Hence, based on previous findings of the literature and knowledge gained during the preceding trial, this study aimed to adapt and simplify the IT-based intervention of the first funding phase further to improve usability and develop a simple print-based intervention that initially inactive participants with little affinity to technology find easy to interact with; investigate the implementation, feasibility, and use of 2 interventions (IT-based vs print-based) as well as changes in PA among older adults (aged ≥60 years) in a randomized trial with a crossover design over the course of 9 months; examine the role of personal preferences for different delivery modes with regard to intervention effectiveness; and explore associations between changes in PA and possible changes in physical fitness and cognitive capacity in a pooled sample of participants from both phases of funding.

We selected 14 community districts in 2 geographically different regions (northwest and northeast) of the city of Bremen, Germany: Burgdamm, Lesum, St. Magnus, Vegesack, Schönebeck, Aumund-Hammersbeck, Rönnebeck, Radio Bremen, Riensberg, Gartenstadt Vahr, Neue Vahr Südwest, Oberneuland, Ellener Feld, and Blockdiek. They were chosen because there are study centers in these districts that can be easily reached by the target group and the project team had already established prior liaisons with stakeholders in these areas of the city who could facilitate community involvement in the implementation of the intervention.

Names and addresses of men and women aged ≥60 years residing in the chosen community districts were drawn from the records of the residents’ registration office. Subsequently, they were invited to participate in the study via mail. Reminders were sent out in case of no response after 2 weeks. The study was also publicized in local newspaper articles and mentioned during talks that targeted older adults by researchers in the team. Individuals made aware of the study through this channel could contact the research team directly to be screened for eligibility. Eligibility for study participation was determined during computer-assisted telephone interviews with trained study nurses based on the inclusion and exclusion criteria.

Men and women were eligible for study participation if they were aged ≥60 years (there was no upper age limit), lived independently (ie, in own apartment or room without assisted living or regular home nursing), and provided informed consent to participate in the study. Additional inclusion criteria were basic knowledge of German, the ability to walk without a walking aid and participate in study assessments and weekly group meetings without external assistance, and no planned long absence (ie, for more than 2 weeks). Another precondition was the availability of a device with internet access in the household, regular access to the device, and the ability to use it.

Individuals were excluded from the study if they reported that they had already been regularly physically active for more than 1 year and had reached the intervention target (ie, the recommended 150 minutes of moderate-intensity endurance training per week). They were also excluded if they reported having participated in the preceding study (PROMOTE I), had planned a vacation lasting longer than 2 weeks during the intervention period, had been medically prohibited to be physically active, displayed severe visual or cognitive impairment (Mini-Mental-State Examination Score ≤14) or other impairments (eg, due to a stroke, transient ischemic attack, brain surgery, or neurological diseases, such as Alzheimer’s disease, multiple sclerosis, or Parkinson’s disease), had an implanted device (eg, pacemaker, implanted hearing aid), or experienced occasional syncope. Additional exclusion criteria were a lack of medical clearance because of fractures or surgeries in the past 6 months that could potentially constrain study participation; report of severe diseases of the cardiovascular system (eg, cardiac arrhythmia, heart failure, or nonmedicated hypertension) or respiratory system (eg, COPD, severe asthma); or severe limitations due to arthritis or osteoarthritis in the legs or severe osteoporosis or acute injuries of the spine. Individuals with diabetes diagnosed less than 6 months prior and without medical clearance were also excluded from the study.

After determination of eligibility, study participants were randomized to either an IT-based intervention or a print-based intervention (see Figure 1). In the first group, 30% were randomly selected to receive a PA tracker in addition to access to the website. Weekly time slots were randomly assigned to both intervention groups. Then, we had participants choose appointments according to their personal time constraints without knowing which intervention group they were assigned to. All intervention groups were offered weekly group sessions, and participants were encouraged to attend each of them. Two weeks before the intervention started, participants in all groups attended an introductory event where they were informed about the study background and procedures. On this day, data collection also started, and participants received the activity tracker and questionnaires for the first time (T0). They were also briefed about their randomization to one of the intervention groups and the existence of the other groups. Therefore, they were not blinded. Participants were given the choice to remain in their intervention group or to cross over to the other group after 3 months (and a 10-week intervention period). With that, a crossover design with the following possible combinations was generated: IT-IT, print-print (matched), or IT-print and print-IT. Preferences regarding the intervention material assessed retrospectively at the 9-month follow-up will be taken into account when analyzing intervention effects later in the study. Research staff members conducting the study were not blinded, and the statistician will not be blinded when analyzing the data.

