A pilot usability study of smart virtual assistants to monitor medication adherence in chronic patients

Abstract Background The World Health Organization identifies low medication adherence as one of the main issues for increased morbidity and mortality rates among hypertensive patients. The increased presence of Smart Virtual Assistant (SVA) devices in households enables the possibility of implementing voice-activated interventions to address behaviors related to non-adherence to medications. Purpose The aim of this study was to develop and evaluate the usability of the most commercially available SVA device, as part of the InTakeCare digital platform designed to monitor and support medication adherence in chronic patients. Methods The InTakeCare platform is a modular and scalable platform composed of a cloud database, an online server and a series of modules for user interaction. A skill was developed for commerciallyavailable SVA devices (Amazon Alexa), together with a connected web dashboard that enabled physicians’ access and management of the therapies of their patients. Pilot study included treated hypertensive patients with sufficient technological capacities and a wi-fi connection at home. Participants completed a questionnaire about their clinical sociodemographic characteristics and the eHealth Literacy Scale (IT-eHEALS) questionnaire. Then, the physician created an account for them on the platform and prescribed medications were inserted into the web dashboard, together with their posology and required intake time. This information was automatically linked to SVA devices at home, setting reminders at the defined time and 55 minutes later. Patients had 120 minutes to vocally confirm to the SVA the medicine intake. After seven days, participants participated in a semi-structured interview including the System Usability Scale (SUS) questionnaire. Results Fifteen subjects (11M; 3F, median [IQR] age 67[57-68] years) participated in the study. IT-eHEALS scores ranged from 16 to 40, with a median value of 27 (20-35). Overall, medium adherence to therapy was self-reported before the experiment. Post-study SUS questionnaire score was 60 (52.5-75). Semi-structured interview reported high interest and perceived innovation, together with lack of trust towards the company producing the SVA device and presence of difficulties in the communication with the devices. Conclusions Participants were partly satisfied with the usability of the developed skill, the median SUS score falling below 68 ie the cutoff required for good usability of a solution. Participants perceived the solution as interesting and interactive, demonstrating high interest in vocal interaction. However, difficulties were reported in correctly interfacing with the SVA device vocal system, and the skill was not identified as easy to use. Further studies are required to better understand and quantify the identified issues, with the aim of further improving vocal interaction as a tool to monitor and improve medication adherence in hypertense patients.