Technologies for active lifestyle in later life: design and evaluation
We describe the design of a system for home-based intervention that enables older adults with different abilities and indeed in spite of their different abilities, to engage in group exercises from home, via a tablet, while keeping these differences invisible to the group.
We then report on interventions run with dozens of subjects in different countries, reporting on the outcomes related to physical, psychological and social wellbeing.
Engaging in physical activity can bring multiple benefits to the health and well-being of older adults. It reduces risk of falls, slows progression of degenerative diseases, and even improves cognitive performance and mood .
There are however a variety of barriers that make it difficult for older adults to engage in regular physical activity: lack of adequate facilities and infrastructures, reduced functional abilities, lack of motivation, and in general, the simple reason that it is no longer easy to leave their homes and participate in physical activities on a regular basis. Thus, and in spite of the growing evidence of the benefits of physical activity, as well as the adverse effects of sedentary behavior, physical inactivity is still prevalent in older adults .
Intervention programs to promote physical activity in older adults, either in group or individual (home) settings, have demonstrated the potential to improve health and functional performance . Both types of intervention have shown equivalent health outcomes but with different results when considering adherence. Group-based interventions seem to achieve higher participation in the long-term, while in the short-term the results are comparable or not conclusive. However, there are many factors that can make group exercises a challenging (or infeasible) setting for older adults. A major obstacle, due to the heterogeneity of this broad population, is the big difference in the level of physical abilities between participants: unless the training class is tailored to the needs and abilities of each group, we are likely to see limited effectiveness as well as lack of motivation in performing the exercises , which proves itself difficult in heterogeneous groups. This difference in physical abilities, along with the logistic and practical obstacles that make it more difficult, as we age, to regularly attend a gym and perform group exercises, means that, for some adults, home-based individual intervention is the only viable training option.
In this paper we report on a technology-based physical intervention that enables older adults with different abilities, and indeed in spite of their different abilities, to engage in group exercises from home while keeping these differences invisible to the group. The intervention is based on the OTAGO Exercise Program for fall prevention  and supported by a set of applications that allow older adults to follow virtual training sessions via a tablet device under the supervision of a remote coach. Different interventions were performed, in different contexts, different target groups, and different countries. In this paper we report on a specific internvention, including a total of 37 adults aged between 65 and 87 years old (28 females and 9 males, mean age = 71, sd = 5.8) that followed a personalized home-based strength and balance training plan for eight weeks. The participants performed the exercises autonomously at home using the Gymcentral application. Participants were assigned to two training groups: the Social group used an application with persuasive and social functionalities, while the Control group used a basic version of the service with no persuasive and social features. We further explored the effects of social facilitation, and in particular of virtual social presence, in user participation to training sessions. The main outcome measures were adherence, persistence and co-presence rate, and we also assessed physical and psychological wellbeing as well as behavior change.
Previous studies have demonstrated effectiveness of the OTAGO Exercise Program in reducing falls and fall-related injuries among high risk individuals, and increasing the percentage of older adults who are able to live independently in their community . Although this specific program was designed for home-based training, a meta-analysis including other exercise programs for fall prevention found that combining group-based and home-based exercises is a strategy used in several effective trials, thus recommending both options to be available .
Technology-based interventions have also demonstrated to be effective in increasing and maintaining physical activity . Among the technological components that have been explored we can mention: online newsletters , personalised emails , web-based videos , tablet applications  and video game consoles . However, most of the existing intervention studies have focused on individual training, or provided a social context that was limited to forums or chats . Even exergames, a type of technology that have traditionally provided more immersive experiences (e.g., via MS Kinect and Nintendo Wii), have not been explored in an online group setting.
A Virtual gym for older adults
Gymcentral is a software designed to enable and motivate independent-living older adults - of potentially different abilities - to follow group exercises from home, relying on social interactions both as a motivation and as an end themselves.
The design of Gymcentral is informed by evidence in the literature as well as previous experiences that shaped the current implementation of the application. Active Lifestyle and its successor, the Virtual Social Gym, explored the feasibility of providing a home-based strength and balance exercise program by means of video exercises in a tablet device . In addition, it studied the effects of using individual (e.g., positive and negative reinforcement) and social persuasion strategies (e.g., collaboration and competition) in the adherence to the training programs.
