help

collapse

Press one of the expand buttons to see the full text of an article. Later press collapse to revert to the original form. The buttons below expand or collapse all articles.

expand

collapse

Robo-Mom

October 28, 2009 permalink

In the future, a computer may decide to take your child to a foster home, if the dreams of Dr Ben Reis of Children's Hospital Boston come to fruition.

expand

collapse

Artificial Intelligence Diagnoses Abuse

Eric Bland, Discovery News

Oct. 26, 2009 -- A broken bone one day, a particular infection a few months later and depression the following year may appear to be separate, medical issues.

However, to a new artificial intelligence program developed by Boston doctors, these are all symptoms of domestic abuse.

The new software can identify abuse victims up to six years before these cases would otherwise be found and could eventually be used to diagnose just about any disease or injury.

"It's very difficult to detect domestic abuse because it often happens in the privacy of the home," said Ben Reis, a doctor at Children's Hospital Boston (CHB) who helped develop the program.

"Doctors are often on the front lines of detecting abuse, but so often the doctor is focused on treating the injury, they don't see the context behind it."

Dozens of studies over the last 40 years have correlated various illnesses, injuries and other conditions with abuse. Bruising to the middle of the forearm or the core of the body instead of the elbow or knee can signal abuse. Depression or alcoholism may also be symptoms of this condition.

The new program takes all of these studies and puts them in one place. Before a doctor even consults with a patient, the software will alert the doctor about the likelihood of abuse in a particular case using a color-coded system.

It would then be up to the doctor whether to treat the patient as a victim of abuse.

To create the artificial intelligence program, the CHB doctors used six years of data from more than 560,000 patients ages 18 years and older.

While many studies have correlated certain injuries with abuse, Reis, along with his colleagues Isaac Kohane and Kenneth Mandi, tasked the program with finding its own connections between an abuse diagnosis and records of injuries leading up to that diagnosis.

The program picked up domestic abuse an average of two years before it was diagnosed. In one case, the program detected abuse more than six years before it was diagnosed by a physician.

Robert Sege, a doctor at the Boston Medical Center not involved with the study, is very impressed with the research.

"By the time someone is identified as abused, the abuse has generally gone on for a long time," said Sege. "This is really powerful. If I had this with particular patients, I would be aware of the possibility and ask them more detailed questions or tell them about the kinds of services we provide."

Electronic health records are the lynchpin of the artificial intelligence system, according to both Reis and Sege.

Abuse victims often seek medical treatment at different locations for each injury to avoid detection. Each hospital or doctor keeps separate patient records but often fail to share those records.

Keeping one, consolidated electronic record of a patient activity would enable the software to detect abuse.

The record doesn't even have to be particularly detailed, says Reis. The CHB doctors used only the most basic patient data, "the lowest common denominator," as Reis calls it, to diagnose abuse in patients. Using better data should only increase the program's success at abuse detection.

Better data could also allow the program to diagnose a variety of diseases more accurately as well. Ultimately, Reis and his colleagues want to expand their software to consider every human disease, a project he likens to the Human Genome Project, a multinational, $3 billion effort to sequence the entire human genome.

To support their research Reis was recently granted a four-year, $1.3 million grant to expand their research.

Source: Discovery Channel

Robo-Mom

sequential