Validation of SDM-Q-Doc Questionnaire to measure shared decision-making physician's perspective in oncology practice.

OBJECTIVE: The aim of this study was to analyze the psychometric properties of the Shared Decision-Making Questionnaire-Physician version (SDM-Q-Doc) in a sample of medical oncologists who provide adjuvant treatment to patients with non-metastatic resected cancer and the correlations between the total SDM-Q-Doc score and physician satisfaction with the information provided. METHODS: Prospective, observational and multicenter study in which 32 medical oncologists and 520 patients were recruited. The psychometric properties, dimensionality, and factor structure of the SDM-Q-Doc were assessed. RESULTS: Exploratory factor analyses suggested that the most likely solution was two-dimensional, with two correlated factors: one factor regarding information and another one about treatment. Confirmatory factor analysis based on cross-validation showed that the fitted two-dimensional solution provided the best fit to the data. Reliability analyses revealed good accuracy for the derived scores, both total and sub-scale, with estimates ranging from 0.81 to 0.89. The results revealed significant correlations between the total SDM-Q-Doc score and physician satisfaction with the information provided (p < 0.01); between information sub-scale scores (factor 1) and satisfaction (p < 0.01), and between treatment sub-scale scores (factor 2) and satisfaction (p < 0.01). Medical oncologists of older age and those with more years of experience showed more interest in the patient preferences (p = 0.026 and p = 0.020, respectively). Patient age negatively correlated with SDM information (p < 0.01) and physicians appear to provide more information to young patients. CONCLUSION: SDM-Q-Doc showed good psychometric properties and could be a helpful tool that examines physician's perspective of SDM and as an indicator of quality and satisfaction in patients with cancer.

A nonlinear congeneric model for continuous item responses.

This paper presents a unidimensional item response model intended for personality and attitude items that use a continuous response format. The model's starting point is the linear congeneric model for item scores, but it takes into account the bounded nature of the item responses and assumes that their conditional distributions for a fixed trait level are truncated normal. This assumption leads to nonlinear item-trait regressions and considerably modifies some aspects of the linear model. The linear model is considered as an approximation to the modified version, and an interval is defined in which the approximation is satisfactory. Procedures for estimating the item and subject parameters are described. The applicability of the model is illustrated using real data.

Fitting the Normal-Ogive Factor Analytic Model to Scores on Tests.

This article (a) describes how McDonald's nonlinear factor analytic approach to the normal ogive curve can be used to factor analyse total test scores, (b) discusses the conditions in which this model is more appropriate than the widely used linear model, and (c) illustrates the applicability of both models using an empirical example. The rationale for the described procedure is that the test scores are simple sums of binary item responses whose item characteristic curves are adequately represented by normal ogives. The results obtained in the empirical example are meaningful and informative, and agree with the results obtained at the item level.

Implementing a Test of Underlying Normality for Censored Variables.

This article describes the implementation of a standard Pearson chi-square statistic to test the null hypothesis of bivariate normality for the latent variables in the Type 1 censored model and discussed the need for such test. The behavior of the statistic is assessed by means of a series of simulation studies and illustrated by means of an empirical example. The limitations of the test are also discussed.

Calibration of Invariant Item Parameters in a Continuous Item Response Model Using the Extended Lisrel Measurement Submodel.

This article attempts to study the utility of Sorbom's (1974) Confirmatory Factor Analysis (CFA) with mean structures model, to estimate invariant difficulty and discrimination item parameters on a personality scale. A set of items from the Social Worry Scale, belonging to the "Anxious Thoughts Inventory" (Wells, 1994) are administered using a continuous response format. The general sample is split into two groups according to gender and an invariant CFA model with mean structures is fitted simultaneously in both groups. The results suggest that both the format and the model may be of interest for personality scales.