Promoting physical activity and health literacy: study protocol for a longitudinal, mixed methods evaluation of a cross-provider workplace-related intervention in Germany (The AtRisk study)

The design and research questions of the AtRisk study were guided by the RE-AIM Framework [21]. The RE-AIM Framework is based on system-based and socio-ecological thinking and enables an overall assessment of the impact of the intervention. It has been specifically developed to facilitate the evaluation of interventions in real-world settings. The model identifies five evaluation dimensions: the Reach, the Efficacy or Effectiveness, the Adoption, the Implementation and the Maintenance of a public health intervention. The five RE-AIM dimensions occur at different levels, e.g. the individual, the organizational or the community level. Reach is defined as an individual-level measure of participation and refers to the percentage and risk characteristics of the participants. Efficacy or Effectiveness refers to the impact of an intervention on important intervention-specific outcomes, including potential negative effects, quality of life, and economic outcomes. Adoption refers to the proportion and representativeness of settings, e.g., such as worksites, health providers and/or payers, that adopt a given program or intervention. The Implementation refers to the extent to which a program is delivered as intended. At the setting level, implementation refers to the intervention providers’ fidelity to the various elements of an intervention’s protocol, including consistency of the delivery as intended and the time and cost of the intervention. Finally, the dimension Maintenance refers to the extent to which a program becomes institutionalized or part of the routine organizational practices and policies. Beyond, maintenance also refers to the individual level and has been defined as the long-term effects of a program on outcomes after six or more months after the most recent intervention contact [21]. Table 1 provides a description of the research questions for each of the five RE-AIM dimensions. The research questions will be answered by means of a mixed methods design combining quantitative and qualitative data sources. Quantitative data regarding intervention effectiveness will be collected using various questionnaires as part of a natural experiment. Qualitative data will be collected by means of supplementary interviews, informal field notes and document studies.

The cross-provider intervention is based on the cooperation of the German Pension Fund Rhineland and cooperating German Statutory Health Insurances. It incorporates the cooperating companies and combines two components: A) a behavior-oriented lifestyle intervention (for all participants) and B) assignment of a health coach (optional). The following components are offered within the comprehensive cross-provider health prevention service:

The quantitative study of AtRisk is a non-randomised controlled trial. The controlled trial made use of a natural experiment in which the implementation and delivery of the intervention was not manipulated by researchers. Instead, the allocation to one of the two possible conditions compared in this study (cross-provider intervention: A) and B); single-provider intervention: A)) is given according to the real world condition. If the statutory health insurance of the participant signed cooperation with German Pension Fund Rhineland in regard to preventive health services, the participant was allocated to the cross-provider intervention. Otherwise, the participant was allocated to the control group (single-provider intervention).

The AtRisk study compares the cross-provider intervention with a single provider intervention and encompasses three measuring points: T0 = start of the behavior-oriented lifestyle intervention (baseline); T1 = end of the behavior-oriented lifestyle intervention; T2 = 6 month follow-up (see Fig. 2).

The study was approved by the German Sport University Cologne Ethics Committee (reference number: 93/2015) and registered in the German Clinical Trials Register (ID: DRKS00010693). Written informed consent is taken from each participant. The recruitment of participants started in July 2016 and will be completed in August 2017. Six months follow-up (T2) will be completed in January 2018.

The study population comprises employees with health related risk factors. Eligible participants were invited from the company doctor to participate in the behavior-related intervention (component A). The company doctor conducts the medical entry examination and the participant makes an application for the preventive measure at the German Pension Fund (§ 31 Social Code Book VI). The corresponding forms for the participant (G0180 and G0185) and the medical report of the company doctor (G0190) can be downloaded [23]. The application is approved by the German Pension Fund Rhineland. Eligibility criteria for participating in the present study is (1) a formally approved application for a preventive health service by the German Pension Fund, (2) age 18 to 65 years, (3) First health impairments (of the musculoskeletal system, internal organs or mental impairments) (4) written informed consent to participate in the study. Exclusion criteria are: (1) The indication for a rehabilitative treatment; (2) the need for acute care; (4) lack of understanding the German language.

