Article which I read recently says "Furthermore there are no formulas for calculating sample size (or statistical power) in CCO designs.(Carracedo-Martínez, E., Taracido, M., Tobias, A., Saez, M., & Figueiras, A. (2010). Case-crossover analysis of air pollution health effects: a systematic review of methodology and application. Environmental health perspectives, 1173-1182.). How can I decide on the adequacy of the sample.
As I understand your question, the power calculation for a CCO design is the same as for a sign, signed-rank, or paired t-test.
303
The reason why the paper says that there are no power calculation methods is because as far as I can see there aren't. The various power calculation methods advocated here have been advocated by people who may not have looked at the particular case you are interested in. The case-crossover study design is a relatively new analytical epidemiological approach, and is unique in that the case serves as his/her own control and is used to investigate the transient effects of an intermittent exposure on the onset of acute outcomes. There's an interesting introduction here: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2831419/pdf/aasm.33.3.283.pdf
I've had a look, and since the publication of the paper you cite, I can find nothing that gives a power calculation approach for this design. Given the complexity of the data, power calculations would be extremely sensitive to prior assumptions.
If that is the case, one way forward is to simulate data of a similar form to what you plan for a range of different sizes; then specify a range of effect sizes, simulate the outcome data and see if you can recover the effects you have specified.
If you have a look at this software and accompanying manual you will see the ideas in practice
http://www.bristol.ac.uk/cmm/software/mlpowsim/
http://www.bristol.ac.uk/media-library/sites/cmm/migrated/documents/mlpowsim-manual.pdf
and here is a Powerpoint presentation
http://www.bristol.ac.uk/cmm/media/software/mlpowsim/ssize08.ppt
Dear all,
Thanks for all your comments. All comments were really helpful
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis