Feasibility study by a single-blind randomized controlled trial of self-management of mobility with a gait-speed feedback device by older persons at risk for falling

ABSTRACT

This single-blind randomized pilot study explored feasibility and safety of a self-management fall prevention program, hypothesizing that older persons can comply with this program, while it does not result in more (injurious) falls, or a decrease in mental wellbeing as an adverse effect of being focused on falls prevention. Eighty-six persons, community-dwelling or home for the aged (mean age 80.3 years [SD: 6.3], 56 women (65.1%)) participated. The intervention group measured their gait speed by using the Mobility Feedback Device (MFD) weekly for 6 months. The control group was monitored for the outcomes without an intervention. Change scores involving health perception and mental wellbeing (Medical Outcomes Study 20-item short form (MOS-20)) were compared between groups. Feasibility was assessed by drop-out rate and compliance to measure gait speed. Safety was assessed by fall incidence during follow-up. MOS-20 decreased significantly in the control group (p = 0.024) but remained stable in the intervention group. Drop-out rate was low (9.3%), and compliance was good. Fall incidence was the same for both groups (p = 0.155). The self-management fall prevention program is feasible and safe in a community-dwelling and home for the aged population, making it worthwhile to further explore self-management fall-prevention studies.

KEYWORDS:

• community-dwelling older persons
• falls
• home for the aged
• prevention
• telemonitoring

Introduction

In today’s aging population, falling is a major problem, causing physical and social impairments and increasing health-related costs (Bertera & Bertera, 2008; Medell & Alexander, 2000; Robinovitch et al., 2013). To reach sustainable healthcare, policymakers are encouraging older persons to be more actively involved in their own healthcare. Researchers are beginning to follow this lead on self-management, also advocated by the chronic care model (Barlow, Wrigth & Sheasby, 2002; Wagner et al., 2001; Quinn, Toms, Anderson & Clare, 2016). The ability of older persons to quickly identify changes in their mobility could give them a tool to take responsibility for their own mobility-related well-being. Previous results of the Senior Step Study showed that gait speed as a single-item tool has the potential to be a self-management tool for mobility and fall risk and showed that older people are willing to engage in their own healthcare using this tool (Bongers, Schoon, Graauwmans, Hoogsteen-Ossewaarde, & Olde Rikkert, 2015; Bongers, Schoon, & Olde Rikkert, 2016). Furthermore, Robinson et al. studied the views of older people on self-management and adherence with exercise-based programs and found that older people are open to this (Robinson, Newton, Jones, & Dawson, 2014). However, further research is needed to study whether self-management of one’s own fall risk is technically feasible and safe in older populations. Regular self-management should not decrease mental wellbeing and the older person should actually use the intervention. To the best of the author’s knowledge, the current study is the first to explore the feasibility and safety of the self-management fall prevention program, which included self-monitoring of mobility and an exercise program for older persons, who are community-dwelling at a home for the aged and regularly visit a community center, in a randomized controlled pilot study. Therefore, we aimed to investigate whether compliance to the self-management intervention can be good, without the adverse effects of a decline in mental wellbeing, an increase in (injurious) falls, or fear of falling (FoF). Second, we investigated whether self-managing influences fall incidence compared to older persons not engaged in their own mobility-related healthcare in a preliminary analysis, which gave the first results of the effectiveness of our self-management fall prevention program.

Methods

Participants

Older persons from the region of Nijmegen, The Netherlands, were asked to participate in the Senior Step Study (ClinicalTrials.gov identifier: NCT01792180). Eligible individuals were aged 70 years or older, fell at least once in the previous year and were able to walk (with or without a walking aid). Not speaking and understanding Dutch, not being able to answer the falls telephone (FT), and not having an informal caregiver who could answer the FT for them were exclusion criteria. The FT system (ASK Community Systems, Rotterdam, The Netherlands) is a computerized system that automatically contacts participants by telephone and was found to be feasible, reliable, and valid for assessing falls and fall related injuries in older people (Bongers, Schoon & Olde Rikkert, 2016; Marck vd, Overeem, Klok, Bloem, & Munneke, 2011; Reelick, Faes, Lenferink, Esselink, & Olde Rikkert, 2011).

Participants were recruited from three settings: community-dwelling, homes for the aged (seven recruitment homes), and community centers (six recruitment community centers). Participants were recruited in several ways, including local senior organizations, home care workers, caregivers of homes for the aged, local newspapers, and supervisors in community centers (Bongers, Schoon, & Olde Rikkert, 2018). All participants provided written informed consent. The local medical ethics committee approved the study (approval number 2012–300).

Design

At the start of this single-blind pilot study, participants were randomized to the 6-months intervention period or the control condition. Randomization took place on the individual level in the community-dwelling setting and on the cluster level in the other two settings. To enable blockwise randomization a research assistant rolled a dice to allocate the randomization (even for intervention, uneven for control condition) for all potential participants per location. In the community-dwelling setting, this was done separately for males and females. The allocations were filled in on an allocation form and put in closed envelopes. During participant enrolment, a research assistant took the first envelope and assigned the participant, according to gender in the community-dwelling setting, to the group as stated in the envelope. When the first participant from a specific home for the aged or the community center enrolled, the first randomization envelope for the study locations was used for the allocation.

Assessments at baseline (T0) and after 6 months (T6) took place at the participants’ home. To preserve the blinding of the researcher collecting the outcome data, research assistants explained the intervention separately from the baseline measurements (T0) done by a blinded researcher. At the end of the study, research assistants visited the participants to remove the intervention device from the participants’ homes and instructed the participants not to mention to the researcher whether they used the intervention or not during the final outcome assessment. The researcher did the final assessments (T6) similarly during a subsequent visit (Figure 1). Monthly telephone calls made by the research assistants were used to measure adherence and monitor safety measures beside the FT.

Figure 1. Participant time line and progress through the study. T0 = baseline assessment; MFD = Mobility Feedback Device; T6 = end of study assessment after 6 months.

Intervention

Participants in the intervention group followed our self-management fall prevention program and measured their usual gait speed weekly for 6 months using the Mobility Feedback Device (MFD) (2M Engineering Ltd., Valkenswaard, The Netherlands). The MFD is a telemonitoring instrument embedded in a picture frame that was placed in the participants’ homes (community-dwelling setting) or in an easily available common room (other two settings), and measured gait speed with two infrared sensors when participants walked past. The sensors were placed at approximately four meters apart from each other. At the bottom of the frame, a display showed the gait speed in kilometers per hour. Participants monitored their gait speed twice per week. The MFD of community-dwelling participants had an internal alarm to remind participants of the next planned measurement. Participants from the other two settings were reminded during a weekly open session with the research assistant.

An instruction book with everyday exercises designed for this study was given to participants in the intervention group in case participants wanted to improve mobility based on their gait speed measurement. The instruction book consisted of low-to-medium intensity exercises and simple-to-complex exercises. Exercises were tailored to the participants’ physical conditions by using the MFD’s feedback. For instance, participants with a low gait speed were recommended to perform simple and low-intensity exercises. The participants tailored the exercises themselves after the first study visit from the research assistant, using the feedback from the MFD and several other clues (e.g., use of walking aid inside or outside the house), which were all explained in the exercise book. Use of the instruction book was not mandatory. Participants were asked to register the type and duration of the exercises, if they performed them, in an activity diary. The control group had no intervention besides the weekly telephone calls from the FT.

All participants were told about the connection between gait speed and falls and reducing fall risk by exercising.

Outcomes

Mental wellbeing

Subjective general health and mental wellbeing, measured with the Medical Outcomes Study 20-item short form (MOS-20), were studied as proxies of subjective acceptability of the self-management program. The MOS-20, assessed at T0 and T6, covers six dimensions: physical functioning, role functioning, social functioning, mental health, current health perceptions, and (physical) pain (Kempen, 1992; Stewart, Hays, & Ware, 1988).

Feasibility, compliance, and safety

The outcome for objective feasibility was compliance to the MFD. Compliance was assessed as the number of weekly measurements and detailed by the reasons for not having a measurement.

Outcomes for safety were falls experienced during the study, injuries obtained during a fall and FoF. A fall was defined as “an unexpected event in which the participant comes to rest on the ground, floor, or lower level” (Lamb, Jorstad-Stein, Hauer, & Becker, 2005). During the study, fall incidents were monitored by the FT. All participants were instructed about the FT during the first visit of the research assistant. An injurious fall was defined as a fall that resulted in an injury (i.e., fracture, head injury, soft tissue injury or other injuries).

FoF was assessed at T0 and T6 using the Falls Efficacy Scale-International (FES-I) (Yardley et al., 2005).

Effectiveness of self-management fall prevention program (secondary outcomes)

To study the first effects of our self-management fall prevention program, we chose outcomes that represent areas on which our program can have an effect, i.e., mobility, daily physical activity, self-management abilities, and disabilities. During visits of the researcher at T0 and T6, mobility was assessed using the Short Physical Performance Battery (SPPB) (Guralnik et al., 1994) and Timed Up and Go (TUG) (Mathias, Nayak, & Isaacs, 1986; Podsiadlo & Richardson, 1991). Daily physical activity was assessed using the LASA Physical Activity Questionnaire (LAPAQ) (Stel et al., 2004). The Self Management Ability Scale (SMAS-30) was used to evaluate self-management (Schuurmans et al., 2005), and disability was evaluated with the (modified) Katz scale (Katz-15) (Weinberger et al., 1992).

At baseline disease burden was assessed using the Cumulative Illness Rating Scale for Geriatrics (CIRS-G) (Miller et al., 1992), cognitive status was assessed using the Mini Mental State Examination (MMSE) (Folstein, Folstein, & McHugh, 1975), and frailty was assessed with the Fried criteria (Fried et al., 2001).

Statistical analysis

Baseline characteristics of the intervention and control groups were compared using independent samples t-test for continuous variables and Chi-square tests for categorical variables. Baseline characteristics of the three settings were compared using univariate analysis of variance. Reasons for drop-out were registered. A participant was called a drop-out when the participant dropped out of the study after the study visit of the research assistant during which the MFD and FT was explained. Baseline characteristics of the participants who completed the study were compared to those who dropped out using independent-sample t-tests for continuous variables and Chi-square tests for categorical variables.

Mental wellbeing

Change scores (T6 minus T0) were calculated for the six dimensions of the MOS-20. Change scores were compared for all participants who completed the study between the intervention and control group using analyses of covariance with setting, age and baseline scores as covariates.

Feasibility, compliance, and safety

Reasons for not having a weekly MFD measurement were categorized as follows: permissible reasons (vacation, sickness), wrong use of the MFD, impersonal reasons (e.g., MFD not functioning), drop-out, other reasons, and unknown. The total number of measurements was corrected for the impersonal and permissible reasons. Compliance to the MFD was assessed for all participants who completed the study by dividing the number of weekly measurements by the total number of possible measurements. Less than a 15% drop-out rate and at least 80% compliance were judged as objective criteria for the feasibility of the self-monitoring set up.

Safety was analyzed by comparing fall incidence and injurious falls between the intervention and control group using independent samples t-tests. The percentage of fallers during the study was compared between both groups using Chi-square tests. Change scores (T6 minus T0) were calculated for FoF and compared between both groups by analysis of covariance with baseline FoF, age and setting as covariates. Safety was defined as an occurrence of less or equal number of falls in the intervention group, less or equal injurious falls and not getting more FoF compared to that of the control group.

Post hoc, the abovementioned analyses were also performed within each setting.

