Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

PG mentioned cancer 6 times in:

7. Ongoing Diagnosis

This is a real problem hackers could solve: early detection of health problems which leads to preventative treatments.

Although I lack any type of expertise outside of building a health startup I think this is probably one of the biggest ways we could reduce the impact of cancer on society... given the technology we have right now.

It's not a matter of medical break-through though, but about building the right monitoring tools, creating meaning out of data, getting people to use them and the greatest challenge for startups: dealing with significant government regulation.



The sensitivity-specificity curve is going to be a roadblock. If your False Positive rate is 0.5%, True Positive rate is 100% and your incidence of disease is 0.1%, the probability that a positive result is indicative of disease is only 17%.

Take a look at mammography, the false negative rate is estimated as high as 30% and the false positive rate of 65%.

If you can stratify your population (ie. increase the disease incidence), you'll do a bit better, but it's an uphill battle. Sure, late detection is a factor in the median life expectancy with pancreatic cancer. However, I suspect that if tumors were detected 6 months earlier, it would move the median survival time by 6 months and make no dent in the 5-year survival rate.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: