Those of us uninitiated into the patois of medical statistics are prone to misunderstandings. If these misunderstandings were “academic” then we might approach them with a leisurely attitude. But often decisions are thrust upon us without warning and the medical literature and physicians do not commonly translate the statistics in a readily comprehensible way.  Our decisions may reflect regrettable choices with haunting consequences.

We have two threshold issues: (1) the reliability of the statistics in light of their purposes and (2) the degree of clarity in the presentation of the statistics.

As to the first point, let’s take as a given the soundness of the underlying studies and data collections. This leap glosses over critical issues but these are not our immediate concerns.  One caveat: even the most elaborate and precise data sets, presented in elegant math,  rest on simple premises that are commonly presented as givens and as non-controversial –  in other words, biases and choices cloaked in the (falsely) reassuring neutrality  of numbers.

We transition.

How often are averages presented with no articulation as to whether they represent the median, mode or mean? Large issue loom in the choice. One outlier skews the mean. Scatter leaves the median, without more, meaningless. And so forth. Opaqueness lurks around risk statistics: are they measuring absolute or relative risk? Is a 50% risk reduction a reduction of an underlying risk of .00001% or a “real” reduction? Should adjustments be made for “priors” within a Bayesian methodology? We could go on and on.

Even laymen with moderate understandings of statistics cannot divine clarity where the assumptions and methodology are unarticulated.

Just as data collections grow exponentially so does public access to the data. But without a plain English explanation of what the statistics mean, we risk costly and regrettable opportunities to guide our choices and the patient, family and society pay dearly for this lack of clarity.