Effectiveness of self-management fall prevention program

Change scores were also calculated for SPPB, TUG, LAPAQ, SMAS-30, and Katz-15. Again, analyses of covariance were used to compare the change score between both groups, with setting, age and baseline scores as covariates.

Significance was set at p < 0.05. Bonferroni corrections were used for multiple testing of significant differences. All statistical analyses were performed using IBM SPSS Statistics 22 (SPSS, Chicago, Illinois, USA).

Results

Because the intervention in the home for the aged and community center settings was shaped as a group intervention for all participants present in the homes or community centers at that timeslot, all individuals participating in these groups were allowed to participate in the study. However, we started the analyses only with persons fulfilling all inclusion criteria. One hundred and five participants consented to participate in the study, of whom 86 met all inclusion criteria.

Of the 86 participants (mean age 80.3 years [SD: 6.3], 56 were female (65.1%)), 58 participants were community-dwelling, 20 participants lived in a home for the aged, and 8 participants regularly visited a community center. Participants from the home for the aged setting were significantly older and more frail compared to the participants from the other two settings (Table 1).

Table 1. Baseline characteristics of participants with a recent fall in the community-dwelling, home for the aged and community center settings.

Participants included (N = 86)
	Setting				Randomization
	Community-dwelling (N = 58)	Home for the aged (N = 20)	Community centre (N = 8)	p	Intervention group (N = 43)	Control group (N = 43)	p
Age, years	79.1 (6.1)	84.3 (5.3)	79.0 (6.4)	.004	79.9 (5.5)	80.9 (7.0)	.373
Female, N(%)	37 (63.8)	12 (60.0)	7 (87.5)	.374	29 (67.4)	27 (62.8)	.651
MOS-20, physical functioning	46.3 (32.3)	35.8 (33.5)	29.2 (26.4)	.226	39.9 (31.9)	44.6 (32.9)	.508
MOS-20, role functioning	43.1 (46.3)	32.5 (43.8)	37.5 (51.8)	.668	27.9 (41.3)	52.3 (47.5)	.013
MOS-20, social functioning	73.1 (28.2)	47.0 (32.0)	62.5 (36.2)	.005	60.0 (33.5)	72.1 (28.3)	.074
MOS-20, mental health	76.3 (18.9)	60.4 (22.7)	64.5 (18.9)	.007	70.2 (21.0)	72.7 (20.8)	.579
MOS-20, current health perception	62.8 (23.5)	48.3 (23.2)	58.1 (22.7)	.062	58.8 (25.0)	59.1 (23.0)	.964
MOS-20, pain	39.2 (34.4)	57.9 (32.3)	53.1 (24.8)	.085	41.7 (34.8)	47.7 (33.1)	.417
SPPB	8.9 (2.8)	6.4 (3.2)	8.4 (3.6)	.007	8.5 (3.1)	8.16 (3.2)	.495
LAPAQ, minutes/day	118.5 (64.6)	78.9 (70.4)	147.4 (119.5)	.042	111.8 (71.2)	112.1 (77.7)	.987
SMAS-30	63.6 (12.4)	53.6 (7.8)	61.3 (14.5)	.006	60.9 (12.7)	61.3 (12.0)	.884
Katz-15	1.8 (1.9)	4.1 (2.7)	4.0 (3.4)	<.001	2.8 (2.8)	2.2 (2.2)	.244
FES-I	30.7 (9.4)	34.5 (11.0)	34.0 (16.4)	.341	32.8 (11.4)	31.0 (9.7)	.430
CIRS-G	9.5 (5.2)	13.9 (4.9)	9.9 (5.0)	.027	10.3 (5.3)	10.2 (5.4)	.929
MMSE	28.1 (1.9)	26.0 (3.5)	27.8 (3.2)	.005	27.7 (2.7)	27.4 (2.6)	.626
Frail, yes (%)	6 (10.3)	9 (45.0)	1 (14.3)	.003	9 (20.9)	7 (16.3)	.579
Setting, N(%) – Community dwelling – Home for the aged – Community centre					1.00
					29 (67.4)	29 (67.4)
					10 (23.3)	10 (23.3)
					4 (9.3)	4 (9.3)

