Effects of Exergaming on Physical and Cognitive Outcomes of Older Adults Living in Long-Term Care Homes: A Systematic Review

As the global older adult (OA) population increases rapidly in the coming decades, the demand for residential long-term care (LTC) is expected to rise [1]. OAs residing in residential LTC homes (also called nursing homes) are characterized by increasing functional decline [2], often measured by deterioration in mobility and activities of daily living (ADLs). This population is complex, as functional decline can be brought on by a myriad of factors including physical aging, comorbidities [3, 4], cognitive impairment (e.g., dementia due to Alzheimer’s or cerebrovascular disease) [5], and reductions in vision, hearing, and proprioceptive senses [6, 7]. The sedentary nature of life in LTC homes [8] also significantly increases the likelihood of further declines in cognitive and functional abilities soon after admission [9]. The consequences of these declines are reflected in the experiences of OAs living in LTC, wherein poor quality of life (QoL) and neglect are often reported [10, 11].

Among various interventions that have been designed to address, maintain, and improve the physical and cognitive health of LTC residents, exergaming holds promise by harnessing novel technologies to improve activity levels [12]. Exergaming is described as interactive exercise-based games whereby players engage in physical and cognitive activities played on a technology-based gaming system. Current literature suggests that exergames have positive social, cognitive, and physical effects [13, 14], but vary markedly in intervention type and outcome data collected [15]. The mechanism of the benefits of exergaming is supported by the Cognitive Enrichment hypothesis [16] which states that the collective behaviors of an individual have a meaningful positive impact on cognitive and functional ability in old age. Another systematic review established the positive effects of exergaming on cognitive function among OAs living with mild cognitive impairment or dementia from a variety of settings, which include the community, hospitals, rehabilitation wards, and nursing homes [17]; however, it is unclear how exergaming interventions impact the physical and cognitive health of OAs residing in LTC. Thus, we conducted a systematic review of quantitative studies to summarize the effects of exergaming interventions on physical, cognitive, and QoL outcomes of OAs (>65 years of age) living in LTC.

Our systematic review was guided by the PRISMA statement [18].

We searched six databases: CINAHL, PubMed, Web of Science, PsycINFO, ScienceDirect, and Cochrane. Searches were not limited by date, with the last search conducted in July 2020. The search strategy was developed for CINAHL in line with the Population, Intervention, Comparator, Outcome, and Study type framework; search terms were selected to capture four concepts: OAs, technology-based games or exergames, physical activity (PA), and LTC. The full CINAHL search strategy is included in the online supplementary material (see www.karger.com/doi/10.1159/000521832 for all online suppl. material). The search was translated into the other databases with the appropriate syntax and index terms for each database. Reference lists of included articles were hand-searched to identify additional records.

Inclusion and exclusion criteria were specified using the Population, Intervention, Comparator, Outcome, and Study type framework as outlined in Table 1. The results of our search were uploaded to Covidence, a web-based platform which facilitates the screening and data extraction process. Following exclusion of duplicate records, title and abstract screening were completed in duplicate (two teams of two) by 4 reviewers (A.Q., A.S., A.K., and A.Z.). Full-text review followed. Disagreements over inclusion of studies for data extraction were resolved through discussion or feedback from the senior author (C.C.).

The aforementioned reviewers (A.Q., A.S., A.K., and A.Z.) conducted data extraction independently using a predesigned data extraction form. Data were collected regarding study setting, country and design, participant characteristics, sample size, intervention characteristics (duration, frequency), gaming system used (e.g., Nintendo Wii, Xbox Kinect), adherence, physical and cognitive outcomes, and QoL. Data on comorbidities, namely cognitive disorders, were also collected.

The quality of the included studies was assessed using the appropriate tools based on study design: the Cochrane Risk of Bias tool (RoB-2) for randomized trials or the Risk of Bias in Non-randomized Studies (ROBINS-I) [19, 20]. Using the RoB-2 tool, risk of bias was classified as “high,” “low,” or “unclear” based on seven items: sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and other biases. Similarly, the ROBINS-I tool assesses risk of bias based on seven items: confounding, selection bias, bias in classification of interventions, bias due to deviations from intended interventions, bias due to data missingness, bias in measurement of outcomes, and bias in selection of the reported results.

