Introduction

Fall sequels are the fifth cause of death in the elderly [1]. In nursing homes and long term-facilities, number of falls is double in comparison to non-institutionalized older adults [2, 3]. Amount of literature has underlined that this phenomenon is also associated with elevated rate of mortality [4], deeply affects families, and increases society financial burden [5, 6]. Most of the recognized fall-risk factors (FRFs) are usually higher in elderly hospitalized patients [7–10]. For instance, reduced gait and balance performance, asthenia, cognitive impairment, and augmented dependence in the execution of activities of daily living (ADL) are recognized FRFs [2, 11–13]. Age-related pathologies such as Alzheimer’s disease, Parkinson’s disease, stroke, and arthritis seem equally associated with this clinical phenomenon [3, 11, 14]. Despite the recognition of this problem in the clinical milieu and several interventional studies aimed to prevent falls in care homes, the literature provides inconsistent results on FRFs. This heterogeneity is likely associated with a large difference in interventions and levels of patient’s dependency [15, 16].

Besides the objective of interventional studies, research needs also to investigate which are the potential detectable FRFs in a multifaceted clinical setting, such as the standard nursing homes. In particular, the early recognition of patients of high risk of falls would be of crucial importance to start specific treatments. In this scenario, adapted physical activity (APA) is becoming a useful means by which institutionalized elderly individuals may gain muscular strength, enhance ADL performance, and reduce the natural progression of cognitive and physical impairment [17–27]. However, the role of APA in the incidence of falls in nursing home residents is still a matter of debate [15]. Moreover, the risk of falling associated with the recognized FRFs in the specific time frame starting from the admission of the patient to the nursing home is not fully elucidated.

Therefore, the aim of this study was to clarify whether some demographic factors such as sex, age, stature, body mass, fall recidivism, gait and balance performance, cognitive impairment, ADL level, Parkinson’s disease, utilization of gait assistant devices, and the employment of a program of APA are associated with the risk of falls in institutionalized elderly individuals.

Methods

Study design

This observational prospective-cohort study compared fallen and not fallen patients, stratified by their demographic, physical, and pathological characteristics. Kaplan–Meier estimate was used to measure the fraction of subjects not incurred in a fall the first 6 months from the admission to a nursing home. The time from admission (days) to an occurrence of the fall, or to the end of follow-up (censored), were the main variables. Moreover, to compare potential difference among sub-groups of the selected patients a Kaplan–Meier estimate was performed by stratifying the group by sex, age, stature, body mass, recidivism, gait and balance performance, cognitive impairment, ADL levels, Parkinson’s disease, utilization of gait assistant device, and the employment of a program of APA during the observed 6 months.

The study was carried out in accordance with the Declaration of Helsinki (2013) of the World Medical Association for experiments involving humans.

Participants

Eligible volunteers were selected at the admission to the geriatrics care units at the Mons. Mazzali Geriatric Institute in Mantua, Italy. Inclusion criteria were an age greater than 65 years old. No exclusion criteria related to chronic diseases were applied.

Screening

128 patients (37 men and 91 women) were selected for the study. Baseline characteristics of the participants are reported in Table 1.

Table 1 Demographic and clinical characteristics of the study participants at baseline
Full size table

Test procedures

Number of fallen were recorded during the first 6 months from the admission of the patients to the geriatric care unit. Potential co-variants were assessed at the admission to the geriatric care unit.