Effects of time and group (main outcomes: subjective and objective measures of PA) will be examined in multivariate analyses. In addition, covariates, such as gender, age, and preferences regarding intervention material, and individual contextual factors (eg, living or social environment) will be taken into account. Furthermore, we will analyze how the intervention enables study participants to change their behavior by running mediation analyses. The estimated target sample size of 390 participants for this study is based on the results of 2 recently published studies on PA in adults [27,28] that reported a mean 111 weekly minutes (SD 116.8 weekly minutes) of MVPA and an intervention effect of an increase in MVPA of 77 minutes per week. However, in a conservative (or realistic) approach, we assumed that the mean increase in MVPA in our study will reach only 40 minutes/week. To detect such a difference between the intervention and control groups with a power of 80% at an alpha of .05 (two-sided), a net sample size of 150 subjects for each group is required. Assuming a loss to follow-up from T0 to T1 of 20-25% (ie, ca. 23%), it is necessary to initially recruit 390 study participants (ie, 195 per group at baseline). Assuming another 20% loss to follow-up from T1 to T2, we expect that approximately 240 individuals at T2 will self-select to one of the 4 groups: IT-IT, print-print, IT-print, or print-IT. Should participants self-select equally into each group (n=60), we will still be able to detect the mean difference of 77 minutes/week [28] between each group, by means of an appropriate post hoc test that is adjusted for multiple comparisons.

The primary outcome is change in PA, which is assessed at baseline, 3 months, and 9 months using triaxial accelerometers worn at the right hip during the day over the course of 7 days and a self-administered questionnaire. Participants received the self-administered questionnaire and accelerometer at their introductory event at baseline (T0), after 3 months (T1) at their tenth group meeting, and after 9 months (T2) at their last monthly meeting. Further, due to cost issues, a random subsample of 114 participants (equally distributed across the intervention groups) underwent additional anthropometric, physical fitness, and cognitive tests to assess secondary outcomes at all 3 time points. Physical fitness was assessed using gait speed via a 4-meter walk test [29], handgrip strength using a dynamometer [30], and cardiovascular fitness using the 2-minute step test [31]. Additionally, participants’ weight was assessed using a scale. Attention and inhibition were assessed using a cognitive test (Simon task) [32].

Data from the objective measures of physical fitness and cognitive performance (Mini Mental State Examination and executive function test) will be preprocessed using customized MATLAB routines and analyzed using multivariate generalized linear models. Data for physical fitness and cognitive performance assessed at baseline and after the 10-week intervention during the first funding phase and in the subsample of participants during the second funding phase will be pooled to increase power for data analysis and to meet sample heterogeneity. Further, the larger sample size will enable us to conduct subgroup analyses with respect to motivational stage, activity behavior, and social engagement.

All participants underwent a short version of the Mini Mental State Examination during their first weekly group meeting [33]. Motivational stages to engage in recommended PA (endurance, strength, flexibility, and balance training), social-cognitive predictors for behavior change, and psychosocial factors (eg, quality of life, depression) were assessed in the self-administered questionnaire (for further detail on the instruments included in the questionnaire, see Table 1).

Anthropometric information, as self-reported in the questionnaire, included height in cm (only T0) and weight in kg (all time points). Sociodemographic information was assessed at baseline only, using items of the German Health Interview and Examination Survey for Adults [63] for date and country of birth, gender, native language, family and relationship status, living situation, education, employment status, and monthly net household income. Employment was assessed using one item of a questionnaire for assessing seniors’ demographic and sociostructural data in Germany [64]. In addition, the self-administered questionnaire at T1 and T2 includes self-generated items regarding adherence to the PA recommendations (weekly minutes of endurance, balance, and flexibility training and weekly units of muscle strength training per muscle group) and attendance of the offered group sessions. Furthermore, use and satisfaction with the intervention material and content are assessed with the following self-generated items: frequency of general use; use of different intervention components reporting using a 5-point Likert scale from “never” to “daily”; perceived helpfulness of intervention components, intervention content, and structure of weekly meetings rated on a 5-point Likert scale ranging from “not helpful at all” to “very helpful”; potential knowledge gained and perceived benefits of PA using “yes” and “no” response options; and whether participants would recommend the intervention to family and friends, answered using open-ended questions. Data on the reasons for dropping out of the study and the reasons for crossing over or not crossing over to the respective other intervention group were also collected.

