Every patient is unique. Future doctors will wonder why we didn’t treat them that way | Opinion

The 20th century was all about mass production and one-size-fits-all. But the 21st century is about diversity and the uniqueness of the individual. The COVID-19 pandemic demonstrated this uniqueness very clearly with some people being ravaged by the disease and quickly dying, others lingering with the effects of the disease for months and possibly years, and still others fending off the illness with only mild symptoms. Even among those who took the vaccine, the outcomes varied a lot. It worked really well for some individuals, and not as well for others.

In the 22nd century, people will be shocked to learn how long it took the medical profession today to catch on to the importance of treating each individual as a unique human being. For the most part, medicine still relies on reports from randomized trials that determine how the average person responds to a drug or other medical treatment. For any known or unknown characteristic, the randomization process puts roughly half of drug study participants with that characteristic in the control group, who get a fake pill, and half in the treatment group, who get the real one. The difference between how the average person in the treatment group responds to the real drug and how the average person in the control group responds to the fake pill is reported as the effectiveness of the drug. We learn all about Joe Average — but not about the unique individual the doctor is treating.

Even more appalling is doctors’ use of anecdotal experiences that they have accumulated over the years. Yes, that drug worked well for that young man with high blood pressure. But that does not mean it is going to work well for the old woman with low blood pressure, or for the young, pregnant woman. Psychologists have long noted that too often people focus on a few memorable outcomes that nominally appear important but are really quite misleading when applied to the current problem at hand. Too often doctors are influenced by this bias, which is also known as the representativeness heuristic in psychology.

By taking a DNA sample, the online health and ancestry service 23andMe can track down relatives you didn’t even know you had. Police departments can use those results to track down criminals who didn’t submit a DNA sample to 23andMe, but their relatives did. “We know your type” has gone from a vague generalization to a method of precise identification. Is your doctor using your DNA sample to help decide what treatment will work best for you?

What if the drug works great for young nonsmoking men in their 20s, but is harmful to older women smokers in their 60s? Ignoring individual differences doesn’t cut it. By using a previously estimated regression equation, the effects of each of our unique characteristics in predicting the effect of a medical treatment on us as individuals can be determined. The values of a particular person’s characteristics are plugged into the regression equation and the predicted outcome, such as disease intensity (for example, prostate specific antigen score for men) is displayed. With all the computer programming capability and computer power we now possess, we could easily use regression analysis to predict how a particular person will respond to a drug or other medical treatment.

Improving randomization, protecting privacy

What about the data from randomized trials? Back in the day, randomized trials were invented to deal with samples that were too small to support extensive regression analysis. A randomized trial implicitly controls for a host of known and unknown confounding factors that would otherwise distort the results through inadvertent biased sample selection. Just as the law of gravity always seems to work, the randomization procedure in randomized trials usually, more or less, evenly divides any subgroup such as smokers between the treatment group and the control group. This randomization process works not only for known characteristics such as smoking, but implicitly works for any and all unknown characteristics that then are implicitly evenly distributed between the treatment and control groups.

Hospitals are now working together across the United States to carry out large-scale randomized trials to produce datasets with large sample sizes. If the individual characteristics of individual patients are also collected, the randomized trial information can be incorporated into the regression equation as a binary variable that takes on the value of one if the person got the real drug, or a value of zero if they got the fake placebo pill. This binary variable can also be additionally represented in interaction terms with other variables such as a person’s age, gender or blood pressure, to name a few.

Research studies need to collect information such as the gender, age, height, weight, blood chemistry, personal medical history, family medical history and DNA of the participants. To ensure privacy, that information needs to go into an anonymous data bank that leaves out the names and addresses of the research participants — with no link to their identity — but that uses their characteristics in a regression analysis to identify how each characteristic interacts with other characteristics in predicting the effect of a drug or other medical treatment.

These studies would need to have large numbers of participants and will be expensive. However, once these regression equations are accurately estimated, your doctor will be able to insert your individual data into the appropriate regression equation to much more accurately predict how you as an individual will react to a drug or other medical treatment. The medical profession will finally be able to treat us as individuals and leave the old one-size-fits-all 20th century behind.

Lawrence C. Marsh taught at the University of Notre Dame, the University of Chicago and in the psychology department at Avila University. He regularly participates in monthly seminars that focus on the use of statistical methods in medical research. He is a former Kansas City Star Midwest Voices opinion columnist and published 81 of those columns in his book “Brain on Fire” in 2011. He lives in Kansas City.