]. Screening tools for identifying elderly people at risk of falling are the Timed Up and Go (TUG) test and Tinetti Performance Oriented Mobility Assessment (POMA) [16]. It is worthwhile to use the Falls Efficacy Scale (FES) at the same time, which measures fear of falling and is suitable for use in research and clinical practice [17].

However, researchers agree that none of these tests separately can predict future falls in older people and correctly identify those who fall [16]. Moreover, there are no clear guidelines on how and with which tests to assess risk and prevent falls [5].

To date, there have been a small number of studies conducted among populations of older adults living in assisted living facilities (ALFs). Therefore, this study aimed to evaluate the relationship between physical activity, physical performance, falls, and cognitive functioning at 1 year in 40 older adults living in 2 residential ALFs in 2 communities in Wisconsin, USA.

Material and Methods

ETHICAL STATEMENT:

The research was done under the ethical and legal supervision of the ethics committee (reference No. 35/2018) and accordance with USA/State of Wisconsin Physical Therapy standards of practice. The study was conducted following the Helsinki Declaration. All participants were informed about the aim and methods of the study and the procedures used. All participants signed a document of voluntary and informed consent.

STUDY DESIGN:

This observational study was conducted in 2 communities of ALFs: Rennes Marinette and Rennes De Pere in Wisconsin, USA, with the Rennes organization in 2019–2020. The patients’ mental state was examined by a therapist licensed to conduct the MoCA test; functional tests and physical assessment were conducted by the same physiotherapist and with the doctor’s approval.

The management of the ALFs consented to the research.

PARTICIPANTS:

The study group consisted of ALF residents who fulfilled the following inclusion criteria: age ≥ 65 years, written informed consent to participate in the study, satisfactory cognitive status (result of MoCA >17), ALF residency status for minimum of 1 year, and medical and functional status allowing participation in physical tests and assessments. The exclusion criteria were aphasia, refusal to participate at any stage of the study, and presence of serious mental disorders or physical disabilities preventing testing procedures.

The data used for statistical analysis was only from participants who were involved in the initial assessment and the 2 additional assessments, after a 6-month and 12-month period.

Of 132 ALF residents, 40 participated in the initial study (T1), including 25 women and 15 men, with a mean age of 86.6 (±6.3) years. After 6 months (T2) and after 12 months (T3), a total number of 27 out of 40 (initial) residents participated (Table 1, Figure 1).

METHODS:

The psycho-cognitive tests and questionnaires used were the MoCA to assess cognitive state, HADS to assess depression and anxiety, and FES to assess fear of falling. Physical performance tests included the 10MWT, 2MST, and lower extremity strength and hand grip strength tests using a dynamometer. Postural stability and balance assessment included posturography, using a portable Wii platform, the TUG test, and POMA. Additionally, sociodemographic, clinical, and fall-related data of the past 6 months were collected based on interview and medical documentation review.

MONTREAL COGNITIVE ASSESSMENT: The MoCA is a screening tool to assess cognitive status and help determine level of impairment from mild cognitive impairment to dementia. The MoCA was designed to test visuospatial and executive function, naming, short term memory, attention, language, abstraction, delayed recall, and orientation. The MoCA scale ranges from 0 to 30 points. The following interpretation of the results was adopted for this study: 26–30, normal status; 18–25, mild cognitive impairment; 11–17, mild dementia; 6–10, moderate dementia; and 0–5, severe dementia [18].

HOSPITAL ANXIETY AND DEPRESSION SCALE: The HADS is a self-assessment scale detecting depression and anxiety in the setting of a medical institution. Residents completed a questionnaire composed of 14 statements relevant to either generalized anxiety or depression. Each item is answered by choosing the most appropriate option reflecting the participants’ current state of emotions on a 4–point (0–3) scale. The scores range from 0 to 21 for anxiety and 0 to 21 for depression. The score of 0 to 7 for either subscale is the normal range, a score of 8 to 10 is borderline abnormal, and a score of 11 or higher indicates the presence of anxiety or depression [19].

FALLS EFFICACY SCALE: The FES is a measure of fear of falling or concerns about falling during the performance of activities of daily living, including taking a bath or shower, reaching into cabinets or closets, walking around the house, preparing meals not requiring carrying heavy or hot objects, getting in and out of bed, answering the door or telephone, getting in and out of a chair, getting dressed and undressed, personal grooming (eg, washing the face), and getting on and off the toilet. Each activity is associated with a scale of 1 to 10, with 1 representing no fear/concern of falling, and 10 representing serious fear/concern of falling. A total score greater than 70 indicates that the person experiences a fear of falling [20].

