The study was performed to compare two groups using different self-managed exercise programmes (DP or PB). Furthermore, to find out if the new digital programme was feasible for self-management, among seniors accustomed to the use of apps, in a forthcoming large RCT. All participants chose their preferred type of programme, based on their personal preferences and access to technology, no technical devices were provided in this study. After an introduction meeting with pre-assessments the participants exercised independently over four months and self-reported their exercise in an exercise diary. The intervention was completed with a final meeting after four months. Follow-up 12 months after study start was completed using a postal survey. An overview of the study can be seen in Figure 1. The study is registered in ClinTrial: NCT02916849.
Inclusion criteria were: ≥70 years old, living independently, able to rise from a chair and stand without support, experiences of deterioration in balance OR need to be more careful not to lose balance OR have experienced a fall the past year. Exclusion criteria were: doing physical exercise more than 3 hours/week, self-reported progressive disease that was likely to influence mobility, and cognitive difficulties. Status of cognitive condition was judged during the screening interview, if the person was able to answer questions satisfactorily, and able to converse about matters regarding the study, they were considered suitable to take part in the study. Participants were recruited at four different senior citizen organisations and at a health care centre. At the senior organizations, information about the project was presented by members of the research team, and contact details of seniors interested in taking part were collected. A research assistant then phoned them for a short interview to screen for suitability to take part. Recruitment at the health care centre was undertaken by a physiotherapist, occupational therapist, nurse, medical doctor or nurse’s aide, who had received an introduction about the programmes sufficient to give information to the potential participants. Participation in the study was voluntary and did not influence the further care of the patient. All participants chose their programme during the first recruitment contact.
Independent of choice of programme, all participants attended an introduction meeting lasting about two hours, including a short presentation about accidental falls and fall prevention, introduction of the exercise programme, and pre-assessments. The respective programme’s main structure, how to select exercises and fill out the exercise diary, and safety aspects during the sessions at home was explained, with opportunity to try some exercises and ask questions. The digital programme group got additional information about the log-in procedure, and how to use parts of the behaviour change support available in the application. Two physiotherapists from the research group led the meetings (LLO, MS), both with experience from the field of fall prevention exercise programmes. Thirteen groups of maximum eight persons met for the introduction meeting. Seven groups from senior citizen organisations had the introduction meeting at the university campus for participants and six were held at the health care centre. The majority of introduction meetings were separate for the DP and PB participants. However, from the health care centre few participants chose the DP, so 2/3 of these meetings were mixed, but the introduction of the actual exercise programme (DP and PB) was kept separate.
Scheduled interaction with the participants during the study was limited as the study focused on self-management of the programmes. A phone interview with all participants was done a few weeks after study start to identify any problems with the programme at an early stage. A help-line phone number was provided in case of encountering any problems while using the programme during the intervention. In order to monitor technical support for the digital program a record of contacts from DP users was kept. Observations by a physiotherapist and a human computer interaction engineer (LM, RJ) were performed with six participants using the DP in their home after approximately eight weeks. A monthly peer-mentor group meeting was held with half of the participants from DP group recruited at senior citizen organisations, this was also by self-selected choice. These meetings were led by two seniors with a mentor role, together with one of the researchers (MS). Three different topics were discussed, one at each meeting: (1) Initial experiences, (2) Motivation for exercise, and (3) Establishment of lasting exercise routines. The researcher’s role at these meetings did not aim to give extra technical support.
Eight final meetings were held at the university campus and another three at the health care centre. Also participants that had withdrawn from the intervention, by notifying that they stopped exercising, were invited to attend the final meeting to give feedback on the programme.
Both programmes were based on exercises from the Otago Exercise Programme  but to provide a variety of exercises at diverse levels, the DP was enriched with both easier and more challenging exercises mainly inspired by the Falls Management Exercise Programme (FaME) . The application for the DP was developed in co-creation with older adults taking their needs and preferences into account. Thus, the exercises are instructed in short video clips imaging older persons doing the exercises, and the user-interface is clean and uncomplicated .
In the DP (Safe Step v1) the user builds his or her own exercise programme by selecting one exercise from each of ten predetermined groups of exercises to improve strength, balance and gait/step ability. Each exercise group had several variants of exercises with different levels of difficulty provided by video clips with verbal instructions. The application also included behaviour change support with written motivational feedback from a virtual physiotherapist (computer generated pre-written messages, delivered according to the participants’ reported exercise), exercise planning and possibility to review the exercise diary, as well as examples on how to integrate exercises into daily activities and practice outdoors.
The PB contained the Otago exercises with drawings and written instructions. In order to help the participants build their programme, the exercises were divided in two sections with strength or balance exercises. Each section was further arranged into three different levels of difficulty with was a modification from the Otago Home Exercise Programme Booklet. Participants were instructed to select five exercises from each section to build a programme of ten exercises. Additional exercises for warm-up and stretching included in the booklet were not considered part of the programme’s ten exercises.
In the Supplementary Table 1, Template for Intervention Description and Replication (TIDieR) checklist, a more detailed description of the interventions can be found, Additional File 1.
