HPN Wants Better Predictive Health Care a.drian via Flickr

It’s almost as if Heritage Provider Network set out to create the perfect PopSci story by mashing up all of our favorite things: clever algorithms, a multi-million dollar intellectual competition, and the future. The California-based health care provider has put up a purse of $3 million for the person or group who can come up with a predictive algorithm that accurately identifies people at risk for hospitalization in the next year, thus encouraging predictive medical measures and reducing unnecessary hospital stays.

The prize, officially launching on Monday, aims to reduce the $30 billion that was spent on unnecessary hospital admissions in 2006 (The last year data was available? We’re quoting HPN here). To quote HPN further:

The winning Team will create a predictive algorithm that can identify patients who are at risk for hospital admissions. Once known, health care providers can develop new care plans and strategies to reach patients before emergencies occur, thereby reducing the number of unnecessary hospitalizations. This will result in increasing the health of patients while decreasing the cost of care.

It also results in the world’s largest predictive modeling contest that we know of, paying roughly twice the recent payouts for the Nobel Prize in medicine. HPN is also offering progress prizes along the way totaling another $230,000 to help leading contenders continue their research. Each team will be provided with a bunch of de-identified medical records on which to test their models. The first team to devise a working algorithm that achieves the required level of predictive accuracy wins.

The competition is expected to run for two years.

[Heritage Health Prize via Network World]

12 Comments

This would be worth a LOT more then $3 million to health care companies... the ability to see that your going to need hospitalization is a great excuse to boost the cost of service right now. Like it isn't bad enough as it is...

Sounds like "Minority report" for meds

and of course, big hospitals may abuse such an algorithm, as well as life insurance companies.

It simply would be a way to sift through potential clients, rejecting the "future ill".

It would be great if they provided test data so I can teach my predictive algorithm how to learn. This will be a cake walk. Easiest $3 million dollars I've never made.

Pioneer10 said: "It simply would be a way to sift through potential clients, rejecting the "future ill"."

What I forsee is that if insurance companies have access to this they will offer a 'basic' policy that covers you for everything that you are not predisposed for, and more expensive policies for anything you're likely to get according to this algorithm. Perhaps costing more per percentage likelyhood?

Two points -
1. 3 mill would be chump change compared to the worth of this app -- even if it was used beningnly as they are suggesting

2. Why does the headline say 15 years while the text discusses a single year.

And when will they be able to say that the algorithm works. I mean really works, on patients, not on selected past records.
In 15 years? Or more?
Otherwise they couldn't tell how accurate the algorithm is on healthy patients and if the algorithm worths the 3million.
Somehow the movie Gattaca comes to my mind...

Do programmers and other non-health related people have access to anonymous health data? or this is gonna promote health data insecurity even further? I agree with concerns regarding the misuse potential of this algorithm, it has to be kept in mind. On the other hand if it is reliable it can be very useful in disease prevention.

It is all the matter of legislation. In my country, health insurance companies cannot screen you this way. They must insure anyone, or they or out of business. Health insurance is mandatory in my country and it is counte as fixed a percentage of the salary. But health insurance companies are private.

This would be near impossible. there are waaaaay too many factors. have they ever heard of chaos theory??

I would love to help these people out but my price range would be in the billions not millions.
Jg16477 said that it would be nearly impossible, not really.
As a matter of fact, predictive algorithm can possibly be predetermined just by genetic characteristics even before a person is born.
I‘ll give you a clue on an example, everyone has a twin, a look-a-like, not a biological twin at birth but a genetically identical twin.
Oddly whenever twins come together, they exhibit the same characteristics in viewpoints, clothing style, likes and dislikes and even physical ailments, etc.
And this happens when in total isolation from each other and in never meeting.
Genetically are the reasons that whenever you see someone familiar looking they remind you of someone that you know.
The most daunting task and maybe a near impossibility to an accurate predictive algorithm will be the collecting of information and a compilation of the data.



June 2013: American Energy Independence

Five amazing, clean technologies that will set us free, in this month's energy-focused issue. Also: how to build a better bomb detector, the robotic toys that are raising your children, a human catapult, the world's smallest arcade, and much more.


Online Content Director: Suzanne LaBarre | Email
Senior Editor: Paul Adams | Email
Associate Editor: Dan Nosowitz | Email
Assistant Editor: Colin Lecher | Email
Assistant Editor: Rose Pastore | Email

Contributing Writers:

Kelsey D. Atherton | Email
Francie Diep | Email
Shaunacy Ferro | Email

circ-top-header.gif
circ-cover.gif
bmxmag-ps