Step 1: Selection of materials and process coordinators.

We searched online throughout professional support groups, such as NSW Therapeutic Advisory Group [16], the Canadian Deprescribing Network [17], Primary Health Tasmania [18], and the US Deprescribing Research Network [19], to find free access online web-based tools for the identification of potentially inappropriate medications identified by the STOPP/START criteria and the Beers criteria. We identified the MedStopper® tool [14].

We contacted one of the original authors [JM which is also an author of this protocol], and the adaptation of the tool for the Portuguese population was agreed. We will review the original version of the tool and, in cooperation with its developers; we will identify the tool’s main elements. We will identify the potentially inappropriate medications, a user-friendly format [icons that instantly give feedback], a rationale and summary of the STOPP/START, and Beers criteria and a deprescribing plan. LM will be the process coordinator.

Step 2: Early review.

The translation will only take place after the revision of the tool by the process coordinator [LM] and by a linguistic expert native speaker of English and fluent speaker of Portuguese.

The linguistic expert will do the translation after contact with one of the developers of the original tool. This collaboration will make sure that difficult concepts will be identified and alternative ways to express the information are found.

Examples of predictable changes after this early review include incorporating recommendations of Portuguese Directorate General of Health; and adding a calendar that visually shows the daily dose of the medicines during the chosen deprescribing process. Also, in 2019, the Beers criteria were reviewed, therefore, we will update the MedStopper® tool accordingly.

Step 3: Translation and back-translation.

This step will not be just a literal translation of the content; in this third step, we will make the translation, quality check, and an independent review. We will then have two forward translations done by two professional translators, expert native speakers of English and fluent speakers of Portuguese. From these two translations, we will obtain a consensual translated version considering the debate between the translators and the research team.

The next step will then be to obtain a back-translated version by a professional translator, native speaker of English, fluent in Portuguese. An independent expert, who also understands the original language, will review the final version of the tool.

Step 4: Comprehension testing.

To guarantee that the adapted and translated decision tool is clear for the Portuguese doctors, we will conduct this step. According to the ECDC [15], one of the possible approaches is to have individual interviews. We will perform interviews in which participants are asked, while using the web-based decision-aid, to express their opinion about the content and usability. All participants will be asked to apply the MedStopper® tool to the fictitious clinical case that was previously tested by family doctors with a special interest in elderly care: First clinical case, with a frail elderly patient taking alprazolam for insomnia, omeprazole for heartburn, gliclazide for type 2 diabetes, naproxen for osteoarthritis and aspirin for the previous stroke; second clinical case, a frail elderly patient with hypertension and diabetes taking furosemide, ramipril/amlodipine, metformin, atorvastatin, trospium chloride and recently medicated with donepezil.

Eligible participants family doctors will be recruited through an invitation to participate by mailing lists and social media. Candidates who respond to the email will be invited for individual semi-structured interviews to take place in the Faculty of Medicine of the University of Porto or another place of convenience for the participant, provided that confidentiality is assured. All participants will receive written information about the study and will sign individual consent forms. We plan a convenience sample of participants until data saturation is reached, which means the collection of qualitative data will be made to the point of closure is attained because new data is considered redundant information [20]. To ensure confidentiality, each interview will be given an alphanumeric code.

Three researchers will conduct the interviews following an interview guide. We will conduct a pilot test to make sure that the interview guide is clear. After gathering the written consent by the participants, they will be asked to give demographic and professional data such as age, number of clinical years, sex, and whether they are interns or specialists in family medicine.

The individual interviews will be audio-recorded, and the researchers will apply cognitive methods such as think aloud, and paraphrasing techniques [21].

Each participating doctor will have a computer with access to the web-based format of the decision aid tool—MedStopper®. They will apply the tool to the same clinical cases presented to the participants. For each clinical scenario, a list of medications and diseases for fiction elderly patients will be presented, and the participants will be encouraged to apply the MedStopper® tool.

Participants will be encouraged to share their opinion aloud while they are going through the tool to identify potential issues in the format and content [22]. At the end of the interview, we intend to use a probing technique retrospectively with questions to elicit how doctors felt throughout the process.

Data obtained after verbatim transcription of the audio files will be analyzed in a personal computer [without a network connection] belonging to one of the researchers.

One month after data analysis, the audio files will be destroyed. Using Ligre® Software, we will conduct a thematic categorical analysis.

Two authors will independently analyze the transcripts and fill a content analysis form. The participants’ feedback will be categorized into six main themes: navigation and usability of the interface, content comprehension and completeness, length of the decision aid tool and amount of information, use of figures/illustrations, time of completion [23], and other issues [e.g., suggestions to improve the decision aid tool]. We will include participants’ proposed solutions or revisions.

Step 5: Proofreading.

Two Portuguese native speaker doctors, selected by the process coordinator, will conduct the final step to check for minor errors which have been missed during the translation process.