To what extent the use of Computerized Diagnostic Decision
Support Systems (CDDS) actually improves the quality of medical diagnoses and
consequential health outcome of the individual patient remains unknown. One
highly specialized algorithm recently outperformed even the best dermatologist
in diagnosing skin lesions. However, such success has so far only been achieved
in visual tasks where there is little to no variation in information format or
content and where human error is already quite low. How CDDS performs on more
“messy” and less well structured tasks that involve human interactions, such as
evaluation of common symptoms like fever, abdominal pain or syncope remains
unknown.
We conduct a single-blinded, cross sectional, multi-center, four-period cross-over
controlled cluster-randomized clinical trial in Swiss Emergency Rooms using both
qualitative and quantitative methods and analysis to evalute the effect of CDDS
on the diagnostic processes and patient outcomes. We will focus on three common
and frequently misdiagnosed symptoms: fever, abdominal pain and syncope.
Patients will be followed-up for 28 days to evaluate outcomes.