Given the heterogeneous nature of study designs and outcomes presented in the selected studies, a descriptive synthesis was conducted. Descriptive statistics across studies regarding age, number of sessions, and the volume of therapy (number of sessions × duration) were reported in terms of mean ± standard deviation. In addition, we report results of null hypothesis significance tests of within- and between-group differences and effect sizes where applicable.

The database and reference list search produced 480 records (Fig. 1). One-hundred and seventy-four duplicate records were removed. After screening of the titles and abstracts of 306 records, 192 records were deemed to be irrelevant based on our inclusion and exclusion criteria. With the remaining 114 records, full-text screening was conducted to assess for eligibility, wherein 93 records were excluded. Twenty-one studies were ultimately included for data extraction. Interrater reliability of trial selection ranged from 0.85 to 1.0 across pairs of reviewers. The included studies were published between 2009 and 2020. Included studies were conducted in North America (n = 3), Europe (n = 10), Oceania (n = 3), South America (n = 1), and Asia (n = 4). With respect to study design, studies included nine randomized control trials [21-29], two nonrandomized control study [30, 31], four before-and-after studies [32-35], five quasi-experimental studies [36-40], and one interrupted time series study [41]. Details of the included studies are presented in Table 2, and summarized results from studies with a control group are provided in Table 3.

A total of 657 OAs participated in the included 21 studies, of which 275 participants (42%) were controls. Group sample size ranged from 5 to 32 participants across studies. Both males and females were included across all studies. Not all studies reported on the gender composition of participants, however, the majority of studies (n = 17) had a higher proportion of women in comparison to men. The average age ranged from 70.1 to 90.4 years of age. Common study inclusion criteria were: (1) cognitive ability to understand the game and instructions given, (2) ability to stand or walk independently with or without aid, (3) ability to communicate based on researcher judgment, and (4) absence of cognitive impairment based on Mini-Mental Status Examination (MMSE) score [cutoffs differed across studies, with ≤15 being the lowest range point for cognitive impairment and ≤27 being the highest possible cutoff for cognitive impairment [32]. However, studies differed by the degree of functionality and assessment scores at baseline, as well as measures used to assess.

Participants generally were ambulatory. Baseline mobility characteristics of participants were reported in 52% (n = 11/21) of studies, with mobility challenges frequently reported, marked by the use of a walking aid, wheelchair, or falls in the last 6 months or year. Other physical characteristics were inconsistently reported between studies; e.g., separate studies reported on BMI [29] and hypertension [41]. Five studies identified included participants with cognitive impairment, with 3 studies noting inclusion of participants with cognitive impairment based on MMSE or Montreal Cognitive Assessment (MoCA) scores [21, 22, 37], and another 2 studies that enrolled participants with mild or moderate dementia [33, 36].

Participant withdrawal was only reported on in 11 of the 21 included studies; 6 studies reported withdrawal from both intervention and control groups [24, 29, 35, 37-39], 2 studies reported withdrawals from only the control group [25, 36], and 3 studies reported withdrawals from only the intervention group [23, 30, 31]. Overall, intervention adherence ranged from 55% to 100%. Most studies did not find pretest differences between the intervention and control groups. Eight studies reported on adverse events [21, 23-25, 30, 31, 35, 39] wherein adverse events were either reported to not have occurred or included medical conditions unrelated to the program, such as serious illness, osteoarthritis, hip fracture, musculoskeletal pain, and death; no studies reported exergaming-related adverse events.

Intervention sessions occurred at least twice a week for 76% (n = 16/21) of studies (range 1–5 sessions per week), and most intervention durations were more than 4 weeks long (n = 13/21, 62%; range = 1–24 weeks). One study measured acute changes following a single session [26]. Length of exergaming sessions was variable across studies with the most common duration being approximately 30-min sessions (n = 7/21; 33%) ranging from 5 to 120 min. Eighty-one percent of studies (n = 17/21) used commercially available exergaming hardware (i.e., Nintendo Wii, Xbox Kinect). However, not all studies used the commercially available corresponding software (i.e., Wii Fit); some exergaming interventions involved noncommercially available games such as StepMania [29] and Jintronix [41] which were designed with OA end users in mind. Only 1 study used virtual-reality hardware and software [33]. All the intervention/exergaming sessions were supervised by either a member of the research staff (n = 7), a physical therapist (n = 5), an exercise therapist (n = 3), nursing staff (n = 2), or volunteer (n = 1).

