*2.1. Data Collection*

In July 2017, *N* = 906 enterprises in the county of Reutlingen in southwestern Germany were determined as potential respondents of the survey. This number contained all of the enterprises in the county, except for the very small ones: we excluded craft businesses with less than 10 employees and non-craft enterprises with less than 20 employees.

At the end of July and the beginning of August 2017, we sent a standardized questionnaire to these 906 enterprises. Craft businesses received our letter via the local chamber of crafts, which supported the survey; non-craft enterprises received the questionnaire directly from our institute, as we were able to use the complete address data record of the county's enterprises that was available from a marketing agency (Creditreform [29]). An enclosed leaflet included the request to hand out the questionnaire either to the managing director or to a member of the personnel department. Fourteen days after the first invitation to participate in the survey, a reminder was sent to all of the potential participants, regardless of whether some of them had already returned the questionnaire.

A formal ethical approval from the ethical committee at the University Hospital Tübingen was not required. Study participants were informed that the study was voluntary, and that all of the data were analyzed anonymously.

#### *2.2. Questionnaire*

The questionnaire was based on previous studies and current literature [24,30–33]. It was developed, discussed, and formulated in a multidisciplinary team consisting of a specialist in occupational medicine (MAR), a sociologist and public health researcher (AS), and a medical student (AH). After a pretest with *N* = 24 participants (senior employees of the personnel departments of different enterprises of the metal and electrical industry in southwestern Germany), the questionnaire was partially modified and supplemented to ensure good comprehensibility.

Based on a self-developed questionnaire for a similar survey of companies in Constance County that was conducted in 2015 [34], questions covered the implementation status of four categories of health-related measures within the enterprise, referring to the above-mentioned four components of workplace health management. Each category was assessed by several items depicting typical measures (cf., Table 1). Answers regarding the implementation of individual measures in the company within the last two years could be given on a three-point Likert scale (zero = 'no', one = 'no, but in concrete planning', two = 'yes'). Hereby, the order of the four categories was as following: workplace health promotion (six items and one possibility for free-text indication), occupational health and safety (seven items and one free-text indication), personnel development (five items and one free-text indication), and reintegration management for employees on long-term sickness absence (eight items and one free-text indication) (cf., Table 1 for all items).



Explications regarding Table 1: Fields in italics: in general legally required according to German laws. The question in the questionnaire had read: 'Which of the listed measures have taken place in your company in the last two years? (Please also take into account offers that took place outside the company but were (co)financed by the company.)'. Answers were given on a 3-point Likert scale: 0 = 'no', 1 = 'no, but in concrete planning', 2 = 'yes'.

Next, the participants were asked about their satisfaction with the current implementation of the four categories of measures. Here, a four-point Likert scale was used (zero = 'very dissatisfied', one = 'rather dissatisfied', two = 'rather satisfied', and three = 'very satisfied').

At the end of the questionnaire, sociodemographic data of the respondents and company characteristics were gathered (branch, number of employees, availability of occupational health and safety experts, and number of employees addressed in reintegration management during the last two years).