Notes. All data presented as means with standard deviation within parentheses unless stated otherwise. MOS-20: Medical Outcomes Study 20-item short form (6 dimensions: physical functioning, role functioning, social functioning, mental health, current health perceptions, and (physical) pain (each dimension a score range 0–100, higher scores indicating better health)); SPPB: Short Physical Performance Battery (summary performance score (range 0–12) for three tasks, higher scores indicating better lower body function); TUG: Timed Up & Go (higher scores indicating lower mobility); LAPAQ: LASA Physical Activity Questionnaire (minutes per day for all physical activities); SMAS-30: Self-Management Ability Scale (30 items, score range 0–100, higher scores indicating better self-management); Katz-15: Katz-15 item scale (basic and instrumental activities of daily living, with a score range of 0–15, lower scores indicating better functional performance); FES-I: Falls Efficacy Scale-International (scale from 1 to 4 for each of 16 items, total scores ranging from 16 to 64, higher scores indicating more fear of falling); CIRS-G: Cumulative Illness Rating Scale for Geriatrics (scale from 0–4 for each of 14 categories, higher scores reflecting more comorbidity); MMSE: Mini Mental State Examination (range 0–30, higher scores indicating better cognitive status); Frail: Fried Frailty score (range 0–5; 0–2: not frail, ≥3 frail) based on the Fried criteria: unintentional weight loss, self-reported exhaustion, low energy expenditure, slow gait speed, and weak grip strength.

Forty-three participants were randomized to the intervention group, and 43 were randomized to the control group. The intervention group had a significantly worse subjective health level on the MOS subscale of role functioning compared to that of the control group (27.9 (41.3) versus 52.3 (47.5) t(82.4) = 2.5, p = 0.013) (Table 1). Within each setting there were no significant differences at baseline between both groups (Table 1).

Of the 86 participants, eight participants dropped out (9.3%): six community-dwelling participants (five from the intervention group, one from the control group) and two participants who lived at a home for the aged (both from the intervention group). Four participants dropped out because they felt that participation was too demanding. Other drop-out reasons were as follows: participation was too time consuming, participants were withdrawn from the study by their children, participants found that participating was not necessary anymore, and a few participants were not capable of walking anymore. Differences between drop-outs and participants who completed the study are shown in Table 2. Drop-outs showed lower role functioning, social functioning and mental health on the MOS-20 subscales and less functional performance on the Katz-15 (Table 2).

Table 2. Baseline characteristics of the participants who completed the study and those who dropped-out.

	Participants (n = 78)	Drop-outs (n = 8)	p
Age, years	80.3 (6.4)	80.9 (5.3)	.796
Female, N(%)	52 (66.7)	4 (50.0)	.346
MOS-20,physical functioning	43.2 (32.5)	33.3 (30.9)	.416
MOS-20, role functioning	42.9 (46.8)	12.5 (23.1)	.008
MOS-20, social functioning	68.5 (30.3)	42.5 (34.5)	.025
MOS-20, mental health	72.9 (20.5)	57.5 (19.9)	.045
MOS-20, current health perception	60.1 (23.8)	48.1 (23.7)	.180
MOS-20, pain	44.2 (34.6)	50.0 (26.7)	.645
SPPB	8.3 (3.1)	8.3 (4.1)	.978
TUG, seconds	12.3 (7.7)	14.6 (13.7)	.451
LAPAQ, minutes/day	113.4 (75.7)	97.8 (58.4)	.573
SMAS-30	61.4 (12.4)	57.9 (12.1)	.447
Katz-15	2.3 (2.3)	4.8 (3.3)	.007
FES-I	31.7 (10.8)	34.1 (8.6)	.536
CIRS-G	10.3 (5.4)	10.8 (5.0)	.824
MMSE	27.7 (2.4)	26.0 (4.3)	.298
Frail, yes (%)	13 (16.7)	3 (37.5)	.149