Twenty (95%) of the 21 included studies [21-25, 27-31, 33-41] cite reported physical outcome measures.

Eight of the included studies [21, 23, 24, 31, 35, 37, 40, 41] measured mobility and endurance. Across functional tests, a medium to large effect size (d = 0.55–1.01) was reported in favor of the exergaming intervention. These studies included Hsieh et al. [37] who reported significant improvements in the intervention compared to a control group for the 6-min walk test (p = 0.01, d = 0.55), measuring mobility and functionality, and the 30-s sit-to-stand test (p = 0.002), which measures endurance. Three studies reported statistically significant (p < 0.05) improvements in mobility in the exergaming group compared to a control group when using a standard clinical test – the timed up and go (TUG) test [23, 31, 35]. Two of these studies included conventional exercise comparison groups: Cicek et al. [31] noted mobility improvements for exergaming above the effects of conventional exercise, while Yesilprak et al. [23] reported moderate to large (0.69–1.03) within-group effect sizes for both types of exercise, which did not differ from each other. Delbroek et al. [21] also reported significant improvements within the intervention group for TUG duration and the turn-to-sit duration during single-task walking, indicating improved mobility (p = 0.02). In the interrupted time series study by Valiani et al. [41], there was a significant improvement (p < 0.05) in 4-meter walking speed and a reduction in time in completing the sit-to-stand test, indicating an improvement in mobility and endurance. Lastly, Taylor et al. [24] reported improved mobility among residents who played exergames, as measured by the de Morton Mobility Index (DEMMI), however this did not reach significance (p = 0.06).

Eight studies [23, 27, 30, 31, 34, 35, 37, 40] reported balance outcomes with mixed results. When administering the Berg Balance Scale, Cicek et al. [31] reported a statistically significant (p < 0.05) improvement among both exergaming and conventional exercise groups compared to a usual care control group. Yeşilyaprak et al. [23] reported improvements among both exergaming and conventional balance groups, thus, exergaming was not more effective at improving balance compared to the control group (see also Janssen et al. [30] for null findings compared to the control). Portela et al. [27] reported improved balance only in a control group with traditional exercise (p = 0.006), and Ellmers et al. [34] reported a stronger alignment between postural control and balance capabilities perceived by participants post-intervention (p < 0.001). Hsieh et al. [37] reported significantly improved (p < 0.05) functional reach among the exergaming group at the 3-month and 6-month time points; effect sizes of 0.5 (3-month) and 1.01 (6-month) favored the intervention over the control. Wu et al. [35] reported no difference in balance post-intervention, as indicated by Berg Balance Scale scores, and Keogh et al. [39] reported no significant quantitative improvements in dynamic balance as indicated by the Four Square Step Test.

Three of the included studies reported on gait [29, 37, 38], and significant improvements in gait characteristics were reported in each. Pichierri et al. [29] reported greater gait velocity and improved single support time during a fast walking dual-task condition for the intervention group relative to baseline and a control group. Ogawa et al. [38] reported modest, statistically significant (p < 0.05) improvements in single-task gait measures but not dual-task gait measures in the intervention group. Hsieh et al. [37] reported a significant improvement (p < 0.05) in 5-m gait speed of the exergaming group; effect sizes of 0.23 and 0.60 at 3- and 6-month timepoints represent improvement over time and favor the intervention group over the control group.

Reaction time was evaluated using the Vienna test [36], and the Grasping ruler test [22]. Of these 2 studies, only Liu et al. [22] reported a significant improvement (p < 0.05) in reaction time among the exergaming intervention group but did not analyze whether this change was significant compared to the control group. Chiang et al. [36] found significantly better post-test reaction time scores in the intervention group.

General level of PA was discussed in 5 of the included studies [25, 28, 30, 39, 41]. The studies collectively reported statistically significant improvements in physical well-being and PA levels according to questionnaire-based measures such as the Rapid Assessment of Physical Activity (RAPA) [39, 41], the LAPAQ (LASA Physical Activity Questionnaire) [30], and the Short Physical Performance Battery Test [25, 28, 41].

