Skip to main content
NCBI home page
Sage Choice logoLink to Sage Choice
. 2023 Feb 2;42(7):1456–1465. doi: 10.1177/07334648231152856

Pragmatic Evaluation of a Low-Threshold Sports Program for Older Adults in Group Homes

Amanda Burke 1,✉, Andy Jones 1
PMCID: PMC10262321  PMID: 36732945

Abstract

Physical activity (PA) reduces with older age, ill health, and disability. For these groups, guidance recognizes the benefit of small increases in PA and low-intensity PA. This study evaluated a low-threshold intervention that addressed known barriers to older people’s participation in PA in residential care and sheltered housing. Ten, competitive sport sessions were delivered by coaches at 49 sites with the aim that they be sustained in-house. Using quasi-experimental methods, participants reported reduced sitting time, increased moderate/vigorous PA, increased participation in sports and improved scores for both health quality of life and fear of falling at 6 months. The program engaged 29% of residents and was sustained at 50% of sites at 8 months. The findings suggest that low-threshold sports programs that overcome known barriers to older people’s participation in PA have the potential to provide a gateway to increased PA in group homes and to be sustained in-house.

Keywords: low-intensity; participation; disability, sheltered housing; care; activities in group homes; evaluation; evidence-based practice, exercise; falls; health behaviors; low-intensity physical activity; participation in physical activity

What this paper adds

  • • Low-intensity interventions have the potential to give older, inactive participants in poor health a gateway into physical activity (PA).

  • • PA interventions can be developed to maximize engagement with older, disabled participants and can include competitive sports.

  • • Low-threshold PA programs have a role in the maintenance of good health in group homes because they can be sustained with few resources.

Applications of study findings

  • • Researchers should report on participation rates and on the sustainability of physical activity programs, as well as outcomes.

  • • Public policy should encourage the provision of appropriate physical activity programs in group homes.

Introduction

Increasing age is associated with an increased prevalence of long-term health conditions such as cardiovascular disease, cognitive impairments and osteoporosis, yet there is good evidence that physical activity (PA) can not only improve older people’s health, but also reduce social isolation, decrease falls, and increase independence (Department for Health and Social Care [DHSC], 2019; McPhee et al., 2016; World Health Organization [WHO], 2020). Despite this, PA levels decrease with age and disability (National Health Service [NHS] Digital, 2019).

In common with the WHO (WHO, 2020), the United Kingdom (UK) government’s recommendations for moderate and vigorous activity for people aged 65+ are the same as for younger adults. However, changes to guidance for older adults issued in 2019 (DHSC, 2019) recognized the benefits of small increases in PA, and of low-intensity PA. Additionally, the WHO have recommended further research around the health benefits of breaking up sedentary time with low-intensity PA (DiPietro et al., 2020).

While the benefits of PA are well understood, it is less clear how to increase it in older people, particularly those that are inactive, disabled, frail, or in poor health. Methods that work for younger people may not be appropriate. For example, there is mixed evidence about the effectiveness of behavior change approaches for older people (French et al., 2014; Zubala et al., 2017). For many older people, drivers of participation include social interaction, enjoyment, belonging, independence, and health (Goodwin et al., 2017; Zubala et al., 2017). A thematic synthesis of the literature (Franco et al., 2015) identified that many older people believed that PA was unnecessary or even potentially harmful. Others, who recognized the benefits of PA, faced barriers to participation that included disability, a fear of falling, access, and affordability.

Competitive sport is one way of increasing PA (Eime et al., 2015). Drivers for sport participation in older people are similar to those for PA more generally, with the additional drivers of competition and a sense of achievement (Jenkin et al., 2017). As with other forms of PA, poor health is a barrier to sport participation in older age (Jenkin et al., 2017).

PA interventions in older adults typically focus on moderate to vigorous activity delivered in structured sessions by professionals. There is less evidence around ‘low-threshold’ programs (Cichocki et al., 2015) that are low-intensity, minimally structured and require few resources to deliver. For example, a scoping review of reablement strategies in residential care (Lewis et al., 2021) found that PA interventions were most often conducted two to three times a week, lasted between three and 4 months, included two or more components of activity such as aerobic, strength and balance exercise, and were physiotherapist-led. However, low-threshold interventions may hold great potential to engage older people in poor health in PA in group homes as they have low barriers to entry and require few resources to deliver.

This study evaluates the implementation, uptake, and effectiveness of “Mobile Me,” a low-threshold sports intervention that addressed known barriers and drivers to older people’s participation in PA. The intervention was low-threshold because it placed minimal demands on participants, many of whom were disabled, frail, or in ill health. It also placed minimal demand on the group housing providers, many of whom had limited resources to deliver physical activities. This was an important consideration, given it was intended that the PA activities would be sustained in-house in the longer term.

