Web-based interventions in multiple sclerosis: the potential of tele-rehabilitation

Abstract

The World Wide Web is increasingly used in therapeutic settings. In this regard, internet-based interventions have proven effective in ameliorating several health behaviors, amongst them physical activity behavior. Internet-delivered interventions have shown positive effects on physical activity and physical function in persons with MS (pwMS). In this review we give an overview on several online exercise programs for pwMS and discuss the advantages and drawbacks of web-based interventions. Although participants of online exercise programs reported a high acceptance and satisfaction with the intervention, decreasing compliance was a major issue. A possible remedy might be the implementation of game-design elements to increase compliance and long-term adherence to internet-delivered interventions. In addition we believe that the integration of social networks seems to be a promising strategy.

Introduction

The positive effects of physical activity and exercise for persons with multiple sclerosis (pwMS) are well documented. However, exercise-related studies with pwMS were predominantly conducted with small sample sizes and relatively short intervention duration. Modern communication technologies via the internet could enable more economic and individual coaching than traditional interventions. Furthermore, a large target group could be addressed over a wide geographical range [Marcus et al. 2000].

Internet and internet-based interventions

Since the appearance of the Internet in the 1970s in the United States, it has continually been evolving. Especially in the past 10–15 years, a rapid growth of several hundred or even thousand percent has taken place (www.internetworldstats.com/stats.htm, accessed 28 October 2015). The penetration (percentage of the population which uses the Internet) is highest in North America with 78.6%; Europe ranks third with 63.2%. The technical possibilities of the digital age offer a variety of opportunities. Thus, the term ‘web-based intervention’ includes numerous concepts with varying technical solutions that have constantly been refined. Here, the range goes from unidirectional delivery of text-based materials and websites using interactive technology to the use of mobile devices [Neville et al. 2009]. In terms of intervention types, Barak and colleagues [Barak et al. 2009] differentiate web-based interventions (education interventions, human-supported or self-guided therapeutic interventions), online counseling and therapy (individual or group-based, using synchronous or asynchronous communication), internet-operated therapeutic software (e.g. expert systems for assessment, treatment selection and progress monitoring) and other online activities (e.g. chats, wikis, social networks). Those interventions may be used in tele-rehabilitation, which is the provision of rehabilitation services at a distance using communication and information technology [Kairy et al. 2009].

Internet use in pwMS

According to an American study, 93% of pwMS use the Internet in North America; this portion is even higher than in the healthy population [Weiss, 2007]. In Germany, Haase and colleagues [Haase et al. 2012] conducted a survey on 586 pwMS between the ages of 17 and 73 who were recruited via multiple outpatient rehabilitation centers. According to that study, 75% owned a computer with internet access, and of those, 70.8% used the computer at least once a day. According to another American study, more than 73% of pwMS had a large interest in using online healthcare services and 83% had an interest in obtaining disease-related information and recommendations online [Wardell et al. 2009]. In a study by Weiss [Weiss, 2007], every other pwMS regarded the Internet as a suitable medium to obtain disease-related information. It had also already been shown that internet-delivered education programs are practicable and well accepted in pwMS [Finkelstein et al. 2004]. Several studies have shown that fatigue management programs for pwMS could be effectively implemented via the Internet [Finlayson et al. 2011; Moss-Morris et al. 2012]. However, an information deficit exists specifically with regard to physical activity and exercise. Somerset and colleagues [Somerset et al. 2001] determined that this applies to 41% of the surveyed pwMS. The desire to get more information on this topic was even more pronounced than for on topics such as immune medication, nutrition or alternative therapies.

