“Statistical Significance ≠ Clinical Significance”
This was from a session today at the Southern California Health IT Summit organized by Healthcare Innovation and co-promoted with SOHL. Though the message here was directed at the technologists in the room—”relying strictly on statistical models, many solutions lack clinical credibility”—it’s a sober reminder for both clinicians and technologists alike. (There indeed are a lot of nonsensical conclusions in the medical literature as well. The p-value is a near-sighted tyrant.)
A related idea is to be “technically innovative but clinically inconsequential.” I appreciate this phrase that Donald W. Light, writing in Health Affairs, uses to describe the practices of many drug companies. He talks about their metered release of “minor variations on existing drugs” to maintain market share and maximize profits. Though the changes are “technically innovative,” they often make little difference to patients. This, of course, isn’t limited to drug development.
For those of us working in healthcare, let us be driven to truly innovate, to bring about solutions that tangibly improve the lives of patients and professionals who rely on them. As beauty is in the eye of the beholder, value must be judged not in p-values and dollar signs but by the end users.
Christopher K. Lee, MPH, CPHQ is Director of Marketing at TEKEZE Solutions, Inc.