Five of the 21 studies [21, 26, 27, 37, 38] reported on cognitive outcomes that were measured through validated measures such as MMSE, MoCA, the Verbal Fluency Test, Digit Span Forward, Digit Span Backward, Trail Making Test (TMT) A and B, and the Cognitive Abilities Screening Instrument (CASI). Ogawa et al. [38] reported better performance following exergaming for MMSE and TMT-B scores from pre-to post-test (p < 0.05), and change scores that favored the intervention group for TMT-A psychomotor speed. Hsieh et al. [37] reported no significant improvements in overall cognitive abilities as measured by the CASI or subscales examining specific cognitive abilities, with the exception of improvements on one abstract reasoning subscale (p = 0.002) at 6 months. Monteiro-Junior et al. [26] reported significant within-subject improvement on the Verbal Fluency Test after a single exergaming session, and effect size between groups (d = 0.63) indicated moderate effect of the exergame. Other studies found no significant group differences or pre-post changes in MMSE [27] or MoCA [21].

QoL was reported on in 6 studies with varying results [22, 27, 31, 39, 40, 42]. The World Health Organization Quality of Life (WHOQOL) Instrument was used in 3 studies: one reported significant improvements in the intervention group compared to the control group for psychological QoL only [39], but another study reported no significant improvements in QoL in any of their three intervention groups [32]. Similarly, Cicek et al. [31] administered the WHOQOL-BREF, and found no statistically significant within-group or between-group differences. Using the SF-8 health-related QoL questionnaire, Liu et al. [22] reported no significant improvements in health-related QoL in the exergaming intervention group. In contrast, Portela et al. [27] used the SF-36 health-related QoL questionnaire and observed significantly improved vitality (p = 0.007) and mental health-related QoL (p = 0.023) in the unsupervised exergaming group, as well as improved physical functioning (p = 0.024) in the supervised exergaming group. In contrast, Keogh et al. [40] found no significant within-group change in the SF-36 QOL scores.

Only 17% (n = 3) of studies mentioned cost of the exergaming intervention: these studies used commercially available products (i.e., Wii, Xbox Kinect, Oculus Rift products) and referred to such technology as “low cost” or “affordable” [21, 33, 39]. Cited barriers to the implementation of exergaming technology in LTC included a lack of space in the LTC home [23, 39] and the time required for staff to introduce a new activity to residents who are often unfamiliar with the technology [23, 41]. One study listed resident-reported barriers to participation in exergaming interventions which included: self-reported “laziness,” lack of interest, feeling it was a “hassle,” and physical health problems (p. 149) [35].

Nine studies were assessed using the RoB-2 tool [21-29]. Figure 2 provides details of each study’s assessments by RoB domain. Generally, there was a low risk of bias for sequence generation, allocation concealment, and data missingness. Given the nature of exergaming interventions, there was a high risk of bias for blinding of study participants and outcomes assessors. There was also a high risk of bias due to selective outcome reporting by one-third of the evaluated studies.

Twelve studies were assessed using the ROBINS-I tool [30-41]. Figure 3 provides details of each study’s assessment by ROBINS-I domain. In general, there was a low risk of bias for selection of study participants, classification of interventions, data missingness, and selection of the reported results. As expected, there was a moderate to serious risk of bias across all studies for measurement of outcomes as participants were not, and generally could not be, blinded due to the nature of exergaming interventions. The ROBINS-I tool detailed guide indicates that if the study is judged to be at serious risk of bias in at least one domain, the study’s risk of bias overall should be judged as serious risk. As such, 75% of studies (n = 9 [31-36, 39-41] were judged as having an overall serious risk of bias.

