Hacker News new | past | comments | ask | show | jobs | submit login
How statistics can help cure cancer (bbc.co.uk)
19 points by stehat on Nov 24, 2013 | hide | past | favorite | 7 comments



Never work for Childhood Cancer: It is to "rare." With 12,000 a year on average for all childhood cancers it is statistically to small of a number.

How about this: Spend money on childhood cancer research! It truly feels that no one cares about a cure.

US Federal Cancer Funding 4% goes to childhood cancer research.

American Cancer Society is worse 1% goes to childhood cancer research.

There has not been a new chemotherapy medicine introduced for almost 25 years!


There has not been a new chemotherapy medicine introduced for almost 25 years!

SJG-136 has finished phase II, it is looking good. The developer of SJG-136 has told me (in a personal conversation) the side effects are very very very mild.

Shameless plug: I'm also looking to work towards bringing a new chemotherapy into the market. (http://indysci.org/projectmarilyn)


Won't be for pediatric cancer till after this has gone through what 5-10 years of a trial stages and then used on adults for several years.

Serious hope this project works BUT very sceptically it can make it to market. Seriously be a game changer if something goes directly to generic.


It comes down to economics and politics.

Suppose for every $1000,000 you spend you could

a) save one childhood cancer patient

b) save 20 patients with more common cancers

c) save a thousand starving African children

To the family of the child with cancer, (a) is the preferred option. To most Americans, (b) is the preferred option. To the starving people in Africa, (c) is the preferred option.


It cost much more than $100,000 to treat my son. Would have better if they could stop the process of the cancer spreading.

I lost my son at 12.5 years old in April to bone cancer (no change in over 25 years in survival rates). To bad the math doesn't add up to saving 20 patients with the money. I also lost my 15 year old sister to brain tumor in 1996. Both not genetic.

a) 20 adult patients (sure) 20 patients who are 65 years old gain 10-15 years (maybe) b) Save a 5 year old gain 70-75 years of a life.

Also the c) comment is very belittling and unbecoming.

By the way I have two family members who work in Africa helping these kids. I give money to help kids in India monthly. Do you give money to any of these thousands of starving kids?


The WSJ had a great article along these lines last week.

http://on.wsj.com/1durT8S


So these general ideas (plus machine learning, which is really advanced statistical calculation) was what Anil Potti tried to do. I believe that he had good intentions going into it. But it's a really really difficult problem, and he maybe oversold the idea. The end result was not pretty.




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

Search: