Is the SF‐12 version 2 Health Survey a valid and equivalent substitute for the SF‐36 version 2 Health Survey for the Chinese?

Abstract

The scoring algorithm of the 12-item Short-Form Health Survey (SF-12) was revised in the second version (SF-12v2), but information on its psychometric properties is lacking. This study determined whether the SF-12v2 was a valid and equivalent substitute for the SF-36v2 Health Survey (version 2) for the Chinese. A total of 2410 Chinese adults in Hong Kong completed the SF-36 Health Survey by telephone. The SF-12v2 data were extracted from the SF-36 data. Internal consistency was assessed by Cronbach's alpha, and test-retest reliabilities were evaluated by intraclass correlation. Criterion validity and equivalence were assessed using the SF-36v2 scores as a gold standard. Construct validity and sensitivity were assessed by known-group comparison. Internal consistency and test-retest reliabilities were good (range 0.67-0.82) for all except three scales. The SF-12v2 summary scores explained >80% of the total variances of the SF-36v2 summary scores. Construct validity and sensitivity were confirmed by significantly lower SF-12v2 scores in people with chronic diseases than those without. Effect size differences were less than 0.3 and relative validities were greater than 0.7 between SF-12v2 and SF-36v2 scores for different groups. The SF-12v2 was valid, reliable and sensitive for the Chinese. It is an equivalent substitute for the SF-36v2 for the summary scales.