ABC | Volume 110, Nº1, January 2018

Original Article Stephan et al Mobile health and atrial fibrillation Arq Bras Cardiol. 2018; 110(1):7-15 We opted for a clean and clear design, with a color code for the risks and the use of graphical information whenever possible to complement the written information. Terminology was adapted to the target users. Personal health information were protected by unique identification and secured by cryptography. A privacy policy was presented prior to the use. Study design Intervention study in patients diagnosed with AF in the anticoagulation outpatient service at Porto Alegre Institute of Cardiology, in April and May 2016. Population and sample characteristics The study population comprised patients attending the anticoagulation outpatient service, and the study was carried out during patients’ waiting time for the prothrombin time (PT) test. Before starting their treatment at the anticoagulation outpatient service, each patient receives instructions about AF, the use of OAC, as well as appropriate dose adjustment every 1-3 months. In ten random mornings, all AF patients attending the outpatient center for the PT test were invited for the study and all of them agreed to participate. There was no patient with severe visual disorder, hearing loss or cognitive disorder that would impair patient’s interaction with the app. Patients with one of these conditions would be excluded from the study. Pilot study and sample calculation In the pilot phase, the beta version of the app was used with 10 patients, who gave their feedback to questions about usability, written and visual language, understanding of information, design and adequacy of time for scrolling the screens. Before the appointment, a questionnaire developed by the investigators was administered to measure the mean level of knowledge about AF in this population. This questionnaire sought to evaluate the minimum essential information required for the patient to understand their condition and adhere to the treatment. Patients were asked to answer each of eight statements with “true”, “false” or “don’t know”. All statements were true. Mean number of correct answers was 5.9 ± 1.37 (73% of correct answers). In a previous study conducted at the same service, 64% of patients had adequate knowledge about the therapy. 19 Considering that the number of correct answers was estimated to increase to 8 (100% of correct answers) after the explanatory intervention, 18 patients were required for a 5% alpha error and a beta error of 90%. Outcome measures After adjustments made after the feedback of the pilot study patients, the app was tested in a sample of 20 patients. As the primary outcome, we analyzed the scores obtained by the patients in the AF knowledge questionnaire before and after the interaction with the app. As secondary outcome, we evaluated patients’ scores in the Decisional Conflict Scale in Health (DCSH) by O’Connor, 20 used to evaluate strategies for shared decision‑making in health care. 20,21 DCSH was validated in Portuguese in 2013 by Martinho et al. 22 and included questions on uncertainties, knowledge, values and provided support. The total score varied from 0 (no decisional conflict) to 100 (extremely high decisional conflict). Also as a secondary outcome, we analyzed the perceived risk of stroke and bleeding with the use of OAC. Patients were asked if they believed they had a low, moderate, or high risk of each event. This question was repeated after the interaction with the app, and results were compared with the “real” risk, calculated by the CHA 2 DS 2 ‑VASc and HAS-BLED scores. Data analysis Data were analyzed using the SPSS software version 20.0. Tables of absolute and relative frequencies were used for sample characterization. Normality of data was tested by the Shapiro-Wilk test. Continuous variables with normal distribution were expressed as mean and standard deviation, and those with non-normal distribution as median and interquartile ranges. Mean knowledge scores about the disease, before and after the intervention, were compared using the paired Student’s t-test and risk perception was compared with the Wilcoxon test. The level of significance was set at p < 0.05. Ethical considerations The study was approved by the Research Ethics Committee of the Institute of Cardiology University Foundation. Privacy, anonymity and confidentiality of data collected were guaranteed, and informed consent form was presented to the patients. Results Mean age of the 20 patients studied was 67.7 years; most patients were men (60.0%), white (83.3%) and lived with their relatives (53.3%). Self-reported educational level was some secondary education in 73.3% of patients, and 33.3% studied less than 4 years. Family income was lower than 2 minimum wages in 53.3% of patients. Most patients (66.7%) used anticoagulants for at least one year. Table 1 summarizes the socioeconomic characteristics and clotting time of the study population. Table 2 shows the prevalence of the main risk factors included in the CHA2DS2-VASc, HAS-BLED and SAMe-TT2R2 scores, and the mean ratings obtained by the patients in these scores. The most prevalent comorbidities were arterial hypertension (80%), diabetes mellitus (30%), and heart failure (30%). With respect to other factors that may influence the risk of bleeding and anticoagulation, the most common factors were the use of medications that interact with coumarins (43.3%), and the use of antiplatelet or anti-inflammatory drugs (26.7%). Most patients (86.6%) had a CHA2DS2-VASc score equal to or greater than 2 and 76.6% of patients had a SAMe-TT2R2 score equal to or greater than 2. The number of correct answers in the disease knowledge questionnaire significantly increased after the interaction with 9