Page 112 - The value of total hip and knee arthroplasties for patients

Chapter 6
distinguished between measurement instruments that measured the importance of expectations and measurement instruments that measured the probability that certain events would happen.This review found that measurement instruments that measured probabilities showed better associations with outcomes. It seems like the construct patients’ expectations has multiple dimensions that can be measured. In current measurement instruments different combinations of these dimensions are incorporated. Our study assessed whether variations in only one dimension of expectations (specificity of the expected outcome) accounted for differences in predictive values. Further research is needed to identify which dimensions of the construct ‘patients’ outcome expectations’ need to be included in the optimal measurement instrument.
Strengths of the study
This study has several strengths. Firstly, all questionnaires used in this study are well known validated measures that are used in research as well as clinical practice. Secondly, patients were recruited consecutively from one general hospital in the Netherlands. The latter is the setting were most TKA and THA surgeries are performed.The characteristics of our sample are not only comparable to theTHA and TKA population of the larger VESPA study) but also very similar to the overall Dutch population of TKA and THA patients in 2011 and 2012 registered in the Dutch Arthroplasty Registry [50] ensuring generalizability of our results.
We chose our candidate variables for the multivariate models based on two criteria, variables had to be associated with outcomes of TKA or THA in previous studies, furthermore they (which is recommended also by several authorities in orthopedics) had to be simple and reliable measures that are already commonly used in clinical practice.
Limitations of the study
A limitation of this study is that because of the sample size, the number of candidate variables that could be examined was limited. It may therefore be that we have missed important predictive variables. A strength of this study is the use of a continuous predictors and outcome measures. Although some may argue that for clinical practice it is more useful to use dichotomous outcomes and define cut off values for the predictors in the study several methodological studies also have
110