Materials and Methods

The Intervention

Mobile Me was a 3-year sport program for older people living in group homes in Norfolk, a county in England. The program was developed and delivered by Active Norfolk, a UK Government funded organization that aims to increase participation in sport and PA. It was developed in close co-operation with the management of three large housing providers. The intervention at each group home consisted of ten, two-hour, weekly sessions of sports (i.e., competitive activities) led by a specially trained coach. After 10 weeks, the program would be sustained in-house by the group home. Key features of the sport sessions were that they were free, delivered on-site, fun, and social. The estimated cost was £347 (approximately 409 USD) per participant starting the program (n = 595). This includes development and marketing costs, in addition to some support for the sustainability of the program, such as equipment for group homes.

The sporting activities were suitable for people of all fitness levels, adaptable for those with disabilities and suitable for relatively small spaces such as communal lounges. They did not require a trained facilitator, so they could be sustained free of charge. Sports fitting these criteria were several forms of indoor bowling, i.e., short mat bowls, new age kurling and boccia, along with table tennis. Adaptations for physical disability included the provision of bowling ramps. Residents were able to play seated, but were encouraged to stand if possible, holding support such as a chair back, if needed. Some chair-based exercises were also delivered at the request of residents.

The two categories of group home were “sheltered housing,” where typically residents live independently in apartments with a manager who checks they are well, and “residential care” where residents are not able to live independently, have sleeping/living space only, and require help with most, or all, activities of daily living. Sheltered housing was run by the local government authority (“LA”), or by housing associations and charities (“non-LA”).

The program was delivered in 49 group homes. Forty-two were run by the three large housing provider providers, a further five housing providers received the intervention at an additional seven sites in total. The program was promoted through the distribution of printed materials, but also through mobilizing housing staff to encourage residents to take part. At the request of one of the housing providers 8 months into the program, the program deliverer ran a training session for staff to help them do this.

Study Design

This was a clustered, nonrandomized, waiting-list controlled study that formed part of a mixed-methods, pragmatic evaluation. Clusters were by group home. The evaluation was pragmatic, or “real world” research (Crane et al., 2019; Robson, 2002), because it was designed in response to planned program delivery and was influenced by practical considerations that might not be anticipated in a formal clinical trial or other experimental design. The evaluation protocol was developed by a university research team, and data was collected by the program deliverer, who developed the intervention. Analysis and reporting has been guided by the Transparent Reporting of Evaluations with Nonrandomized Designs (TREND) checklist (Des Jarlais et al., 2004). The primary research question was: “How effective is the provision of a program at reducing the prevalence of inactivity amongst residents of supported housing?” The secondary research question addressed improvements in other outcomes described in “outcome measures,” for example, health quality of life.

Sample Population

The program was delivered in eight waves between 2015 and 2017 with approximately six group homes takings taking part in each. Residents were invited to an activity session (intervention) or information session (waiting-list control) and given the option of participating in the evaluation. As this was a pragmatic evaluation of an intervention that was being delivered independent of the research study, a-priori sample size/power calculations were not carried out.

The order in which delivery took place was decided by the program deliverer in discussion with housing providers. Waiting list control sites were selected on the basis that they had been earmarked to receive the intervention later in the program. The ratio of intervention to control sites for the three main housing providers were non-LA Sheltered Housing 18 intervention, 5 control; LA Sheltered Housing 6 intervention, 14 control; and Care 8 intervention, 4 control.

Data were collected by program delivery staff at four time-points: the first session (baseline), last session (10 weeks), and 6 months and 12 months. For the control group, pragmatic considerations around staff availability meant data were collected at three time-points only: at baseline, 10 weeks (delivery waves 1–4) or 6 months (delivery waves 5–8), and at 12 months.

All participants in the evaluation provided written, informed consent. Ethical approval was obtained through the University of East Anglia’s Faculty of Medicine and Health Science Research Ethics Committee (ref: 20152016-11 SE).

Outcome Measures

The principal data collection tool was a self-completion questionnaire; most residents had assistance from staff when completing this. The following measures were used:

  • • Inactivity (measured through minutes sitting), as well as minutes of moderate and vigorous activity, and walking, were collected using the International PA Questionnaire for the Elderly (IPAQ-E; Hurtig-Wennlf et al., 2010). Metabolic Equivalents (METS) were calculated from moderate and vigorous activity and minutes walking.

  • • Fear of falling was measured through a single-item visual analogue scale (Scheffer et al., 2010; a maximum score of 10 being most fearful).

  • • Mental wellbeing by the short-form Warwick Edinburgh Mental Well-Being Scale (Tennant et al., 2007; a maximum score of 35 indicating higher wellbeing).

  • • Health-related quality of life using the EQ-5D-5l (Herdman et al., 2011; a maximum score of 25 is indicative of having poorer health QOL), and general health using the EQ VAS, a maximum score of 100 indicative of the “best health you can imagine.”

  • • Minutes of sport in the last 7 days using a Sport England question derived from the IPAQ-E.

Reliability-validity studies existed for all measures except for the Sport England measure, which was included as a requirement by the funder.

