November 2005
Reviewer: Jeff C. Huffman, MD
Somatization increases medical utilization and costs independent of psychiatric and medical comorbidity
Summary: Primary care patients with high scores on somatization questionnaires have significantly high rates of health care utilization, and such increases in health care costs appear to be relatively independent of comorbid mood disorders, anxiety disorders, or medical illness. This suggests that somatization itself, rather than somatic sympoms associated with depression or medical illness, is a syndrome that drives maladaptive use of health care resources.
Background/Methods: Patients with high rates of unexplained physical symptoms are difficult patients for primary care physicians and mental health specialists alike. Such patients return frequently to their general medical physicians and receive multiple negative workups for (often) a pattern of diffuse and somewhat vague symptoms. This utilization of health care does not seem to reduce their symptoms or their distress, and such patients (despite frequently having comorbid mood or anxiety disorders) often decline mental health evaluation and treatment. In this study, questionnaires were distributed to consecutive patients on random days at two primary care clinics affiliated with an academic teaching hospital. These included measures of somatization (Somatic Symptom Inventory and somatiform disorder module of the Patient Health Questionnaire; PHQ) and anxiety/depression (also using PHQ), role impairment, and medical morbidity. Approximately 2600 patients were approached with approximately 1500 questionnaires returned and completed in a valid manner. Utilization of health care was measured in the preceding 12 months via the hospital's computerized database.
Results: Approximately 20% of patients were given provisional diagnoses of somatization disorder based on their responses. These patients had high rates of comorbid mood and anxiety disorders (58% with MDD/anxiety disorder vs. 14% of non-somatization disorder patients with MDD/anxiety disorder). Somatizers had significantly higher rates of hospitalization, emergency room visits, and inpatient/outpatient costs (overall costs nearly double); of note the only form of health care not elevated in the somatization group was utilization of mental health care. Patients with somatization and MDD/anxiety disorder had significantly more utilization in all domains than non-somatizers with MDD/anxiety disorder. Somatization contributed significantly more to variance in utilization than did mood or anxiety disorders.
Comment: The strengths of this study are its large sample size, multisite nature, and use of standardized instruments. It does suffer from potential sampling bias (located in academic hospital rather than community clinic; further, ~40% of participants did not return valid questionnaire), as well as the use of self-report questionnaire to diagnose somatization with no individual medical evaluation to rule out medical causes of somatic symptoms.
The study contains both surprising and unsurprising results. Unsurprising is the finding that somatization in primary care leads to significant utilization of health care services and is frequently associated with mood and anxiety disorders. However, the relatively small impact of mood and anxiety disorders—and higher impact of somatization alone—on health care utilization is surprising and somewhat discouraging. It could have been hoped that much of the distress and utilization associated with somatization might be to some degree an epiphenomenon related to mood and anxiety disorders, and that treatment of these disorders would then reduce or erase the associated somatization/utilization. However, this study suggests that somatization itself is a syndrome that independently drives utilization and that while treatment of comorbid mood/anxiety in the roughly 60% of patients in whom this occurs is important, such treatment may not erase the somatization and cannot explain the profound impact of somatization in the 40% without comorbid mood or anxiety disorders. This gels with our clinical experience in psychosomatic medicine that somatization disorder is indeed a distinct phenomenon; unfortunately, effective treatments for somatization have generally been more elusive than for mood or anxiety disorders.
Overall, the study suggests that somatization is a distinct phenomenon that drives health care utilization in a manner at least moderately independent of mood and anxiety disorders, and emphasizes a significant need to develop more effective approaches to and treatments for somatization.