The effectiveness of the community-based intervention for the promotion of PA is examined over the course of 1 year in a controlled intervention study. The PA behavior of individuals aged 25 to 75 years residing in the intervention districts of the city of Duesseldorf will be compared with that of individuals residing in the respective control districts of the city of Wuppertal at baseline and 1 year. Random samples were drawn based on data from the respective residents’ registration offices. As the size of the cities in North Rhine-Westphalia did not permit an assessment of all residents living in both cities, as was done in the original study [14], adjacent city districts were selected that encompass a similar number of residents as the cities included in the original study. Publicly available information on the levels of socio-spatial deprivation of the districts of Duesseldorf and Wuppertal was used [28,29] during the selection process to ensure a composition of districts balanced by the proportions of residents living in them. The socio-spatial deprivation index is composed of indicators, such as welfare benefits, living space per person, and migrant population. In Duesseldorf, the index ranges from 1 (no deprivation) to 5 (high deprivation), and in Wuppertal, the index ranges from 1 (no deprivation) to 4 (high deprivation). Corresponding to the heterogeneous population of the original study, the Duesseldorf districts of Flingern Nord, Flingern Sued, Oberbilk, Friedrichstadt, Gerresheim, Eller, and Wersten and the Wuppertal districts of Elberfeld and Barmen were chosen. The selected areas and districts in the 2 cities are shown in Figure 1 [30].

On March 1, 2021, a random sample of 2500 residents (25-75 years) based on postal codes (500 individuals per age group: 25-35, 36-45, 46-55, 56-65, and 66-75 years) was drawn for the Duesseldorf districts of Flingern Nord, Flingern Sued, Oberbilk, Friedrichstadt, Gerresheim, Eller, and Wersten, and 2500 residents, using the same distribution, were sampled for the Wuppertal districts of Elberfeld and Barmen. Equal proportions of men and women were included. Thus, the overarching setting in this study is the respective city district in both cities, and the subordinate settings for the implementation of the complex intervention are the organizations located in these districts (eg, senior citizen associations, sports clubs, and companies), as well as the public spaces located in them (eg, parks and green spaces).

Recruitment for the study was conducted from mid-May to the end of December 2021. Individuals were informed about the study and invited to participate in it via mail. The study information, including information on data protection and the consent form (Multimedia Appendix 1), was provided in the mailed letters. Individuals interested in participating in the study were asked to return the signed consent form (and keep one for their records) and a form, including their telephone number, to the study team in the mail. Alternatively, they could contact the study team directly to make an appointment for the baseline interview and return the signed consent form after the completion of the telephone interview. However, in that case, verbal consent was provided by participants at the beginning of the telephone-based interview. In total, 3 invitation letters were sent to potential participants, and 2 reminder letters were sent 3 and 6 weeks after the first invitation letter. Once contact information was provided by a potential participant, up to 4 phone calls were made to reach the person and schedule an appointment for the baseline interview.

Owing to the COVID-19 pandemic, we expected the response rate to be extraordinarily low. To ensure that we reached the required sample sizes in both cities, a second random sample of another 4000 residents (2,000 in each city) was randomly drawn in both cities and with the same age and gender composition on July 29, 2021. During the second wave of recruitment, only 1 invitation and 1 reminder letter 4 weeks later were sent. Individuals were included in the study if they were (1) between the ages of 25 and 75 years, (2) able to understand German, and (3) residing in Duesseldorf or Wuppertal.

The baseline assessments started at the end of May 2021 and were completed at the end of March 2022. Baseline assessment entails a questionnaire completed by participants over the phone. Two study nurses and 4 student assistants underwent structured training to conduct the 1-hour telephone interviews. Data entry is done simultaneously to the interview over the secured web application, REDCap (Research Electronic Data Capture; Vanderbilt University). The respective REDCap study database is hosted at the Institute for Biometrics and Epidemiology at the German Diabetes Center Duesseldorf and is part of the regular Institute for Biometrics and Epidemiology at the German Diabetes Center IT environment with respect to data security and backups. After the interview, study participants are asked to track their PA over 7 consecutive days using the YAMAX EX210 pedometer, which records daily number of steps. In addition, they are asked to complete a wear-time diary to document the steps walked every day, the times the pedometer was worn, and the possible reasons for not wearing it. After 7 days, participants return the pedometer to the study team via mail, using a stamped return envelope provided by the study team. To validate the pedometer data, a subsample of approximately 30% of study participants receives an accelerometer (ActiGraph wCT3X-BT; ActiGraph LLC) in addition to the pedometer, which they are asked to return in a stamped return envelope after 7 days of wear time. The intervention is being implemented in Duesseldorf for 1 year since April 2022, and follow-up assessments will be conducted, starting in May 2023. The same participants who participated in the baseline assessments will be contacted again via mail, invited to participate in the telephone-based interviews again, and asked to wear pedometers (and accelerometers) for 1 week afterward. To ensure confidentiality, all person-related data assessed in this study will be pseudonymized; that is, all study participants will be assigned an ID number. The roster with the names and ID numbers will be saved in a password-protected file on the server of the Institute of Medical Sociology. The data will be stored for 10 years after the completion of the trial and deleted afterward.

