Comparison of Physician-Based and Patient-Based Criteria for the Diagnosis of Fibromyalgia

The American College of Rheumatology (ACR) 2010 preliminary fibromyalgia diagnostic criteria require symptom ascertainment by physicians. The 2011 survey or research modified ACR criteria use only patient self-report. We compared physician-based (MD) (2010) and patient-based (PT) (2011) criteria and criteria components to determine the degree of agreement between criteria methodology.

We studied prospectively collected, previously unreported rheumatology practice data from 514 patients and 30 physicians in the ACR 2010 study. We evaluated the widespread pain index, polysymptomatic distress (PSD) scale, tender point count (TPC), and fibromyalgia diagnosis using 2010 and 2011 rules. Bland-Altman 95% limits of agreement (LOA), kappa statistic, Lin's concordance coefficient, and the area under the receiver operating curve (ROC) were used to measure agreement and discrimination.

MD and PT diagnostic agreement was substantial (83.4%, κ = 0.67). PSD scores differed slightly (12.3 MD, 12.8 PT; P = 0.213). LOA for PSD were -8.5 and 7.7, with bias of -0.42. The TPC was strongly associated with both the MD (r = 0.779) and PT PSD scales (r = 0.702).

There was good agreement in MD and PT fibromyalgia diagnosis and other measures among rheumatology patients. Low bias scores indicate consistent results for physician and patient measures, but large values for LOA indicate many widely discordant pairs. There is acceptable agreement in diagnosis and PSD for research, but insufficient agreement for clinical decisions and diagnosis. We suggest adjudication of symptom data by patients and physicians, as recommended by the 2010 ACR criteria.