Objective measurement data were collected from subgroups as follows: functional fitness was assessed using the Fullerton Functional Fitness Test (6 tests, for which, except the “timed up and go”, higher scores indicate better functional fitness) and standing balance using the Nintendo Wii-Fit balance board (Huurnink et al., 2013), with Balancia software (Park & Lee, 2014). Average path length was measured, with a shorter path indicating improved standing balance as standing balance assesses stability in the standing position. Measurements were taken with feet apart for 30 seconds eyes open, and 30 seconds eyes closed.

Process Measures

Sports coaches kept attendance records. Summary data from these administrative records were used with information about the average number of residents per site from the three main housing providers to estimate levels of engagement in the intervention. For those that consented to participate in the evaluation, levels of attendance were linked to questionnaire data to explore potential predictors for engagement in the program. Additionally, the program deliverer contacted housing providers to establish whether PA sessions had been sustained 8 months after the funding ceased.

Analytic Plan

Comparisons to assess equivalence between groups were carried out using Chi-squared analysis, the Mann–Whitney U Test and t-tests. Statistical modeling was carried out using linear mixed effects regression to account for non-independence of participants clustered within the group homes and for the repeated measures for participants. This approach has the advantage that it models data regardless of missing data at any time point (O’Connell & McCoach, 2004); this was an important consideration due to the manner in which control group data was collected either at 10 weeks or 6 months. Data was analyzed on an intention to treat basis so that any individual that completed a baseline and at least one follow-up was included irrespective of the number of sessions attended.

For the outcome analysis, the data was modeled with random intercepts for group homes (level 3), participants (level 2), and measurement occasion (level 1). The dependent variable was the outcome measurement at all follow-up time points with baseline scores as a covariate (ANCOVA). In addition to a variable indicating whether participants were in the control or intervention group, we included the following covariates: the number of weeks since baseline, group home type, age at baseline, and gender. All continuous variables in regression models were centered at the grand mean. To understand potential intervention effects at 1 year, we carried out an additional analysis using the final follow-up (12 month) data only, with the same covariates.

Confidence intervals (95%) around the point estimate in the original units are provided to enable one estimate of effect size. In addition, standardized effect sizes are provided to enable comparisons between different outcomes. These were calculated using the difference between the adjusted means for the control and intervention groups over pooled standard deviation at baseline for each outcome (Morris, 2008), that is, Cohen’s d using pre-test variance only in the denominator. Data for three outcome measures were found to be zero-inflated and regressions resulted in non-normal residuals (METS, sport minutes per week and moderate and vigorous minutes per week) and heterogenous variance (sport minutes per week). The bootstrap method has been shown to be an appropriate approach in dealing with such data (Paneru et al., 2018; Waguespack et al., 2020) and has the advantage of producing estimates in original units which helps interpretation. For consistency, confidence intervals for all measures were produced using bootstrap with 1000 repeats. A confidence interval that does not contain zero enables the recognition of a statistically significant finding. We have not adjusted p-values for secondary outcomes due to the relatively low sample sizes for sub-samples and consequent risk of type II errors (Feise, 2002) in this pragmatic exploration of a behavioral intervention.

To assess which characteristics may be associated with levels of participation (in the initial 10 weeks of delivered activity), we included covariates that the literature suggests may predict participation in PA and sport i.e., age, gender, health-related quality of life through the EQ-5D-5 L (as this contains measures for disability and pain), and existing levels of sports participation, in a mixed effect regression. The nested structure for the data was modeled with group homes at the top level (level 2), within which participants were nested (level 1). Mixed effects regression was carried out in R version 4.1.1. (R Core Team, 2021) using lme4 (Bates et al., 2015) and bootstrapped confidence intervals were produced using lmeresampler (Loy et al., 2021).

Results

Evaluation Participants Samples

Recruitment to, and participation in, the evaluation is shown in Figure 1. All participants completed a questionnaire, and a subsample of these were recruited for functional fitness and standing balance tests. In addition to the 282 participants completing at least one baseline questionnaire and follow-up (the analysis sample), subgroup functional fitness test data was collected from 96 (34%) of participants at baseline and at least one follow-up, and standing balance data from 40 (14%) of participants at baseline and at least one follow-up.

Figure 1.

Figure 1.

Open in a new tab

Consort diagram showing participation in the quasi-experiment.

Questionnaire baseline measures were compared for those retained in the study and those lost to follow-up. Those retained reported more minutes of weekly sport (18.5 vs. 15.4 min, p = .009), better health quality of life (EQ-5D-5 L) scores (10.5 vs. 10.8, p = .038) and higher general health (EQ VAS) scores (67.4 vs. 60.6, p = .038). However, the magnitudes of differences were relatively small. When subgroup participants were compared to the rest of the sample, those in the standing balance subgroup has lower EQ-5D-5 L scores that those that were not (9.1 vs. 10.5, p = .038). No other statistically significant differences were found.

Differences in Outcome Measures Between Intervention and Control Groups

Test for equivalence between the intervention and control groups at baseline showed that there were less males in the control group than the intervention group (6.7% vs. 30.8% p = .024) and different proportions of participants in “group home type” between the control and intervention (non-LA sheltered = 13.2%/86.8%, LA sheltered = 74.7%/25.3%, care = 25.0%/74.0%, P ≤ .001). No other statistically significant differences were found.

