Page 86 - The value of total hip and knee arthroplasties for patients

Chapter 5
Discussion
This study examined whether the existing instruments for optimism, pessimism, hope, treatment credibility and treatment expectancy measure distinct psychological constructs in patients undergoing TKA or THA. Because it was not our purpose to develop new instruments or to revise the existing ones, we chose a confirmatory approach (CFA) in all our analyses instead of an exploratory approach (EFA). Moreover, we aimed to use all instruments in the same way as they are currently utilized in research and practice and therefore did not delete items with low factor loadings.
The results of the theory driven CFA showed that a five factor model in which optimism (LOT-R subscale optimism), pessimism (LOT-R subscale pessimism), hope (HS), treatment credibility (CEQ subscale credibility) and treatment expectancy (CEQ subscale expectancy) had the most optimal fit. However, there were two interesting observations. First, a strong correlation (r = 0.82) was observed between expectancy and credibility.Therefore a post-hoc analysis was performed in which a four factor model in which expectancy and credibility were forced to load on one factor was tested. Although fit indices were very similar as the five factor model, the Δχ2 test indicated the five factor model was the preferred model. Earlier studies found moderate to very high correlations between expectancy and credibility (r = 0.56 36 , r = 0.68 14 and r = 0.83 14), though exploratory as well as confirmatory factor analyses suggest expectancy and credibility are two separate factors 14;36.Although the current study cannot provide the definite answer regarding the distinctiveness of the constructs treatment expectancy and treatment credibility in patients undergoing THA or TKA, it seems reasonable that the a-priori, theory driven 5 factor model is preferred.Additionally the five factor model is also supported by the results of those earlier studies on the constructs expectancy and credibility. Nevertheless, future studies should investigate the factorial structure of the CEQ and the distinctiveness of the constructs treatment credibility and treatment expectancy to determine if our findings are replicable or unique to our study sample.
Secondly, a strong correlation was also observed between optimism and hope (r = 0.79). A priori we hypothesized that these two factors would be correlated but still distinct because both are defined as general future oriented constructs
84