Growth Curve Models for Longitudinal Data: Application for Psychiatric Research Público

Chiodi, Sarah Nicole (2013)

Permanent URL: https://etd.library.emory.edu/concern/etds/m900nt951?locale=es
Published

Abstract

Background: Psychiatric researchers commonly use repeated measures ANOVAs to analyze longitudinal data with repeated measures, however growth curves provide different tests of efficacy that may be more relevant to study goals. This thesis compares these various analytical methods for establishing efficacy in the same dataset. Methods: We utilized both complete case and available case ANOVA, as well as linear mixed model growth curves and piece-wise growth curves to determine efficacy of randomly assigned treatment. Results: The results exhibited no significant group differences using ANOVA methods. In contrast, a significant difference was demonstrated in the added CBASP group in both the rate of change in phase two as well as the difference in slope between phases one and two, indicating both a more rapid decrease in symptomatology and a less significant slowing of the rate from phase one to phase two. This result was found when looking at only the second phase and when looking at both phases using a piece-wise growth curve model. Due to the CBASP and medications group being significantly different in both phases, one may believe it could be due to poor randomization rather than the efficacy of added CBASP. Conclusions: Growth curve models, when accommodated (i.e. after demonstrating simple linear relationships) provide an advantage and should be the predominantly used method on longitudinal repeated measures data. When using multiple phases in a trial, piece-wise growth curve models should be the model of choice. Keywords: Depression, Longitudinal, MMRM, repeated measures, growth curves

Table of Contents

Table of Contents

1 Introduction…………………………………………………………………………..........1

1.1 Background……………………………………………………………………………………1

2 Review of Literature……………………………………………………………………....3

2.1 The Statistical Problem………………………………….………………….………..3

2.1.1 Repeated Measures Analysis of Variance……………..………….....…3

2.1.2 Mixed Model Repeated Measures Analysis of Variance……….....5

2.1.3 Growth Curve………………………………….……………………………..........6

2.1.4 Growth Curve with Piecewise Linear Fit....………………………......7

2.2 The Clinical Problem……………………………………….….…………………...…8

2.2.1 Disease, Symptoms, Causes………………….....………………….….....8

2.2.2 Medication………………………………………………..….…………........…….9

2.2.3 Counseling……………………………….……………………..…………........…10

2.2.4 Combination Therapy....…………….……………………..…………......…11

3 Methodology…………………………….………………………………….…..........…12

3.1 Clinical Methodology……………………………………………………………....…12

3.2 Modeling……………………………………………………….…………….…………...14

4 Results……………………………………………………..………………..….........….16

5 Conclusions………………………………………………..…………………..........…18

Bibliography………………………………………………..……………………….........…20

Tables and Figures….…………………………………..…………………………........25

Appendix….………………………………………………..…………………....………......30

List of Tables

1 Phase 2 HAM-D Scores…….……….………………………........25

2 Model Results on Phase 2 Data….………………………….......27

3 Phase 1 HAM-D Scores…….…………………………………......28

4 Phase 2 HAM-D Scores (Using Combined Phase Models)....28

5 Piece-wise Growth Curve Analysis Results……………….....29

List of Figures

1 Mean Scores by Week For Each Treatment Group

(Phase 1) ………………………………………………………………..................…26

2 Mean Scores Per Week Per Treatment Group

(Both Phases)…………………………………………………………....................29


About this Master's Thesis

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
School
Department
Subfield / Discipline
Degree
Submission
Language
  • English
Research field
Palabra Clave
Committee Chair / Thesis Advisor
Committee Members
Partnering Agencies
Última modificación

Primary PDF

Supplemental Files