Participants composed their own programme and exercised independently throughout the four months intervention, directed by material in the programmes and information given at the introduction meeting. Independent of which programme, all participants were asked to choose exercises that they experienced challenging but not too difficult to perform. For balance exercises this meant feeling unstable but without losing balance, and for strength exercises feeling a strain in the muscle but still able to complete the suggested number of repetitions. They were also advised to select new exercises to progress when an exercise became too easy, or to modify if they felt that the exercise they chose became too challenging. The recommendations were to exercise 30 minutes at least three times per week, according to instructions in the Otago Exercise Programme .
Baseline information to describe participants was collected at the introduction meeting with a study specific questionnaire about: age, sex, living condition, education level, fall history, use of walking-aids, self-reported health, and access to technology devices. Activity level was measured with the Saltin-Grimby Physical Activity Level Scale (SGPALS) that also assimilates household activities . Assessment of balance and functional strength was completed using the Short Physical Performance Battery (SPPB) with a maximum score of 12 for the best performance , assessed by a physiotherapist blinded to group allocation. Self-rated balance confidence was measured using a translation of the Activities-specific Balance Confidence Scale (ABC), rating from 0-100% for 16 activities  and a higher score means better confidence. Attitudes to Falls Related Interventions (AFRIS) was determined by a form with six translated statements about the attitude to the programme, to grade if agreed or not on a scale 1-7 to each statement , a higher score means a more positive attitude.
Exercise diaries were filled out by the participants over the four months intervention. The exercise diary for the DP allowed self-reporting of: date, which of the predetermined exercises were done and time spent on the practice. The digital diary allowed self-report of exercise once per day, information was stored in a database, from which researchers received data electronically on a monthly basis. The exercise diary for the PB group consisted of a monthly paper sheet, with rows for daily exercise reports, containing the same information as in the DP diary. It was returned in pre-paid envelopes at the end of each month. All diaries were reviewed monthly by the first author (LM), and if there was no data or the data was uncertain the participant was contacted by phone.
A questionnaire developed for this study was answered by the participants at the final meeting. The questionnaire dealt with their experience of using the programme and perceived effects. It had three parts: (1) eleven statements where participants were asked to answer on a Likert type scale from 1 = strongly disagree to 5 = strongly agree (as example “I’m satisfied with the programme I used” or “I notice improved strength in my legs”), (2) two multi-answer questions about positive and negative effects, and (3) further questions about any falls while performing the exercises, if they would recommend the programme to others and if they were going to continue with the programme. If participants did not attend the final meeting the questionnaire was sent out by mail with a pre-paid envelope as their opinions were considered important. Participants that withdrew from the intervention were presented with the option to take part in this questionnaire.
Finally, 12 months after study start a short survey was sent out with a pre-paid envelope to the participants that completed the study and took part in the final meeting (n=45). The aim was to investigate if they continued with the programme, or if not, the reasons why and if they planned to restart.
Differences between groups (based on choice of programme) for baseline characteristics were analysed using Chi-square test (Fisher’s exact test if expected count were <5), Student’s t-test or Mann-Whitney U-test depending on variable. The activity level of the SGPALS was dichotomized into groups of being inactive (level 1-2) or active (level 3-6) using the same method as Äijö et al. . Withdrawal was noted when participants informed that they stopped exercising with the programme and attrition rate was defined as the proportion of participants that withdraw.
For adherence analyses, the first 16 weeks of self-reported exercise were used. Adherence was described according to guidelines by Hawley-Hague et al.  for older adults participating in exercise classes. Their recommendation is to report four types of adherence: completion, attendance, duration, and intensity. We considered two of those to be relevant to our self-management exercise programme: completion, and exercise duration. Four subgroups were created to compare adherence for the DP and the PB with the following definitions:
- Enrolled, everyone that started the intervention.
- Completed study, all participants that did not explicitly withdraw from the exercise programme, independent of the degree of participation.
- Exercise completion ≥75% of the weeks, participants that self-reported exercise at least one session per week for 12 of the 16 weeks.
- Exercise duration ≥75%, participants that self-reported at least 75% of the recommended 90 minutes of exercise per week (at least in total 1080 minutes over 16 weeks).
For each participant, the mean number of minutes and sessions exercised per week were calculated for each week, until the participants stopped reporting to allow for short lapses during the intervention.
Many studies report adherence as percentage of the intended number of sessions over the intervention period, independent on how long time is spent within a session. With the purpose to be able to compare our study with others we also reported these numbers for adherence, recommended number of sessions were 48 over 16 weeks. Outcomes for adherence was analysed with Mann-Whitney U-test. All data were analysed using IBM Corp. Released 2016. IBM SPSS Statistics for Macintosh, Version 24.0. Armonk, NY: IBM Corp.
The study was approved by The Regional Ethical Review Board in Umeå (Dnr 2016/106-31). All participants got written and verbal information about the study and gave written informed consent. Concerns about safety to prevent falls during exercising was considered, it was stressed both during the introduction as well as in the information given in both programmes. Exercise was preferably done close to a wall, sturdy furniture or surface for support, and adapted to the participants’ functional level. Material in the programmes was clear and tailored to the age group to ensure good understanding and reduce any possible risks.