Based on this early work and the feedback we obtained, we developed a software application that enables home-based training and that i) includes tailored exercise programs, ii) includes the possibility of interaction with a human coach, and iii) performs remote monitoring by employing an activity monitor in the context of a home-based physical intervention. In addition, we wanted to experiment the possibility of older adults of different abilities, and indeed despite this difference, to engage in (virtual) group exercises from home while keeping disparities invisible to the group. Thus, we explored different design alternatives to realise the group exercising (see Figure 1c.), from simply indicating that another trainee was also training (online status) to having a real-time motion and visualisation (3D and motion), each alternative with a different level of immersion, feedback and requirements in terms of technology.
The design alternative materialised in the current version of the tool (Figure 1d), relies on the virtual environments, which have been shown to increase the sense of presence, or psychological immersion. In addition, social presence, along with user embodiments (avatars), helps to reduce physical barriers and get users more engaged in the activities while preserving their privacy. Avatars however do not follow the actual trainees movement but predefined movements. This was both a practical constraint (i.e., to keep the technological requirement to a minimum) and a design constraint (i.e., to keep the specifics of the exercise performed hidden from others).
Figure 1. Training applications. a) Active Lifestyle app, exploring the use of individual and social persuasion strategies; b) Virtual Social Gym, exploring the use of activity monitors in home-based interventions; c) Gymcentral early design alternatives; d) Gymcentral application in its current form.
The Gymcentral platform is organised in two main applications that serve the needs of both trainees and the coach. Older adults can follow the training programs by using a tablet-based application and sensors, and the coach can monitor the entire coaching process from a web-based application (Figure 2).
Figure 2. Gymcentral application workflow
Design and Setup
The main objective of the study was to understand if social presence increases adherence to an exercise program. Secondary objectives were to assess effect on physical wellbeing, social wellbeing and behavior change. The study followed a framework for the design and evaluation of complex interventions in health settings [Campbell et al., 2000]. Using a random assignment procedure, participants were assigned to an experimental (or social) condition and to a control condition, considering age and participants' frailty level as the random assignment variables.
Participants in the social group were given a version of the Trainee App that included the personalised training program, social environment for group exercising, messaging and persuasion features. Participants were aware that they were exercising together and they could choose to do so. In the control condition, participants received a version of the application that focused only on the home-based program, delivering the personalised training but without social or individual persuasion features. Participants were offered three training modules (~1.5 hours each) focusing on operating the tablet, the use of the main applications and the Gymcentral app.
All participants underwent physical assessment with a personal trainer before starting the exercise program. This allowed for personalised tailoring of exercise type and intensity, and for personalisation of the starting level of each participant. Enjoyment of physical activity was also measured, using the Physical Activity Enjoyment Scale [PACES; 4], to test for differences that might favor the participation in one of the groups. An analysis of variance however showed no significant main effect for the groups (p = .477).
The study took place over a period of 10 weeks. The first week was devoted to technical deployment and application testing, followed by 8 weeks of training and 1 week of post-training measurements. The training program was supervised by a training coach, who could intervene to advance trainees in the exercise program, and to provide technical support upon request. No differences in support related to the app were registered.
The exercise program implemented in this study was developed on the basis of the OTAGO Exercise Program , and was adapted with by a professional personal trainer in order to fit the original program into 10 levels of increasing difficulty. The study received ethical approval from the CREATE-NET Ethics Committee.
Following several threads of previous work, we measured:
- Physical ability based on commonly adopted tests such as 30-second chair stand and get up and go, to assess muscle strength and gait speed
- Enjoyment, using the PACES scale 
- Wellbeing, using the Wellbeing scale of the Multidimensional Personality Questionnaire [MPQ; 15]
- Processes of behavior change. One theory that has been recently applied to the study of persuasive technologies is the Trans Theoretical Model of Behavior Change [TTM; 16].
Attrition. The intervention resulted in a 7.5% attrition rate, measured in terms of the proportion of participants lost at the end of the study. Reasons behind the withdrawal of these participants were related to unexpected health and family problems or, in one case, because of Internet connection issues which could not be solved.
