Repeated measures: Approaches to Analysis Peter T. Donnan Professor of Epidemiology and Biostatistics.

Slides:



Advertisements
Similar presentations
MANOVA: Multivariate Analysis of Variance
Advertisements

Independent t -test Features: One Independent Variable Two Groups, or Levels of the Independent Variable Independent Samples (Between-Groups): the two.
Departments of Medicine and Biostatistics
Factorial and Mixed Factor ANOVA and ANCOVA
Simple Repeated measures Peter T. Donnan Professor of Epidemiology and Biostatistics Statistics for Health Research.
SPSS Series 3: Repeated Measures ANOVA and MANOVA
Analysis of variance (ANOVA)-the General Linear Model (GLM)
January 7, afternoon session 1 Multi-factor ANOVA and Multiple Regression January 5-9, 2008 Beth Ayers.
MSc Applied Psychology PYM403 Research Methods Quantitative Methods I.
Clustered or Multilevel Data
PSYC512: Research Methods PSYC512: Research Methods Lecture 19 Brian P. Dyre University of Idaho.
13-1 Designing Engineering Experiments Every experiment involves a sequence of activities: Conjecture – the original hypothesis that motivates the.
13 Design and Analysis of Single-Factor Experiments:
Repeated Measures ANOVA Used when the research design contains one factor on which participants are measured more than twice (dependent, or within- groups.
Assessing Survival: Cox Proportional Hazards Model Peter T. Donnan Professor of Epidemiology and Biostatistics Statistics for Health Research.
6.1 - One Sample One Sample  Mean μ, Variance σ 2, Proportion π Two Samples Two Samples  Means, Variances, Proportions μ 1 vs. μ 2.
© Copyright 2000, Julia Hartman 1 An Interactive Tutorial for SPSS 10.0 for Windows © Analysis of Covariance (GLM Approach) by Julia Hartman Next.
Professor of Epidemiology and Biostatistics
Practical statistics for Neuroscience miniprojects Steven Kiddle Slides & data :
ANCOVA Lecture 9 Andrew Ainsworth. What is ANCOVA?
LEARNING PROGRAMME Hypothesis testing Intermediate Training in Quantitative Analysis Bangkok November 2007.
Lab 2: repeated measures ANOVA 1. Inferior parietal involvement in long term memory There is a hypothesis that different brain regions are recruited during.
Biostatistics Case Studies 2015 Youngju Pak, PhD. Biostatistician Session 4: Regression Models and Multivariate Analyses.
Assessing Survival: Cox Proportional Hazards Model
TAUCHI – Tampere Unit for Computer-Human Interaction ERIT 2015: Data analysis and interpretation (1 & 2) Hanna Venesvirta Tampere Unit for Computer-Human.
Practical Missing Data Analysis in SPSS (v17 onwards) Peter T. Donnan Professor of Epidemiology and Biostatistics.
CPSY 501: Class 8, Oct. 26 Review & questions from last class; ANCOVA; correction note for Field; … Intro to Factorial ANOVA Doing Factorial ANOVA in SPSS.
Biostatistics Case Studies 2007 Peter D. Christenson Biostatistician Session 3: Incomplete Data in Longitudinal Studies.
Biostatistics Case Studies 2008 Peter D. Christenson Biostatistician Session 5: Choices for Longitudinal Data Analysis.
Simple Repeated measures Peter T. Donnan Professor of Epidemiology and Biostatistics Statistics for Health Research.
Introduction to sample size and power calculations Afshin Ostovar Bushehr University of Medical Sciences.
Statistical Inference for more than two groups Peter T. Donnan Professor of Epidemiology and Biostatistics Statistics for Health Research.
Analysis of Variance 1 Dr. Mohammed Alahmed Ph.D. in BioStatistics (011)
Mixed-Design ANOVA 5 Nov 2010 CPSY501 Dr. Sean Ho Trinity Western University Please download: treatment5.sav.
Analysis of Variance (ANOVA) Brian Healy, PhD BIO203.
ANOVA: Analysis of Variance.
PSY2004 Research Methods PSY2005 Applied Research Methods Week Six.
Entering Multidimensional Space: Multiple Regression Peter T. Donnan Professor of Epidemiology and Biostatistics Statistics for Health Research.
Three Statistical Issues (1) Observational Study (2) Multiple Comparisons (3) Censoring Definitions.
EDCI 696 Dr. D. Brown Presented by: Kim Bassa. Targeted Topics Analysis of dependent variables and different types of data Selecting the appropriate statistic.
Analysis of Covariance Combines linear regression and ANOVA Can be used to compare g treatments, after controlling for quantitative factor believed to.
Analysis of covariance. When… ANCOVA is an ‘extra’ on an ANOVA ANOVA outcome = number of words learned IV = Sex ANCOVA adds a covariate covariate = size.
Smoking Data The investigation was based on examining the effectiveness of smoking cessation programs among heavy smokers who are also recovering alcoholics.
ANCOVA. What is Analysis of Covariance? When you think of Ancova, you should think of sequential regression, because really that’s all it is Covariate(s)
1 Mohamed Alosh, Ph.D. Kathleen Fritsch, Ph.D. Shiowjen Lee, Ph.D. DBIII, OB, CDER, FDA Efficacy Evaluation in Acne Clinical Trials.
Stats Lunch: Day 8 Repeated-Measures ANOVA and Analyzing Trends (It’s Hot)
Overview and Common Pitfalls in Statistics and How to Avoid Them
ONE-WAY BETWEEN-GROUPS ANOVA Psyc 301-SPSS Spring 2014.
Sample Size Determination
Biostatistics Case Studies 2006 Peter D. Christenson Biostatistician Session 3: An Alternative to Last-Observation-Carried-Forward:
Handout Twelve: Design & Analysis of Covariance
One-way ANOVA Example Analysis of Variance Hypotheses Model & Assumptions Analysis of Variance Multiple Comparisons Checking Assumptions.
Tutorial I: Missing Value Analysis
Session 6: Other Analysis Issues In this session, we consider various analysis issues that occur in practice: Incomplete Data: –Subjects drop-out, do not.
Handout Ten: Mixed Design Analysis of Variance EPSE 592 Experimental Designs and Analysis in Educational Research Instructor: Dr. Amery Wu Handout Ten:
1 Bandit Thinkhamrop, PhD.(Statistics) Dept. of Biostatistics & Demography Faculty of Public Health Khon Kaen University Overview and Common Pitfalls in.
ANCOVA.
Biostatistics Case Studies Peter D. Christenson Biostatistician Session 3: Missing Data in Longitudinal Studies.
ANOVA EDL 714, Fall Analysis of variance  ANOVA  An omninbus procedure that performs the same task as running multiple t-tests between all groups.
Analysis of variance Tron Anders Moger
Differences Among Groups
Mixed-Design ANOVA 13 Nov 2009 CPSY501 Dr. Sean Ho Trinity Western University Please download: treatment5.sav.
Analysis of Variance (ANOVA) Scott Harris October 2009.
Repeated measures: Approaches to Analysis
Don't Sweat the Simple Stuff (But it's not all Simple Stuff)
Sample Size Determination
2nd Level Analysis Methods for Dummies 2010/11 - 2nd Feb 2011
Statistical Inference for more than two groups
An Introductory Tutorial
Analysis of covariance
Presentation transcript:

Repeated measures: Approaches to Analysis Peter T. Donnan Professor of Epidemiology and Biostatistics

Objectives of session Understand what is meant by repeated measures Understand what is meant by repeated measures Be able to set out data in required format Be able to set out data in required format Carry out mixed model analyses with continuous outcome in SPSS Carry out mixed model analyses with continuous outcome in SPSS Interpret the output Interpret the output

Repeated Measures Repeated Measures arise when: In trials where baseline and several measurement of primary outcome In trials where baseline and several measurement of primary outcome Example - Trial of Chronic Rhinosinusitis Example - Trial of Chronic Rhinosinusitis Treatment usual care vs 2 weeks oral steroids Treatment usual care vs 2 weeks oral steroids Measurements at 0, 2, 10, 28 weeks Measurements at 0, 2, 10, 28 weeks

General Principles Battery of methods to analyse Repeated Measures: Repeated use of significance testing at multiple time points Repeated use of significance testing at multiple time points ANOVA - ANOVA - ‘a dangerously wrong method’ - David Finney MANOVA MANOVA Multi-level models / mixed models Multi-level models / mixed models

Significance testing at all time points Probably most common – multiple t-tests Probably most common – multiple t-tests Least valid! Least valid! Sometimes account for multiple testing by adjusting p-value i.e. 0.05/k with k tests Sometimes account for multiple testing by adjusting p-value i.e. 0.05/k with k tests Assumes that aim of study is to show significant difference at every time point Assumes that aim of study is to show significant difference at every time point Most studies aim to show OVERALL difference between treatments and /or reaching therapeutic target quicker Most studies aim to show OVERALL difference between treatments and /or reaching therapeutic target quicker PRIMARY HYPOTHESIS IS GLOBAL PRIMARY HYPOTHESIS IS GLOBAL

Repeated Measures: Summary Measures Post treatment means Post treatment means Mean change (post – baseline) Mean change (post – baseline) ANCOVA or Multiple regression account for baseline as covariate ANCOVA or Multiple regression account for baseline as covariate Slope of change Slope of change Maximum value – with multiple endpoints select highest value and compare across treatments Maximum value – with multiple endpoints select highest value and compare across treatments Area under the curve – difference Area under the curve – difference Time to reach a target or peak Time to reach a target or peak

Type of Analyses – Compare Slopes Compare slopes which summarise change Activity Baseline3-months Difference in slopes as summary measure e.g. β 1 -β 2 Advice only Pedometer Controls β1β1 β2β2 β3β3