Web-based physical activity interventions in pwMS

In 2010 Motl and colleagues published the first study results on a web-based 3-month program for physical activity promotion in pwMS [Motl et al. 2011]. The social cognitive theory was selected as a theoretical framework. This concept has already proven to be effective regarding physical activity promotion in a previous study in a traditional ‘face-to-face’ approach with pwMS [McAuley et al. 2007]. The study participants were recruited per email from the pool of an existing study; 54 ambulatory but physically inactive test persons with relapsing–remitting disease course participated (average age: 46 years). They were paired on the basis of physical activity and neurological function and were randomized either to the intervention group or to a waiting control group. The primary endpoint physical activity was measured with the Godin Leisure Time Exercise Questionnaire (GLTEQ) [Godin and Shephard, 1985]. The intervention was conveyed via a website and provided text-based information supported by multimedia content (video data). In addition, a chat session was completed twice a week and a discussion forum was offered via the website. The intervention group significantly increased physical activity compared with the waiting control group (p = 0.01, d = 0.72). An increase in physical function was not reported. The intervention appeared to be particularly effective in pwMS who were very physically inactive and exhibited only minor functional limitations. The acceptance of the website was very good in the first week (96% of participants logged in) and then decreased steadily (approximately 52% logged in weekly in the last weeks; average 71%).

In an additional study, the effectiveness of this concept was reproduced with the help of objective measurement (accelerometry) [Dlugonski et al. 2011]. For this study the waiting control group from the previous study was recruited. Interestingly, the study showed a correlation between usage of the website and the objective parameters of physical activity (activity counts, number of steps).

Finally in a third study [Dlugonski et al. 2012], the existing intervention concept was extended with seven individual behavior-related online-coaching sessions (video conferencing, ‘one-on-one’). This lead to an improved website usage compared with the previous study. During the 12-week intervention, physical activity increased within the first 6 weeks and then remained constant until the end of the intervention and up until follow up 3 months after the intervention. The objectively measured physical activity increased significantly (1533 steps per day). The intervention was very well accepted according to a participant survey. However, no effect on health-related quality of life (MSIS-29) [Hobart et al. 2001] or the ability to walk (Multiple Sclerosis Walking Scale, MSWS-12) [Hobart et al. 2003] was observed.

In their study, Finkelstein and colleagues selected a web-based therapeutic software for their intervention [Finkelstein et al. 2008; Finkelstein and Wood, 2012]. Their ‘home-based physical tele-rehabilitation’ included a 3 months of home training led by a physiotherapist. The starting point for the training schedule was a stay in the hospital, during which the concept and the exercises were explained. The intervention consisted of strengthening and balance exercises (a catalog consisting of 47 predefined exercises was used). The technical realization was performed with an individually programmed platform. Therapist and patient, respectively, received a laptop with software that transferred training data and the training schedule. The system was installed during an initial visit by the therapist to the participant’s home. An exact description of technology and software can be found in the literature [Finkelstein and Wood, 2012]. Each participant received a personal exercise list with prescribed repetitions as well as text-based and audiovisual instructions (video), accessible via the laptop. After training, the participant documented the completed exercises and the therapist adapted the training schedule. A total of 12 pwMS participated in the 12-week intervention study; a control group did not exist. The study showed significant positive effects concerning walking ability: timed 25-foot walk [Fischer et al. 1999], 6-minute walk test [Crapo et al. 2002] and balance (Berg Balance Scale) [Berg et al. 1989]. However, no changes in health-related quality of life (MSQoL-54) [Vickrey et al. 1995] and self-efficacy (MSSE) [Rigby et al. 2003] were observed [Finkelstein et al. 2008].

MS-Intact program

Our own research on web-based exercise began in 2009 with the MS-Intact program (internet-based physical activity promotion for persons with multiple sclerosis, 2009–2011). The Hertie Foundation, the German foundation for neurology, and Bayer HealthCare supported the research. The core content of the intervention was an individualized prescription of strengthening and endurance training via the internet (e-training) as well as general physical activity promotion.

The MS-Intact intervention used a browser-based software (motionNet e-Training, motionNet Systems Ltd, Nuremberg, Germany) with separate therapist (back end) and participant interface (front end). No additional devices or software installation were needed. In a randomized controlled study design, ambulatory patients (able to walk at least 500 m) were stratified according to neurological function and aerobic capacity. They were then either assigned into the intervention group (6 months e-training) or into a waiting control group (3 months waiting, then 3 months e-training).