This systematic review aimed to summarize evidence regarding the effectiveness of exergaming interventions on the physical, cognitive, and QoL outcomes of OAs (>65 years of age) living in LTC. To the best of our knowledge, this systematic review is the first to consider the overall effect of exergaming interventions on the physical health, cognitive abilities, and QoL of this complex population. The analysis combined 21 studies involving a total of 657 participants. We observed wide variability of exergaming interventions, which promoted diverse types of PA (e.g., low intensity, coordination, dance) on different gaming systems. The majority (66%, n = 14/21) of studies were associated with a serious risk of bias due to lack of randomization, uncontrolled study designs, and lack of blinding due to the nature of the intervention. Only nine of the 21 studies included were randomized control trials, and four of the studies [32-34, 41] lacked a control group. The sample sizes within the studies were also relatively small (ranging from 8 to 65 participants). These limitations across research designs may be because many of the studies identified themselves as feasibility or pilot investigations mostly lasting 2 months or less (15/21 studies; 71%) with the goal of providing preliminary evidence. Taken together, the current research base lacks conclusive evidence of a positive benefit of exergaming on physical, cognitive, and QoL outcomes for OAs in LTC but does suggest promise in particular domains.

Preliminary evidence that supports possible physical health benefits of exergaming was found in 18 out of 21 studies [21, 22, 24, 25, 27-31, 33-41]: these included observational improvements within a group of exergamers (no control group) and between-group differences in favor of the exergaming group over a control group. Among the studies that measured mobility and endurance outcomes (n = 8), 88% reported statistically significant improvements within the exergaming group, indicating a potential effect of exergaming interventions on OA’s mobility and endurance. In addition, all the included studies that measured gait (n = 3) also reported statistically significant improvements among the exergaming intervention group. This preliminary evidence for physical benefits of exergaming in LTC residents is consistent with stronger evidence of exergaming benefits among broader populations of OAs [42], but evidence from higher quality studies is needed.

Some additional caution is required to interpret the effects of exergaming on cognitive and QoL outcomes, which were more inconsistent across studies. Five studies with a control group included cognitive outcomes but only two reported improvements relative to the control group, and these benefits were limited to specific cognitive domains (i.e., psychomotor speed, abstract reasoning) [38]. In addition to high risk of bias across studies, the use of different cognitive measures (e.g., MMSE, CASI) hindered our ability to conclude the effect of exergaming on cognitive effects in this population. Effects on QoL were similarly inconsistent across studies and measures; of the 2 studies that tested between-group differences, one found improvements for exergaming relative to a control condition for psychological QoL only [40], and the other found no difference [23].

Despite incomplete evidence for positive improvements related to exergaming, it is notable that no exergaming-related adverse events and injuries were reported. Negative effects on an exergaming group were limited to a measure of QoL (Emotional Performance) in a single study at high risk of bias [27] and may have reflected decline over time that is often seen in this population [10]. No other studies reported negative impacts of exergaming on physical, cognitive, or QoL outcomes. Studies either reported no change or an improvement in physical and cognitive performance of residents, suggesting the potential for exergaming to mitigate cognitive and functional declines that are common after admission to LTC [9]. The safety of exergaming among OAs has been reported in other literature reviews [13, 43].

Based on the results of this review, it is evident that further investigation into the effects of exergaming on physical, cognitive, and QoL outcomes for LTC home residents is required. In order to reach more definitive conclusions, it is recommended that future studies employ research designs associated with reduced bias, including RCTs, and recruit large enough samples of residents to provide sufficient statistical power. Future research would benefit from the use of standardized tests and tools to evaluate physical (e.g., TUG, 3MWT, 5-meter gait speed), cognitive (e.g., MMSE), and QoL (e.g., WHOQoL) outcomes, which would permit quantitative synthesis of findings using meta-analysis. The addition of objective measures of PA levels, such as step counts based on a Fitbit activity tracker may also be useful to evaluate whether exergaming helps reduce sedentary time for LTC residents. In addition, evaluations of exergame interventions with longer sessions and increased frequency of sessions per week (i.e., more than 2 per week) may be beneficial to demonstrate more robust effects. Finally, it is recommended that future studies consider investigating moderators of exergame intervention efficacy on OAs (see also [44]). Loneliness and depression are highly prevalent among LTC residents [45], but these were not commonly included in the studies and were only measured by Hsieh et al. [37]. Loneliness, depression, and anxiety can negatively impact resident QoL and have been associated with poor physical functional ability [46]. Future studies should consider these conditions and how they may moderate the effect of exergaming on physical, cognitive, and QoL outcomes.

