We designed a mixed methods intervention trial including three study modules. The study is located in the Bremen-Oldenburg metropolitan region. The region is a geographically and politically defined area in the Northwest of Germany comprising about 150 rural and urban municipalities. The study started in February 2015 and will last until January 2018.
Community readiness assessment
In the first study module, a cross-sectional community readiness assessment for PA interventions in older adults (65–75 years) is conducted in a sample of municipalities (settlements and districts) within the Bremen-Oldenburg metropolitan region. Municipalities are selected if they already display a comparably high proportion of older adults or if they expect a high increase in the proportion of older adults over the next decade. Overall, 24 municipalities (12 rural, 12 urban) are included in this assessment.
The community readiness assessment is based on a structured interview administered to key informants in each of the selected communities. Key informants are, for example, representatives from local public authorities, senior citizen organisations, sports clubs, for-profit PA providers, faith-based organisations, and members of the target group (i.e., older adults living in the municipality). Strategies to identify key informant include internet searches and snowball sampling via interviewees. The interview comprises 36 semi-structured questions covering the five key dimension of community readiness: (1) community knowledge of the issue, (2) community knowledge of the efforts, (3) community climate, (4) leadership, and (5) resources [20]. In each municipality five key informant interviews are conducted [18]. Interviews are administered either face-to-face or as telephone interviews. Transcripts of the responses are analysed by two independent raters and scored from 1 to 9 for each readiness dimension, according to rating criteria described in the community readiness assessment manual [20]. The raters agree upon a final score for each municipality.
Effects of strategies to increase community readiness
In the second study module, eight municipalities identified as having the lowest community readiness score in module 1 are randomly assigned to one of two study groups: active enhancement of community readiness (intervention) versus no enhancement (control; see Fig. 1).
Treatment conditions
In the intervention group strategies are applied to increase community readiness for PA interventions in older adults. The community readiness concept suggests selecting strategies according to the stage of readiness a community has attained. Practical strategies for the lower readiness stages include (a) the identification of local key stakeholders via one-on-one visits with community leaders, (b) information campaigns to increase problem awareness in the community (articles to local newsletters, presentation of in-depth local statistics), (c) the establishment of a community working group on PA in older adults, and (d) the recruitment of older adults for the PA intervention via local health practitioners [19]. Intervention materials are developed to support the information campaign and the installation of the working group. Furthermore, technology-based interactive tools are provided to engage stakeholders and citizens in PA, such as public displays to visualize a group’s activity, or an online social network for uploading and commenting on photos of places and facilities in the community where one can be physically active.
In the control group no strategies are applied to increase community readiness.
After setting up the local working group in the intervention group, older adults aged 65 to 75 years are recruited for participation in a PA intervention in both study groups. The PA intervention is based on regular group meetings and a simple counselling aid to facilitate self-regulation of PA behaviour. The recruitment process comprises standard recruitment strategies that are applied in both study groups, such as newspaper articles, public service advertising, and direct mailing obtained from the local residents’ registration offices. In addition to the standard procedures, the built-up local networks in the intervention communities are involved in the recruitment process using their personal or organisational capacities to reach older adults for the intervention.
Outcomes
Reach of the PA intervention is chosen as a main outcome criterion for evaluating the effects of strategies to enhance community readiness. According to the RE-AIM model (Reach Efficacy Adoption Implementation Maintenance), programme reach is defined as the number of program participants divided by the number of the target population in the municipalities [27]. Administrative data is used to determine the size of the target population for the denominator. Apart from reach, access equality is evaluated as another outcome criterion. We define access equality as number and proportion of vulnerable older adults among the participants of the PA Intervention. Vulnerability is defined as having at least one socio-demographic risk factor (low socioeconomic status or migrant background or male) and one health-related risk factor (low level of physical activity or being obese). Information on the participant characteristics is gathered during baseline assessment for the PA intervention. From these variables, an index is constructed to distinguish between vulnerable and non-vulnerable older adults. Cost-effectiveness is used as the third outcome variable. The recruitment costs per person are calculated for the two study groups. Recruitment costs for the intervention group also include per capita shares of the overall costs of the community readiness assessment and the community readiness strategies described above.
Statistical analysis
To assess the effects of the community readiness enhancing strategies, odds ratios (OR) are calculated (reached versus non-reached in intervention and control group in a 2 × 2 table). As the size of the target population varies between the municipalities and because the control and intervention municipalities are selected at a later stage of the study, exact determination of power and sample size is not possible. However, a population size of at least 2,000 older adults aged 65 to 75 years in each of the eight municipalities would allow us to detect differences in programme reach at the level of OR = 1.5 (α = 0.05, power = 0.80), stratifying for region (rural/urban) and assuming that small proportions of the population will be reached (3% reached in the intervention group vs. 2% reached in the control group). A population size of at least 2,000 older adults is a realistic assumption as almost all of the 24 municipalities in the community readiness assessment have even more inhabitants in this age group.
For the analysis of the costs, the difference in recruitment costs is divided by the difference in effects to obtain the incremental cost-effectiveness rate (ICER). The effect is measured in terms of participation rates and the rate of participants from vulnerable groups.
Reasons for non-participation in PA interventions
The third module focusses on biographical and cultural reasons why vulnerable older adults participate or do not participate in PA interventions, and which special recruitment strategies can be identified for underrepresented target groups. For this purpose, 10 to 15 qualitative interviews with participants and non-participants of the PA intervention are conducted (ratio 1:3). Potential interviewees are identified throughout the recruitment process in module 2. Grounded theory methodology is used for sampling and analysing the interview material [28].
Ethics statement and consent
The study protocol was approved by the Ethics Committee of the University Bremen on February 11, 2015. All participants in the community readiness interviews, the PA intervention, and the qualitative interviews receive written and oral information about the study. All interviewees give informed consent for their data to be used.