MEETING OF THE PRIMARY HEALTH CARE STUDY GROUP

Monday 26th March 2012
Royal Statistical Society
Errol Street, London
2-5pm

All welcome. No pre-registration is necessary

THEME: SELECTING THE DESIGN: HOW IS THE CHOICE MADE?

Speakers:
Tim J Peters (School of Clinical Sciences, University of Bristol) Design, analysis and reporting of 2×2 factorial trials.
Reporting standards for factorial trials remain to be established. In this context my talk will introduce factorial designs including their advantages and disadvantages, indications and contra-indications, and key aspects of their design and analysis. Drawing from published guidance and reviews, and from examples of personal experience of their application, I will briefly discuss issues such as: interactions; sample size; cluster randomisation in factorial trials; economic evaluations; and reporting issues especially for complex interventions. I may also have time to attempt some rationalisation (of debateable quality but undoubtedly post-hoc) of where in my applied work I have not always followed guidance I have myself published!

Toby Prevost (Dept of Primary Care and Public Health Sciences, King's College, London) An example of a simple adaptive design.
Validation studies and internal pilot studies are examples of types of study that may be undertaken in two recruitment stages. In this talk I will describe a clinical study with a two-stage design in which multiple candidate predictors of response to treatment (predictive biomarkers) are assessed part-way through the study recruitment, with the worst-performing candidates then discarded from further investigation. Special rules are required for valid joint analysis of the data from both stages. I will show how the sample size was set to provide acceptable interim and final error rates for a range of effect sizes. The design is compared to a similar one used in a primary care study that had the aim of developing and then validating multiple family health-history questions which could predict the risk of common health conditions. The two-stage design is a simple adaptive design which could be more commonly considered in primary care research.

Celia Taylor (Medical School, University of Birmingham) The stepped wedge randomised controlled trial study design.
In a stepped wedge randomized controlled trial, individuals or clusters are randomly allocated to the order in which they will receive an intervention, with one individual/cluster receiving the intervention at the beginning of each study period (step).  Therefore by the end of the study all participants will have received the intervention. This paper will detail such a study design and discuss when it might be particularly appropriate.  The strengths and weaknesses of the design will then be considered.  The results of a systematic review of the application of the stepped wedge cluster randomized controlled trial design will also be presented.  Twenty five studies are included in the review, which have been categorised according to the motivations for using the design, the context of the intervention, methods of data analysis and quality of reporting. 

Followed by Group Discussion 

5pm Close

Meeting contact: Maurice Marchant ([log in to unmask])

You may leave the list at any time by sending the command

SIGNOFF allstat

to [log in to unmask], leaving the subject line blank.