Ethics approval to conduct the study was obtained from the Ethics Committee of the Medical Faculty of the Heinrich Heine University Duesseldorf (reference number: 2021-1364; April 6, 2021). The study was registered in the German Clinical Trials Register on April 21, 2021 (trial number: DRKS00024873). All study participants are fully informed about the study and are requested to provide informed consent.

A number of 1000 steps were assumed as the minimal relevant effect. According to the previous study by De Cocker et al [14], the SD of the number of steps is ≤4500. This ensures a power of 80% for a 2-sided t test at the 5% level with a total of 638 patients (SAS, version 9.4, PROC POWER, TWOSAMPLEMEANS statement). A total of 798 participants were included in the study to compensate for an expected dropout rate of 25%. It should be considered that the power calculated here represents a lower limit because a noticeable gain in power is expected in the planned analysis owing to the adjustment for baseline characteristics of participants. Figure 2 shows the study design.

We will use the propensity score method for statistical analysis to allow for a nonrandomized design of the study. Specifically, the primary outcome will be estimated in a weighted 2-group comparison (intervention: yes vs no). For weighting, we will use matching weights [31] that are derived from a propensity score model for the intervention effect and an exhaustive list of covariates (Multimedia Appendix 2). The analysis will be blinded for the intervention effect and performed according to the intention-to-treat principle; missing values will be replaced via multiple imputation.

Sociodemographic characteristics of study participants, such as age, migration background, marital status, and socioeconomic status (including education, employment status, occupation, and income), as well as history of chronic disease and current diagnoses are assessed during the telephone interview.

The primary outcome is the number of steps assessed via pedometers. Participants are asked to track their steps using the YAMAX EX210 pedometer, which records the daily number of steps. They are also requested to complete a wear-time diary to document the steps walked every day, the times the pedometer is worn, and the possible reasons for not wearing it. PA is also assessed using the Global Physical Activity Questionnaire in the telephone-administered questionnaire [32]. It is one of the most frequently used questionnaires for measuring PA with high levels of reliability and validity reported when comparing it with other PA questionnaires (International Physical Activity Questionnaire—Short Form and European Health Interview Survey—Physical Activity Questionnaire) [33]. On the basis of the Global Physical Activity Questionnaire, the number of minutes spent on MVPA and different types of movement will be determined (eg, for transport, see Table 1). This questionnaire was chosen because it is also used in the National Cohort (NAKO) health study [34], a large German national cohort, allowing for a comparison of PA levels of our sample to a large population-based sample.

Furthermore, a subsample of our study participants also receives an accelerometer (wCT3X-BT) to track habitual PA. Participants are instructed to position the device at their nondominant wrist during the day to capture triaxial acceleration at a sample rate of 30 Hz. Data will be processed with ActiLife (version 6.13.4) software (ActiGraph LLC), using the algorithm by Choi et al [38] to derive valid wear times with at least 4 days of a minimum 8 hours (480 minutes) wear time. Within 10-second epochs, data are analyzed using cut-off points for MVPA, energy expenditure, and metabolic equivalent of tasks, according to Freedson et al [39].

Secondary outcomes assessed at baseline and follow-up in the telephone-administered interview are described in detail in Table 2.

Briefly, at the individual level, height and weight, general self-rated health, and health behavior (smoking status, fruits and vegetables consumption, and alcohol use) are assessed. In addition, the determinants of PA are assessed (eg, social support). At the environmental level, perceptions of characteristics of the physical environment relevant for engaging in PA are assessed. To assess knowledge and awareness of intervention messages and activities spread and implemented in the intervention city districts, self-generated items are included in the follow-up questionnaire. Study participants are also asked about health care use (physician’s visits, contacts with therapists, hospital stays, and rehabilitation) retrospectively for the previous 6 months and costs pertaining to devices bought to support engagement in PA or membership fees (eg, gym membership), as well as sick leave days.

The aim of the process evaluation is to document the frequency and intensity of the implemented intervention activities at every level of the complex intervention over the course of 1 year. This ensures intervention fidelity, meaning that it will be possible to assess whether the implementation of the intervention in Duesseldorf, Germany, was comparable with the implementation of the original intervention in Ghent. According to the RE-AIM framework [54], reach (the extent to which the program reaches the intended target group), effectiveness, adoption, implementation, and maintenance (of individual behavior change of the target group as well as long-term decision-making behavior of relevant stakeholders to consolidate intervention activities) will be assessed. A similar approach will be followed as developed by Van Acker et al [51].

For this study, the five components of the framework will be determined based on various data sources: (1) the commercial register of companies located in the intervention districts or the register of the Chamber of Commerce and Industry, (2) the individual-level data assessed in the telephone-administered interviews, and (3) a web-based questionnaire that heads of participating organizations (eg, companies) will be invited to participate in. The web-based questionnaire includes four thematic item blocks regarding (1) the characteristics of the respective organization (eg, company size) and knowledge of the intervention, (2) the adoption of the intervention and reasons for or against adoption, (3) the implementation of the intervention (eg, of the different intervention components, frequency and duration, and resources needed to implement the intervention or raise awareness), and (4) long-term intervention maintenance (reasons for or against it). Examples of the items in the web-based survey are shown in Textbox 1.

To estimate reach, first, the number of organizations potentially participating in the program in each intervention district will be determined based on the commercial register or by the Chamber of Commerce and Industry. Second, the number of organizations participating in the program will be assessed and surveyed using the web-based questionnaire. Furthermore, the proportion of individuals who were aware of intervention activities will be calculated, and to determine representativeness, age, gender, level of education, and occupation of those aware of the intervention will be compared with those unaware of the intervention. Effectiveness will be determined by comparing the 4 domains of active living (PA for transport, to work, in the household, and during leisure time activities) among individuals with and without awareness of the intervention. Adoption will be determined by estimating the proportion and representativeness of organizations implementing the intervention (compared with those not implementing the intervention). The implementation of the different intervention components (website, print-based materials, and initiation of partnerships) will be assessed, and an implementation score ranging from 0 to 100 will be calculated (see the study by Van Acker et al [51] for more details regarding the calculation of the score). To estimate maintenance, the proportion of organizations that voiced the intention to maintain the intervention will be determined.

The health economic evaluation will include a cost-effectiveness analysis (CEA) and a cost-utility analysis from a societal perspective. An incremental cost-efficacy ratio (ICER; additional costs per additional person walking 10,000 steps/day) and an incremental cost-utility ratio (ICUR; additional cost for an additional QALY gained) will be determined. The outcome of the CEA is based on the primary outcome of the intervention and will be taken from the step counter measurement. QALYs will be calculated based on the EQ-5D-5L [44], which is a widely used standardized instrument to assess health-related quality of life. The EQ-5D-5L is evaluated using a German tariff to obtain preference weights [55]. As the intervention might influence people beyond health, we will also consider capability well-being using the ICEpop CAPability measure for adults (ICECAP-A) [48]. We will use a UK scoring tariff because a German tariff does not exist yet [56] and compare utilities based on capabilities with those based on QALYs.

The costs considered in the health economic evaluation comprise costs related to the development and provision of the intervention (eg, print-based materials, website, and time associated with the development of the intervention as opportunity costs), health care use (physician’s visits, contact with therapists, hospital stays, and rehabilitation), and costs associated with PA behavior (devices to support PA, membership fees [eg, gym membership], disability to work, and patient time). Productivity loss due to disability to work is calculated based on the human capital approach [57]. In a sensitivity analysis, patient time associated with PA, evaluated via opportunity costs, will also be examined. A discounting of costs and outcomes is not planned because of the short study duration.

The analysis will be performed according to the intention-to-treat principle. In the base case analysis, missing values will be imputed via multiple imputation. To make the random samples from the intervention and control region comparable, propensity score matching in the form of a weighted regression model will be applied after the imputation and for each imputed data set as suggested by Al-Janabi et al [58]. Covariates used for the matching will be the number of steps taken at baseline, age, gender, level of education, and household income.

Mean incremental costs and the mean incremental outcome of the CEA will be estimated by a using a Generalized Linear Model with gamma distribution and log link function. To estimate the mean incremental outcome of the CEA (an additional person who walks 10,000 steps/day), a Generalized Linear Model with binomial distribution and logit link function will be estimated. Bootstrap procedures will be used to calculate 95% significance intervals for the ICER and ICUR [59,60]. To investigate uncertainty surrounding the ICER and ICUR, cost-effectiveness and cost-utility planes will be generated [61,62].

