This is a bit out of topic and I hope that "savemylife" will find out more about his/her problems. I find that a "crowdsourced diagnosis" approach may have a r

This is a bit out of topic and I hope that "savemylife" will find out more about his/her problems. I find that a "crowdsourced diagnosis" approach may have a revolutionary impact on the way doctors and patients look at diseases. Recently I have lost a close family member because the doctor who first visited him overlooked some symptoms that could potentially lead to the discovery of his disease (a cancer). Time passed and the cancer developed to a stage that was impossible to cure. Following this experience, I've started playing with the idea of a world-wide user generated symptoms database, managed and moderated by real doctors and experts. The main issue with such database would be the enormous amount of "false positive fright" triggered by casual users just checking out basic symptoms (for instance, a sudden high fever can be a symptom of a cancer initiated infection http://www.ncbi.nlm.nih.gov/pubmed/16257792 http://www.ncbi.nlm.nih.gov/pubmed/16257792, but can also be a symptom of many other less life threatening diseases). I'm not a doctor so I'm not aware if something similar already exist - probably it does, on a country specific basis maybe? I know for sure that some doctors use Google regularly to double check their diagnoses. In most European countries, the majority of initial diagnoses are done by General Practitioners, who are doctors assigned to a patient by the national healthcare system. These doctors actually redirect patients to specialists or to hospitals for specific analysis based on - often very quick - assessments of the patient's symptoms. I'm not sure how the healthcare system works in other countries, but I believe that there is always some kind of initial medical routing based on symptoms. I wonder if the precision of this initial assessment could be improved through this database and some statistical algorithms? Any doctor, biologist (and of course, regular HNers) care to share his/her catch on this idea?