Persistence (fig 3). In the social group the persistence rate was 85%, while in the control group it was 64%. A t-test for independent samples showed that the social group had a significantly higher persistence rate (M=85.4%, SD=16.1%) compared to the persistence rate of the control group (M=64.2%, SD=24.1%), t(35)=3.18, p=.003).
Strength and speed (fig 4). Both groups showing a significant improvement in the test assessing leg muscle strength,(F(1, 33)= 19.275, p < .001, partial eta squared = .369), with with participants in the social group performing significantly better despite non-significant differences in prior conditions. In terms of speed, multiple comparison tests with Bonferroni correction showed that gait speed significantly improved for the Social group (p = .002) but not for the Control group (p = .285).
Enjoyment and wellbeing both improved, but we did not see any significant difference among the groups.
Figure 3: adhrence to the program.
Figure 4: muscle and speed improvements.
Participants in the Control group have experienced an increased usage of some of the processes of change such as counterconditioning, reinforcement management, self-liberation, social liberation, which can be attributed to an increased awareness about the benefits of physical
activity as a result of the intervention. In the Social group, participants have also increased in consciousness raising, dramatic relief and self-reevaluation.
The main finding we take home from the study is the impact on group exercising, both in terms of adherence and results. Furthermore, participants did prefer to train in group, as we witnessed a clear tendency for people to train (virtually) together when possible.
We also observe that virtual training programs, and specifically group-based ones, are effective in triggering the process of change. One of the most striking examples of this is that months after the end of the trial the participants came back to ask to reactivate the platform permanently, which we did, and bought their own tablets and internet access to keep exercising.
Longer studies with more participants are instead needed to assess the potential for effects on wellbeing and differences among the groups.
 Landi et al. Moving against frailty: does physical activity matter? Biogerontology 11, 5 (2010).
 Harvey et al. Prevalence of sedentary behavior in older adults: a systematic review. International journal of environmental research and public health 10, 12 (2013).
 El-Khoury et al. The effect of fall prevention exercise programmes on fall induced injuries in community dwelling older adults: systematic review and meta-analysis of randomised controlled trials. Bmj, 347, f6234.
 Carraro et al. A contribution to the validation of the physical activity enjoyment scale in an Italian sample. Social Behavior and Personality: an international journal, 36(7).
 de Groot and Fagerstro. Older adults’ motivating factors and barriers to exercise to prevent falls. Scandinavian Journal of Occupational Therapy 18, 2 (2011)
 Gardner et al. Practical implementation of an exercise-based falls prevention programme. Age and ageing 30, 1 (2001).
 Sherrington et al. Exercise to prevent falls in older adults: an updated meta-analysis and best practice recommendations. New South Wales public health bulletin, 22(4), 2011.
 Müller et al. Non-face-to-face physical activity interventions in older adults: a systematic review. Int J Behav Nutr Phys Act,11(1), 35. 2014.
 Hageman et al. Tailored versus standard internet‐delivered interventions to promote physical activity in older women. Journal of Geriatric Physical Therapy, 28(1). 2005.
 Ferney et al. Randomized trial of a neighborhood environment-focused physical activity website intervention. Preventive medicine, 48(2), 2009.
 Benavent-Caballer et al. The effectiveness of a video-supported group-based Otago exercise programme on physical performance in community-dwelling older adults: A preliminary study. Physiotherapy, 9/2015
 Silveira et al. Motivating and assisting physical exercise in independently living older adults: a pilot study. International journal of medical informatics 82, 2013.
 Jorgensen et al. Efficacy of Nintendo Wii training on mechanical leg muscle function and postural balance in community-dwelling older adults: a randomized controlled trial. The Journals of Gerontology Series A: Biological Sciences and Medical Sciences, Oct 2012.
 Aalbers et al. Characteristics of effective Internet-mediated interventions to change lifestyle in people aged 50 and older: a systematic review. Ageing research reviews 10, 4 (2011)
 Tellegen et al. Exploring personality through test construction: Development of the Multidimensional Personality Questionnaire. The SAGE handbook of personality theory and assessment, 2, 2008.
 Velicer et al. Smoking cessation and stress management: Applications of the transtheoretical model of behavior change. Homeostasis in Health and Disease, 38(5-6), 1998