In healthcare, or anything, that is an important question. It comes up in almost every healthcare encounter.

If the number of variables and their combinations were not so huge, it would be an easier question to answer.

We have individual genomes with their own proclivities. We have a dozen different prescription drugs or classes of drugs for the same condition. Then, we also have endoscopic and surgical interventions as well as medical devices, sometimes all for the same condition.

Millennia ago, the answer was easy: ask the shaman. The shaman was a religious leader with access to all hidden knowledge. In the West the first physicians were in Egypt and Babylonia, but still with heavy spiritual overtones. Later medicine seems to have split definitively from religion during the classical Graeco-Roman period, with physicians such as Hippocrates and Galen.

Knowledge began with spirit, continued with logic from the classical period to the medieval, and gave way to empirical methods in the Renaissance, formalized in the science of the modern period.

How we know something is the scope of the epistemology, a branch of philosophy. So, why am I waxing philosophical? Two events within the last month.

First, I went to my primary care physician for a preventive checkup. In discussing my medications to lower cholesterol, and I take more than one, he pointed out that studies had shown that one did not lower LDL and that another did not decrease mortality by raising HDL. And, we discussed our differing views of the matter.

Second, I read a blog discussion today about anti-depressants. A polymath named Irving Kirsch is a recognized expert on the placebo effect and questions the efficacy of antidepressants, most notably in a 2009 book The Emperor’s New Drugs.

During the discussion the name Karl Popper was mentioned. Popper was a 20th century philosopher credited with the concept of “falsification.” In the simplest terms, and not being a philosopher, I may distort his views, propositions cannot be established but only be proved false, or fail to be proved false. Furthermore, to be scientific a proposition must be falsifiable, so saying there is or is not a Deity is not a falsifiable proposition.

Let’s get back to healthcare: if I say that a drug for hypercholesteremia or for depression is no different than a placebo, then it is falsifiable. I can gather data and compare the effect of the medicine to the effect of a placebo. If they diverge by a sufficient amount statistically, then the idea that there is no difference has been falsified.

Even so, questions can always be posed about the methodology of any study, and the ensuing debate begins to resemble the medieval religious debates we thought we left behind.

Furthermore, not every variable that can cloud a conclusion can be eliminated. Two drugs may have negligible effects alone but work powerfully in tandem. A drug that works on one individual may not work on another. A drug may not work on 99 individuals and still be valuable for number 100.

Finally, even if every question of efficacy was answered, the questions of values remain. Those questions are most visible in the public arena with respect to abortion, birth control, end-of-life decisions, and the availability of certain drugs. Less visible but just as important is the following scenario: a patient has a painful condition for which there is an effective intervention. Unfortunately the odds are 60% complete cure, 20% no effect, 15% debilitating injury that is worse than the initial condition, and 5% death. Even if these percentages are accurate and cut in stone with no chance of distinguishing a priori in which basket a given patient will fall, then the clinician can recommend but not decide. The 20% aggregate possibility of a worse outcome than doing nothing has to be left to the patient.

We can produce rules for tests to be made, pharmaceuticals and other interventions to be offered, but for the moment the number of possible interventions and the nearly infinite possibilities of human diversity mean our rules will be imperfect, no matter how well researched, studied, and reviewed.

Having read a bit about Kirsch, I doubt if he is right, but even if he is right, the 25% impact of pharmaceuticals on depression may make them worthwhile even if most of their impact is placebo. I am still taking the same cholesterol medications as I have found through experimentation that removing one raises my cholesterol and retaking it lowers it again, and I am alive 20 plus years beyond the life span of my father and his father.

That is how it should be. The important decisions in life (where to live, who to live with, where to work) are all made based upon imperfect information. So are our health care decisions, but we make them as patients and clinicians with the best knowledge available and hope to look back upon our having been correct.