It is worth noting that the participants enrolled in the included studies were higher functioning physically and cognitively compared to the general population of OAs living in LTC. For example, in Ontario, Canada, 86% of OAs in LTC are dependent (e.g., require extensive help with activities of daily living such as getting out of bed, eating, or toileting) and 64% have dementia [47]. Therefore, results of exergame trials may not be generalizable to the broader population of LTC home residents and there may be issues related to the implementation and adoption of exergames with residents who have lower levels of mobility and cognitive function. Of the studies that reported adherence (n = 12), adherence ranged from 55% to 100%; 67% of these studies (n = 8/12) reported at least 75% adherence. Where reported, common reasons for missing exergaming sessions were tiredness, loss of interest, and conflicting schedules [24]. Poor adherence and attrition can lead to nongeneralizable conclusions because the participants did not receive the intended intensity or dose of the intervention [48]. Usability of exergames influences their implementation, but only 2 studies reported on system usability [32, 49]. Only 1 study reported, using the System Usability Scale, that users were satisfied with the usability of the system [32]. However, results from Gerling et al. [49] indicated a large difference in perceived usability between exergaming experienced and inexperienced groups, thus indicating that previous exergaming experience may influence LTC residents’ willingness to participate in exergaming-delivered PA.

Enjoyable and motivating exergame platforms are important to increase transfer of results from exergaming studies into the LTC home context. There is a need to develop fun and engaging PA interventions since residents have reported negative perceptions and viewed exercise sessions as “boring” or “monotonous” with little interest in participating [50]. This suggests that currently available PA interventions do not take residents’ interests and needs into consideration. Exergaming interventions provide the opportunity to implement engaging PA in LTC; user-centered design is important for the design of accepted, useable and thus successful products. Domains for consideration in the design process include user enjoyment, accessibility, and usability. In addition, the majority of studies included in this review used commercially available exergames (i.e., Nintendo Wii, Xbox Kinect) [51], and while these gaming systems were accessible for research study purposes, some of these systems are discontinued, no longer in production, or not widely commercially available. Future exergame interventions developed for this population should employ the principles of user centered design [52], an iterative process that would allow the needs of LTC residents to be captured throughout the design process, and use validated measures such as the System Usability Scale [53], to measure residents’ perceptions of exergaming intervention usability. Systems developed through user centered design processes such as the MouvMat have been recently studied but are not yet commercially available [54].

This systematic literature review has some limitations. The studies included within this review were highly heterogeneous with respect to the console used, game software, impact of the exergame, and the outcomes assessed. For this reason, we were also unable to conduct meta-analyses. Therefore, conclusions reflect exergames for OAs in LTC more generally, and not a single exergaming intervention or training design.

Our review indicated that exergaming for residential LTC homes is a growing field of research, but existing evidence should be interpreted cautiously due to the heterogeneous nature of the interventions, uncontrolled designs, and small samples. Larger and more methodologically robust evaluations are required to mature the evidence base. Exergames might be a promising intervention to benefit the physical health of OAs (>65 years) living in LTC, but more research is required to determine the effects of exergaming on physical health, as well as cognitive and QoL outcomes. Cost-effectiveness analyses are also warranted as cost was identified as a barrier to exergaming systems implementation. There is also room for future co-design and development of exergaming technologies and systems that will consider the interests as well as the physical and cognitive needs of OAs to facilitate uptake and implementation.

We thank Atefeh Zare for her assistance in the search and screening process.

An ethics statement was not required for this study type, no human or animal subjects or materials were used.

The authors have no conflicts of interest to disclose.

Dr. Chu is supported by the Alzheimer Society of Canada New Investigator Award. This study was financed in part by grants held by Drs. Chu and Biss from the New Frontiers Research Fund (00693) and the Center for Aging and Brain Health Innovation (00279).

C.C., A.Q., and R.B. participated in all stages of this systematic literature review, from the design, extraction and interpretation of data, and final writing. A.S., A.K., and A.Q. extracted and analyzed the data. All authors participated in the writing and review of the final manuscript.

All data generated or analyzed during this study are included in this article and/or its online supplementary material files. Further enquiries can be directed to the corresponding author.