Research reportPatterns of early change and their relationship to outcome and follow-up among patients with major depressive disorders
Introduction
Different types of psychotherapy – such as, e.g., cognitive behavioral therapy (CBT; Dobson, 1989, Gloaguen et al., 1998, Hollon and Beck, 2004, Scott, 2001) or interpersonal psychotherapy (IPT; de Mello et al., 2005, Reynolds et al., 1999) – as well as pharmacotherapy with several antidepressant medications (Kornstein et al., 2006, Thase et al., 2007) have been demonstrated to work for patients with major depressive disorders (MDD). However, studies comparing the efficacy of different types of psychological and pharmacological treatments for MDD often failed to find differences between treatment approaches (de Mello et al., 2005, Keller et al., 2000, Luborsky et al., 2002, Wampold et al., 1997). Also, the findings of the National Institute of Mental Health (NIMH) Treatment of Depression Collaborative Research Program (TDCRP; Elkin et al., 1989) often are interpreted as supporting the idea of treatment non-specificity for MDD (e.g., Lambert and Ogles, 2004). Whereas in this randomized controlled trial all four treatment approaches (CBT, IPT, psychopharmacological treatment with imipramine plus clinical management (IMI-CM), pill-placebo plus clinical management (PLA-CM)) resulted in a meaningful reduction of depressive symptoms until treatment termination, there were significant differences between these interventions only in a few of the outcome measures used to assess depression severity at treatment termination and follow-up (Elkin et al., 1989).
As in the NIMH TDCRP, studies which compare different types of treatment usually rely on comparisons of patient samples in terms of their average outcomes or their mean pre-to-post treatment change. The often occurring lack of differential effects then is explained mainly by nonspecific treatment factors that are common to all interventions, such as the expectation of the patient that the treatment will help him or her to overcome the presenting problems (Lambert and Ogles, 2004).
However, similar mean effects of various psychological and pharmacological treatments do not necessarily imply that there are no identifiable differences in effects on individuals or subpopulations within a sample (Cuijpers et al., 2005). Findings from comparisons of aggregated pre-to-post treatment change between patient samples, relying on the assumption of linear and steady change over the course of treatment, may mask potentially meaningful differences in individual treatment courses (Krause et al., 1998) or different patterns of change that are shared by many individual patients (Barkham et al., 1993, Stulz and Lutz, 2007). Therefore, patient-focused (psychotherapy) research is concerned with the monitoring, modeling and prediction of individual treatment progress and the feedback of this information to therapists (and patients) in a timely manner during the ongoing course of treatment (e.g., Howard et al., 1996, Lambert, 2007, Lutz, 2002). Therapists can then use this information to early identify patients at risk of treatment failure and to optimize their treatment strategy, if necessary (Lambert, 2007, Lambert et al., 2003, Lambert et al., 2001; e.g., Lueger, 1998, Lutz et al., 2006, Lutz et al., 2009, Lutz and Stulz, 2008).
For instance, random coefficient regression models (Raudenbush and Bryk, 2002) and an extended growth curve methodology that employs nearest neighbor (NN) techniques derived from alpine avalanche occurrence research (e.g., Brabec and Meister, 2001) have been used in samples of psychotherapy outpatients with repeated outcome assessments during treatment to identify client intake characteristics (e.g., treatment expectations) which allow prediction of individual treatment progress (e.g., Lutz et al., 2005, 1999). These latent growth curve modeling techniques bring new possibilities for the analysis of individual change. The growth curve approaches assume an underlying path or trajectory of change over time and typically describe change in treatment by a log-linear function of the number of sessions (Howard et al., 1996, Lutz et al., 1999). However, while original analyses were based on the implicit assumption that the relationship between predictors and treatment progress is the same in all patients (prediction weights are calculated from the whole sample), it has also been noted that particular predictors may work best for restricted subsets of patients (Krause et al., 1998). To address this problem, a nearest neighbor (NN) techniques was recently introduced to predict individual treatment courses (e.g., Lutz et al., 2005). This NN approach identifies those already-treated clients in the reference group who most closely match the target client (hence “nearest neighbors”) on intake variables. It then uses this homogeneous subgroup of already-treated patients to generate predictions for the target client. This concept has successfully been used to provide feedback early in treatment to therapists based on comparisons between the actual and the expected early response in treatment and to develop decision rules to support clinical practice (Lambert et al., 2003Lambert et al., 2001, Lutz et al., 2006, Lutz and Stulz, 2008).
