Cambridge Neuroscience Event

This event profile is in the events archive.


WelcomePlenaryProgramme
RegistrationDirectionsSponsors
ExhibitingContact Us


Meta-Analysis in Mental Health Research

When

April 27th, May 4th and May 11th

Where

Physiology/Psychiatry

Description

Target audience: Researchers in Psychology/Psychiatry/Clinical Neurosciences/ Neuroscience who are undertaking or will undertake systematic reviews with a view to meta-analysis. We will assume only basic statistical knowledge: participants should be familiar with the concepts of means, standard deviations, standard errors, odds ratios, risk ratios, significance tests and confidence intervals.

Plan (Three half-day sessions, April 27th, May 4th and May 11th)

 

Session 1 (April 27th, 14:00-17:00) – Extracting outcome data for use in meta-analysis – how to produce an estimate and standard error for a range of types of outcome data.

1.5-hour lecture covering the following types of outcome data:

  • Non-comparative data: Log odds, mean.
  • Comparative data: odds ratio, risk ratio, risk difference, mean difference, standardised mean difference.
  • Incidence and prevalence rates
  • Change scores (assuming correlation between pre-and post-scores known and no drop out)

In each case (where possible), we will describe what you need to extract from a publication, give a formula, and give an example.

1-hour pen-and-paper practical: we will give the participants a published paper and ask them to extract the relevant data. We’ll also discuss study quality.

Optional 0.5-hour question-and-answer session based around participants’ current systematic reviews.

 

Session 2 (May 4th, 14:00-17:00) – Performing a meta-analysis

1.5-hour lecture covering the following topics:

  • Fixed-effect and random-effects models
  • Fitting these models in R and Stata.
  • What does the R/Stata output mean – and an indication of how the calculations were performed.
  • Approximations required – large studies and for RE meta-analysis a large number of studies.
  • Cochran’s Q statistic, the standard test for the presence of heterogeneity, DerSimonian and Laird’s estimate of the between-study variance, I-squared.

1-hour practical in which we give participants some examples of meta-analyses – say, a forest plot including fixed-effect and/or random-effects summaries and I-squared – and ask them to explore what has been shown, what it means, how reliable it is.

Optional 0.5-hour question-and-answer session based around participants’ current meta-analyses.

 

Session 3 (May 11th, 14:00-17:00) – Advanced topics in meta-analysis

1.5-hour lecture covering the following topics:

  • Publication bias: funnel plots, Egger’s test, [trim and fill?]
  • Moderator variables: subgroup analysis, meta-regression
  • Quality of studies
  • Pointers to other methods: individual participant data meta-analysis; multivariate meta-analysis; network meta-analysis (giving references so that participants can know where to look if they need these topics)

1-hour practical in which we give participants some examples of these outputs.

Optional 0.5-hour question-and-answer session based around participants’ current meta-analyses.

 

Back to top

Plenaries

/

Back to top

Programme at a Glance

/

Back to top

Directions

Session 1 (April 27th, 14:00-17:00)

Bryan Matthews Seminar Room, Physiology Building, Downing Site, CB2 3EG

Session 2 (May 4th, 14:00-17:00)

Bryan Matthews Seminar Room, Physiology Building, Downing Site, CB2 3EG

Session 3 (May 11th, 14:00-17:00)

Seminar Room, Herchel Smith Building, Department of Psychiatry, Robinson Way, Cambridge CB2 0SZ

Back to top

Registration

Registration for this event is now closed.Back to top

Sponsors

/

Back to top

Exhibiting

/

Back to top

Contact

Please contact Dr Dervila Glynn if you require any further information.



Further events

Go to the events index page.