Taking questionnaire results, when compared to the control group and over the three follow-ups, there were statistically significant improvements for the intervention group in measures for sitting minutes per day, as well as fear of falling, moderate and vigorous minutes of activity in the last 7 days, minutes of sport in the last 7 days and health quality of life (EQ-5D-5 L). Table 1(a) shows the intercept, or mean, for each outcome measure when categorical variables are zero and the continuous variables are at their mean (i.e., for 78-year-old males in the intervention group living in sheltered housing at 6 months from baseline). Beta shows the mean difference between control and intervention group in original units, along with 95% confidence intervals. Cohen’s d is beta standardized, and N represents the number of individuals (and time points for repeated measures).

Table 1.

Questionnaire Results.

Outcome Intercept Beta [95% CI] Cohen’s d N
a) Repeated measures modeled over all three follow-ups (46 sites)
  Sitting mins (daily) 595.37 −75.26 [−132.12, −18.48] a 0.35 266/492
  Fear of falling 5.81 −1.08 [−1.88, −0.33] a 0.34 280/521
  METS (weekly) 1246.14 452.64 [−137.91, 1034.71] 0.25 279/514
  Mod/Vig mins (weekly) 106.01 91.62 [8.93, 174.13] a 0.33 279/519
  Sport mins (weekly) 40.70 71.88 [37.36, 104.70] a 0.97 277/514
  Mental wellbeing 23.62 0.28 [−0.87, 1.35] 0.07 279/519
  EQ-5D-5L 10.46 −0.83 [−1.59, −0.08] a 0.20 276/512
  EQ VAS 68.37 −0.09 [−4.73, 4.48] 0.00 280/521
b) Final (12 months) follow-up (37 sites)
  Sitting mins (daily) 577.20 −73.30 [−174.99, 31.31] 0.34 139
  Fear of falling 5.93 −1.33 [−2.61, −0.04] a 0.42 144
  METS (weekly) 1390.90 261.25 [−817.40, 1306.37] 0.15 139
  Mod/Vig mins (weekly) 99.24 44.13 [−80.61, 171.63] 0.16 142
  Sport mins (weekly) 29.08 68.96 [22.14, 118.70] a 0.93 141
  Mental wellbeing 23.90 0.38 [−1.26, 1.99] 0.09 142
  EQ-5D-5L 10.31 −0.75 [−1.80, 0.38] 0.18 140
  EQ VAS 75.82 −5.96 [−14.14, 1.96] 0.30 143
Open in a new tab

aconfidence interval (CI) does not contain zero.

Table 1(b) shows results at final follow-up (12 months) where there were statistically significant differences for fear of falling and sport minutes in the last 7 days. Effect sizes remained similar to those for repeated measures for sitting minutes for day, sport minutes per week, mental wellbeing, and health quality of life (EQ-5D-5 L); effect sizes were larger at final follow-up for fear of falling and lower for both METS and moderate and vigorous minutes of activity per week indicating that, except for sport, the difference between the intervention and control group for PA measures reduced over time.

Differences shown in Table 1 between the intervention and control at follow-up were all in a “beneficial” direction for the intervention group, except for health-related quality of life (EQ-5D-VAS).

There were improvements on all six functional fitness measures in the intervention when compared to the control for repeated measures and final follow-up at 12 months (Table 2). However, this difference was only statistically significant for arm curls (repeated measures) over all three follow-ups for this smaller subsample. There were no significant differences for standing balance.

Table 2.

Function Fitness and Standing Balance Results.

Outcome Intercept Beta [95% CI] Cohen’s d N
a) Repeated measures modeled over all three follow-ups (24 sites for FFT, 13 for SBT)
  FFT: Arm curls 13.18 2.75 [0.43, 5.16] a 0.39 77/147
  FFT: Chair stand 9.05 0.29 [−1.59, 2.13] 0.08 65/108
  FFT: Two-minute step 47.58 5.00 [−14.70, 25.02] 0.16 41/65
  FFT: Sit and reach −7.90 3.20 [−1.67, 8.17] 0.33 84/157
  FFT: Back scratch −14.24 2.09 [−1.89, 6.15] 0.20 71/134
  FFT: Up and go 12.85 −0.87 [−2.27, 0.49] 0.14 81/148
  SBT: Pathlength EO 84.27 −7.51 [−20.47, 6.21] 0.05 50/81
  SBT: Pathlength ES 103.73 −6.98 [−21.76, 7.42] 0.05 47/76
b) Final (12 months) follow-up (16 sites for FFT, 8 for SBT)
  FFT: Arm curls 14.72 2.88 [−1.46, 7.07] 0.41 38
  FFT: Chair stand 9.46 0.71 [−2.89, 4.32] 0.19 25
  FFT: Two-minute step 64.11 2.81 [−33.92, 40.47] 0.09 13
  FFT: Sit and reach 1.36 3.02 [−4.32, 10.35] 0.31 38
  FFT: Back scratch −13.13 3.97 [−4.83, 12.99] 0.37 36
  FFT: Up and go 12.64 −1.39 [−2.86, 0.11] 0.22 36
  SBT: Pathlength EO 78.33 −1.59 [−18.16, 15.15] 0.01 28
  SBT: Pathlength ES 100.92 3.89 [−14.91, 22.42] 0.02 27
Open in a new tab

FFT: Function Fitness; SBT: Standing Balance.

aconfidence interval (CI) does not contain zero.

