The reduced expenditures of the physicians with higher proportions of recorded mental health diagnoses may result from 2 pathways: less medical workup of unexplained medical symptoms and more efficient treatment of mental health disorders. These physicians may recognize when unexplained medical symptoms are the result of emotional distress or a mental health diagnosis. They may be less likely to order diagnostic tests, refer to specialists, or admit patients to the hospital. For example, the young patient with chest pain and a normal electrocardiogram may be recognized as having panic disorder and not hospitalized.
Also, the higher proportion of recorded mental health diagnoses likely results in more efficient treatment for these patients. Studies have demonstrated that the effective treatment of depression is associated with improved health status and reduced health care use and costs.21,22 Mental health diagnoses, particularly depression, are associated with poor outcomes from chronic physical disorders, and effective treatment of depression is likely to improve the medical outcome of these illnesses.
In our study, the physicians with higher proportions of recorded mental health diagnoses had lower diagnostic testing expenses, suggesting that some of their patients may have received less medical workup. The lower inpatient expenses may have resulted from avoidance of hospitalization of patients with unexplained symptoms that are attributable to a mental health disorder or to the improved health of patients whose mental health problems were treated. The costs associated with visits to a primary care physician were the same in all groups and may be the least likely to be affected by mental health diagnoses. Any reduction in outpatient medical visits for physical symptoms may be offset by visits to treat mental health problems.
Previous studies reveal little about the impact of physicians’ recording of mental health diagnoses on health care costs. A few studies have suggested that appropriate identification and treatment of somatization and mental health disorders are cost-effective and may reduce medical costs. Smith and coworkers have demonstrated that psychiatric consultation23,24 and group therapy25 for patients with somatization disorders can improve medical outcomes and substantially reduce health care costs. Two randomized controlled trials demonstrated that the treatment of depression in primary care is cost-effective, and patients who receive adequate antidepressant treatment have lower overall medical costs.13,26 A technically oriented physician practice style has been associated with higher medical costs.27
We found that physicians in the highest quartile of recorded mental health diagnoses may be less likely to have their patients admitted for an avoidable hospitalization condition, but the effect was not statistically significant. This finding suggests that the medical care of these physicians for these conditions was at least as appropriate as more biomedically focused physicians. If the physicians with higher rates of mental health diagnoses were paying inadequate attention to biomedical problems, higher rates of avoidable hospitalizations might be expected. In a related study, Bertakis and colleagues28 found that a practice style that emphasized the psychosocial aspects of care (including discussing the current emotional state of the patient) was predictive of improvement in the patient’s health status after 1 year.
It is important to note that even the physicians in the highest quartile of proportion of recorded mental health diagnoses made the diagnosis of a mental health disorder in only 9% of their patients and were probably underrecording mental health diagnoses. Epidemiologic studies suggest that 25% or more of primary care patients have a diagnosable mental health disorder.2,3 This is consistent with other studies that have shown that primary care physicians underdiagnose and undertreat mental health disorders but tend to identify and treat patients with more severe mental health disorders that are associated with significant functional impairments.29,30
Limitations
There are several limitations to our study. This is an ecologic study in which it is not possible to link the mental health diagnoses in particular patients with lower health care costs. We cannot determine whether the lower health care costs occur only in patients who received a mental health diagnosis or occur for all patients of physicians in the highest quartile. It is possible that these findings are the result of a physician trait or style, such as patient-centeredness, that is associated with both higher proportions of mental health diagnoses and lower health care costs.
It is also possible that physicians with higher proportions of recorded mental health diagnoses had more patients with more mental health problems but overall lower severity of illness. Although the analysis of health care expenditures is adjusted for case mix using the ambulatory care groups (ACG) system, this adjustment may not have adequately controlled for the severity of illness in the panels of the physicians with lower proportions of mental health diagnoses. However, the physicians with the highest proportion of recorded mental health diagnoses also had the highest expected expenditures, a measure of the illness severity of their patients. Thus, inadequate adjustment for case-mix severity should result in higher expenditures for the physicians with higher proportions of recorded mental health diagnoses. Conversely, the ACG system may have resulted in overadjustment for case mix: Physicians with a greater proportion of recorded mental health diagnoses may also have coded the other problems with greater severity, making their patients appear more ill.