How should importance and severity ratings be combined for item reduction in the development of health status instruments?

Patients' ratings of the severity and importance of items are often used to select items for health status instruments. The purpose of this study was to compare six different methods of combining severity-importance ratings. Two different patient groups separately rated the importance and severity o f their complaints; (i) 76 patients with upper extremity disorders rated 70 upper extremity-related questions; and (ii) 86 patients with hip arthrosis rated 22 questions relating to their hip problem. The rank ordering of the items using the six different methods in the two populations were very sim ilar (tau(bi) = 0.91 and 0.87, respectively). Furthermore, the six methods when used to choose 30 upper-extremity items shared 25 items in common and shared 9 (of 10) hip items in the second group. In conclusion, the results of item reduction were not affected by the method of creating importance-se verity ratings. J CLIN EPIDEMIOL 52;3:193-197, 1999. (C) 1999 Elsevier Scie nce Inc.