An Easy Way for Clinical Validation of the Pharmacoeconomic Model in Dupuytren Disease

Fig. 37.1

Model 1 (simple model)

Fig. 37.2

Model 2 (including details on complications)

37.3 Results

The questionnaire to assess the models was answered by 27 orthopedic surgeons. Only five residents responded to the survey, and the rest of the surgeons had an average length of practice as specialists of 10.1 (minimum: 0; maximum: 30) years (Fig. 37.3).

Fig. 37.3

Level of care of the hospital of the orthopedic surgeons

The total score obtained was 35.49 (CI 95 %: 32.33–38.64) for model 1 and 38.72 (CI 95 %: 35.78–41.65) for model 2 (the difference was not statistically significant). Upon analyzing the pattern of the responses for the items, no floor or ceiling effects were observed. The standard error of measurement (systematic and random error of a participant’s score that is not attributable to true changes in the construct to be measured (Mokkink et al. 2010)) was calculated at 0.796, which is 8.0 % with respect to the global for the scale. The minimal detectable change (magnitude of change required to be 95 % confident that the observed change between the 2 measures reflects real change and not just measurement error (Haley and Fragala-Pinkham 2006)) was 2.21 (Fig. 37.4).

Fig. 37.4

Mean score

Corrected item-total correlation for the two pharmacoeconomic models shows a linear relationship between structural simplicity and comprehensibility. The total Cronbach’s alpha was 0.803 for model 1 and 0.805 for model 2.

Globally, there were no significant differences between the scores given to both models. No different results were found relative to the level of care at the hospitals to which the surgeons belonged or their experience as practitioners. The internal consistency of the scores was moderately high; Cronbach’s alpha based on standardized items was 0.741. The intraclass correlation coefficient (average measures) to assess the agreement between the two models was 0.614, and for consistency, it was 0.727, with both being statistically significant (P < 0.05).

37.4 Discussion

We have used two statistical analyses: a descriptive analysis applied to the results obtained from the questionnaire and a factorial analysis for the assessment of the tool’s reliability index through calculation of Cronbach’s alpha (Cronbach 1951). Two decision-making trees were used, since they explain acute surgical and medical events, such as DC, in a better way than other methods (such as Markov models). Furthermore, these provide the advantage of maximum flexibility in the design, as well as a greater interpretability by clinicians (Carrera-Hueso and Ramon 2011). Our results do not demonstrate the superiority of one pharmacoeconomic model over the other in terms of the total scores obtained from the survey. However, there are significant differences in the scores awarded to both on different scales. We found no differences in scores in terms of the characteristics of the physicians surveyed or with regard to their experience or the level of healthcare of the hospital. Model 1, the simpler one, had better scores for structural simplicity and comprehensibility. This model has been used in other pharmacoeconomic studies published regarding DC (Chen et al. 2011; Baltzer and Binhammer 2013). Model 2 had better scores for adaptability and reliability, apparently because it better reflects actual clinical situations.

Only gold members can continue reading. Log In or Register to continue