Type of Analyses – Area under the curve Activity Baseline3-months Difference in Area between treatment slopes as summary measure Advice only Pedometer Controls 6-months

Simple approach Basically just an extension of analysis of variance (ANOVA) Basically just an extension of analysis of variance (ANOVA) Pairing or matching of measurements on same unit needs to be taken into account Pairing or matching of measurements on same unit needs to be taken into account Method is General Linear Model for continuous measures and adjusts tests for correlation Method is General Linear Model for continuous measures and adjusts tests for correlation

Simple approach But simple approach can only use COMPLETE CASE analysis where say wk 0 50, wk 2 47, wk10 36, wk But simple approach can only use COMPLETE CASE analysis where say wk 0 50, wk 2 47, wk10 36, wk Then analysis is on 30 Then analysis is on 30 Assumes data is MCAR Assumes data is MCAR Better approach is MIXED MODEL which only assumes MAR and uses all data Better approach is MIXED MODEL which only assumes MAR and uses all data

Organisation of data (Simple Approach) Generally each unit in one row and repeated measures in separate columns Unit Score 1Score2Score ….

Repeated Measures in SPSS: Set factor and number of levels Within subject factor Within subject factor levels Within subject factor name

Repeated Measures in SPSS: Enter columns of repeated measures Use arrow to enter each repeated measure column Between subject factor column

Repeated Measures in SPSS: Select options Use arrow to select display of means and Bonferroni corrected comparisons Select other options

Select a plot of means of each within subject treatment Repeated Measures in SPSS: Select options

Repeated Measures in SPSS: Output - Mean glucose uptake Means for four treatments and 95% CI 1 = Basal; 2 = Insulin; 3 = Palmitate; 4 = Insulin+Palmitate

Basal Insulin Palmitate Insulin+Palmitate Repeated Measures in SPSS: Output – Plot of Mean glucose uptake

Repeated Measures in SPSS: Output – Comparisons of Mean glucose uptake Comparison of means with Bonferroni correction 1 = Basal; 2 = Insulin; 3 = Palmitate; 4 = Insulin+Palmitate

Repeated Measures: Conclusion Energy intake significantly higher with insulin compared to all other treatments Energy intake significantly higher with insulin compared to all other treatments Addition of palmitate removes this effect Addition of palmitate removes this effect

Organisation of data (Mixed Model) Note most other programs and Mixed Model analyses require ONE row per measurement UnitScore Etc…….

Repeated Measures in SPSS Mixed Model in SPSS is: Mixed Model in SPSS is: Mixed Model Mixed ModelLinear Hence can ONLY be used for continuous outcomes. Hence can ONLY be used for continuous outcomes. For binary need other Software e.g. SAS For binary need other Software e.g. SAS

Repeated Measures in SPSS: Mixed: Set within subject factor Repeated Within subject factor Within subject factor name

Repeated Measures in SPSS: Enter columns of repeated measures Use arrow to enter subjects and repeated measure column Choose covariance type = AR (1)

Repeated Measures in SPSS: Select options Add dependent Treatment factor And covariates Select other options

Add effects as fixed And Main Effects Repeated Measures in SPSS: Select options

Repeated Measures in SPSS: Output - Overall test for treatment p = 0.024

Repeated Measures in SPSS: Output –

Mixed Model Repeated Measures:Conclusion Use of Mixed Models ensures all data used assuming data is MAR and so more efficient in presence of missing data (if MAR) than the simple repeated measures Use of Mixed Models ensures all data used assuming data is MAR and so more efficient in presence of missing data (if MAR) than the simple repeated measures Other software e.g. SAS can also handle binary outcome data Other software e.g. SAS can also handle binary outcome data

Sample size for repeated Measures Number in each arm = Where r = number of post treatment measures p = number of pre-treatment measures often 1 Frison&Pocock Stats in Med1992; 11:

Sample size for repeated Measures Number in each arm = Where σ = between treatment variance δ = difference in treatment means ρ = pairwise correlation (often 0.5 – 0.7)

Sample size for repeated Measures Efficiency increase with number of measurements (r) (z α +z β ) 2 = 7.84 for 5% sig and 80% power Methods assumes compound symmetry – often wrong but reasonable for sample size

Example: Sample size for repeated Measures For r = 3 post-measures, correlation=0.7, p=1, (z α +z β ) 2 = 7.84 for 5% sig. and 80% power Say δ=0.5σ then…..

Example: Sample size for repeated Measures Which gives n = 19 in each arm with 80% power and 5% significance level

References Repeated Measures in Clinical Trials: Analysis using mean summary statistics and its implications for design. Statist Med 1992; 11: Field A. A bluffers guide to …Sphericity. J Educational Statistics 13(3): Pallant J. SPSS Survival Manual 3 rd ed, Open University Press, Field A. Discovering Statistics using SPSS for Windows. Sage publications, London, Puri BK. SPSS in practice. An illustrated guide. Arnold, London, 2002.

Thank you forlistening!