Process Measures

Overall, 595 residents attended at least one session during the 10-week delivered program. The percentage of residents engaging in at least one session of the delivered program at the three main housing providers was 33.1% for the non-LA Sheltered Housing provider, 21.5% for LA Sheltered Housing provider and 26.4% for the Care housing provider. Numbers engaging at individual group homes ranged between 2.8% and 85.7%. Overall, 28.9% of residents took part in at least one session.

The percentage inactive at baseline in the sample (<30 MVPA per week) was higher than national data (Sport England, 2017) for age groups 65–75 (46.3% vs. 28.9%) and 75–85 (56.3% vs. 47.2%), and similar for those 85+ (66.7% vs. 70.2%). Participants had poorer health-related quality of life scores (EQ-5D-5 L) at baseline in this study than national data, with differences being greater for those 65–74 than for those 75 (NHS Digital, 2019).

Residents in the intervention group completed on average 6.35 sessions. Table 3 shows the results of a regression analysis for the number of sessions attended. Participation in sport at baseline was a statistically significant predictor of participation in the program, with each additional hour of sport at baseline associated with attendance at 0.41 additional sessions (minutes of sport have been converted to hours to ease interpretation).

Table 3.

Results for the Number of Sessions Attended by the Intervention Group (n: 242).

Variable Beta [95% CI]
Intercept (adjusted mean) 6.40 [5.35, 7.43]
Gender 0.13 [−0.70, 0.98]
Winter delivery −0.64 [−2.03, 0.76]
Age 0.03 [−0.01, 0.06]
Hours of sport at baseline 0.41 [0.07, 0.76] a
Health quality of life (EQ-5D) at baseline −0.07 [−0.16, 0.01]
Care home −1.26 [−4.25, 1.58]
LA sheltered housing −3.34 [−6.98, 0.12]
Open in a new tab

aconfidence interval (CI) does not contain zero.

Eight months after the program ceased (and therefore 2 years and 8 months from its commencement), the program deliverer reported that sports sessions continued regularly as follows: Non-LA Sheltered Housing 57% (sustained by residents), LA Sheltered Housing 13% and Care 100% (sustained by staff). This equates to 50% of sites overall. In addition to this, some other sites continued to offer PA sessions sporadically.

Discussion

Mobile Me was developed by Active Norfolk to engage inactive, older people with health conditions living in group homes in PA and to reduce sedentary behavior. This was a low-threshold intervention (Cichocki et al., 2015) that involved low-intensity PA and that was intended to be self-sustaining within the group homes after the 10-week intervention.

Statistically significant findings were that being in the intervention group was associated with improved results at 6 months for the primary outcome of time spent sitting, as well as fear of falling, participation in sport, moderate/vigorous PA, health quality of life, and arm curls. The EQ-5D-DL questionnaire which is used to measure health quality of life includes domains are relevant to independent living such as mobility; this is important within group homes as it is likely to be associated not only with the individual’s quality of life, but with the need for support (Singh et al., 2018). From 6 months to 12 months, there was reduction in the size of the difference between the groups for moderate/vigorous PA and METS, but the differences between the control and intervention groups for other measures remained relatively stable, although only statistically significant for sport and fear of falling.

The largest effect size was for minutes of weekly sports, which might be expected given the nature of the intervention which involved 2 hours of bowling, or other low-intensity PA per week. The IPAQ-E which was used in this study to measure PA does not record low-intensity PA even though it is intended for use in elderly populations. Without the inclusion of a Sport England question which enabled low-intensity sporting activity to be captured, we are likely to have under-estimated change in participation in sport in this population. It is not uncommon for PA self-report tools to not include low-intensity PA, for example, Sport England Active Lives Survey which is used for surveillance of PA in England (Ipsos MORI, 2021). This may be a concern where national guidance suggests that low-intensity PA is of benefit to some adults. It is recognized that low-intensity activity can be challenging to measure using self-report, although attempts using a daily logging approach have shown good validity when compared to device-based measures (Barwais et al., 2014).

For the 10-week, delivered program, levels of attendance were not statistically significantly associated with age, gender, season or health quality of life at baseline. This suggests that activities were accessible to those with mobility problems, or poor mental, or physical health. Participation was statistically significantly associated with sport at baseline which indicates that those individuals with an existing interest, or habit of engaging in sporting activities were more likely to attend more sessions; indeed a history of participation in sport has previously been found to predict current participation in sport for older adults (Jenkin et al., 2017). We do not have information to assess whether those that did not access the program had different characteristics to those that attended; collecting data from individuals who do not wish to engage in a program is challenging.

Much of the literature on PA interventions for the elderly focus on outcome measures, such as increases in PA levels, rather than on how successfully such interventions recruit their target audience in real-world circumstances (Olanrewaju et al., 2016). However, without engaging target audiences change cannot be achieved (Allen, 2018) therefore we have reported levels of participation, alongside outcome measures. Overall, around 29% of residents from the three main housing providers participated in the program. However, levels of participation varied not only by housing provider, but also by individual site. This suggests that factors other than program design influenced residents’ participation; this could be for example, the way housing sites are staffed and managed. While, overall, levels of participation in the program appear high, a lack of similar data from other studies prevented comparisons from being made.

We also found that the program successfully engaged with the target population, as participants were found to be in poor health and relatively inactive compared to the national population, particularly those in “younger older age” (<85 years). It seems likely that interventions in group homes have greater potential to reach target audiences than those in the community because the target population is less dispersed, because housing staff can support engagement, and also because practical barriers such as transport (Franco et al., 2015) can be removed.

The selection of a low-threshold approach (Cichocki et al., 2015), whereby activities did not require delivery by a professional and could take place in a typical communal space in a group home, enabled the activities to be sustained by residents in sheltered housing sites. The program was sustained in 57% of non-LA social housing sites and in 13% of LA sheltered housing sites 8 months after the rolling program ended. As with levels of participation in the program, differences in how the program was sustained are likely to be related to issues such as staffing levels and organizational priorities. We know anecdotally for example that staffing levels within LA sheltered housing were lower than in non-LA sheltered housing. These were reduced further in both types of provider part-way through the program due to changes to funding, demonstrating how real-world interventions may be affected by external factors outside of the control of deliverers.

The program was sustained at all sites with the residential care provider. A contributing factor for this was that further funding allowed the residential care provider to bring about organizational change and embed PA provision within day-to-day practice. This approach was more feasible within a residential care setting due to higher levels of staffing and higher levels of involvement in residents’ lives.

In terms of the evaluation, poor health, frailty, and disability within the target audience presented challenges when collecting data. Some participants were not able to perform all, or some of, the Fullerton Functional Fitness Tests, or the standing balance tests, due to issues with their mobility or confidence. Some residents lacked capacity to consent and could not therefore be included in the main evaluation; however, a separate study was carried out with these individuals using structured observation (Burke et al., 2021). Further limitations were a lack of randomization to the clusters, the inclusion of two rather than three follow-ups in the control group, the fact that the behavioral outcome measures were based on self-report, and small sample sizes for subgroup measures. These limitations are due to this being a pragmatic evaluation which was designed to fit around the realities of program delivery and with limited resources. Nevertheless, a strength of pragmatic evaluations of real-world interventions, developed and run by organizations that normally undertake this type of activity, are that they have greater claim to ecological validity (Robson, 2002) when compared to evidence generated within the controlled conditions of a clinical trial. Further limitations are than findings may not be generalizable across different populations and cultures.

Conclusion

Mobile Me demonstrates how a PA intervention can be developed to maximize the participation of older participants in group housing if known drivers and barriers to engagement are addressed, that is, where activities are sociable, inclusive to people of different abilities and health conditions, affordable or free, easy to travel to, perceived to be safe, and where housing staff are mobilized to encourage residents to take part. It shows that competitive sports should not be overlooked as a method of engagement, even for frail participants, and for the potential of low-threshold PA interventions to be sustained in group homes in the longer term. By low-threshold, we mean activities that not only place minimal demands on participants, but also on housing staff because they require minimal resources and expertise to deliver. In such cases, interventions such as the one described have the potential to offer inactive residents in poor health, who may be reluctant or concerned about participating, a “gateway” into PA. The evaluation also suggests that organizational culture and practice in group homes influences participation in, and the longer term sustainability of PA activities regardless of the intervention’s design.

The program did not aim to break up sedentary time and emerging evidence suggests that this may be important for health (DHSC, 2019); future programs could consider this, and explore how, once recruited, participants can either be supported to maintain PA levels, or be progressed to higher intensity activities where appropriate.

Acknowledgments

We are grateful to all those that helped with the evaluation, especially residents and staff at group homes, the Active Norfolk team that developed the intervention and collected and inputted data (particularly Ryan Hughes, Stephen Hitcham and Lucy Butcher), and the Mobile Me Steering group. We also acknowledge Sport England who funding this intervention and its evaluation through the ‘Get Healthy, Get Active’ funding stream.

Footnotes

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding: The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Sport England under the ‘Get Healthy, Get Active’ funding stream [grant number R21986].

