The Darker Side of Data Mining Ask Your Doctor

I’m a digital marketer. We mine data from the Internet all day long in order to sell more shoes. I have no problem with that. People like shoes, they want shoes. If they spend so much money on shoes and can’t pay their rent because of a campaign I designed, that’s perfectly OK with me. It’s called a free market.

What digital marketing does well is collect (mine) all the data floating around the Internet -- things like tweets, likes, shares, webpage visits, downloads, shopping cart activity and purchases -- in order to find out what experiences with the brand drive the most desired outcomes. We call it “actionable insight” -- data that reveals how we might drive better outcomes. The outcomes we drive are always one of two things: reduced cost or increased revenue. 

But here’s a conversation I had with my sister recently about the way data mining is used now in medicine that I found disturbing.

Data Collection, For What Outcome? 

My sister is a family practice physician and a college of medicine faculty member at a large state university. Recently, management spent upwards of $130 million on a new record-keeping system that has made her life miserable.

Where she once spoke for about a minute into a little tape recorder dictating patient notes after each appointment, under the new system, her notes must be entered through an online form. She must tab and click from this field to that, use pull-down menus a mile long for diagnoses, procedure codes used, medications proscribed. Where once test results were simply faxed over from the lab and clipped directly into patient charts, my sister is now is responsible for putting test results into the proper fields when they arrive. In other words, she’s now a doctor plus a data entry clerk. Her new data entry responsibilities have added hours to each and every workday.

As a business person who works with data every day, I’m thinking, “Well, digitized data is always more useful than hard-copy documentation because you can look at it, manipulate it, model it, find trends and uncover actionable insight. And structuring data on the way in to a database makes it much, much easier to use.” But as a business person I’m thinking, “What actionable insight is this system designed to uncover?” So I asked her.

“They want to know which doctors have bad outcomes so they can align compensation accordingly,” she said.

“Excuse me?” I asked.

“For instance,” she explained, “they want be able to say, ‘Dr. Jones, the average blood sugar levels across the diabetics you care for didn’t come down significantly this year. Your compensation will be adjusted down next year.'”

In other words, not only did my sister acquire 12 years of specialized education to become a part-time data entry clerk (in her gobs of spare time, of course), she went to medical school so she could cede control over her income to her patients.

What Medicine Is, What Medicine Isn't

Here’s the thing: Making digital marketers accountable for outcomes is good idea because the business of business is to make money. Make someone accountable for the amount of money they save and/or bring in, and they usually save and/or bring in more money.

But the business of medicine is to offer suffering people a form of relief that has provided relief for others like them in the past. That’s it. The business of medicine is not “to heal.” Because whether or not healing actually takes place is 100 percent dependent on the patient’s particular physiology and behavior. Doctors can’t force their patients to take their meds, lose weight, exercise or quit smoking. Doctors have no control over whether 20 really, really sick people walk in the door each day or just two. That an accepted form of treatment has worked well for thousands of previous patients doesn’t mean it will work well for any particular one of us. As all doctors will attest, all humans are different.

The most disturbing thing of all about making doctors responsible for patient outcomes? As a digital marketer, I can tell you that when we don’t like the outcomes we’re getting from a particular channel, campaign or tactic, we drop it. We minimize the downside and move our attention and investments to the places where the payoff is known to be good.

In medicine, minimizing bad outcomes means getting rid of the least healthy patients. Lest we think this beyond the realm of possibility, what do we think motivated health insurers to refuse to insure those with previously diagnosed conditions? Yep -- risk mitigation.

Here’s a sad tale that illustrates the very dark side of motivating physicians to mitigate risk. A few years ago, my sister’s hometown of 750,000+ lost its only pediatric cardiac surgeon. The surgeon’s malpractice insurance dropped her because she had too many bad outcomes in a row. Too many really, really sick babies had been born that year.

Now, I ask you: If your little girl was just born with a large ventricular septal defect (a hole between her ventricles) that required surgery, would you like her tiny blue body (because she’s not getting enough oxygen) to be medevaced out to the next city over because yours no longer has a pediatric cardiac surgeon of its own? Would you like the surgeon in the next city to refuse to take her because surgery for her particular condition was only 20 percent likely to save her and he was only taking cases right now with a 60 percent chance of a positive outcome?

People do what they are incentivized to do. People stop doing what they are penalized for doing. Doctors are people. Give them incentives for taking care of healthy patients, penalize them for taking care of unhealthy patients, and very soon the unhealthy among us will have no where to turn.  

Title image by Official US Navy Page (Flickr) via a CC BY 2.0 license