New York clinical criteria for ankylosing spondylitis. A statistical evaluation

Abstract

The New York clinical criteria for ankylosing spondylitis have been evaluated. In a study of 412 subjects, clinical examination was carried out subjectively and objectively. Criterion evaluation was applied to both subjective and objective data. The following points emerged from the investigation: (a) No overall difference was found between subjective and objective results; (b) Calculated in terms of frequency of positive and false nagative tests, sensitivity, and the Youden index (sensitivity + specificity - 100), criteria were listed in the following order of value: (i) Thoraco lumbar pain, (ii) Limited chest expansion, (iii) Limited back movement. However, in terms of false positive tests and specificity, the reverse order was obtained. It was calculated that individual criteria in their present form fail to provide a satisfactory diagnostic index of ankylosing spondylitis. Thoraco lumbar pain is too sensitive and too non specific, and limited chest and back mobility are too insensitive and too specific. In order to obviate this problem it is suggested that criteria should be numerically weighted.