Guarding Against Overtesting, Overdiagnosis, and Overtreatment of Older Adults: Thinking Beyond Imaging and Injuries to Weigh Harms and Benefits

EDITORIAL Guarding Against Overtesting, Overdiagnosis, and Overtreatment of Older Adults: Thinking Beyond Imaging and Injuries to Weigh Harms and Benefits M ost of us assume that “knowledge is power” is so obviously axiomatic that it has to be true. But right at the start of “the information age,” more than 20 years ago and thus well before our current era of big data, Ken Ringel wrote a prescient warning in the lay press about unintended negative consequences likely to derive from the wholesale acquisition of data . . . the meaning of which we do not always understand. 1 Ringle furthermore stressed the importance of distinguishing between data (a collection of isolated facts), information (recognition of the pattern that such data implies), knowledge (an understanding of what that information means), and wisdom (knowing how to apply knowledge in a way that improves outcomes). In this issue of the Journal of the American Geriatrics Society, Jawa and colleagues present their evaluation of the information contained in a large dataset in which they found that, of the subset of older adults hospitalized after a ground-level fall, many had a second spinal injury found on advanced imaging. They go on to speculate about missed injuries that may have been present in individuals who did not have equivalent imaging and ultimately sug- gest a need for greatly enhanced diligence with all older adults who present after a fall. 2 We have three areas of concern with regard to this anal- ysis. The first, and least important, has to do with the “in- formation” they believe they were able to glean from their data. Both the reliability and accuracy of large data sets like theirs have been appropriately challenged, 3,4 and the fact that this particular data set “has previously been used for research” should provide no reassurance in this regard. Fur- thermore, findings in admitted patients who underwent extensive imaging—ostensibly because of a clinical indica- tion in at least some of them—should never be generalized to individuals in whom providers felt no need to pursue such imaging (not to mention those who were evaluated but had nothing felt to require hospitalization). Finding additional fractures in the former group may not mean that older adults routinely harbor occult fractures after a ground-level fall but merely that clinicians appropriately evaluated the subgroup of those who had relevant symptoms. Still, we are willing to believe that the authors’ conclu- sions about what they found are qualitatively, if not pre- cisely quantitatively, accurate; they seem reasonable and are consistent with results of many studies in other groups of individuals with spinal injury. 5–8 If young, healthy DOI: 10.1111/jgs.14737 JAGS 2017 © 2017, Copyright the Authors Journal compilation © 2017, The American Geriatrics Society individuals with a primary spinal fracture often harbor a second, less clinically obvious fracture, it only makes sense that the same might be true in older, frail individuals. A much larger concern relates to the “knowledge” that supposedly derives from these findings, in particular the assumed benefit of seeking and finding every occult fracture. It is critically important to distinguish between disease-oriented outcomes (DOOs), which are typically “objective” findings that are relatively easy to measure, and person-oriented outcomes (POOs), which are effects that people actually experience. The former are often assumed to be surrogate markers for the latter, and fur- thermore, “fixing” the former is often believed to be important to prevent, or treat, the latter. Nevertheless, although doctors can follow an individual’s glycosylated hemoglobin, and even “treat” it, it is far more important to know whether doing so changes the likelihood that the individuals will experience preventable diabetes mellitus– related morbidity—such as crushing chest pain, nausea and vomiting due to kidney failure, or inability to speak and move one side of the body—or even to die (the ulti- mate POO). Challenges to traditional received wisdom about the value of following—no less attempting to nor- malize—many DOOs—bone density, cholesterol, blood pressure in the face of intracranial hemorrhage, prostate- specific antigen levels, among many others—are increas- ingly recognized as valid and important. 9–12 Similarly, there are calls to stop using the word “cancer” to describe cells that are unlikely to produce any clinical harm, even though they are in fact “neoplastic” under the micro- scope. 13 Thus it is important to distinguish finding a fracture from finding a fracture that will lead to some intervention that would improve some POO. It seems highly likely that the physicians who treated the individuals included in this study did something about at least some of those extra fractures that were discovered, but surely this by no means implies that most (or even any) of the individuals so trea- ted benefited from such action. Geriatricians know better than anyone that many spinal fractures in elderly adults are asymptomatic, and not only would routine interven- tion not lead to benefit, it would almost certainly produce substantial (and avoidable) harm for many of these people. 14–16 This leads to our most important concern, about whether and to what extent we can derive any degree of