Further developments to analyze the course of patient progress within the paradigm of patient-focused research also showed that the shape of early response to treatment (Haas et al., 2002, Lutz et al., 2007, Stulz et al., 2007) and clear-cut changes between two subsequent treatment sessions in the form of sudden gains (Stiles et al., 2003, Tang and DeRubeis, 1999) are associated with different final treatment outcomes. Such findings support the identification of patients at risk of treatment failure at an early stage in treatment.
In this study on shapes of early change we make use of a new technique for the analysis of longitudinal data, which allows the identification of distinct subgroups of individuals, differing in the initial level and course of an outcome variable, through the empirical identification of developmental trajectories (Muthén, 2001, Muthén and Muthén, 2000). By allowing identification of patient subgroups with different mean growth curves through the use of latent class analysis techniques, the Growth Mixture Models (GMM) go beyond approaches so far used to analyze the phenomenon of early change by not assuming that all individuals vary around one common average latent growth curve (see Methods section).
All these studies described above used patient progress information from large naturalistic datasets, where no control of the treatment protocol was conducted. So far it remains unclear if such subgroups of early change can be identified under the condition of a controlled clinical experiment, where treatments follow a standardized protocol and the heterogeneity of early change patterns might be influenced by the protocol, e.g. by less variability within specific protocols.
Following up on this question, in this study the data of the NIMH TDCRP was reanalyzed using longitudinal data analysis techniques such as the above mentioned for an in-depth exploration of the change patterns early in treatment under the conditions of a clinical experiment. Specifically, we aimed at answering the following research questions: first, is it possible to identify distinct patterns of early change of depression severity among patients with MDD? Second, are these patterns of early change related to the type of treatment protocol? Third, is there an association between these patterns of early change and outcome at termination and follow-up? And fourth, are the patterns of early change also predictive of treatment outcome at termination and follow-up when controlling for treatment conditions (CBT, IPT, IMI-CM, or PLA-CM) and symptom severity at intake?
Section snippets
Design
Details on the design and procedures of the NIMH TDCRP have already been described elsewhere (Elkin et al., 1985, Elkin et al., 1989). Therefore, we will provide only a brief overview. The TDCRP was a collaborative randomized controlled clinical trial comparing four treatments for MDD at three research sites (George Washington University, University of Pittsburgh, and University of Oklahoma). The 250 outpatients who gave informed consent were randomly assigned to one of four treatment
Patterns of early change
In the following analysis, the number of distinct patterns of early change was determined by means of GMM (Muthén, 2004, Muthén and Muthén, 2000). Starting with one latent class (i.e., with a conventional LGM), additional latent classes were entered into the GMM until the optimal number of latent classes was found under the condition of fixing parameter (co-)variances of growth to be equal across latent classes. The fit criteria to determine the optimal number of latent classes (patterns of
Discussion
The goal of this study was to identify typical patterns of early change over the first 8 weeks of treatment in depression and to use those patterns to predict final (i.e., after 16 weeks of treatment) as well as follow-up outcome (6, 12, and 18 months after treatment). These patterns were controlled for the type of treatment protocol conducted (CBT, IPT, IMI-CM, and PLA-CM), and for the severity of overall symptoms at intake. Using a GMM approach, we identified three subgroups of patients on
Role of funding source
This work was partially supported by grants from the Swiss National Science Foundation (SNF), Nr. PP001-102651 (Wolfgang Lutz). The SNF had no further role in study design, analysis and interpretation of data, in writing the manuscript and in the decision to submit the paper for publication.
