Has growth mixture modeling improved our understanding of how early change predicts psychotherapy outcome?

OBJECTIVE: Early change in psychotherapy predicts outcome. Seven studies have used growth mixture modeling [GMM; Muthén, B. (2001). Second-generation structural equation modeling with a combination of categorical and continuous latent variables: New opportunities for latent class-latent growth modeling. In L. M. Collins & A. G. Sawyers (Eds.), New methods for the analysis of change (pp. 291-322). Washington, DC: American Psychological Association] to identify patient classes based on early change but have yielded conflicting results. Here, we review the earlier studies and apply GMM to a new data set. METHOD: In a university-based training clinic, 251 patients were administered the Outcome Questionnaire-45 [Lambert, M. J., Hansen, N. B., Umphress, V., Lunnen, K., Okiishi, J., Burlingame, G., … Reisinger, C. W. (1996). Administration and scoring manual for the Outcome Questionnaire (OQ 45.2). Wilmington, DE: American Professional Credentialing Services] at each psychotherapy session. We used GMM to identify class structure based on change in the first six sessions and examined trajectories as predictors of outcome. RESULTS: The sample was best described as a single class. There was no evidence of autoregressive trends in the data. We achieved better fit to the data by permitting latent variables some degree of kurtosis, rather than to assume multivariate normality. Treatment outcome was predicted by the amount of early improvement, regardless of initial level of distress. The presence of sudden early gains or losses did not further improve outcome prediction. CONCLUSIONS: Early improvement is an easily computed, powerful predictor of psychotherapy outcome. The use of GMM to investigate the relationship between change and outcome is technically complex and computationally intensive. To date, it has not been particularly informative.

Evaluation of the Bifactor Nominal Response Model Analysis of a Health Efficacy Measure.

The bifactor nominal response item response theory (IRT) model, proposed by Cai, Yang and Hansen (2011), provides an extension of Bock's (1972, 1997) unidimensional nominal response model to multidimensional IRT. This model has not been utilized in any published studies since its original development. In this study, the model was applied to data from a sample of college students (N = 799) to evaluate the psychometric properties of a health efficacy measure. The nominal response model has the unique capability to estimate the functioning of each single response category, and higher response categories were found to have better functioning in this study. Poor-functioning categories were identified and combined into their adjacent categories. Items with revised response format showed improved functioning. The bifactor nominal response model is a useful tool for evaluation of bifactor scales with ordered while non-equivalently functioning categories.

Network or regression-based methods for disease discrimination: a comparison study.

BACKGROUND: In stark contrast to network-centric view for complex disease, regression-based methods are preferred in disease prediction, especially for epidemiologists and clinical professionals. It remains a controversy whether the network-based methods have advantageous performance than regression-based methods, and to what extent do they outperform. METHODS: Simulations under different scenarios (the input variables are independent or in network relationship) as well as an application were conducted to assess the prediction performance of four typical methods including Bayesian network, neural network, logistic regression and regression splines. RESULTS: The simulation results reveal that Bayesian network showed a better performance when the variables were in a network relationship or in a chain structure. For the special wheel network structure, logistic regression had a considerable performance compared to others. Further application on GWAS of leprosy show Bayesian network still outperforms other methods. CONCLUSION: Although regression-based methods are still popular and widely used, network-based approaches should be paid more attention, since they capture the complex relationship between variables.

A note on testing for intervention effects on binary responses.

OBJECTIVES: In some circumstances controlled trials are not feasible and treatments can only be evaluated using clinical databases. Here we consider the situation where treatment is introduced at a particular calendar time and can only be evaluated by comparison with historical controls. In these circumstances Heuer and Abel recommended using change-point methods to search for change in characteristics over the whole study period rather than simply comparing treated and untreated patients. Their recommendation is to only conclude that the intervention had an effect if a change-point could be demonstrated close in time to the introduction of the new treatment. This reduces the risk of false positives caused by confounding changes in population characteristics or changes in patient management. For binary data we develop a method that follows their philosophy and apply it to an observational study in the treatment of pin sites after orthopaedic surgery. METHODS: Tests for change in binomial probabilities based on Brownian bridge and Hansen's approximation for maximally selected chi(2) statistics are compared to an exact test by Worsley. The approximate method is generalized to logistic regression models allowing for covariates. RESULTS: The agreement of the exact and approximate method is good for sample sizes of 100 or more. The actual test size of the Hansen approximate test allowing for covariates is close to the nominal level, whereas the Brownian bridge approximation is slightly conservative. The change in pin site treatment significantly reduces the risk of infection for both adults and children. CONCLUSIONS: We consider the Hansen approximation to provide a very good and very simple method for obtaining the p-value when testing for a change in binary data event probabilities, with or without covariates.