gms | German Medical Science

Introduction: Patient-reported outcomes (PROs) are an unique tool to evaluate the patients' perceived health status and health related quality of life. They allow clinicians and researchers to gather data from a different point of view and therefore present a way towards patient centred medicine [1]. PROs can be collected as classic paper questionnaires or in electronic data capture (EDC) systems. The latter holds some interesting benefits as they could provide a platform for bidirectional communication between patients and clinicians or researchers. They also simplify the access and analysis of the collected data. Despite the potential of PROs their successful usage inherently depends on the patients' participation. Surveys showed that some potential barriers are the time required, difficulties to use electronic devices, and the lack of perceived value [2]. To overcome some of the reported barriers we developed a prototypical interface between an EDC system and a virtual assistant to fill in PRO forms via verbal communication.

Methods: We based our development on two well established products. As an EDC system we selected the Open Source Registry System for Rare Diseases (OSSE) as it supports all needed prerequisites to handle patient user accounts and PROs and as an open source software it is easily extendable [3]. For the speech assistant we used Amazons Alexa system which provides a development platform and software libraries to define possible conversations and transfer information between the user and a remote system.

Results: As a proof of concept we created an exemplary PRO to document medication intake. To interact with this PRO we developed a conversation flow to ask and tell Alexa which medication was taken. To associate an Alexa user with its OSSE account we implemented the requirements to use Amazons account linking process. Finally we implemented an API to receive requests from Alexa and read or write the needed information to OSSE's data structure.

Discussion: Our approach could lower the threshold for patients to fill in PROs by eliminating the barrier of typing into an EDC. Furthermore, verbal communication improves the overall quality of data [4] and in combination with the possibility for spoken feedback presents an opportunity to emphasise the value of PROs to the patient.

Besides these benefits we identified two critical points in our concept, patient safety and acceptance. Patient safety plays a role as speech recognition can be an additional source of incorrect entries. To minimize such risks, carefully designed conversation flows are crucial. Acceptance is a controversial topic in itself as our method holds potential to lower the efforts to collect PROs at the cost of adding a third party. This rises concerns towards data protection and privacy, which are also identified barriers regarding PROs [2]. In the future this conflict could be mitigated by replacing Alexa with open-source alternatives.

Conclusion: In summary we developed a prototypical connection between OSSE and Alexa to fill and read PROs. This can lower some of the barriers that prevent patients from filling in PROs, but also rise questions regarding safety and acceptance.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.