Confronting the Geriatrician's Nightmare

Imagine that you are a cardiologist and you read a scientific paper that says that cardiologists have no better outcomes in managing heart failure than other physicians. After challenging the study's methods, denying that the finding could be valid, and probably being very angry for a while, you might begin to have doubts about your discipline's ability to do what it is supposed to do. The study published by Phelan et al.1 could be expected to have the same effect on geriatricians (at least one of them) whose raison d'être is to prevent functional decline in older persons. Any way you parse the results, geriatricians, physicians with considerable experience in caring for older persons, and those who appeared to practice according to geriatric principles fared no better than other physicians at preventing decline in activities of daily living (ADLs). So how can these findings be reconciled with the strong conviction held by most geriatricians that they are actually making a difference in the functioning of their patients? First, the study's methodology needs to be examined closely to determine whether the findings are valid. Indeed, the study's observational design and several aspects of implementation raise methodological concerns. In addition to the substantial attrition of the study sample (approximately 28%) over the 4 years of follow-up, one must question whether the almost 30% of patients who changed primary care physicians during the 2 years after enrollment in the study should be included in the assessment of study outcomes. There were only a dozen geriatricians providing care to study participants, and none had been formally trained. The baseline rate of ADL disability was very high, with a peculiar pattern (transferring difficulty occurring at more than three times the rate of difficulty in bathing or dressing). The number of events (incident disability), particularly at 2 years, was relatively small. As a result, the power of the study was limited. In fact, the adjusted odds ratio (0.72) for 2-year risk of incident ADL disability in participants cared for by a certified geriatrician compared with those receiving care from a physician who was not certified might be considered impressive, although it was not statistically significant. Finally, the process measures of care for geriatric syndromes were crude (i.e., whether the provider acted in any way to address or manage), and better process measures2 might have yielded different results. To their credit, the authors recognize many of these issues. Despite these methodological questions, we are still confronted with the possibility that these findings are the truth. What then? I propose that we begin by examining three aspects of geriatricians' practices: who they see, what they do, what they do not do. Older persons with complicated biopsychosocial problems gravitate to geriatricians, a bias that Phelan et al. refer to as confounding by indication. In an observational study, attempts can be made to adjust for this bias in the analysis, but it cannot be eliminated. Frequently, geriatricians care for patients who are on a relentless path toward functional decline and death because current therapy is not powerful enough to change the course of their illnesses or because patients have chosen not to fight the progression of functional decline (e.g., by selecting hospice care). Thus, it is likely that the geriatricians in this study faced a more difficult challenge in attempting to prevent functional decline than did the comparison physicians. Moreover, a fundamental question must be raised as to how sensitive ADL function is to provider practice. The best evidence supporting mutability of ADLs by provider practice has been in trials conducted in hospital settings using comprehensive geriatric assessment,3 acute care of elders units,4 and Department of Veterans Affairs geriatric evaluation and management units.5 Changing provider practice in outpatient settings (e.g., through comprehensive geriatric assessment6 and geriatric team management7) has failed to affect patients' ADL functioning, although functional decline measured using more sensitive instruments has been delayed. So it is possible that geriatric care provided in this study may have affected function that was not captured in the outcome measure that was selected for this study. As mentioned, geriatricians commonly manage complicated patients. As medical, psychological, and social problems increase, fewer strategies are available that provide benefit without the risk of incurring harm. Consider the patient whose symptomatic heart failure requires diuretic therapy that aggravates urinary incontinence. Threading the narrow needle of optimal therapy is difficult, and choices must be made as to which toxicity is the lesser evil. Geriatricians negotiate with patients and obtain their preferences, which may lead to individual outcomes that differ from those that are desired for a population. A frail patient may prefer using a wheelchair rather than physical therapy with daily practice at home for months to regain mobility. In other words, sometimes some important clinical outcomes, such as functional status, must be sacrificed to preserve others, such as a quality of life that avoids any discomfort. Although formal studies have not been conducted comparing geriatricians' management of older patients' medical conditions, there is no evidence to suggest that geriatricians do any better job at managing some of the key medical contributors to functional decline (e.g., heart failure, stroke, diabetes mellitus, chronic obstructive pulmonary disease) than other physicians. In fact, it could be argued that it is more difficult for geriatricians to provide high quality of care for these conditions because of conflicting guidelines for coexisting diseases, drug-disease and drug-drug interactions and the lack of appropriateness of some quality processes8 for those who have multiple comorbidities. The absence of healthcare delivery systems that would make it easy to do it right compounds these obstacles.9 If the generalist who is attempting to manage a patient with one or two chronic diseases is barely staying afloat, the geriatrician is likely to be drowning in his or her attempts to provide high-quality care for a half dozen conditions. I suspect that a careful audit of most geriatricians' records would find gaping holes in the quality of care for many medical conditions. In the end, the readers of the Journal should keep the findings of this study in perspective. It is only one study, and at least one other study10 with a similar design presents a more optimistic view. In that study, physicians with geriatric credentials were better at recognizing cognitive impairment.10 Nevertheless, observational studies are limited. Note the rise and fall of combined hormone therapy for the prevention of heart disease based on observational studies and clinical trials, respectively.11–13 Although no discipline has scrutinized its value more than geriatrics, it is unlikely that a trial comparing geriatricians with other physicians will (or should) ever be mounted. Nevertheless, Phelan et al. remind us that certification, experience, and practicing a few of the principles of geriatric care may not be enough to stem the tide of functional decline. While recognizing some unalterable factors and patient preferences as being important contributors to the progression of functional decline, it is critical to focus on improving what we can. In addition to seeking new and better treatments for the disorders that affect older persons, it is critically important to improve the quality of care that geriatricians provide to patients, not only for geriatric syndromes but also for all conditions that affect older patients. The gap between knowledge and practice must be closed. Doing so will require systems-level interventions and practice redesign. As noted by the Institute of Medicine, “simply trying harder will not work.”14 Accordingly, geriatricians must take the lead in devising systems that insure that a major mutable contributor to functional decline, quality of care, does not fall through the cracks.