Risk factors of symptom underestimation by physicians

The aims of this study were to characterize patient-physician agreement on various psychological and somatic symptoms in internal medicine inpatients and to identify predictors of symptom severity underestimation by physicians.

Consecutive adult inpatients of two internal medicine wards of a university hospital completed visual analogue scales (VASs) for severity of disability, anxiety, depression, somatic symptoms, and pain at the time of admission (n=639, participation rate=70%) and 5 days thereafter (n=401, participation rate=82%). In addition, the Patient Health Questionnaire 9 (PHQ-9) depression scale and the Clinical Global Impression Scale--Revised were used. At the same time, the six treating physicians independently rated the complaints of their patients using the same VAS. Rates of overestimation, concordance, and underestimation of symptom severity were analyzed. Logistic regression analyses were used to identify predictors of symptom underestimation by physicians.

Concordance between patients and physicians regarding disability, anxiety, depression, somatic symptom severity, and pain reached 50-60%. Symptom severity of patients suffering from major depressive episode was significantly more often underestimated than that of nondepressed patients (all P<.01). Of all the variables, greater depression on the PHQ depression score was the most important risk factor for symptom underestimation by physicians (OR ranging from 1.29 to 1.57; all P<.05, except underestimation of disability). Symptom underestimation of pain severity was also associated with panic disorder symptoms (odds ratio, 2.44; P=.01).

Depressed patients seem to be at greater risk of symptom underestimation by their physicians--a finding bearing implications for avoidance of underdiagnosis and insufficient treatment. Mutual understanding could be improved by better doctor-patient communication skills. Brief self-report depression screeners might help to reliably identify patients at risk for symptom underestimation by physicians.