ORCID iD

Amanda Burke https://orcid.org/0000-0002-7905-8947

References

  1. Allen B. (2018, June 21) What matters in wellbeing: Engagement vs outcomes. Oomph Wellness. https://oomph-wellness.com/2018/06/21/what-matters-in-wellbeing-engagement-vs-outcomes/ [Google Scholar]
  2. Barwais F. A., Cuddihy T. F., Washington T., Tomson L. M., Brymer E. (2014). Development and validation of a new self-report instrument for measuring sedentary behaviors and light-intensity physical activity in adults. Journal of Physical Activity & Health, 11(6), 1097–1104. 10.1123/jpah.2012-0118 [DOI] [PubMed] [Google Scholar]
  3. Bates D., Mächler M., Bolker B., Walker S. (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67(1), 1–48. 10.18637/jss.v067.i01 [DOI] [Google Scholar]
  4. Burke A., Jones A., Hughes R., Player E. (2021). From evidence to practice: Developing best practice guidelines for the delivery of activities to people living with moderate to advanced dementia using a pragmatic observational study. Dementia, 20(5), 1604–1616. 10.1177/1471301220957805 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Cichocki M., Quehenberger V., Zeiler M., Adamcik T., Manousek M., Stamm T., Krajic K. (2015). Effectiveness of a low-threshold physical activity intervention in residential aged care – results of a randomized controlled trial. Clinical Interventions in Aging, 10, 885–895. 10.2147/CIA.S79360 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Crane M., Bauman A., Lloyd B., McGill B., Rissel C., Grunseit A. (2019). Applying pragmatic approaches to complex program evaluation: A case study of implementation of the New South Wales get healthy at work program. Health Promotion Journal of Australia, 30(3), 422–432. 10.1002/hpja.239 [DOI] [PubMed] [Google Scholar]
  7. Des Jarlais D. C., Lyles C., Crepaz N., the TREND Group. (2004). Improving the reporting quality of nonrandomized evaluations of behavioral and public health interventions: The TREND statement. American Journal of Public Health, 94(3), 361–366. 10.2105/AJPH.94.3.361 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. DHSC . (2019). UK chief medical officers’ physical activity guidelines. [Google Scholar]
  9. DiPietro L., Al-Ansari S. S., Biddle S. J., Borodulin K., Bull F. C., Buman M. P., Cardon G., Carty C., Chaput J.-P., Chastin S., Chou R., Dempsey P. C., Ekelund U., Firth J., Friedenreich C. M., Garcia L., Gichu M., Jago R., Katzmarzyk P. T., Willumsen J. F. (2020). Advancing the global physical activity agenda: Recommendations for future research by the 2020 WHO physical activity and sedentary behavior guidelines development group. International Journal of Behavioral Nutrition and Physical Activity, 17(1), 1–11. 10.1186/s12966-020-01042-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Eime R., Harvey J., Charity M., Casey M., van Uffelen J., Payne W. (2015). The contribution of sport participation to overall health enhancing physical activity levels in Australia: A population-based study. BMC Public Health, 15(1), 806. 10.1186/s12889-015-2156-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Feise R. J. (2002). Do multiple outcome measures require p-value adjustment? BMC Medical Research Methodology, 2(1), 8. 10.1186/1471-2288-2-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Franco M. R., Tong A., Howard K., Sherrington C., Ferreira P. H., Pinto R. Z., Ferreira M. L. (2015). Older people’s perspectives on participation in physical activity: A systematic review and thematic synthesis of qualitative literature. British Journal of Sports Medicine, 49(19), 1268–1276. 10.1136/bjsports-2014-094015 [DOI] [PubMed] [Google Scholar]
  13. French D. P., Olander E. K., Chisholm A., Mc Sharry J. (2014). Which behaviour change techniques are most effective at increasing older adults’ self-efficacy and physical activity behaviour? A systematic review. Annals of Behavioral Medicine, 48(2), 225–234. 10.1007/s12160-014-9593-z [DOI] [PubMed] [Google Scholar]
  14. Goodwin V., Orr N., Sharpe R., Smith J., Phoenix C., Lang I., Garside R. (2017). Physical activity: How well does systematic review evidence on physical activity interventions reflect the expressed views of older people? Physiotherapy, 103, 22–23. 10.1016/j.physio.2017.11.179 [DOI] [Google Scholar]
  15. Herdman M., Gudex C., Lloyd A., Janssen M., Kind P., Parkin D., Bonsel G., Badia X. (2011). Development and preliminary testing of the new five-level version of EQ-5D (EQ-5D-5L). Quality of Life Research, 20(10), 1727–1736. 10.1007/s11136-011-9903-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hurtig-Wennlf A., Hagstrmer M., Olsson L. A. (2010). The international physical activity questionnaire modified for the elderly: Aspects of validity and feasibility. Public Health Nutrition, 13(11), 1847–1854. 10.1017/S1368980010000157 [DOI] [PubMed] [Google Scholar]
  17. Huurnink A., Fransz D. P., Kingma I., van Dieën J. H. (2013). Comparison of a laboratory grade force platform with a Nintendo Wii Balance Board on measurement of postural control in single-leg stance balance tasks. Journal of Biomechanics, 46(7), 1392–1395. 10.1016/j.jbiomech.2013.02.018 [DOI] [PubMed] [Google Scholar]