2-MINUTE STEP TEST: The 2MST is a test for aerobic endurance, an assessment component of the Rikli & Jones Test (Senior Fitness Test). This test is performed as an alternative to the 6-minute walk test, for older people. The participant marches in place for 2 min, lifting the knees to the height of the mark on the wall. The total number of times (n) the right knee reaches the tape level in 2 min is recorded. The final score is compared with the average normative number of repetitions established for each age and sex group [21].

10-METER WALK TEST: The 10MWT is a performance measure used to assess walking speed in meters per second over a 10-meter distance. The test includes trials of walking at normal walking speed and fast-paced walking speed, tested 2 times for each speed. The mean average of the 2 walks for both trials were used in this analysis. A cut-off value of <0.7 m/s, indicating increased risk of falling for the population in this study, was adopted [22].

TIMED UP AND GO TEST: The TUG is a test of dynamic balance, which is commonly used to examine risk of falling and functional mobility. The TUG test measures time of getting up from the chair, walking distance of 3 m marked by a cone, turning around the cone (180 degrees), returning to the chair, and sitting down again. The test was performed twice, and the better of the 2 results was considered the final score. A result ≤13.5 s, indicating high risk of falling, was adopted [23].

PERFORMANCE ORIENTED MOBILITY ASSESSMENT: The POMA (Tinetti test) measures an older adult’s balance and gait abilities and identifies fall risk. The examination is subdivided into 2 assessments – balance and gait – and the total score is 28. A score less than or equal to 18 indicates a high risk for fall incidents. A score between 19 and 23 indicates a moderate fall risk, and a score greater than or equal to 24 indicates a low fall risk [24,25].

MUSCLE STRENGTH, LOWER EXTREMITY: Muscle strength of the lower extremity was assessed using a hydraulic dynamometer (Baseline Evaluation Instruments, USA). The device recorded force in kilograms and was calibrated once, at the start of the study. Hip flexors were tested in the sitting position with back support reclined at 15 degrees, with the hip joint flexed to 75 degrees, and knee joint flexed to 90 degrees. The knee flexors and extensors were tested in the sitting position without back support and with the hip joint flexed to 90 degrees, and the knee joint flexed to mid-point between maximal knee flexion and extension at approximately 60 degrees of flexion. The test was performed twice for each muscle group, alternating the left and right lower extremity with 2 min rest time between tests for the same leg knee flexors and extensors. The result of the test was calculated as the average of the 2 tests for each muscle group [26].

HAND GRIP STRENGTH: Hand grip strength was measured using a handheld Jamar hydraulic dynamometer, with the participant in a seated position with the arm rested at 90 degrees. The device recorded force in kilograms and was calibrated once, at the start of the study. The patients squeezed the dynamometer for 5 to 6 s, generating maximal isometric contraction. The procedure was performed twice, with 1 min rest. The average grip strength of the dominant hand was used in this analysis. Although hand strength does not play a direct role in maintaining balance in the standing position, many studies have shown correlation with the decline of generalized muscle strength and increased risk of falling in older adults [27,28].

ROMBERG TEST: The Romberg test was performed on a Wii posturographic platform. It consisted of two 40-s trials, with participants standing on a hard surface with the eyes open and standing on a hard surface with the eyes closed. The result of the test was presented in the form of a graph of the center of gravity deviation and the designated area of the center of gravity deviation at 40 s. In case of complete loss of balance requiring additional support of assistive device or opening eyes during trials with eyes closed (positive Romberg), the posturography assessment was considered invalid, and data from the center of pressure sway area was not used to calculate outcomes. The use of the Wii platform together with the “Romberg Lab” software, both in the form of freeware and commercial applications, has been validated as a statistically reliable testing method, comparable to commonly used, established diagnostic platforms [29,30].

NUMBER OF FALLS:

Participating residents and care-giving staff were educated in the definition of a fall and then they were asked to report all qualifying incidents to facility administration. All reported data were used to calculate frequency of falls in 6 month increments starting 6 months prior to initial assessment and throughout the study.

The above studies were performed at 3 measurement points over 1 year: an initial survey (T1), survey after 6 months (T2), and survey after 12 months (T3). All residents underwent the same battery of tests and surveys. The measurements were conducted by the same researcher.

