Page 113 - The value of total hip and knee arthroplasties for patients

Patients’ pre-operative general and specific outcome expectations predict postoperative pain and function after total knee and total hip arthroplasties
suggested that it is better to not dichotomize in prediction studies as continuous variables contain more information and model fit generally is better with continuous variables[51;52]. Further, because patient acceptable symptom states have not been established yet for the HOOS and KOOS measures and therefore any cut off point for the outcomes used in this study would be arbitrary. Lastly, to answer the second research question it was necessary to calculate a summary score for the pain and function expectation items of the HSS expectation surveys. However, these questionnaires were developed for the use of the individual item scores, and although in literature all items have been summed before into one total score, factor structures have not been developed officially yet. We therefore did exploratory factor analyses to derive comprehensive ‘expectations about function’ factors. As only one (THA) or two (TKA) items are about pain, we did not run a factor analyses for those items but dichotomized them to get a proper distribution of answers.
Conclusion 6 In conclusion, 1) patients’ outcome expectations were consistently part of
the combination of variables that best predicted function and pain 12 months postoperatively for both TKA and THA. However, the amount of variance explained
by the expectation measures alone was limited. 2) The CEQ expectancy subscale predicted outcomes slightly better as compared to the HSS expectation surveys, but differences in predictive value of the two measurements were too small to recommend the use of one of the two for prediction purposes.
Implications
Given the observed importance of patients’ outcome expectations, we suggest that in planning surgical treatment orthopedic surgeons should take these, in addition to a broader range of variables, into account of which the patient’s expectations about outcome of surgery is one.
Because differences in predictive value of the CEQ expectancy subscale and HSS expectations surveys measurements were very small, future studies are advised to replicate the findings and externally validate the models presented.
111