In this issue of JAMA Internal Medicine, West et al1 report results of a randomized clinical trial of an intervention to improve physicians’ psychological well-being. The intervention—a series of small-group discussions—showed success in 2 indicators of distress. But the results also constitute a warning about the growing demoralization of physicians.

The participants were 74 internal medicine physicians at the Mayo Clinic. West et al1 studied whether facilitated discussions in small groups, related to physician well-being and work experience, could reduce burnout and influence other measures of psychological health. (Burnout, characterized by a loss of enthusiasm for work, cynicism, and feelings of low accomplishment, is reported to affect almost half of US physicians.2) The physicians from both the intervention and control arms received 1 hour of paid time every 2 weeks, taken from clinical activities, to use for the discussions—or, in the control group, for any purpose they chose. The intervention lasted 9 months, with 1 year of follow-up.

Compared with the control group, the intervention group showed a substantial and sustained reduction in depersonalization (one of the subscales assessing burnout), which is manifested as a sense of alienation from patients, and better scores on the Empowerment at Work Scale, which measures a sense of control, participation, and meaning. No significant difference was observed in the other measures tested.

Why was the intervention successful, albeit narrowly? There is some evidence that fostering reflection and self-awareness can improve physicians’ sense of well-being.3 In addition, collegiality is associated with physicians’ professional satisfaction, and these small groups may have created an important sense of community.4 In any case, the small-group sessions appear to have been effective in reconnecting many physicians to their patients, and to the worth of their work.

It is interesting to consider the differences between the measures of well-being that did improve compared with the control arm, and the measures that did not. Depersonalization was ascertained by responses to statements such as, “I’ve become more callous toward people since I took this job,” and “I feel I treat some recipients as if they were impersonal objects.” The Empowerment at Work Scale consists of statements such as, “The work I do is meaningful to me.” Both measures could be construed as representing fairly downstream effects of the work environment. In contrast, measures such as emotional exhaustion (which did not differ between the study groups), assessed with statements such as “I feel I’m working too hard on my job,” may more directly represent structural factors such as high patient volume or long hours. If so, small-group discussions may be most useful in modulating physicians’ adaptive responses to stress rather than the stress itself. These downstream effects are important: a sense of alienation from patients and lack of meaning are surely destructive to good care.

But the shadow of the upstream stress haunts this study, mostly in the findings in the control group and the nonparticipants. The study was timed with annual departmental surveys so that on 2 occasions, 4 measures of well-being—meaning at work, depersonalization, emotional exhaustion, and quality of life—were determined in both study participants (intervention and control groups) and nonparticipants. The control group did much better than the nonparticipants, suggesting that benefit was derived just from having paid protected time away from clinical responsibilities. Can it be that faculty are so overloaded that taking 1 hour every 2 weeks has such an effect?

Of more concern, the 350 physicians who did not participate in the study but responded to annual departmental surveys appear to have had an alarming decrease in all 4 measures of well-being in just 1 year. The exact numbers are not given, but Figure 2 suggests a reduction in finding meaning in work and large increases in emotional exhaustion, depersonalization, and overall burnout; reporting poor quality of life also rose. In addition, when the same authors surveyed Mayo internal medicine faculty members for another study just 3 years earlier, 34% met criteria for burnout5—far less than the 47% at this study’s baseline.

It is apparent that faculty physicians’ morale is under assault, and there is no reason to think the situation is worse at the Mayo Clinic than elsewhere (and some reason to think it might be better). How can this be explained? The years during which this study took place saw an acceleration in trends affecting hospitals: (1) Hospitals have increasingly emphasized clinical productivity as a means to stay afloat financially. According to annual reports, the Mayo Clinic increased inpatient and outpatient volume during the study. (2) Resident duty-hour restrictions have shifted clinical workload to faculty at the expense of teaching and research,6 with duty-hour restrictions tightened in July 2011. (3) Obtaining funding for research has become more competitive as federal support dwindles, and shrinkage was at its highest during this study. (4) Quality of care has come under increasing scrutiny, and physicians and hospitals are subject to a growing array of externally imposed standards and reporting requirements. The Hospital Value-Based Purchasing Program, a large pay-for-performance program from the Centers for Medicare & Medicaid Services, began performance evaluation during the study. (5) As hospitals attempt to recover from the economic crisis of 2008 and from public payment shortfalls, physicians are increasingly asked to prioritize the financial health of their organizations, contributing to the phenomenon of the “physician as double agent.”7

These factors combine to create a higher clinical workload, threaten teaching and research activities, increase oversight and reporting requirements, and dilute physician advocacy for patients, all of which could certainly affect physician well-being.

Most important, I suspect that physician well-being is a powerful correlate of patient well-being and attitudes. There are many reasons to become a physician, but for most, the first is that we want to help people. Physician satisfaction is associated with the perception of being able to provide high-quality care—and there is a strong correlation between physician and patient satisfaction.8 This is almost always explained as a causal relationship in one direction: more satisfied physicians provide better care and, therefore, have more satisfied patients. But I am certain that when factors beyond the control of either physicians or patients reduce the time and attention physicians can give to their patients, or make patients resentful or distrustful, that is profoundly demoralizing to physicians.

In my experience, physicians today feel guilty that they cannot spend enough time with each patient. They are embarrassed that short shifts and multiple providers threaten the continuity of care and that their patients often are not even sure who their physician is. They are dismayed that, based on insurance, some patients have too little medical care, while others are burdened with overtreatment and iatrogenesis. They are appalled by the financial burden their patients and their families experience. They, like their patients, have a sense that our titanic health care system no longer primarily supports the physician-patient dyad but serves myriad external interests.

The study by West et al1 demonstrates that a well-designed, institutionally supported curriculum can ameliorate those downstream symptoms of physician distress that are perhaps most detrimental to patient care. But we would do well not to forget to look upstream.