Notes. All data presented as means with standard deviation within parentheses unless stated otherwise. MOS-20: Medical Outcomes Study 20-item short form (6 dimensions: physical functioning, role functioning, social functioning, mental health, current health perceptions, and (physical) pain (each dimension a score range 0–100, higher scores indicating better health)); SPPB: Short Physical Performance Battery (summary performance score (range 0–12) for three tasks, higher scores indicating better lower body function); TUG: Timed Up & Go (higher scores indicating lower mobility); LAPAQ: LASA Physical Activity Questionnaire (minutes per day for all physical activities); SMAS-30: Self-Management Ability Scale (30 items, score range 0–100, higher scores indicating better self-management); Katz-15: Katz-15 item scale (basic and instrumental activities of daily living, with a score range of 0–15, lower scores indicating better functional performance); FES-I: Falls Efficacy Scale-International (scale from 1 to 4 for each of 16 items, total scores ranging from 16 to 64, higher scores indicating more fear of falling); CIRS-G: Cumulative Illness Rating Scale for Geriatrics (scale from 0 to 4 for each of 14 categories, higher scores reflecting more comorbidity); MMSE: Mini Mental State Examination (range 0–30, higher scores indicating better cognitive status); Frail: Fried Frailty score (range 0–5; 0–2: not frail, ≥3 frail) based on the Fried criteria: unintentional weight loss, self-reported exhaustion, low energy expenditure, slow gait speed, and weak grip strength.

Mental wellbeing

An analysis of the change scores yielded a significant difference in current health perception (MOS-20) between the intervention and control group (F(1,70) = 5.3, p = 0.024) in that the current health perception of the control group decreased during the 6-month follow-up, while the current health perception of the intervention group did not change (Table 3).

Table 3. Changes in outcomes of the intervention and control group following 6 months self-monitoring of gait speed.

	Intervention group (N = 36)	Control group (N = 42)	p
MOS-20,physical functioning	−2.8 (25.0)	−1.2 (26.4)	.411
MOS-20, role functioning	11.1 (38.0)	−4.8 (39.5)	.776
MOS-20, social functioning	−3.9 (22.3)	−1.4 (25.9)	.640
MOS-20, mental health	−2.0 (11.3)	−0.8 (11.4)	.496
MOS-20, current health perception	−1.3 (17.1)	−3.9 (16.4)	.024
MOS-20, pain	7.1 (26.8)	−2.4 (40.5)	.587
SPPB	0.3 (1.5)	0.5 (2.0)	.268
TUG, seconds	0.4 (3.6)	−0.2 (5.3)	.697
LAPAQ, minutes/day	−5.3 (61.0)	0.9 (55.6)	.880
SMAS-30	−1.0 (7.2)	−2.3 (6.7)	.378
Katz-15	0.1 (1.3)	−.02 (1.3)	.523
FES-I	2.1 (6.9)	−1.6 (7.0)	.110

Notes. All data presented as means with standard deviation within parentheses unless stated otherwise. Bonferroni corrections were used for multiple testing of significant differences. MOS-20: Medical Outcomes Study 20-item short form (6 dimensions: physical functioning, role functioning, social functioning, mental health, current health perceptions, and (physical) pain (each dimension a score range 0–100, higher scores indicating better health)); SPPB: Short Physical Performance Battery (summary performance score (range 0–12) for three tasks, higher scores indicating better lower body function); TUG: Timed Up & Go (higher scores indicating lower mobility); LAPAQ: LASA Physical Activity Questionnaire (minutes per day for all physical activities); SMAS-30: Self-Management Ability Scale (30 items, score range 0–100, higher scores indicating better self-management); Katz-15: Katz-15 item scale (basic and instrumental activities of daily living, with a score range of 0–15, lower scores indicating better functional performance); FES-I: Falls Efficacy Scale-International (scale from 1–4 for each of 16 items, total scores ranging from 16 to 64, higher scores indicating more fear of falling).