At the beginning and after 3 and 6 months physical activity (Baecke Questionnaire) [Wagner and Singer, 2003], fatigue (WEIMuS) [Flachenecker et al. 2008] and quality of life (HaQuaMS) [Gold et al. 2001] were assessed. In addition, muscle strength (static maximum strength of knee extensors and flexors and trunk extensors and flexors), lung function and aerobic capacity via bicycle spiroergometry were assessed. The e-training intervention began with a 2-day, onsite training seminar on the content and procedures of the e-training.

• Knowledge of action and effect: basic information on MS and exercise, strength training/endurance training with MS; dose and periodization of training.

• Practice and training/experience: execution and discussion of basic strength exercises; body awareness exercises; heart rate measurement and endurance training.

• Motivation and volition: motivational help and advice regarding action planning and barrier management for the implementation of exercise in everyday life.

• Familiarization with the software: documentation of training, handling of physical activity diary.

Communication during the intervention took place asynchronously via a messaging service in the software and, when required, by email and telephone. A social network for the participants to use amongst themselves was not set up. A resistance training plan contained five to eight exercises, especially for the lower extremities and trunk muscles (stomach, lower back, upper back/arms, quadriceps, abductors). No special equipment was necessary except an elastic exercise band or a large gymnastic ball. The number of series and repetitions to be completed for each exercise were prescribed by the participant’s therapist and adjusted according to a standardized progression plan. To ensure training overload and progression, every exercise was prescribed and adjusted from 2 sets of at least 6 repetitions to 3 sets of 20 repetitions. Training intensity was regulated by the participant’s perceived exertion. Participants were asked to rate subjective effort on the 6 to 20 BORG Scale [Borg, 1982]. Therapists governed exercise intensity aiming at eliciting a Borg Feedback of between 11 and 16. Between sets and exercises, a rest period of approximately 1–2 minutes was recommended. In addition to the strength training, endurance training was to be carried out once a week. Based on the spiroergometric evaluation, recommendations regarding the intensity of jogging, walking, cycling and swimming were made. The form of activity for the endurance training was freely selected. In contrast to the strength training, the endurance training was, however, not systematically controlled and progressed. All the training sessions could be planned and documented in an exercise journal, which could also be seen by the therapist. Every training session, including all the exercises and associated training load parameters, was automatically stored electronically.

The preliminary data from the MS-Intact program have been presented at scientific meetings [Tallner et al. 2012]. We included 126 pwMS. The dropout rate after 3 months was 14%, which is very low. Over the course of the study, a total of 3639 training sessions were documented and feedback was provided on 21,566 individual exercises. A training session included an average of 5.9 strengthening exercises and had an average duration of 29.0 ± 15.4 minutes. Figure 1 shows the training frequency of all participants, which amounted to 6.82 training units per month. In the online activity diary, a total of 8548 physical activities were entered. Apart from the strength training sessions, the most frequent were cycling (984 entries), Nordic walking (780), walking the dog (605), running/jogging (459), gym (381), swimming (333), walking (209) and cross-trainer (110). Although the e-training resulted in positive and significant effects on muscle strength and lung function compared with the control group, no positive effects on fatigue or quality of life could be identified.

Figure 1.

Number of strength training sessions per month.

Documented strength training sessions of the participants over the course of the study. The control group did not start training until after a 3-month waiting period.

Discussion

The introduced studies have shown that the Internet is an effective medium for the delivery of exercise interventions for pwMS. A big advantage is the enormous range of the Internet, which allows the supervision of large training groups. An additional advantage of web-based interventions is electronic data acquisition, which means that a higher amount of data can be recorded automatically and more economically than in traditional interventions. A disadvantage is the lack of direct supervision of the intervention or training program. The introduced interventions did not include options to directly monitor the exercises via cameras or motion detection systems. The results of the MS-Intact study, significant increase in muscle strength, physical activity and lung function suggest that the training was implemented effectively. The increase in strength was comparable with supervised training interventions for pwMS [de Souza-Teixera et al. 2009].

Effects of training on fatigue or quality of life seem, however, to be more difficult to achieve. Missing effects on fatigue with the MS-Intact study can be based on the low fatigue baseline level of the participants. The baseline was around 10 points below the fatigue threshold of WEIMuS (32 points).

