**2. Methods**

### *2.1. Likert Scale*

The original questionnaire used an incomplete Likert scale with an even number of answers to the questions, while it is recommended that questionnaires use an odd number of answers [18]. Furthermore, there were more positive options than negative in the first two questions. We decided to add an extra possible response to the first four questions in order to have an odd number of options and therefore obtain a complete Likert scale.

#### *2.2. Translation into Catalan and Data Collection*

To achieve the higher content validity, a translated and back-translated methodology was used [19,20]. The original English version of the questionnaire (Appendix A) was translated into Catalan independently by two authors who are native Catalan speakers and who are both proficient in English. An agreed Catalan version of the questionnaire was obtained following several drafts (Appendix B).

In order to validate the new Catalan version of the Health Optimum questionnaire, the Google Forms tool was used to send it to primary healthcare professionals in the Catalan central region who had used telemedicine services in the past. WhatsApp health professional groups were used to disseminate the questionnaire. Members were asked to answer the questionnaire and to resend it to other potential respondents. The twitter account of the principal investigator (@jvalaball, >10K followers) was also used for further dissemination. Additional information regarding the respondents' basic characteristics was added to the questionnaire: age, sex and professional role and the kind of telemedicine services available in the Catalan central region which they had used (teledermatology, teleulcers or teleaudiometries). Non health professionals were asked not to answer the questionnaire.

The questionnaire called "Questionnaire to assess healthcare professionals' perceptions of asynchronous telemedicine services" (Qüestionari per avaluar la percepció dels professionals sanitaris amb els serveis de telemedicina asíncrona) had 8 questions with a complete Likert scale of 5 answers the first 4 questions and 3 answers the last 4 questions. It was first sent at the beginning of April 2018 and it was closed 3 d later. It was resent 2 weeks later, asking participants to answer it again in order to check for consistency. The questionnaire was closed definitively 5 d later. Completing the questionnaire was considered as an indication of consent to participate in the study. The study protocol was approved by the University Institute for Primary Care Research (IDIAP) Jordi Gol Health Care Ethics Committee (Code P16/046).

#### *2.3. Scale Level Descriptive Analysis*

Following Argimon et al.'s methodology [21], we have assessed the variability in responses to the questionnaire calculating the average and the standard deviation (SD) and calculated the frequencies to check for floor and ceiling e ffects [22]. These e ffects are important as they can influence the validity, reliability and responsiveness of a questionnaire and they are used to check the percentage of participants with very low and very high scores. We have taken this e ffect to exist when 15% or more of the responses are found in the higher or lower values [23].

We checked the discriminating capacity of the items using the discriminative rate, which compares the responses in the two extreme groups (individuals who have obtained a total score below the 33rd percentile and individuals who have scored above the 66th percentile). Discriminative rates above 0 indicate discriminating capacity of the items [22].