Feasibility, compliance, and safety

Reasons for not having a weekly measurement were permissible reasons (14.5%), wrong use of the MFD (2.2%), impersonal reasons (30.9%), drop-out (19.5%), others (9.9%), and unknown (23.0%). After correcting the total number of measurements for the impersonal and permissible reasons, the compliance of self-monitoring gait speed using the MFD was good, with a median (25% and 75% centiles) percentage of measurements at 82.1% (57.6% and 100%) for participants who completed the study (n = 78).

There was a significant difference between the three settings (F(2,33) = 4.9, p = 0.014) in that the community center participants had significantly lower compliance compared to that of the community-dwelling participants.

In total 40 falls were registered in the intervention group, and 82 falls were registered in the control group. Fall incidence did not differ between the intervention (mean incidence 0.9 [SD 1.8]) and the control group (mean 1.9 [SD 4.1]) (t(57.7) = 1.4, p = 0.155). Additionally, the number of fallers in the intervention group (n = 17 (39.5%)) did not differ from the control group (n = 19 (44.2%)) (χ²(2, n = 86) = 0.2, p = 0.662). The incidence of injurious falls did not differ between the intervention (mean incidence 1.4 [SD 1.5]) and the control group (mean 1.7 [SD 2.4]) (t(34) = 0.5, p = 0.637). FoF did not differ between the intervention and control group (F(1,69) = 2.6, p = 0.110, Table 3). There were no differences in FoF within each setting.

Effectiveness of self-management fall prevention program

Analysis of the change scores gave no significant differences in the effectiveness of the fall prevention program between the intervention and control group (Table 3).

Discussion

The current study showed that a self-management fall prevention program, including self-management of mobility and an exercise program, gave good technical feasibility and compliance, but it did not yet show an overall positive effect involving a decline in fall rate, although there appears to be a trend toward a beneficial effect. Additionally, no effect of participating in our program was found on mobility, daily physical activity, self-management abilities, and disabilities.

Mental wellbeing is important to older persons and the current study showed that the self-management fall prevention program did not change subjective general health and mental wellbeing. Moreover, current health perception decreased in the control group, suggesting a net positive effect of self-managing on mental wellbeing, although we did not see this change in the other dimensions of subjective general health. Drop-out rate (9.3%) was low for a study requiring participants’ responsibility when compared to most other intervention studies (18.5–23%) (Nagayama et al., 2016; Waters, Hale, Robertson, Hale, & Herbison, 2011). However, in terms of technological acceptance according to the Technology Acceptance Model (TAM), we only studied the usage behavior by looking at our drop-outs (Venkatesh & Davis, 2000). For a full study into the acceptability of our self-management fall prevention program, we also needed to study the other three constructs of TAM, i.e., perceived usefulness, perceived ease of use, and intention to use. The low drop-out rate of our study does show a good usage behavior. Unfortunately, it is not clear whether this was caused by the MFD, the exercise book or the study visits and contact moment. The reasons for drop-out, “Participation was too demanding” or “Participation was too time consuming”, did not give insight into this phenomenon. Therefore, technological acceptance needs to be studied in future research. Compliance amounted to more than 80% for participants who completed the study and was comparable to our previous study based on self-management (Bongers et al., 2016). The current research method is safe because participants in the intervention group did not fall more, and they were not more fearful of falling. However, the study sample was too small for a powered analysis of fall frequency changes, although such an analysis was not the study’s primary aim.