None of the web-based interventions could show an effect on quality of life for pwMS. This may reflect a central disadvantage of web-based interventions. Nevertheless, effects of traditional, face-to-face exercise interventions on quality of life in pwMS have been shown in some reviews and meta-analyses [Motl and Gosney, 2008; Kuspinar et al. 2012], but not in all [Latimer-Cheung et al. 2013]. Quality of life may only be improved in specific exercise setups and settings. The studies that were included in meta-analysis by Motl and Gosney took place in a group setting [Motl and Gosney, 2008], which could have played a central role for the effects on quality of life. In their qualitative analysis of a group exercise intervention for pwMS, Aubrey and Demain identified social support as an important element of illness related self-management [Aubrey and Demain, 2012]. In the study by Learmonth and colleagues, social benefits were documented as the central motivation for pwMS to participate in an exercise program [Learmonth et al. 2013]. Therefore, web-based interventions should focus on social components of training and should offer social networks for participants. Richardson and colleagues could show in a randomized controlled study that an online community as a supplement to a web-based intervention resulted in a lower dropout rate, but effects on physical activity were not observed [Richardson et al. 2010]. Nevertheless, social networks are rarely used in health-related interventions; mainly commercial providers like Facebook have recognized the potential of social networks [Gold et al. 2012].

Compliance in the MS-Intact study was below the target of 2 units per week (average 6.82 training units per month). This reflects another potential disadvantage of web-based interventions: a relatively low degree of institutionalization (i.e. no fixed (group) appointments) leads to heterogeneous individual exposure to the intervention. The training frequency for participants in the intervention group (6 months training) had a range of between 22 and 69. This depends on physical activity habits existing prior to the study; (active) participants were explicitly informed that they should not alter their existing physical activity habits.

Another problem in internet-based interventions is the decline of training frequency over time. In the MS-Intact study, the target of two strengthening units per month was maintained perfectly during the first month, but already after the second month, a decline was noticeable (Figure 1). In the last three training months, the participants trained at a relatively constant level of about once a week. It is not clear if this training frequency is sufficient to maintain the initial training success. Explanations for the decline of training frequency remain speculative; the decline can be based on the genuine frequency of exercise sessions, but also on a sinking motivation to log into the computer and document them. Steadily decreasing website usage is a general problem of web-based interventions [Davies et al. 2012b; Norman et al. 2007]. This holds especially true for interventions using automated, interactive websites [Schulz et al. 2012]. Online support through a personal coach might result in a more intensive use. Even the high internet affinity of pwMS does not protect against declining login frequency.

In general, the effectiveness of an intervention is dependent on the frequency of use [Davies et al. 2012a]. Aalbers and colleagues determined that the effectiveness of an intervention increases when the intended login frequency can be maintained over 10 weeks [Aalbers et al. 2011]. Therefore the frequency of use is of central importance, but this component is often not identified or stated in studies [Norman et al. 2007]. Future studies should take this into consideration and should report these data.

Another drawback of internet-delivered interventions might be participants’ neurological symptoms. The latter can produce cognitive or motor challenges (handling of technical devices and platforms, intelligibility of manuals, requirement on executive functions) that can limit program usage [Torsney, 2003].

In conclusion, the Internet with its enormous technical possibilities and global accessibility has the potential to make therapeutic interventions widely available and effective for pwMS. However, we believe that web-based interventions cannot substitute traditional interventions though they can perfectly complement them. Web-based interventions seem promising for physical activity promotion after inpatient rehabilitation programs. The success of such an intervention is heavily dependent on the quality of its content, especially when no face-to-face contact takes place.

Future prospects

The following strategies might be suitable to increase the effectiveness of an intervention.

Intervention content

• Existence of a theoretical basis for behavior change: theory of reasoned action/planned behavior, trans-theoretical model, social cognitive theory [Webb et al. 2010].

• Integration of techniques for behavior modification (coping/barrier management, social comparison, goal setting, action planning, feedback) [Webb et al. 2010].

