**3. Results**

### *3.1. Phase 1: Semi-Structured Interviews*

A total of 26 people were interviewed, nine men and 17 women, between the ages of 26 and 49. Twenty-two interviewees identified themselves as smokers, while four were non-smokers or ex-smokers but lived together with a smoking partner. Eighteen of them were born in Germany, the others had lived in Germany for an average of 15 years.

In the mixed inductive-deductive analysis of the interviews [34], a lack of knowledge was identified with respect to the definition of SHS exposure and its potential health consequences, especially less severe conditions such as otitis media or tooth decay. For a potential m-health campaign, participants asked for concrete information about health consequences and on prevention measures to effectively reduce SHS exposure, especially in difficult contexts (e.g., the absence of a balcony or garden so that smoking outside would mean leaving the children alone at home). The participants also suggested that messages should be created from the perspective of the children as victims, not from the perspective of the parents.

At the formal level, messages should be (audio)visual, simple, concise, and with a positive appeal. The participants highlighted that the campaign should not only target women but also men as they are less frequently addressed by typical information providers such as gynaecologists or paediatricians. Online social networks, especially Facebook, were classified as the most suitable access paths to the target population and thus the best way to distribute the information.

### *3.2. Phase 2: Focus Group Discussions*

Twenty of the 26 individuals from phase 1 participated in the focus groups. Regarding the design of the illustrations, the participants were in favour of the photo campaign developed by the 3D artist, especially because of the adoption of the children's perspective. However, they suggested to present the key messages more concisely and to have a uniform structure for all illustrations. In addition, illustrations were considered to be even more interesting if they were animated and included a voiceover by a child.

With respect to the content of the illustrations, the focus group participants selected four illustrations with key messages as the most suitable ones. Based on these suggestions, texts were optimised, pictures animated and a voiceover for the first part of the text added (Figure 1). The final illustrations addressed two health effects of SHS exposure in children (otitis media, asthma) and two social effects (children of smokers start smoking at a younger age, smell of SHS in clothes). Each of them were combined with up to two of the following simple and effective measures to reduce SHS exposure: Not smoking inside the house/apartment, not smoking in front of children and not smoking in cars. Finally, participants considered it important that the campaign would be distributed in online social networks by a trustworthy source, e.g., physicians.


**Figure 1.** Pictures with key messages of the final campaign (Original in German with English translation provided below). MP4 files with the voiceover can be found at http://www.klinikum.uni-muenchen.de/Institut-und-Poliklinik-fuer-Arbeits-Sozial-und-Umweltmedizin/de/forschung/arbeitsgruppen/Prof\_\_Radon/aktuelles/Passivrauchkampagne.

### *3.3. Phase 3: Quantitative Assessment*

Of the 121 participants in the online survey, 45 were women, 41 were men and the remaining 35 did not indicate their sex. The mean age was 35 years (range 20–56 years). All participants were smokers with at least one child below the age of six in the household. About half of the study population did not complete high school and 27% were migrants (either themselves or their parents were not born in Germany).

With respect to attitude towards SHS, participants were mostly aware that SHS has adverse health effects. However, many did not support a restriction on smoking in cars. More than half of the participants agreed that a campaign about SHS and its effects on children would be useful (Figure 2).

**Figure 2.** Attitude towards second hand smoke of the 121 participants in the quantitative evaluation of the campaign (\* Original item formulated negatively; SHS = Second Hand Smoke).

### *3.4. Evaluation of the Individual Campaign Illustrations*

Concerning the first impression of the illustrations, they were evaluated as very poor by one-third of the participants, while one-third found them (very) appealing without statistically significant differences between the single illustrations (Table 1). The overall quality of the illustrations and their key messages were positively evaluated with small but sometimes statistically significant differences between them (Table 2). The intention to share the pictures was moderate (mean value between 3.14 and 3.21 for the individual illustrations on a Likert scale from 1 = very likely to 5 = very unlikely).



ANOVA = Analysis of Variance.


**Table 2.** Evaluation of the quality of the illustrations and their key messages by the 121 participants in the quantitative evaluation.

ANOVA = Analysis of Variance; \* Dunn-Bonferroni post hoc Test *p* < 0.05 between illustration 1 and 3; \*\* Dunn-Bonferroni post hoc Test *p* < 0.05 between illustration 1 and 4.

### *3.5. Overall Evaluation of the Campaign*

Participants agreed that it would be good to include information about SHS in social media (mean 4.36; SD 0.94 on a scale from 1 'I fully disagree' to 5 'I fully agree'). However, on average they rated the topic as too sensitive to be shared in social media (mean 3.98; SD 1.29 on the same Likert scale).

The content of the key messages was evaluated very positively. The recommended measures were rated as easy to implement (mean 4.32; SD 1.06 on a scale from 1 'I do not at all agree' to 5 'I fully agree') and reasonable (mean 4.56; SD 0.80). The presented consequences of SHS were rated as realistic (mean 4.39; SD 0.92).