  18. Ipsos MORI . (2021). Active lives survey: Mid-year 6 technical note. https://sportengland-production-files.s3.eu-west-2.amazonaws.com/s3fs-public/2021-10/Active-Lives-Survey-2020-2021-Year-6-%28mid-year%29-technical-note..pdf?VersionId=K1aIIdJ.kD2ZQk3wrJmAUi2crMwGZWdX [Google Scholar]
  19. Jenkin C. R., Eime R. M., Westerbeek H., O’Sullivan G., van Uffelen J. G. Z. (2017). Sport and ageing: A systematic review of the determinants and trends of participation in sport for older adults. BMC Public Health, 17(1), 976. 10.1186/s12889-017-4970-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Lewis L. K., Henwood T., Boylan J., Hunter S., Lange B., Lawless M., Milte R., Petersen J. (2021). Re-Thinking reablement strategies for older adults in residential aged care: A scoping review. BMC Geriatrics, 21(1), 667. 10.1186/s12877-021-02627-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Loy A., Steele S., Korboval J. (2021). lmeresampler: Bootstrap methods for nested linear mixed-effects models. https://CRAN.R-project.org/package=lmeresampler [Google Scholar]
  22. McPhee J. S., French D. P., Jackson D., Nazroo J., Pendleton N., Degens H. (2016). Physical activity in older age: Perspectives for healthy ageing and frailty. Biogerontology, 17(3), 567–580. 10.1007/s10522-016-9641-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Morris S. B. (2008). Estimating effect sizes from pretest-posttest-control group designs. Organizational Research Methods, 11(2), 364–386. 10.1177/1094428106291059 [DOI] [Google Scholar]
  24. NHS Digital . (2019). Health survey for England 2018: Summary report. https://digital.nhs.uk/data-and-information/publications/statistical/health-survey-for-england/2018/summary#adult-physical-activity [Google Scholar]
  25. O’Connell A. A., McCoach D. B. (2004). Applications of hierarchical linear models for evaluations of health interventions: Demystifying the methods and interpretations of multilevel models. Evaluation & the Health Professions, 27(2), 119–151. 10.1177/0163278704264049 [DOI] [PubMed] [Google Scholar]
  26. Olanrewaju O., Kelly S., Cowan A., Brayne C., Lafortune L. (2016). Physical activity in community dwelling older people: A systematic review of reviews of interventions and context. PloS One, 11(12), Article e0168614. 10.1371/journal.pone.0168614 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Paneru K., Padgett R. N., Chen H. (2018). Estimation of zero-inflated population mean: A bootstrapping approach. Journal of Modern Applied Statistical Methods, 17(1), Article eP2557. 10.22237/jmasm/1525133460 [DOI] [Google Scholar]
  28. Park D.-S., Lee G. (2014). Validity and reliability of balance assessment software using the Nintendo Wii balance board: Usability and validation. Journal of Neuroengineering and Rehabilitation, 11(1), 99. 10.1186/1743-0003-11-99 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. R Core Team . (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R-project.org/ [Google Scholar]
  30. Robson C. (2002). Real world research: A resource for social scientists and practitioner-researchers (2nd ed., p. 599). Blackwell Publishers. [Google Scholar]
  31. Scheffer A. C., Schuurmans M. J., VanDijk N., van der Hooft T., de Rooij S. E. (2010). Reliability and validity of the visual analogue scale for fear of falling in older persons. Journal of the American Geriatrics Society, 58(11), 2228–2230. 10.1111/j.1532-5415.2010.03105.x [DOI] [PubMed] [Google Scholar]
  32. Singh J., Pokhrel S., Longworth L. (2018). Can social care needs and well-being be explained by the EQ-5D? Analysis of the health Survey for England. Value in Health, 21(7), 830–838. 10.1016/j.jval.2018.01.002 [DOI] [PubMed] [Google Scholar]
  33. Sport England . (2017). Active Lives data tables: November 2017-18, tables 1-3 levels of activity. https://www.sportengland.org/know-your-audience/data/active-lives/active-lives-data-tables [Google Scholar]
  34. Tennant R., Hiller L., Fishwick R., Platt S., Joseph S., Weich S., Parkinson J., Secker J., Stewart-Brown S. (2007). The Warwick-Edinburgh mental well-being scale (WEMWBS): Development and UK validation. Health and Quality of Life Outcomes, 5(1), 1–13. 10.1186/1477-7525-5-63 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Waguespack D., Krishnamoorthy K., Meesook L. (2020). Tests and confidence intervals for the mean of a zero-Inflated Poisson distribution. American Journal of Mathematical and Management Sciences, 39(4), 383–390. 10.1080/01966324.2020.1777914 [DOI] [Google Scholar]
  36. WHO . (2020). WHO guidelines on physical activity and sedentary behaviour. World Health Organisation. https://www.who.int/publications/i/item/9789240014886 [Google Scholar]
  37. Zubala A., MacGillivray S., Frost H., Kroll T., Skelton D. A., Gavine A., Gray N. M., Toma M., Morris J. (2017). Promotion of physical activity interventions for community dwelling older adults: A systematic review of reviews. Plos One, 12(7), Article e0180902. 10.1371/journal.pone.0180902 [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Journal of Applied Gerontology are provided here courtesy of SAGE Publications

RESOURCES