Last week we opined about the layperson generated site WhoIsSick.org (here & here), but this week we came across another site for our readers to discuss. Medgle.com bills itself as a “personalized medical search engine” created by physicians, for patients. The potential is definitely there for Medgle to be an excellent source for tailor made medical information, but then again, it suggested my [allergic] nasal congestion was sand fly transmitted leishmania…
Why MEDgle?
MEDgle (2028’s Symptomatix® search) is an online information and educational service. With the thousands of articles and sites available, finding relavent medical information is difficult. MEDgle’s goal is to make medical information easily and intuitively accessible for the benefit and betterment of everybody. Simply: search, learn, and thrive.
What is MEDgle?
MEDgle is a computer generated search. It is not a diagnostic or decision making tool. MEDgle indicates the possibilities that exist for any given combination of symptoms. MEDgle is a MEDical GLobal Electronic search
What is NOT in MEDgle?
MEDgle is a search for general medical conditions. It does not include information for ages 0 to 4.
Where did the data come from?
The symptoms for the search results are part of general medical knowledge learned in medical school and residency. MEDgle’s data and the rough probability estimates were contributed by a number of physicians working with MEDgle. Additionally , publicly available information from National Institutes of Health – NIH was used and referenced.
How are results ranked?
MEDgle is based on 2028’s Information Engine which provides the basis for expert systems. The expert system has been seeded with information regarding age, duration, symptoms, and diagnoses. Based on a given search criteria, probabilities are calculated. The potential results are then ranked according to the aggregate proability scores.
What do the blue bars mean?
The blue bars are displayed to give users a visual cue regarding the common occurences of results for a given set of symptoms. A score between 0-100 is calculated based on probablity estimates. The user can thus get an understanding of the relative differences in “commoness” betwenn symptomatix results.
How is frequent, infrequent, and rare defined?
Whenever possible the prevalence/occurrence has been determined by literature review. In the absence of such data prevalence has been estimated from physicians’ daily practice experience.
What do you think? Does this strike an acceptable balance between simplicity and accuracy? If Google is right 58% of the time, how often should Medgle be correct?
Try it for yourself at MEDgle.com…
(hat tip: The Red Ferret Journal)