DATA ANALYSIS:

Descriptive statistics were calculated. Depending on the result of testing the normality of the distribution of quantitative variables (Shapiro-Wilk test), the mean or median was used as a measure of central tendency, and the standard deviation (SD) or interquartile range (IQR) was used as a measure of dispersion. For qualitative variables, counts and percentages of total participants were calculated. Due to the non-normality of the distribution of most of the analyzed characteristics, the significance of differences between repeated measures (T1 vs T2 vs T3) was checked using the Friedman ANOVA and, when necessary, the Dunn Bonferroni Holm post hoc test. Effect size for the Friedman ANOVA was checked with the Kendal W coefficient, whose values range from 0 (no agreement) to 1 (full agreement). Spearman rank correlation was used to assess the relationship between number of falls, POMA, FES, age, BMI, MOCA, HADS, 2MST, 10MWT, TUG, lower extremity strength, hand grip strength, and results of the Wii platform.

The logistic regression model and odds ratio, in which fall risk was a dependent variable, were determined. Calculations were performed using Statistica 13.1 and PQ Stat 1.8.2. P<0.05 was considered as the level of significance.

Results

CHARACTERISTICS OF PSYCHOPHYSICAL PARAMETERS IN T1, T2, AND T3:

Forty ALF residents were eligible for the preliminary study. Sixteen participants (40%) had at least 1 fall in the past 6 months, and 24 participants used a cane or walker for locomotion.

Evaluation of gait and balance performed with the POMA showed a medium risk of falls in 21 participants (45% of the participants), and a high risk in 1 participant (2.5%). The results of measurements and observations of selected psychophysical characteristics are shown in Table 2.

The study was repeated twice, at 6-month intervals. Twenty-seven of the residents participated in all 3 measurements. There was no statistically significant difference between repeated measurements on most of the variables analyzed. The exception was normal gait speed, which was significantly decreased in T3, compared with measurements at T1 and T2. Also in T3, a significantly higher HADS test score was obtained on the depression subscale, indicating an increased risk of depression, compared with the initial measurement (T1). ANOVA also showed a result that was significant in the section on the number of falls. However, due to the small group size, post hoc tests did not confirm differences between measurements (Table 2).

CORRELATION ANALYSIS BETWEEN SELECTED PSYCHOPHYSICAL PARAMETERS:

A correlation analysis of the number of observed falls and tests assessing the risk of fall with measurements of selected psychophysical characteristics was performed. A significant moderate, negative correlation was found between the number of falls and the 2MST test. The summed POMA test score, indicating the risk of falls, showed a statistically significant high positive correlation with gait speed and a negative correlation with the TUG test score. The FES test showed a statistically significant high negative correlation with the 2MST test (Table 3).

Restricting the number of observations only to patients with at least 1 fall in the last 6 months showed a statistically significant high and negative association of the POMA test score with the size of the area delineated for center of pressure oscillations in the open-eye and closed-eye balance tests and with the HADS test (depression section). The FES test, on the other hand, showed a significant positive relationship with the HADS test for both depression and anxiety (Table 4).

RESULTS OF LOGISTIC REGRESSION MODEL:

In addition, the effect of variables on the magnitude of fall risk estimated by the odds ratio was analyzed. Due to the inability to determine a multivariate logistic regression model, resulting from an insufficiently large group of participants, single models were performed. A statistically significant model was obtained only for the number of repetitions performed during the 2MST (Table 5).

Discussion

The aim of this study was to evaluate the relationship between physical activity, physical performance, falls, and cognitive functioning at 1 year in 40 older adults living in 2 residential ALFs in 2 communities in Wisconsin, USA. Our study showed that among ALF residents, a higher risk of falling was associated with decreased gait speed, impaired balance, decreased mood, increased fear of falling, and fewer repetitions performed in the 2MST.

Falls, a major geriatric concern, require a focus on prevention and early identification of risk factors, which are essential in mitigating their often irreversible consequences. An analysis of the literature indicates that there are more than 400 factors that can cause a fall alone or in cooperation with other factors [31]. Analyzing modifiable factors linked to fall risk in elderly residents of nursing facilities is vital for designing effective fall prevention strategies [32]. Residents residing in ALFs are individuals who do not need constant, round-the-clock care or supervision, but rather require assistance and support in complex activities of daily living, which have some risk due to the complexity of their performance. The premise of the present project was to observe and analyze as many factors as possible regarding the emotional-cognitive, somatosensory, and physical spheres of elderly people residing in ALFs.

