*2.3. Statistical Analysis*

For each category of measures a score, ranging from 'zero' to '10', was calculated to represent a standardized implementation grade. This score was only calculated if at most one entry per category was missing. A score of 'zero' points corresponded to no offered measures and no measures in concrete planning, while a score of '10' stood for the complete implementation of all the listed measures in a given category. The legal requirements in a given category were considered 'fulfilled' if all of the legally required measures of that category had been implemented. All seven measures listed in the category 'occupational safety and health' were legally required due to regulations in the "Arbeitsschutzgesetz" ([Act on the Implementation of Measures of Occupational Safety and Health to Encourage Improvements in the Safety and Health Protection of Workers at Work]—ArbSchG (1996) [35]), in the "Verordnung über die arbeitsmedizinische Vorsorge" ([Ordinance on Occupational Health Care]—ArbMedVV (2008) [36]), in the "Arbeitssicherheitsgesetz" ([Act on Occupational Physicians, Safety Engineers, and Other Occupational Safety Specialists]—ASiG (1973) [37]), in the "DGUV Vorschrift 1" ([DGUV Regulation 1 "Principles of prevention"] (2013) [38]) and the "DGUV Vorschrift 2" ([DGUV Regulation 2 "Occupational physicians and OSH professionals"] (2011) [39]), and the first two measures listed in the category 'reintegration management' were legally required due to the respective regulations in "Book Nine of the Social Code" (Sozialgesetzbuch) (§ 167 SGB "Prevention" [40]) (cf., Table 1).

Rank correlation (Spearman's r) coefficients were calculated to analyze relationships between ordinal variables (such as, e.g., satisfaction with a given implementation status) and metrically scaled variables or when metrical variables were not normally distributed. Thus, e.g., Spearman's r was calculated to compare companies of different sizes, which were measured by their number of employees, in terms of adherence of legal requirements (categorized as either 'yes'/'fulfilled' or 'no'/'not fulfilled'). To analyze relationships between metrically scaled and normally distributed variables, we calculated Pearson's correlation coefficients. Coefficients up to 0.3 were classified as low, those between 0.3–0.5 were classified as moderate, and those from 0.5 on were classified as high [41]. The level of significance was set to *p* < 0.05.

As part of a non-responder analysis, responding and non-responding companies were compared concerning their company size. For this purpose, we used an ordinal five-point scale of company size that had been delivered by the Reutlingen Chamber of Crafts for craft enterprises, and thus was available for both responders and non-responders. We proceeded similarly with regard to the non-craft enterprises.

All of the analyses were performed with SPSS, version 24 (IBM Analytics, IBM Corporation, Armonk, NY, USA).

#### **3. Results**

#### *3.1. Participants*

The response rate to the questionnaire was all in all 24.5% (*N* = 222/906). On average, there were less than 5% missing values in each category of the questionnaire. The response was above average in medium-sized companies (cf., Table 2) with 101 to 500 employees (31.5% and 32.3%), whereas it was clearly below average in small enterprises with up to 50 employees (22.1%) and in big companies with more than 500 employees (23.1%). Then, the correlation between response and company size seems to be of the inverted u-shaped type. About half of the companies (48.2%) indicated the availability of an occupational health physician, with a range from 29.4% (small companies with up to 50 employees) to 85.0% (companies with 201 to 500 employees). The presence of an occupational safety engineer was reported by 76.8% of all the participating companies (cf., in greater detail in Table 2).


**Table 2.** Company characteristics of participating companies according to company size.

\* Occupational medical check-ups according to the relevant legal regulation (ArbMedVV [36]) (e.g., screen work, handling of hazardous substances, or noisy work places) have to be available to all employees in Germany. According to another regulation [39], an occupational health physician has to be available in all enterprises with more than 50 employees (in some branches, this limit is lower), and in the smaller enterprises in case the employer feels the need for occupational health counseling (so-called "alternative, demand-based supervision"). ‡ An occupational safety engineer has to be available in all enterprises with more than 50 employees (in some branches, this limit is lower), and in the smaller enterprises in case the employer feels the need for occupational health counseling (so-called "alternative, demand-based supervision"). In small enterprises (max. 50 employees), the employer can receive special training with regard to occupational health and safety by the statutory accident insurance in order to reduce the need for support by occupational safety engineers [39].

In enterprises with up to 50 employees (the maximum number differs between individual branches due to the respective accident prevention regulation of the respective statutory accident insurance), the employer can participate in a specific occupational health and safety training that entitles him to utilize the service of an occupational safety engineer only when necessary (so-called "Unternehmermodell"). This was indicated by 32/127 enterprises (missing *n* = 5) with up to 50 employees.

As to the sectoral affiliation of the participating companies, almost one third (30.2%; *n* = 67) of the participating companies belonged to the manufacturing industry, and 16.7% (*n* = 37) belonged to the construction industry. Another 15.3% (*n* =3 4) and 14.0% (*n* = 31) can be attributed to services and trade, respectively. The remaining 24% of participating companies were distributed among the following sectors: hospitality industry, agriculture and forestry, maintenance and repair, banking and insurance, transport/storage/communication, public administration, mining and quarrying, education, and energy and water supply.

The sociodemographic characteristics of the responding persons in the companies are shown in Table 3. As to the position of the respondents, 52.7% of these were managing directors, 34.7% were from the personnel department, and 11.7% were other employees (cf., Table 3).


**Table 3.** Sociodemographic characteristics of respondents.


**Table 3.** *Cont.*