APA program

During the 6-month of standard geriatric care, a program of APA (5 times/week) was offered to the participants. Noticeably, APA program was not mandatory and employed by the selected patients according their preferences. Specifically, the APA program included circuit-training exercises as described by Venturelli et al. [17]. Each exercise session lasted approximately 60 min: 5 min of warm-up, 30 min of aerobic exercise, 20 min of resistant circuit training, and 5 min of cool-down. Exercising tools used in the study were cycle ergometers (Bike-Race Technogym SpA, Gambettola, Italy), resistance elastic bands, dumbbells from 0.5 to 3 kg, sticks, and sponge balls. The warm-up program comprised active arm and shoulder mobilization by sponge balls and sticks. Aerobic exercise was performed on a cycle ergometer at a workload corresponding to the 60% of the maximal predicted heart rate. Resistance circuit-training exercises for the upper limbs and trunk were performed at 50% of 1RM of the arm curl test, and subsequently enhanced by increasing the number of repetitions and/or load. The initial work load for resistance training by the elastic band was set according to the 1RM arm curl test: subjects with low maximal strength (1st and 2nd quartiles of 1RM) used the yellow Theraband, whereas patients with high maximal strength (3rd and 4th quartiles of 1RM) used the red Theraband. The elastic band work load was increased during the training period only by augmenting the number of repetitions. During the cool-down phase, subjects performed static stretching exercises for legs, arms, shoulders, and neck, keeping the maximal muscle elongation for 30 s, and increasing the range of motion. The exercises were performed under the supervision of a kinesiologist. Adherence to the APA program was recorded into a log.

Baseline assessment

Fasted body mass and stature were measured with a professional mechanical scale fitted with a stadiometer (Seca mod. 713; III-M; Seca Medical Scales and Measuring Systems, Birmingham, UK). History of previous falls was documented with a standard form. Nursing home residents’ gait and balance performance were assessed by POMA scale [28]. The Mini Mental State Examination (MMSE) [29] was utilized to determinate cognitive performance. Independence and level of ADL were evaluated by Barthel Index [30]. Confirmed diagnosis of Parkinson’s disease was also documented during the physical medical evaluation.

Statistical methods

Kaplan-Mayer survival analysis estimated the proportion of not fallen patients from admission to nursing home. The time from assessment (days) to an occurrence of falling, or to the end of follow-up (6 months—censored), was the main variable. The difference in the mean survivor time (MST) between each parameter assessed at the admission of the patient has been tested with a Log Rank (Mantel-Cox) model. Sex, history of previous falling, Parkinson’s disease, and the utilization of gait assistant devices have been treated as dichotomous variables. Age, POMAb, POMAg, MMSE, ADL, stature, body mass, the adherence to the APA program have been distributed according the higher (Ht), middle (Mt), and lower (Lt) tertile.

Results

The participants’ characteristics are summarized in Table 1. Thirty falls have been documented during the 6 months of observation period.

Overall survivor analysis

The proportion of not fallen patients in the 180 days following the admission to the nursing home is represented in Fig. 1. Specifically, the overall MST to a fall event was 154.8 ± 4.6 days.

Fig. 1
figure 1

Kaplan–Meier estimation model of the fraction of not fallen patients (proportion/months)

Full size image

Survivor analysis and demographic factors

Female and male patients exhibited similar MST to a fall event 156.4 ± 5.5 and 150.7 ± 9.1 days (p = 0.309), respectively. No difference was detected in the MST to a fall event among the three age-tertile groups (p = 0.070; Table 2). Patients’ stature and body mass were not associated with an increased risk of falling (p = 0.071 and 0.122, respectively Table 2).

Table 2 Mean survivor time (MST) and range of days to a fall event: Values are expressed as mean ± standard deviation
Full size table

Survivor analysis and FRFs

As expected, a history of previous falls significantly affected the MST to a fall event (Fig. 2e; Table 2). Interestingly, patients with a reduced balance performance, assessed by the POMAb test, did not demonstrate an increased risk of falling (p = 0.661). Contrarily, patients in the Lt of gait capacity (POMAg) showed a significant reduction in MST to a fall event (Fig. 2a; Table 2). Patients in the Lt of cognitive capacity, assessed with the MMSE, demonstrated a worsen MST to a fall event (Fig. 2c; Table 2). Similarly, patients with Parkinson’s disease and lower levels of ADL showed an increased risk of falling (Fig. 2b, f). Interestingly, the utilization of gait assistant devices did not seem to affect MST to a fall event (Table 2). However, in the group that did not attend the proposed APA program (Lt), the risk of falls was significantly increased (Fig. 2d; Table 2).