• Structured information material [Davies et al. 2012b].

• Tailoring of the intervention to the target group [Aalbers et al. 2011].

Communication and feedback strategies

• Use of automatic, context sensitive feedback, SMS/telephone contact; provision of therapists/consultants [Webb et al. 2010].

• Provision of interactive tools for self-monitoring (activity diary, logs, etc.) [van den Berg et al. 2007].

• Enable contact between participants [Norman et al. 2007].

Study design and methods

• Inclusion of physically inactive sample groups [Davies et al. 2012b].

• Choice of appropriate measurement methods for physical activity (mixed methods with subjective and objective measurements) [van den Berg et al. 2007].

Gamification to increase participation

Some of these features are based on the contents of the intervention; others primarily serve to increase the frequency of use. Regarding frequency of use, an excursion into the gaming industry seems promising. The strategies of the gaming industry are very purposeful – and very successful; there were 36 million active gamers in Germany in 2011 and they spent 47 million hours per day (!) gaming online [Newzoo, 2013]. The average age of gamers is 37 years; 29% are older than 50 and the portion of women equals 42% [Herger, 2013]. Therefore, the success factors that bind gamers to playing could be of relevance for health-related web-based interventions as well.

In fact, an intersection already exists between games and exercise related interventions: the so-called ‘exergames’. These are games that use physical activity to influence gameplay. Prominent examples are gaming consoles like the Nintendo Wii or Microsoft Kinect. Exergames have already been successful in scientific studies in the areas of physiotherapy, psychotherapy and physical activity promotion [Primack et al. 2012].

There are possibilities to transfer typical game elements even into a non-game context. This process is called ‘gamification’ [Deterding et al. 2011] and can also add to web-based activity-related interventions. Here, the original goal of an intervention remains the same but attractive interim goals as well as game-like activities and incentives within the existing context are added [cf. Oja and Riekki, 2012]. Typical game mechanisms which can come into use are as follows [Deterding et al. 2011; Kapp, 2012; Koch and Ott, 2012].

• Narration/instruction using avatars.

• Visible player status: points on a visible ranking list, rewards (e.g. higher player ranking or status, ‘badges’) for accomplishment of tasks.

• Quests/challenges: puzzles and/or tasks which are interspersed; competitions with specific rules.

• Feedback: immediate evaluation of game actions, progress display during accomplishment of tasks or the achievement of goals/interim goals/quests.

• Cascading information: to avoid excessive demands, only those parts of information that are necessary at the time are conveyed.

• Epic meaning: superordinate meaning of game actions; the game must include something (in the eyes of the player) great or desirable.

• Community collaboration: certain goals/quests can be achieved only in player groups/teams.

Using these elements, training and training processes can be designed more interestingly and appealing. Through the introduction of competitions or tasks, practicing becomes a sport, and inter and intra-individual comparisons motivate participants [cf. Mueller et al. 2011]. For behavior-oriented exercise therapy [Geidl et al. 2014], techniques of behavior modification to be included in physical activity promotion interventions have been identified [Pfeifer et al. 2011]. Many of these techniques are very similar to the typical game mechanisms (e.g. information provision/instruction, feedback about current training status, agreement on activity goals or behavioral contracts, use rewards, offer of social comparison, etc.). Therefore, it is reasonable to combine both approaches in web-based interventions. An example is virtual rewards in the form of gold, silver and bronze trophies and medals that are granted for the achievement of specified training frequencies or exercise minutes. A medal table or count in the participant interface can visualize the achieved training success. Training progress regarding individual exercises can be displayed graphically.

We hypothesize that the potential of web-based interventions for physical activity promotion can be optimized through targeted coupling of behavior-oriented exercise therapy, technical and communicative (social media) opportunities of the Internet, and game elements and mechanics. However, this remains to be evaluated in future studies. According to Kairy and colleagues, it should not be taken as granted that modern communication systems automatically lead to economization [Kairy et al. 2009]. Thus, additional research needs to be conducted with regard to a health economic evaluation of internet-based interventions [Tate et al. 2009].