Re: Treatment for acute anterior cruciate ligament tear: five year outcome of randomised trial
A brief reply to Prof. Kreiner
We are pleased that Prof. Kreiner explicitly states that his criticism of KOOS is related to "both validity and the fit to the probability functions of the Rasch model" as this clarifies that validity and Rasch model fit are two separate issues. We are also pleased by his statement that item-fit statistics cannot provide evidence of validity.
However, as much as we appreciate Prof. Kreiner's personal assurance of the invalidity of KOOS, we would have preferred evidence of actual bias, its direction and magnitude. Without such information, discussions on whether KOOS systematically understates or exaggerates treatment effects are based on shaky grounds.
Competing interests:
I am a coauthor of the criticized article
26 March 2013
Jonas Ranstam
Professor of medical statistics
Philippe Wagner, MSc
Lund University
Department of Clinical Sciences, Lund, Skane University Hospital, 221 85 Lund, Sweden
Rapid Response:
Re: Treatment for acute anterior cruciate ligament tear: five year outcome of randomised trial
A brief reply to Prof. Kreiner
We are pleased that Prof. Kreiner explicitly states that his criticism of KOOS is related to "both validity and the fit to the probability functions of the Rasch model" as this clarifies that validity and Rasch model fit are two separate issues. We are also pleased by his statement that item-fit statistics cannot provide evidence of validity.
However, as much as we appreciate Prof. Kreiner's personal assurance of the invalidity of KOOS, we would have preferred evidence of actual bias, its direction and magnitude. Without such information, discussions on whether KOOS systematically understates or exaggerates treatment effects are based on shaky grounds.
Competing interests: I am a coauthor of the criticized article