To the best of our knowledge, this is the first study in Germany aimed at replicating the effects of a complex intervention for PA promotion previously shown to be effective in another European country. The study findings will contribute to the growing body of evidence in Germany concerning the role of complex community-based interventions for the promotion of PA in the general population [19-27]. We anticipate to obtain results similar to those reported in the original study, where adults exposed to a complex PA intervention for 1 year displayed a greater increase in daily step count (ie, average of daily steps walked increased by 896 in intervention vs no increase in the comparison city) and were more likely to reach the WHO recommendation for PA than controls (ie, 8% increase in the number of individuals reaching the recommended 10,000 steps) [14]. With regard to the expected results of the process evaluation, we foresee that the processes involved in the implementation of the various intervention components may be somewhat different, as the organizational structures of administrative offices and community organizations in Germany may operate differently compared with Belgian organizations. However, by using a similar approach to monitoring processes, we expect to be able to compare our results with those obtained in the Ghent study. Furthermore, similar to the health economic evaluation of the original intervention indicating that the intervention was cost-saving [63], we expect “10,000 Steps Duesseldorf” to be cost-effective. Other previous health economic evaluations of population-level, community-based interventions addressing PA also demonstrated cost-effectiveness [64]. Finally, previous research suggests that community-based interventions involving local stakeholders in particular may yield intervention effects after the primary evaluation is completed [13,15,16]. Our study was preceded by a 6-month participatory planning phase involving key stakeholders of PA promotion in the city of Duesseldorf. Therefore, it is conceivable that the effects may become visible after the completion of the 1-year follow-up in our study.

As outlined earlier, the primary focus is on the comparison with the Ghent- and Rockhampton-based studies. Furthermore, we expect that the recommendation and use of step counters will be associated with a similar increase in daily steps walked, as was previously reported in several reviews and meta-analyses synthesizing the results of studies examining the effects of population-based PA interventions [11,13]. The analysis of the accelerometer data of the subsample in our study will allow us to analyze weekly minutes spent on MVPA and compare the results to those reported by Wahlich et al [13] in their review and meta-analysis.

A major strength of this study is the use of an innovative methodological approach for examining the effects of a complex community-based PA intervention in Germany and the ability to compare the results obtained in this study to those reported in 2 other countries in 2 continents. To our knowledge, this is the first study to examine the effect of a multistrategy population-level intervention on PA behaviors targeting individuals aged ≥18 years, which was developed based on a participatory approach. Another strength is the use of objective measurements in a subsample of the study, allowing for the validation of the questionnaire data collected in the telephone-based interviews. A limitation is that, unlike in the original study, 2 entire cities could not be compared in terms of PA behavior of participants living in them. However, by balancing the city districts of both cities included in our study using indicators of socio-spatial deprivation, we hope to have obtained a similar sociodemographic composition of our study participants. Another limitation is the ongoing COVID-19 pandemic, which complicated the recruitment of study participants. Adjustments in the number of potential study participants contacted and invited to participate were necessary to offset the difficulties in enrolling participants to the study. Finally, in our study, a trial-based health economic evaluation is conducted that does not include modeling of long-term effects and costs. Nevertheless, the health economic evaluation will provide detailed information to decision makers regarding short-term cost-effectiveness of the intervention.

Should this study reveal positive effects on PA in the intervention compared with the control city districts, conducting a 4-year follow-up to assess the long-term effects of the intervention on PA and quality of life, similar to Cocker et al [15], may be useful to demonstrate long-term effects. Furthermore, comparing the results of the 3 studies conducted in Rockhampton, Ghent, and Duesseldorf may help devise recommendations for adaptation and implementation in other European cities. In Germany, a rollout of the intervention in other cities (possibly starting with the control city, Wuppertal) could be a next step after the completion of the primary evaluation. A plan for scale up or dissemination will be developed at the end of the project. In the original Ghent-based study, the intervention was rolled out in all of Flanders and evaluated based on the RE-AIM framework [14]. A similar approach could be taken after completion of this study (eg, in the entire state of North Rhine-Westphalia or in an additional German state). Following the original study by Dubuy et al [18], combined baseline data from Duesseldorf and Wuppertal from this study could serve as a historical control group. Finally, future research may also include health economic modeling techniques that could shed further light on the long-term effects of intervention participation on cost-effectiveness of this multistrategy approach for PA promotion at the community level.