Analysis of the study, conducted at 3 measurement points, showed a significant deterioration in normal gait speed after 12 months of follow-up, as well as weakness in the knee extensor muscles. In addition, a significant reduction in mood was noted in the study group of residents. These results are not surprising, probably related to age, and related to limited ability to function in complex activities of daily living and falls. A period of 12 months (with an initial examination and follow-up examinations after 6 and 12 months) was not very long for there to have been so much change in the residents’ functional area. Nevertheless, it should be remembered that the observation period was disrupted by the outbreak of the COVID-19 pandemic, which probably had a significant effect on the aforementioned parameters. The need to adapt to numerous restrictions, including free movement and communication, may have caused negative changes not only in the physical state of the participants but also in their emotional state. Disregarding such exceptional circumstances, there are studies in the literature that also reported significant reductions in gait speed measured before a fall [33]. Reduction in gait speed can therefore be an important risk factor for falls among the elderly.

This was also confirmed by correlation analyses. In the entire study group of ALF residents, lower gait speed (at both normal and maximum speed) was associated with an increased risk of falling and poorer balance, as assessed by the POMA test. Similar results were obtained in studies by Park et al, Rivolta et al, Dokuzlar et al, and Chan et al, confirming the strong correlation of gait speed with POMA scores [9,34–36]. There are also some reports that do not emphasize this relationship. For example, the results of Pham et al show that the association between gait speed and risk of falls was weaker than the association between grip strength and risk of falls [27]. Nevertheless, researchers emphasize that gait speed is a key indicator of health – the “sixth vital sign” – and recommend measuring gait speed as a clear indicator to identify the risk of falls [1,5].

A similar correlation analysis conducted among ALF residents who had experienced at least 1 fall in the past 6 months revealed additional factors significantly associated with fall risk, beyond gait speed and POMA scores. These included balance (assessed using the Wii platform), depressive disorders, and fear of falling. These findings underscore the profound impact of falls on elderly individuals and the far-reaching consequences they entail.

Anxiety about falling, or post-fall syndrome, affects 29% to 65% of older people who have experienced a fall but can also occur in people who have never experienced a fall [37,38]. In 40% of elderly people, anxiety about falling leads to secondary impairment of their physical activity, including limitations in activities of daily living [39]. Hallford et al showed that elevated levels of anxiety were associated with a 53% increase in the likelihood of falling [40].

Furthermore, the present study, conducted with individuals who had experienced a fall, revealed significantly lower gait speed in those with a high fear of falling. This finding confirms that fear of falling directly impacts activity levels and participation in daily activities, leading to reduced physical fitness, subsequent gait disturbances, increased fall risk, and loss of independence. These outcomes can also contribute to lower mood and the development of depressive disorders, which are often accompanied by a strong anxiety component [41,42]. On the other hand, the depressive symptoms that occur can lead elderly people to fall due to physiological and cognitive disorders, especially lack of attention [8]. This is a multidirectional relationship that requires better identification of the underlying mechanisms and the use of effective non-pharmacological interventions for primary and secondary prevention [9,43]. Studies have reported an increased likelihood of falling in individuals with depression [7,44]. Also, Gambaro et al, in conducting a systematic review and meta-analysis, showed that depressive symptoms caused a later increase in the number of falls. In addition, they stressed that all the results analyzed supported the hypothesis of an interdependent relationship between the presence of post-fall anxiety, depression, and fall risk, although the high percentage of cross-sectional studies makes it impossible to infer the direction of the relationship [45].

In the present study, the balance test performed on the posturography platform with the eyes open and eyes closed also correlated strongly with fall risk. The worse the test result, in other words, the larger the area of deviation of the center of gravity, and the higher was the risk of falling in the group of residents who had experienced a fall in the past 6 months. Studies indicate that balance assessment based on static posturography is able to show significant differences between elderly fallers and non-fallers and to assess the likelihood of future falls [46,47]. Unfortunately, there are few scientific reports on the use of the posturography platform in care centers. This is probably due to the lack of access in these centers to laboratory equipment, which is most often used only in specialized clinics. This could be an argument for increasing the use of portable, easily accessible posturography platforms, such as the Wii, in such care centers for the elderly.

