Using analytics could be key to curbing medical bias

Today at Cleveland Clinic and HIMSS' Patient Experience Summit, Dr. David Newhouse, formerly of Permanente Medical Group, discussed putting statistics into practice.
By Laura Lovett
03:20 pm

Medical bias has been a hot-button topic this year, as racial disparities continue to loom over the COVID-19 pandemic.

The medical industry has a historic racism problem. In fact, J. Marion Sims, sometimes called the father of modern gynecology" experimented on three Black women, and did not use anesthesia. Nor is this the only example in the history of the U.S.

"This legacy lives on. This is why having good data collection today is critical in trying to undo the damage from the past," Dr. David Newhouse, a retired doctor from The Permanente Medical Group in the San Francisco Bay Area, said during the Cleveland Clinic and HIMSS' Patient Experience Summit today.

Newhouse worked for the Kaiser Health systems and over the years used analytics to help determine where doctors may have unconscious bias and can improve care for patients across demographics.

"What happens with a patient and a physician meet for the first time? Body langue is identified in 115 milliseconds, and a first impression is formed," he said. "What do two people see when they meet for the first time? They see gender, they see skin color, tattoos, what they are wearing how big they are. From those impressions they have a very quick impression made. Is this someone favorable in our mind or not favorable?"

At Kaiser, they developed a more in-depth patient survey in order to look at patient-satisfaction rates.

"Get the good data and break it down so you have a better idea of what is going on. So, we took the basic survey form that Kaiser had and developed a more sophisticated model," Newhouse said.

The survey asked questions about whether or not the last interaction was a return visit or first encounter with a doctor, as well as the person's age, gender, ethnicity and tenure with the health system.

Newhouse, an OBGYN by training, began to notice that initial trends in his data indicate that his scores were much lower with his patients between the ages of 18 to 34 than with older patients. He noted that many of the younger patients requested to see female physicians.

Being able to pinpoint his weakness, Newhouse said he was able to transform how he practiced and took particular care to answer all of his younger patient's questions and concerns.

While the data gave insights into potential bias, there were some possible stumbling blocks along the way.

"There is an impact of demographics on ratios. For example, a newer doctor is going to have more new patients, since return patients tend to give their physicians a higher score. This could negatively impact their overall score," he said.

There are also operational pitfalls. For example, a younger Mandarin-speaking physician had lower than average scores. She also saw a higher-than-average proportion of Mandarin speakers.

"We did not have adequate interpreter services. The survey tool was inaccurately translated. There was no word for excellent for the survey in Chinese. I said, 'But let's look at how her patents are coming back to see her,'" he said.

While the tools aren't perfect, using data to inform practice is key, said Newhouse.

"It allows a physician to see where they do well and focus on the scores that are lower. Doctors are very busy, and this is an efficient use of time in a doctor's schedule," he said.

"Too often they just coach doctors to smile more or do things that may be effective in other ways, but doesn't really drill down to where is the problem at. When a doctor does better in their scores, the morale improves, which turns around to having better service scores. It holds everyone accountable."


The latest news in digital health delivered daily to your inbox.

Thank you for subscribing!
Error! Something went wrong!