A strength of the current study was the prospective follow-up over 6 months of many variables in older persons in three settings. The design of the study (randomized, single-blind) made it possible to thoroughly investigate the effect of self-management on mental wellbeing. When looking at the percentage of fallers and self-management abilities at baseline, our study’s population was comparable to other studies in community-dwelling populations mentioned in the literature (Hausdorff, Rios, & Edelberg, 2001; Kremers, Steverink, Albersnagel, & Slaets, 2006; Schuurmans et al., 2005). Furthermore, participants were older and showed lower overall performance in mobility compared to participants from our previous study (Bongers et al., 2016), and the number of frail participants was comparable to the literature (Bandeen-Roche et al., 2015; Buckinx et al., 2016; Fried et al., 2001), making our results generalizable not only for a community-dwelling population but also for a more frailer population. Interesting findings regarding the results in the home for the aged setting demonstrated equal drop-out rates, fall data and compliance compared to those of the two other settings, suggesting that even with the low inclusion rate, self-management is possible in this older and more frail population. A recent review confirmed that elements of self-management are still possible in older persons with cognitive decline (Quinn, Toms, Anderson, & Clare, 2016). However, the drop-outs from our study showed less daily activity, social functioning, mental health (MOS-20), and lower functional performance (Katz-15) compared to the participants who finished the study. This finding might limit the generalizability of our results, and future research needs to study which population is most suitable for self-management of fall prevention interventions before a full RCT can be conducted. To reduce falls in older adults, Weijer et al. suggest that interventions might be posed to categorize the effectiveness of older populations by both poor physiological and psychological statuses (Weijer, Hoozemans, van Dieën, & Pijnappels, 2018). According to this suggestion, the drop-outs from our study possibly are the persons who would benefit most from our self-management fall prevention program, so further study is needed to explore how drop-outs of these participants can be prevented.

The preliminary results of the effectiveness of our self-management fall prevention program indicate no effect on the secondary outcomes. However, we only studied these outcomes on a group level and it would be very interesting to also test results on an individual level (in a full single patient trial design with repeated measurements), which goes beyond the scope of the current study. Furthermore, our self-management fall prevention program consisted of both the MFD and the exercise book, and this condition makes it difficult to explain which part of the program might create the best effect. Future studies are encouraged to explore this phenomenon by allowing both groups the exercise program, but only the intervention group the use of the MFD to assess the benefit of the MFD. However, this was also beyond the scope of the current study.

Limitations

The small sample size of participants in a home for the aged and those in a community center setting was a limitation of the current study. During recruitment, especially in the community center setting, participants stated that they felt stigmatized when asked to participate because participating meant they had fallen in the previous year and were thus getting old. For this reason, the intervention was shaped in a group intervention in these settings. A further limitation was the lack of qualitative data to explore in depth participants’ acceptability and feasibility, although the study by Robinson et al. did show that older persons are open for this aid (Robinson et al., 2014). This phenomenon needs to be clear before a full RCT can be conducted. Experiences regarding the conduct of the study were not explored, as well as the success of blinding.

Conclusion

In conclusion, the increasing importance of self-management in this aging society warrants feasible and safe self-management research methods that do not decrease an older person’s mental wellbeing. It would be highly interesting to determine whether elements of interventions with proven effectiveness could be transferred to a self-management strategy and remain effective (Mirelman et al., 2016). The current study showed that the self-management fall prevention program including self-monitoring of mobility and an exercise program, in a community-dwelling setting and a home for the aged is feasible and could be the first steps in such experiments. The degree of acceptability of such self-management programs by the older people themselves and their caregivers should best be measured prospectively, both by scientifically sound qualitative observations and quantitative measurements. Mental wellbeing remained the same for a longer period of time, and there is also a trend toward fall prevention in the intervention group. All these outcomes make it very interesting and possible to further study self-management in fall prevention studies in older people.

Acknowledgments

We thank Maartje Graauwmans, Leontien van Nieuwenhuijzen and Naomi Likumahwa for their contribution to the data acquisition.

Disclosure statement

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Additional information

Funding

This work was supported by theNational Programme for Elderly Care [grant number 314050301], which is coordinated and sponsored by ZonMw, The Netherlands, Organisation of Health Research and Development. The study sponsors had no role in study design or conduct; collection, management, analysis, or interpretation of the data; or preparation, review, or approval of the manuscript.