To simplify and further develop the IT-based intervention implemented during the first funding phase and to translate it into a print-based intervention, 9 focus groups consisting of former participants of the previous intervention (n=1), other members of the target group (n=32), and stakeholders in communities such as members of senior citizen organizations and advisory boards (n=12) were held in May and June 2018. Participants of these focus groups were recruited via a press release and already established contacts to the stakeholders. Focus group discussions consisted of two parts. First, intervention materials used in the previous study and materials and health brochures for PA promotion developed by other researchers; federal agencies, such as the Federal Ministry of Health (BMG); or health insurance agencies were discussed with regard to their suitability for PA promotion in inactive older adults. In this context, the participants were invited to vote for the material they liked most and to justify their decision. Second, using the World Café method, participants were asked to discuss the following questions in small groups of 4-5 participants:

The focus group discussions and votes for or against certain intervention materials were protocolled by the moderating researchers, and the resulting recommendations and requirements for the intervention material were applied when refining the existing IT-based intervention and when translating the IT-based intervention into a print-based intervention. Results consisted of recommendations regarding font size and color, amount of text to read, appropriateness of content (eg, inclusion of recommendations regarding diet, sleep hygiene, or pain management), and considerations with regard to the target group (eg, mobility, physical limitations, and the role of social support, loneliness, and wellbeing).

Both interventions are based on self-regulation theory [65,66] and principles of behavior change (eg, shaping knowledge, feedback and monitoring, goals and planning, social support, comparison of behavior, rewards) [67] and are designed to facilitate a physically active lifestyle by promoting regular self-monitoring of PA. Participants of both the IT-based and print-based interventions received PA recommendations based on WHO recommendations for this age group regarding endurance training (at least 150 minutes per week of moderate exercise, at least 75 minutes of intensive exercise, or a combination of both, in bouts of at least 10 minutes), strength training (at least 2 units per week for the 8 main muscle groups on non-consecutive days), and balance and flexibility training (at least 4 units per week for 5 minutes) [2]. Depending on gender, participants were provided with different brochures (online and offline) outlining exercises for different difficulty levels and displaying pictures of male or female older adults modeling these exercises.

They also received a weekly PA diary in the form of a pyramid to subjectively track their PA behavior. Intervention material was printed for the print-based intervention group, while participants in the IT-based group received access to the exercise brochure and diary on a website and were provided the opportunity to download a smartphone app containing similar features. On the website or Android app, as well as in the printed diary, weekly feedback on whether participants reached PA goals was provided by displaying the amount of minutes or units exercised in the according week, as well as units required to reach the goal (ie, WHO recommendations for moderate exercise time, flexibility, and strength training). In addition to the weekly feedback, the website and Android app also provided participants a daily overview of PA. In the daily overview, participants had access to more detailed information about the exercises for the individual muscle groups, whereas in the weekly overview (in the IT-based and print-based interventions), only information about the progress towards the weekly goals was provided (see Figure 2). The app also allowed participants to record exercises without an active internet connection and synchronized recordings with the website, once the device was connected to the internet again. Participants in the IT-based intervention group who used the PA tracker in addition to the website or app also monitored their PA behavior objectively via synchronization of data regarding daily steps with the website. Data from the PA trackers will be analyzed based on methods presented in Meyer et al [68] to gain insights into participants’ PA behavior in everyday life during the time between assessments.

Paralleling the 10-week IT-based and print-based interventions, weekly group sessions with up to 25 participants per group were offered separately to both intervention groups. Trained student assistants led these sessions in which participants performed endurance, strength, balance, and flexibility exercises in groups or went for walks in the communities. Each session was 90 minutes long and designed to include 60 minutes of exercise, 10 minutes of answering questions, and 20 minutes of discussion of topics surrounding healthy aging that changed every week (eg, social support, relaxation, pain management). During their first weekly group meeting, participants received the necessary equipment (printed material or information to access the website and/or a Fitbit) and a comprehensive introduction on how to use the equipment and materials. After the first follow-up, interventions continued for another 6 months. After the initial 10 weeks of the interventions, the weekly group meetings were replaced with monthly events held over the course of the following 6 months. During these events, participants of all intervention groups were offered workshops and lectures on lifestyle-related topics, such as healthy nutrition in older age, overcoming loneliness, and strategies for developing healthy habits.

We will develop a toolbox during this project that will include both IT-based and print-based interventions and all related intervention materials that can be used later for long-term implementation and dissemination via different stakeholder groups (for further detail, please see the Discussion).

Ethical approval was obtained from the Medical Association in Bremen (RA/RE-635, on July 3, 2018). The study was registered with the German Clinical Trials Register on January 10, 2019 – number DRKS00016073. All study participants were fully informed about the study and were requested to give informed consent.