A diagnostic support system in general practice: is it feasible?

Current clinical practice often lacks tools to assist physicians in generating diagnostic hypotheses during consultations. While prior research has shown that decision support systems can improve diagnostic accuracy in specialized settings, their application in general practice remains limited. The challenge lies in integrating these systems into real-time patient interactions without disrupting the workflow. Doctors must balance rapid decision-making with thoroughness, especially in primary care where diagnostic uncertainty is common. Existing systems often fail to match the dynamic nature of consultations. This gap motivated the exploration of a system that could provide timely support while preserving clinical autonomy. No prior work had resolved how to seamlessly embed such systems into routine practice. The feasibility of such a system in general practice remains unclear. This paper addresses that uncertainty by testing a novel approach.

The study aimed to evaluate the feasibility of a diagnostic decision support system (DDSS) in general practice. The specific problem is the need for real-time, user-friendly tools that assist physicians during patient consultations. The motivation stems from the high diagnostic uncertainty in primary care and the potential for computer-aided systems to improve accuracy. The system was designed to generate hypotheses based on patient data during consultations. The goal was to integrate the system into existing workflows without causing delays. The study focused on two key aspects: the system’s ability to support diagnosis and its practical usability. The researchers sought to determine if such a system could be effectively used in real-world settings. The ultimate aim was to assess whether the system could enhance diagnostic accuracy while maintaining clinical independence.

The DDSS was developed with a focus on real-time integration into consultations. It used a knowledge base of symptom configurations to match patient data. Doctors entered clinical items via a user-friendly interface with a comprehensive term set. The system generated ranked diagnostic hypotheses during consultations. The interface allowed doctors to store symptoms with a single keystroke and request support at any time. The system’s performance was tested in a practical setting with 20 general practitioners. Each doctor addressed five patient cases, entering 2000 clinical items in total. Consultation times were monitored to ensure the system did not disrupt workflow. The study evaluated both system performance and physician proficiency in using the tool.

The DDSS demonstrated strong performance in generating accurate hypotheses. In 96% of cases, the correct diagnosis appeared in the differential list. Doctors used the system effectively during consultations without significant delays. The interface allowed rapid entry and retrieval of clinical data. Physicians entered 2000 clinical items across all cases. The system’s ability to rank hypotheses by concordance was a key feature. Doctors’ diagnostic accuracy was reported at 43%, suggesting room for improvement. The system’s usability was praised, with doctors showing high proficiency in its operation. These findings suggest the system can support real-time decision-making in general practice.

The study suggests that a DDSS can be feasible in general practice settings. The system’s ability to generate accurate hypotheses was confirmed in 96% of cases. Doctors used the system effectively without disrupting consultations. The interface design facilitated rapid data entry and retrieval. The system’s performance met expectations in terms of usability and accuracy. The study supports the idea that such systems can be integrated into routine practice. The authors propose that the DDSS can serve as a tool to enhance diagnostic reasoning. The system’s success depends on maintaining clinical autonomy while offering timely support.