Our study results presented here indicate that within the group of individuals who experienced falls, a higher number of falls in the past 6 months was associated with better cognitive status. This result may come as a surprise, especially since most studies indicate a strong association and higher risk of falls among patients with cognitive impairment and dementia [48,49]. Nonetheless, this result reflects the greater activity levels of ALF residents with better cognitive function, who tend to be more engaged in daily life. Their motivation to be self-reliant and independent can at the same time be a situation provoking the occurrence of falls during these activities. Additionally, it is important to note that only individuals who met the inclusion criteria for cognitive performance were included in the study (MoCA >17). Arguably, a more heterogeneous study group in terms of cognitive status and the presence of dementia symptoms would have highlighted this relationship, which has already been confirmed by research.

Another significant factor identified in the overall group was the result of the 2MST, which, along with the HADS score, was the only factor that correlated with fear of falling measured with the FES. The greater the participants’ fear of falling, the fewer repetitions were obtained in the 2MST. This limited the number of repetitions performed in the test and affected as many as 60% of the participants.

In addition, the 2MST was the only factor that correlated with the number of recorded falls in the last 6 months. The lower the number of repetitions on the 2MST, the higher the number of falls in the entire study group of ALF residents. It is important to note that elderly individuals, often diagnosed with osteoarthritis of the lower limbs and spine, can adopt various compensatory biomechanical strategies when walking, such as when walking in place. These alterations in walking patterns can lead to postural instability, difficulty in shifting the center of gravity, and an increased risk of falls [9,50]. This factor was found to be significant among all the factors studied, also in the single regression models that analyzed the impact of these factors on fall risk, which revealed that for each additional repetition in the 2MST, the likelihood of falling decreased by 0.94 times.

Bootsman et al, in an Australian study of 44 residents from 4 residential aged care facilities, found that residents with more physically activity scored higher on balance and dual-task ability, which are key predictors of fall risk. The authors suggest that physical activity programs targeting balance and dual-task ability could help prevent falls in aged care residents [32].

Ikegami et al found that reduced physical performance is related to serious problems in older individuals, specifically cognitive function, increased fall risk, and reduced quality of life [51].

These data, along with the results of the present study, show some existing trends and relationships between psychophysical factors and fall risk. They also signal the need for continued research to increase the reliability and predictability of fall prevention assessments, to identify more relevant factors associated with fall risk, and to build effective intervention strategies for older adults residing in ALFs.

This study had limitations. We initially aimed to recruit approximately 100 participants from 4 ALFs. However, the recruitment process at 2 of the 4 centers was significantly disrupted by the onset of the COVID-19 pandemic and subsequent restrictions. This led to a sharp reduction in the sample size, both at baseline and during follow-up assessments at 6 and 12 months, and may have influenced the psychophysical condition of the residents. In addition, the emotional-cognitive status scales used in the study were of a screening nature and were not used for making a medical diagnosis. Both our own analysis and that of the recent literature suggest that 6 months as an interval between follow-up surveys may not be sufficient to observe the dynamics of change. The consensus of studies targeting dynamic change indicates the need for at least a 1-year interval between assessments.

Monitoring falls in the form of self-reporting by study participants was not an ideal choice; although it was cost-free, it included the risk of participants concealing falls or not remembering them. The use of fall monitoring devices should be considered in the next study. Also, due to the large number of psychophysical factors compared, we did not consider pharmacological factors, polypharmacy, and comorbidities in this study. According to the literature, these factors can influence emotional state and fall risk in the elderly. A meta-analysis by Shao et al suggests those factors have a low to moderate correlation with these outcomes [7]. Finally, this research done among ALF residents does not make it possible to generalize the results to the entire elderly population.

Conclusions

Over the 12-month observation period, participants residing in the ALFs included in this study demonstrated a statistically significant decline in gait velocity and mood parameters, suggesting progressive deterioration in physical mobility and emotional well-being. Across the entire study group, a higher risk of falling was associated with reduced gait speed and a worse score on the TUG test, while fewer repetitions performed on the 2MST was associated with a higher number of falls. About 40% of participants had experienced a minimum of 1 fall in the past 6 months. The 2MST was the only significant parameter affecting the risk of falls in the study group. It is necessary to use systematic screening and further research and analysis of factors that can be related to fall risk, particularly among older adults residing in ALFs.

Data Availability Statement

The data presented in this study are available on request from the corresponding author.