Conflict of interest
All authors declare that they have no conflicts of interest.
Acknowledgements
We express our appreciation to the investigators in the Treatment of Depression Collaborative Research Program (TDCRP), especially Irene Elkin as the Coordinator, for providing access to their data set. Other leading collaborators at the National Institute of Mental Health (NIMH) were M. Tracie Shea (Associate Coordinator), John P. Docherty and Morris B. Parloff. The principal investigators and project coordinators at the three participating research sites were Stuart M. Sotsky and Davis Glass
References (63)
- et al.
Psychometric properties of the Beck Depression Inventory: twenty-five years of evaluation
Clinical Psychology Review
(1988) - et al.
Examining differential effects of psychological treatment of depressive disorder: an application of trajectory analyses
Journal of Affective Disorders
(2005) - et al.
Psychometric properties of the credibility/expectancy questionnaire
Journal of Behavior Therapy and Experimental Psychiatry
(2000) - et al.
A meta-analysis of the effects of cognitive therapy in depressed patients
Journal of Affective Disorders
(1998) - et al.
The Hopkins symptom checklist: factors derived from the HSCL-90
Journal of Affective Disorders
(1979) - et al.
The shape of change in psychotherapy: longitudinal assessment of personal problems
Journal of Consulting and Clinical Psychology
(1993) - et al.
Outcome measures of depression
- et al.
An inventory for measuring depression
Archives of General Psychology
(1961) - et al.
Cognitive therapy of depression
(1979) - et al.
A nearest-neighbor model for regional avalanche forecasting
Annals of Glaciology
(2001)
A finite mixture model of growth trajectories of adolescent alcohol use: predictors and consequences
Journal of Consulting and Clinical Psychology
A systematic review of research findings on the efficacy of interpersonal therapy for depressive disorders
European Archives of Psychiatry and Clinical Neuroscience
SCL-90-R: Administration, scoring and procedures manual - II
A meta-analysis of the efficacy of cognitive therapy for depression
Journal of Consulting and Clinical Psychology
NIMH treatment of depression collaborative research program
Archives of General Psychiatry
National institute of mental health treatment of depression collaborative research program
Archives of General Psychiatry
Clinical management-imipramine/placebo administration manual: NIMH treatment of depression collaborative research program
Pharmacology Bulletin
Some conceptual and statistical issues in analysis of longitudinal psychiatric data. Application to the NIMH treatment of depression collaborative research program dataset
Archives of General Psychiatry
Do early responders to psychotherapy maintain treatment gains?
Journal of Clinical Psychology
Development of a rating scale for primary depressive illness
British Journal of Social and Clinical Psychology
Discontinuities and cognitive changes in an exposure-based cognitive therapy for depression
Journal of Consulting and Clinical Psychology
Cognitive and cognitive behavioral therapies
The dose-effect relationship in psychotherapy
American Psychologist
Evaluation of psychotherapy. Efficacy, effectiveness and patient progress
American Psychologist
Clinical significance: a statistical approach to defining meaningful change in psychotherapy research
Journal of Consulting and Clinical Psychology
A comparison of nefazodone, the cognitive-behavioral analysis system of psychotherapy, an their combination for the treatment of chronic depression
The New England Journal of Medicine
Interpersonal Psychotherapy of Depression
Duloxetine efficacy for major depressive disorder in male vs. female patients: data from 7 randomized, double-blind, placebo-controlled trials
Journal of Clinical Psychiatry
Exploring individual change
Journal of Consulting and Clinical Psychology
What we have learned from a decade of research aimed at improving psychotherapy outcome in routine care
Psychotherapy Research
The efficacy and effectiveness of psychotherapy
Cited by (0)
- 1
Tel.: +49 651 201 2964; fax: +49 651 201 2886.