Big Data Malpractice

Garbage in, garbage out.  Remember that? 

Some related observations from the article we discussed in our last post:

The word “data” connotes fixed numbers inside hard grids of information, and as a result, it is easily mistaken for fact. But including bad product introductions and wars, we have many examples of bad data causing big mistakes.

Big Data raises bigger issues. The term suggests assembling many facts to create greater, previously unseen truths. It suggests the certainty of math.

That promise of certainty has been a hallmark of the technology industry for decades. With Big Data, however, there are even more hazards, some human and some inherent in the technology.

I think that’s a suggestion from an expert that we need to keep some perspective about Big Data.  In particular, note the reference to “bad data.”

For a health care specific example, consider a recent article in The Atlantic.  The article describes how a mistake in a voice recognition dictation system resulted in a patient record that stated the patient had an amputation below the knee.  But:

When the team arrived in the patient’s room, they made a surprising discovery.  The patient had two feet and ten toes.

Garbage in, garbage out.

There’s probably too much of a gold rush attitude about Big Data in general, but in the health care world there is a particular need for caution.  It doesn’t always seems to be there.  For example, consider this observation by Mark Tobias in an a recently published article:

What if “electronic health records” could become “electronic health oracles” – not just recording the past, but helping to predict and influence the future?

Sounds great.  Maybe a little over the top?

In fairness, Mr. Tobias does say this:

The success of prediction depends on the quality of data used in the process and the predictive models that are applied to make sense of that data.

But that is a single cautionary note, made as an aside, in an article that also contains this statement:

Data from social media sources like Facebook and Twitter, as well as wearable sensors and fitness apps can also provide key direction for delivering more personalized, precise health care.

I think that pretty much puts to rest any idea that Mr. Tobias is concerned about data quality.  I don’t know about you, but I don’t I want my “personalized, precise health care” driven by data from Facebook or Twitter.  Perhaps the reason for Mr. Tobias’ enthusiasm is that:

Mark Tobias is president of Pantheon, which combines technology expertise and a deep knowledge of health care, education, and social impact markets to provide online technology solutions for nonprofits, associations, and government.

Unsurprisingly, he promotes his company in his article.

We could go into the history of medical advances that have turned out to be mistakes.  If you watch too much TV, you’ve probably seen commercials for lawyers that are seeking plaintiffs for cases involving such mistakes.  Imagine one that starts:

Did your doctor make a treatment decision using the Big Medical Database?  If so, you might be entitled to compensation.

The software industry has (at least in the past) understood how high risk situations are different, as evidenced by the frequent use of this type of provision in software licenses:

Customer acknowledges that the Software is not designed or licensed for use in or in connection with: on-line control equipment in hazardous environments such as operation of nuclear facilities or aircraft navigation, communication or control; life-support systems or procedures; in medical diagnostic applications; or in surgical or other intrusive procedures or otherwise to implement medical procedures or sustain life.

There will always be people who insist on adopting the latest technology – no matter how unproven and how significant the risk.   Let’s be careful about moving too fast and taking advice from people who are selling something.  If for no other reason, they might not be there when you get into a fix because you that advice and used that something.

This entry was posted in Big Data, Diatribe and tagged , , , . Bookmark the permalink.