Fig. 2
figure 2

Kaplan–Meier estimation models of the fraction of not fallen patients (proportion/months). a–d Represents Kaplan-Meier curves of the recruited 128 new-admitted patients distributed in higher (Ht), middle (Mt), and lower (Lt) tertiles. d, e Represents dichotomous variables (yes–no). POMAg, gait performance test [28]; ADL, Barthel Index [30]; MMSE, the Mini Mental State Examination [29]; APA, adapted physical activity. *MST significant shorter

Full size image

Subject’s characteristics according their adherence of APA program

As reported in Table 3, the adherence to the proposed AFA program varied among the institutionalized patients. Indeed, 30 of the fifty-three patients that did not attended (Lt) the APA program have incurred in a fall during the first 6 months of hospitalization. However, their physical characteristics and medical status tended, without reaching a significant difference, to be lower in comparison to the patients that attended APA.

Table 3 Subjects characteristics according to their adherence to APA
Full size table

Discussion

Although FRFs have been already investigated in clinics, the influence of these factors on the MST to a fall event from the patient’s admission to a nursing home facility has received so far only little attention. The main finding of this study was that history of previous falls, reduced gait capacity, poor cognitive function, limited performance in the ADL, and Parkinson’s disease were associated with an increased risk of falling. Conversely, patients that, during this 6-month observation period, were involved in an APA program exhibited a significant reduction in the risk of falling.

Incidence of falling and demographic risk factors

In a recent study, the incidence of falling among 230,730 individuals newly admitted to nursing homes was examined [3]. The 21% of this cohort had experienced at least one fall. In another cohort study among institutionalized elderly individuals [13], several demographic FRFs have been studied such as age, sex, and body mass index (BMI). Interestingly, only smaller BMI values were associated with an increased risk of falling. Data from the present study are partially in agreement with these previous reports. In fact, the percentage of patients recruited for this study that incurred in a fall during the first 6 months of hospitalization was very similar (about 23%) to that previously reported [3]. Moreover, in agreement with Diaz Gravalos and co-workers [13], age and sex did not appear to be associated with an elevated risk of falls. Noticeably, alterations in the anthropometric parameters, such as lower body mass and stature, were not associated with an increased risk of falling [13].

Fall-risk factors in hospitalized elderly

It has been recently highlighted that most of the recognized FRFs are elevated in hospitalized elderly individuals [7–10]. Therefore, specific investigations are strongly recommended to prevent this clinical phenomenon. Previous studies involving hospitalized elderly people revealed that the most important FRFs were: history of previous falls, reduced gait and balance capacity, reduced muscular force, cognitive impairment, and difficulties in the execution of activities of daily living [2, 11–13]. Moreover, an increased risk of falling is also associated with the severity of pathologies such as Alzheimer and Parkinson’s disease, stroke, and arthritis [3, 11, 14]. Data from the present study are in agreement with this literature as history of previous falls, reduced gait capacity, poor cognitive function, limited performance in the ADL, and Parkinson’s disease were associated with an increased risk of falling in our cohort of institutionalized elderly.

The role of APA in the risk of falling

During the last decade, APA is becoming a useful approach through which institutionalized elderly individuals may gain muscular strength, enhance independence, and contrast cognitive and physical deterioration [17–20, 25–27, 31]. However, it has been recognized that APA interventions for preventing falls in care facilities remain uncertain due to conflicting results, possibly associated with differences in interventions and levels of patient’s dependency [15]. While many co-factors are likely associated with the adherence to an APA program (Table 3), data from the present study suggest that the participation to an APA program may help to decrease the risk of falling among hospitalized elderly individuals. This result is extremely important because it indicates the effectiveness of APA as a successful and tailored treatment for the patients at high risk of falling.

Conclusion

In nursing homes, number of falls is double in comparison to non-institutionalized older adults, causing an elevated rate of mortality and society financial-burden. Results of this study indicate that history of previous falls, reduced gait capacity, poor cognitive function, limited performance in the ADL, and Parkinson’s disease are associated with an increased risk of falling. Therefore, particular attention has to be payed to these fall-predictors in institutionalized elderly. Even though the adherence to a program of APA in the clinical setting is potentially affected by confounding factors such as the cognitive level of the patients, their functional capacity, and health status, our data seem to suggest the effectiveness of an APA program in the prevention of falls.