The participants answer a questionnaire on physical activity, health literacy, health-related quality of life and sociodemographic variables. The baseline questionnaire (T0) is answered at the beginning of the initial phase. The questionnaire at the end of the training phase (T1) and the 6 month follow-up questionnaire (T2) are collected using a postal questionnaire.

We chose leisure physical activity (MET¹-min/week) as our primary outcome. Physical activity was operationalized by the Global Physical Activity Questionnaire [3, 24], which collects information on both physical activity during a typical week within three settings (workplace, transport and leisure time) as well as on sedentary behaviour. Activity specific scores are summed to give the total MET-min/week. Thereby, each minute of vigorous physical activity is multiplied by 8 METs and each minute of moderate physical activity by 4 METs. Bull et al. (2009) showed a moderate to strong positive relationship of the Global Physical Activity Questionnaire with the International Physical Activity Questionnaire (concurrent validity: Spearman’s rho 0.45–0.65), and the reliability was of moderate to substantial strength (kappa 0.67 to 0.73; Spearman’s rho 0.67 to 0.81) [3]. Compared to the accelerometer data, the Global Physical Activity Questionnaire provided low-to-moderate validity and generally acceptable evidence of reliability [25].

The secondary outcome measurement of the AtRisk study is health literacy. We assessed health literacy by the questionnaire from Lenartz [15] that is based on the definition of health literacy, representing “[…] the cognitive and social skills which determine the motivation and ability of individuals to gain access to, understand and use information in ways which promote and maintain good health.” ([26], p. 357). Lenartz (2012) developed a structure model explaining health behavior and health through the influence of individual basic competences (health-related knowledge, health-related skills and beneficial personality triats) and further developed abilities and skills. Based on this model, the health literacy scale from Lenartz comprises 30 items assessing six abilities and skills: self-regulation (5 items), self-control (5 items), self-perception (5 items) and responsibility (5 items), communication and cooperation (5 items) and the handling of health information (5 items) [15]. Each item has four response options”not correct at all”, “rather not correct”, “rather correct”, “correct” (Fig. 3).

Additionally, the domain „Actively managing my health“of the Health Literacy Questionnaire (HLQ) is used [27]. This domain contains five items (“I spend quite a lot of time actively managing my health”; “I make plans for what I need to do to be healthy”; “Despite other things in my life, I make time to be healthy”; “I set my own goals about health and fitness”; “There are things that I do regularly to make myself more healthy”) measured by a four point likert-scale (strongly disagree, disagree, agree and strongly agree). The scale score is devised by summing the item scores and dividing by the number of items in the scale [27].

For obtaining person-related variables, age, sex, height, weight, education level, family status and health-related lifestyle were assessed by unstandardized questions. Beyond, health-related Quality (HRQoL) of life was measured using the EQ-5D-5 l questionnaire [28]. The EQ-5D is a standardized measure of health status and assesses five dimensions of HRQoL: mobility, self-care, daily activities, pain and discomfort, anxiety or depression. EQ-5D-5 L provides five levels for answering each dimension: “no problems”, “slight problems”, “moderate problems”, “severe problems” and “extreme problems”. Descriptive answers were converted into a single summary index [29]. Index value of health status range from ‘1’ (perfect health) to ‘0’ (death). Beyond, workability was assessed by the work ability index (WAI) [30, 31] employability was measured by the 3-item subjective prognosis of gainful employment (SPE-scale) [32‐34]. Overall, the following instruments are used (Table 2).

Descriptive statistics will be used to describe the baseline characteristics of study population and to explore the distributions of the variables at the individual level. Depending on the distributions of the quantitative variables parametric and non-parametric tests are used to compare the intervention and the control group at baseline.

