Technology: Will docs battle Watson for your diagnosis?

Healthcare has always experienced a love-hate relationship with technology.

New technologies have certainly both contributed to costs and reduced costs.

There has been considerable literature on how expensive new technologies become unnecessarily overused, and as the recent showdown with Ontario doctors demonstrates, can also be a catalyst for savings when procedures take less time and resources.

Can new technologies replace doctors altogether?

IBM’s Dr. Martin Cohn says no, although looking at the company’s healthcare applications for Watson, what is sometimes referred to as “IBM’s artificial intelligence computer,” it is not hard to see it becoming a major part of any doctor’s diagnostic process.

Watson is the computer IBM used to defeat two Jeopardy champions in a three-day tournament. It uses natural language processing than can distinguish between a pun and more conventional meaning. Or in the case of Jeopardy, understand the rules of the game and interpret a question from a series of answers. It is not a search engine, but can sort through 200 million pages of health care articles in seconds and provide an ordered list of diagnostic probabilities from the list of symptoms provided it. It also knows what it doesn’t know, and can urge a doctor to look further. Watson also continually evaluates the reliability of its sources.

Watson can also tap into an e-health record and prioritize information in it as part of the process. While electronic health records have been presented as a panacea for doctors, they suffer from the same problems of an overly thick folder of paper records — too much information to reasonably sift through.

Kohn, Associate Director, Healthcare Analytics at IBM Research spoke to a roomful of health care administrators June 26 as part of Longwood’s Breakfast With The Chiefs series.

Kohn points out that we are overwhelmed with information, stating doctors average less than five hours a month reading medical journals.

According to the Harvard Business Review, about 15 per cent of doctors’ medical diagnosis are wrong. That figure may be low. The Canadian Association of Radiologists has suggested that as much as 30 per cent of diagnostic imaging procedures ordered by doctors are unnecessary. Kohn’s own Powerpoint slides suggest the number of wrong diagnosis may be somewhere around one in five.

The tantalizing thing is, could Watson improve that percentage? Would improving the diagnosis better utilize resources, as the radiologists suggest? It is hard not to look at the potential and start estimating its potential impact on health care funding and utilization.

Kohn describes his early internship in a hospital emergency department and how more experienced nurses would discreetly make suggestions in his ear when he was particularly stuck in determining a diagnosis. He says Watson is intended to be like that nurse – providing assistance, but not replacing the doctor’s role.

A question from the audience raised the possibilities that doctors would defer their own judgement to Watson, making it less of a tool and more of a crutch. And how would liability issues work if Watson was making the decision instead of the doctor?

Kohn admitted that Watson was only as good as the information it was processing. It is not intended to be a decision-maker, one of the reason’s IBM doesn’t like the term “artificial intelligience” to describe Watson. However, Kohn does admit that Watson is continually learning.

And Watson is capable of making mistakes. During the three-day Jeopardy tournament, Watson identified Toronto as a U.S. city.

Will patients themselves prefer the diagnosis of Watson over that of their doctor if the two differ?

All-time Jeopardy champion Ken Jennings told the Washington Post that the audience for his three-game tournament against the machine “was an all-IBM crowd: programmers, executives. Stockholders all!” he said. “They wanted human blood. It was gladiatorial out there. The stage had a big Watson logo on it too. This was definitely an away game for humanity.”

What was once science fiction may be a commercial reality within two years as IBM prepares Watson for release.

Unlike Jeopardy, the questions it raises have only just begun.

If you liked this story, pass it on! Click the share button below, post it on your Facebook site, or retweet it to your followers.

One response to “Technology: Will docs battle Watson for your diagnosis?

  1. FYI –
    There are at least two places in USA called Toronto.
    1]Toronto is a city in Woodson County, Kansas, United States, along the Verdigris River. As of the 2010 census, the city population was 281.
    2]Toronto is a town in Deuel County, South Dakota, United States. The population was 212 at the 2010 census. – All text is available under the terms of the Creative Commons Attribution-ShareAlike License‎ – Wikipedia

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s