At Isabel we are constantly building upon and improving our tool, testing the limits and always trying to make the diagnostic process easier and more efficient for healthcare professionals. As part of this continual development, last year we conducted a large internal study, with some outstanding results.
The aims of the study were to determine how well Isabel could provide a differential diagnosis in a ranked list of results. We took 563 cases with a range of specialties and complexities, and placed them through the Isabel system to see where the correct diagnosis would appear on the list. The correct diagnosis was somewhere on the list in all of the cases, which in itself was great, but the overall results for the rankings of the correct diagnosis were incredible. In 98% of cases the correct final diagnosis was listed in Isabel’s top 10 suggestions, and in 74% of cases it was listed in the top 3 suggestions.
The Likelihood Indicator
Following these unprecedented results, we wanted to create an easy-to-use feature on the Isabel tool, to allow users to quickly see which diagnoses are the most likely on the ranked list. We call it the ‘Likelihood Indicator’ and it is a vertical orange-yellow bar alongside results. The darker the shade, the more likely the diagnosis is the correct one.
These groupings are based on the calculations made by Isabel’s sophisticated pattern recognition software working together with numerous other algorithms and uses the percentage figures that were previously available from the ‘what did this diagnosis come up’ function.
Download the Infographic
We’ve created an infographic below detailing the study highlights and the new Likelihood Indicator feature. You can download this infographic by selecting the image below or clicking here.
We’re very pleased with the results of our internal study, and indeed the new Likelihood Indicator, and we’d love your feedback if you do try it out. You can email us on firstname.lastname@example.org or comment on this blogpost below. Watch this space for more updates on the tool and Likelihood Indicator, as we will be releasing some of the individual cases we placed through the system during our study.