Skip to content

AI-powered assistance system for the collection of health data

Project description

Detailed data on the course of medical treatments are needed for the evaluation of proven but especially new treatment methods in everyday clinical practice. A promising method for generating these data are so-called Patient Reported Outcome Measures (PROMs), that is, data on health status and treatment success provided by the patients themselves. They are characterized by assessing the subjective experience of the patient’s own health status and thus form a patient-centered supplement to the diagnostic perspective. For this purpose, concrete questions about health and living conditions are asked both at the beginning and at the end of treatment – similar to medical anamnesis and catamnesis, except that the assessment is not carried out by medical staff, but in the form of a questionnaire. This data can then be used to draw conclusions about the quality of certain forms of treatment and thus improve medical care.

However, one problem that often arises is the so-called non-response, which has a negative impact on the quality of the data: Not all of those treated can or want to fill out the questionnaires, fail due to language barriers or have difficulties in understanding, which reduces the comparability of the data. In particular, this also means that data are missing from those groups of people who are usually already disadvantaged anyway, such as older people or people with disabilities.

The reasons are manifold, sometimes this is due to motivation, but more often it is simply due to a lack of understanding of individual technical terms or contexts – such as the purpose of the survey itself. Therefore, the completion of these questionnaires is often accompanied by clinic staff, which represents an additional personnel burden and is also not foreseen in a standardized way in the process.

The MIA-PROM project aims to address this problem and is developing a [M]ultimodal [I]nteractive [A]ssistance that supports patients in completing the [PROM] questionnaires. The assistance system should be able to provide information on medical terms, respond to follow-up questions, and motivate patients to fill in the questionnaire fully. A central point is also accessibility, which the assistance system should enable by being able to read questions aloud and translate them into simple language if necessary. The assistance system should also be visible in the form of a small agent and thus be more responsive – in this way, we hope to be able to make the assistance system more social and increase acceptance.

Of social scientific interest to us is the question of whether it is advantageous for data collection if the agent is made as a small robot, or whether a virtual form – as an avatar on the screen – has the same effect on the quality of the data.

The project follows a participatory approach – this means that central and sensitive decisions regarding the design of the assistance system are discussed and made in close exchange with an patient advisory board. Four workshop dates are planned for this collaboration, in which the advisory board – methodically guided – will be involved in the decision-making process. The language module of the assistance system will be developed using artificial intelligence methods and will also learn continuously from the interaction with patients during a longer practical test phase.

The project is realized in a consortium consisting of the University of Applied Sciences Munich, the Charité Berlin, the Technical University Berlin, the University of Applied Sciences Hamm-Lippstadt, the two companies Acalta GmbH and dexter health GmbH as well as the two practice partners of the Rehaklinik Seehof and the Center for Outpatient Rehabilitation Berlin.

Are you interested in our project?

We are looking forward to the exchange with you