Walker D J, Stephenson A M, Griffiths I D, Roberts D F
Department of Rheumatology, University of Newcastle-upon-Tyne, England.
Genet Epidemiol. 1987;4(4):249-53. doi: 10.1002/gepi.1370040403.
A discriminant analysis has been applied to physical and laboratory data on a series of 37 patients with rheumatoid arthritis (RA). Good discrimination was found between RA when it was associated with organ-specific autoimmune phenomena and when it was not. When data from other members of the families was included there was consistently better discrimination than when using data from the patients alone. The results suggest that future studies of classification of variable diseases should include study of relatives.