The two study groups will be compared using ANCOVA, with the outcome variables leisure time physical (primary outcome) activity and health literacy (secondary outcome), at 6 month follow-up set as the dependent variable [35]. Baseline values and the rehabilitation center will be used as a covariate to control for imbalance between control and intervention groups at baseline.

Standardized mean-difference effect size will be calculated by using the means of the two groups, mean-square error and the correlation between the covariate and dependent variable.

Intervention outcomes will be evaluated in relation to the base case data but also to the intention to treat principle. The base-case analysis will be performed using data restricted to those patients who replied to the postal 6 month follow-up questionnaire. The intention-to-treat analysis (ITT_LOCF) will be performed assuming “last observation carried forward” since the effectiveness of physical activity promotion is considered controversial [36‐38]: if 6 month follow-up data (T2) were missing, it was assumed that the physical activity data were the same as at baseline (T0). The statistical analyses will be conducted independently of the research team. The data will be analyzed using the statistical software IBM Statistics 23.0.

The calculated total sample size is 237 (ANCOVA: Fixed effects, main effects and interactions; Input: effect size f = 0.25 with 80 % power at p < 0.05; number of groups = 8 (group = 2; ambulatory rehabilitation center = 4); numerator df = 7; number of covariates = 2 (age; baseline value)). Therefore, it is estimated that a sample size of 119 participants per study arm would be required to detect a medium effect between the cross-provider intervention and the single-provider intervention. To allow for missing data and loss to follow up, we aim to recruit 150 participants per study arm.

The qualitative study comprises semi-structured interviews, systematic field notes of stakeholder meetings and document analyses. Semi-structured interviews and field notes will be conducted to explore the questions in regard to the RE-AIM dimensions implementation (“What are the facilitators and barriers for the cross-provider intervention?”; “Is the intervention conducted according to the manual?”; “What are the different stakeholders’ expectations and needs regarding health coaching in a work-related setting?”; “What are costs associated with implementation for the different stakeholders?”) and maintenance (“What structures resources were built in companies, providers and social insurers to implement the cross-provider intervention in daily routine?”) (see Table 1). The different stakeholders involved in the intervention (German Pension Fund Rhineland, cooperating statutory health insurers, providers, health-coaches, participants, company doctors and company management) will participate in the semi structured interviews.

In regard to health-coaching the semi-structured interviews encompass the perspectives of different stakeholders. Amongst others, the aim is to get information on tasks and objectives of health coaching, the necessary qualification, skills and competences of a health coach as well as the working methods and main topics in health coaching. The interview partners will be interviewed personally once, using semi-structured qualitative interviews. Two researchers lead through the semi-structured interviews: one will lead the conversation and the other one will do the documentation. If the participants agree, the interviews will be audio recorded. All semi-structured interviews will be in German. After the interviews are completed, transcriptions will be done according to the rules of Kuckartz [39]. Transcriptions will be double-checked.

The transcripts will be analysed according to the structuring content analysis by Mayring [40], which is similar to the framework approach [41]. For the analysis a qualitative research specialist at German Sport University Cologne, Germany will use the software MAXQDA 12 in order to code the information and look for relevant trends and themes within the data. Common themes and issues will be grouped together in main- and sub-categories and frequencies will be noted in order to highlight key issues. Though this is a qualitative study, the frequencies will be reported. However, direct quotations will be pulled from the session reports when they are illustrative of the issues. Overall, around 15 interviews will be included in this study, approximately two interviews per stakeholder group. Beyond, field notes of protocols of conversations with stakeholders (German Pension Fund Rhineland, statutory health insurances, health coaches, therapists and managers of ambulatory rehabilitation centers) and documents will be considered to gain insight into the dimension adoption (“How many providers offer the cross-provider intervention?”; “How many companies cooperate in the cross-provider intervention?”; “How many statutory social insurers cooperate in the cross-provider intervention?”) and implementation (“Is the intervention conducted according to the manual?”).