I hear a lot about what’s “clinical” or what isn’t, or what’s being done clinically, or clinical assessments. Think for a moment about what this term means to you. It should have a precise meaning, but oftentimes the true meaning is lost amidst a sea of misconception with strong headwinds of nonsense.

Getting back to the “Klinikos”
The first known use of the term “clinical” was c. 1728. It derives from the Greek klīnikós, which means ‘of the bed.’ This later became klinike (the sick bed), and was drawn into Latin as clinicus (physician that visits patients in their beds).

But even those who do this kind of work don’t always get it right. In the Journal of Pediatrics, they have written: “…three of these six foods were actually the cause of symptoms, and upon their elimination, clinical cure was effected.”

There are two huge scientific problems with this statement, and in a reputable journal no less.

First, nothing can ever be proven to be “cured.” This is instead a statistical hypothesis that must be tested against the empirical evidence, in which case the best that can ever be said is that a condition was “resolved” or “abated.”

This is NOT the same thing as “cured” because the null hypothesis is never “proven”—it is either rejected or it fails to be rejected. A medical treatment may be said to be effective if we reject the null hypothesis, which should be a priori written as “there is no difference between the ‘treatment’ group and the ‘no treatment’ group. Falsification of the null hypothesis gives us evidence that there is indeed a difference between the two groups, but can never prove an effect.

Second, “clinical cure” is no different from “regular cure.” Something either works or it doesn’t. In cases of use like these, it’s the author’s choice to weigh down their language with hubris. “Clinical” just sounds so important in that journal quote, but something that appears to have an effect in an actual clinic should be expected to have a similar effect in someone’s home, in a shopping mall, or on Mars.

Here’s a semi-faulty famous use of the term as well:

“Clinical observation supports this contention in many cases.” —Paul Ehrlich, Histology of the Blood, 1900

What were the observations? What data were gathered? The term “supports this contention” should bring to your mind the idea of a null hypothesis. What is the null hypothesis, the contention that’s being investigated? And what does “in many cases” mean? What’s the quantification of “many”?

Because something done in a clinical way necessarily involves independent assessments of singular patients, it is ultra-highly subjective, susceptible to opinion, and anecdotally-based.

Better medicine comes from better data
Good population-based medicine, and good pharmaceutical testing and trials are based on many thousands of data points. So what we typically do in clinical trials is amass many, many investigators’ independent evaluations and pieces of information and combine them together to try to detect a viable signal. This is essentially taking a Grand Mean of many different measured parameters, including subjective ones, and is absolutely fraught with variability. This is the main source of noise in our industry’s trial data. Any good scientist eschews the notion of a qualitative assessment being a good source of data to conduct experimentation.

The more we can use personalized medicine, precision medicine, pharmacogenomics, and better measurement paradigms that take human appraisal variability out of the equations, the better the future of medicine will become. It’s nice to see a clinical effect in one or a small number of patients, but remember that virtually every single drug trial shows a positive effect in the placebo group, and so we see every day that people are showing up to clinical visits reporting positive outcomes from null treatment.

We can treat patients clinically (at the bedside), but we only know what works by using and improving upon our data measurement paradigms. What really matters is whether or not we can demonstrate evidence that a medical treatment works systematically through randomized controlled trials, which are used specifically to try to reduce the impact of chance in the outputs we measure. If we don’t have this degree of confidence in a treatment—supported by data—then we can’t say a treatment ‘works clinically.’ Period.

---

Ben Locwin

Ben Locwin, PhD, MBA, MS, MBB, is a healthcare futurist, award-winning published author, and has been featured in The Wall Street Journal, USA Today, Forbes, and many books and magazine publications. He serves on several industry advisory boards and boards of directors. He was the Keynote for Contract Pharma’s 2018 Annual Conference.