Algorithm: New Statistical Model to Predict future Health

29 year old Sandy Cabrerra is a picture of health--and she wants to keep it that way.

Looking down the family tree keeps her in motion.

"Crime example, I do try to stay healthy, I do try to stay in the gym and exercise as much as possible because I am prone to diabetes,” Sandy said. “I have a family history of diabetes so already knowing that I try to take the proper steps to try to prevent it in the future."

Sandy isn't the only one looking ahead.

Researchers at the University of Washington analyzed the medical records of thousands of patients and came up with a statistical model to predict possible future health issues.

Researchers said an algorithm forecasts the future the same way Netflix recommends movies or websites list other items you might like.

Family practice Dr. Hania Alaidroos said doctors already look into a patient’s past to gauge their future health.

She added that more information is always good--but wondered if it will alter the way doctors approach patient care.

"I don't know if it will change management, in terms of how doctors practice medicine,” Dr. Alaidraoos said. “It may give patients more information but i don't know if it will really affect the doctors."

Researchers looked at 1,800 ailments--from acid reflux to strep throat and found that one condition often leads to another.

If you have stomach issues like epigastric pain--heartburn might be next.

Researchers hope to make the model available doctors and patients. Sandy considers it preventive medicine.

"It would definitely give me more information about how to take care of myself and I would like to be informed to try to prevent any other health issues in the future."

Sandy said the algorithm will allow her to crunch the health number herself.

"I would definitely do some research, yes."

The study appears in the Annals of Applied Statistics





Copyright © 2014, Los Angeles Times