Prospective User-Centered Design of Clinical Decision Support Systems in the Context of Personalized Medicine
About the project
Clinical decision support systems in the context of personalized medicine are seen as having great potential. There are numerous studies showing that such systems can increase diagnostic accuracy, improve clinical outcomes, avoid errors caused by information deficiencies and improve physicians’ decisions through faster and more comprehensive access to relevant data. However, despite all the recent technological advances, many challenges remain: Research shows that clinical decision support systems are insufficiently accepted and used, for example, because physicians fear a reduction of their professional autonomy, users do not trust the system or the systems are not seen as suitable for the respective workflows. Furthermore, such systems can also be associated with unintended negative consequences (provocation of errors, generation of irrelevant alarms). This is mainly due to the fact that no systematic, user-centred development process is followed, which places the users with their tasks and their specific working environment in the centre of the development activities from the very beginning. Thus, successful implementation – in addition to the completeness and accuracy of the evidence base – also depends significantly on the user- and context-adapted design and the integration of the systems into everyday clinical work. The results of the junior research group CDS2USE should help to ensure that the innovative, intelligent solutions developed within the framework of MIRACUM are designed in a user-centred way in order to ensure the broad and sustainable use of such intelligent IT systems in everyday practice.
The CDS2USE junior research group pursues three scientific goals:
Goal 1: Develop algorithms for context-sensitive decision support:
A major challenge in the development of decision support systems is the provision and use of context-sensitive data, as the identification and delineation of different contexts in the medical field is largely unexplored. Therefore, algorithms of the medical informatics initiative are to be adapted, combined and supplemented in such a way that the systems are enabled to present the right (relevant) information, to the right person, at the right time, in the right (relevant) context. In doing so, competing factors and values that influence clinical decision-making must also be adequately considered.
Goal 2: Develop “intelligent explanation tools”/a transparent and comprehensible decision support:
Currently, existing concepts and approaches of human-computer interaction are not sufficient for the user to understand clinical decision support systems. Therefore, concepts are to be developed that present the results to the users in an explainable, transparent and comprehensible way and enable them to gain individual control over the system on the basis of “intelligent explanation tools” and thus to deal with “their” data in a confident manner.
Goal 3: Further develop existing evaluation methods for a fit to clinical decision support systems:
Numerous recommendations exist for user-adapted design and evaluation of interactive systems, but these do not take into account the specific medical context. Existing evaluation methods are therefore to be further developed and adapted or expanded for clinical decision support systems in the context of personalised medicine. In particular, both new methods for evaluation with regard to the comprehensibility of system proposals/results and new methods for testing intelligent visualisations that prepare the information from the complex data sets in a user-friendly way are to be developed.
Three work phases – set-up, implementation and generalization – are planned to achieve the goals:
Set-up phase (year 1-1.5):
Based on the state of the art in science and technology, requirements for the development of algorithms for context-sensitive decision support systems are specified, existing concepts for a transparent presentation of results are identified and possible evaluation methods for a user-centred system design are compiled. The work is closely linked to the MIRACUM use cases.
Implementation phase (years 1.5-3):
The results for the MIRACUM use cases developed in the set-up phase are continued, validated and iteratively refined or adapted. In addition, the results (concepts, methods, tools) are tested for new use cases.
Generalization phase (year 4-5):
The developed methods and tools are transferred and adapted for as many types of decision support systems as possible in the context of personalized medicine.
To achieve its goals, the junior research group places a systems engineering approach at the centre of its research, which takes equal account of the end user (human), the technology used (engineering) and the work processes (organization). In addition, the group conducts “participatory research” involving medical staff, patients and decision-makers (depending on the work in question) to ensure that the results produced take into account the actual needs and specific situation of the people directly affected.
Lead of the Junior Research Group
Dr. Brita Sedlmayr
Lead of the Junior Research Group CDS2USE
Institute for Medical Informatics and Biometry | Technical University Dresden – Medical Faculty Carl Gustav Carus
Ian-Christopher Jung (M. Sc.)
Katharina Schuler (M. Sc.)
Maria Zerlik (M. Sc.)
- Gruendner J, Deppenwiese N, Folz M, Köhler T, Kroll B, Prokosch HU, Rosenau L, Rühle M, Scheidl MA, Schüttler C, Sedlmayr B, Twrdik A, Kiel A, Majeed RW. Architecture for a feasibility query portal for distributed COVID-19 Fast Healthcare Interoperability Resources (FHIR) patient data repositories: Design and Implementation Study. JMIR Med Inform. 2022; 10:e36709. Doi: 10.2196/36709.
- Sedlmayr B, Sedlmayr M, Kroll B, Prokosch HU, Gruendner J, Schüttler C. Improving Coronavirus Disease 2019 (COVID-19) Research of University Hospitals in Germany: Formative Usability Evaluation of the Codex-Feasibility Portal. Appl Clin Inform 2022; 13:400–409. Doi: 10.1055/s-0042-1744549.
- Schüttler C, Prokosch HU, Sedlmayr M, Sedlmayr B. Evaluation of Three Feasibility Tools for Identifying Patient Data and Biospecimen Availability: Comparative Usability Study. JMIR Med Inform. 2021; 9:e25531. Doi:10.2196/25531.
- Sedlmayr B, Ahmadi N, Jung IC, Zerlik M, Sedlmayr M. Prospective user-friendly design of clinical decision support systems – first results of the Junior Research Group CDS2USE. Postervortrag auf der 66. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e. V. (GMDS), 12. Jahreskongress der Technologie- und Methodenplattform für die vernetzte medizinische Forschung e. V. (TMF) 26. – 30.09.2021, Online, https://www.egms.de/static/en/meetings/gmds2021/21gmds047.shtml