4 ms·

I have zero medical expertise, but it sounds like this situation could be improved with software. Given a database that contains an exhaustive list of every ill

I have zero medical expertise, but it sounds like this situation could be improved with software. Given a database that contains an exhaustive list of every illness and symptoms: illness symptom diagnosis_symptom_set ------- ------- ---------------------- id id id name name symptom_id (FK->symptom.id) illness_id (FK->illness.id) (Pseudo SQL) SELECT diagnosis.name as name FROM diagnosis_symptom_set diagnosis_sympt JOIN illness diagnosis ON diagnosis_sympt.illness_id = diagnosis.id WHERE symptom.name in (hives, chills, weight-loss, .....) This will at least give a patient awareness of what has been tried/tested and what remains, possibly saving them months or years.

nicolashahn 9y ago
Much easier said than done. There are several startups out there attempting to achieve something like that. The issue isn't in the technical difficulty, but the bureaucracy and red tape involved in getting the patient data, getting enough illness/symptom data to make diagnoses possible (which is a lot), proving to the hospital that it's worth the effort to adopt the platform, and making a solution agnostic enough to extrapolate to enough hospitals to turn a profit.
- mychael 9y ago
  Agree 100%. I never claimed it would be easy, just that software could help.
- beagle3 9y ago
  The issue is also the technical difficulty. There is huge variance and the number of variables to consider is also great. A computer/database is a better tool for the job, of course, but the balancing act between false negative and false positives is extremely hard. See, e.g. the (very) simplified case of base rate fallacy[0] ; There is an inherent information problem in doing diagnosis, that a large database and fast search do not address. [0] https://en.wikipedia.org/wiki/Base_rate_fallacy#Example_1:_Drunk_drivers https://en.wikipedia.org/wiki/Base_rate_fallacy#Example_1:_D...
always_good 9y ago
Someone give this guy a nobel peace prize!
- mychael 9y ago
  Haha
Terr_ 9y ago
https://xkcd.com/793/ https://xkcd.com/793/
phonon 9y ago
https://www.gideononline.com/about/gideon/ https://www.gideononline.com/about/gideon/ is like that for infectious diseases.

nicolashahn 9y ago

Much easier said than done. There are several startups out there attempting to achieve something like that. The issue isn't in the technical difficulty, but the bureaucracy and red tape involved in getting the patient data, getting enough illness/symptom data to make diagnoses possible (which is a lot), proving to the hospital that it's worth the effort to adopt the platform, and making a solution agnostic enough to extrapolate to enough hospitals to turn a profit.

mychael 9y ago
Agree 100%. I never claimed it would be easy, just that software could help.
beagle3 9y ago
The issue is also the technical difficulty. There is huge variance and the number of variables to consider is also great. A computer/database is a better tool for the job, of course, but the balancing act between false negative and false positives is extremely hard. See, e.g. the (very) simplified case of base rate fallacy[0] ; There is an inherent information problem in doing diagnosis, that a large database and fast search do not address. [0] https://en.wikipedia.org/wiki/Base_rate_fallacy#Example_1:_Drunk_drivers https://en.wikipedia.org/wiki/Base_rate_fallacy#Example_1:_D...

always_good 9y ago

Someone give this guy a nobel peace prize!

mychael 9y ago
Haha

Terr_ 9y ago

https://xkcd.com/793/ https://xkcd.com/793/

phonon 9y ago

https://www.gideononline.com/about/gideon/ https://www.gideononline.com/about/gideon/ is like that for infectious diseases.