ABSTRACT

Perkins, Wright, and Dorsey (2000) illustrated the use of Rasch measurement to produce a clear, simple picture of the relation between laboratory abnormalities

in the diagnosis of gout. Ninety-six sets of medical record observations were submitted for analysis-half for patients diagnosed with gout and half for patients not diagnosed with gout. Each observation had values recorded for indicators routinely considered in the diagnosis of gout: uric acid, gender, age, diabetes, hypertension, weight, height, body surface area, nephrolithiasis, diuretics, cholesterol, triglyceride, blood urea nitrogen, and creatinine. Rasch measurement techniques were used in an attempt to establish a single dimension for diagnosing gout. Because blood chemistries in mg/dl, height, and weight are recorded as continuous variables, each of the physical science data points (X) was rescored linearly to its nearest integer code, using the conversion

Y = (X − Min. Value) / ((Max. Value − Min. Value) / 9).