Objective: The 1982 American College of Rheumatology (ACR) revised criteria for the classification of systemic lupus erythematosus (SLE), updated in 1997, have become the standard for establishing eligibility of subjects for epidemiologic and clinical lupus studies. These criteria may exclude patients with limited disease, restricting the generalizability of research findings. We developed and evaluated the ability of a weighted classification system to identify a broader spectrum of patients with lupus.
Methods: We constructed the Boston Weighted Criteria system for the classification of SLE, updating that developed in 1984. Using a hospital billing database, we identified 27l patients seen in our rheumatology clinic for possible SLE and reviewed medical records for all ACR criteria and the treating rheumatologist's diagnosis. We compared both the Boston Criteria and the treating rheumatologist's diagnosis to the updated 1982 ACR criteria; we also compared the Boston Criteria to the treating rheumatologist's diagnosis.
Results: The Boston Criteria identified 190/271 patients as having SLE, the rheumatologist's diagnosis identified 179/271, and the ACR criteria identified 171/271. The Boston Criteria had a sensitivity of 93% and specificity of 69% compared to the ACR criteria, and would identify 7% more patients.
Conclusion: The Boston Criteria identify a larger number of patients compared with the current ACR criteria, while retaining face validity. This reflects the inclusion of patients with objective findings of SLE but less than 4 ACR criteria. Our Boston Criteria system could minimize selection bias and increase the generalizability of clinical SLE studies.