I asked a statistician involved in evidence-based reviews to comment on
Owen's design. His repsonse is below (with his permission):
Cheers
Sandra
>>This is not a trivial problem. If only the main treatments were
>>allowed in each centre it would be easy. Thus a comparison of (a)
>>against (b) would be equivalent to comparing Treatment A to Treatment
>>B. But because combinations of treatments are allowed, (a) versus (b)
>>then becomes a comparison of A - B + BC - AC yuck!!!!
>>
>>
>>Regards
>>
>>Adrian Esterman
>>Senior Research Fellow in Biostatistics
>>Clinical Epidemiology & Health Outcomes Unit
>>Flinders Medical Centre
>>
>>
Prof Sandra V. Dunn RN PhD FRCNA ph: (618) 8204 5543
Chair in Nursing Practice fax:(618) 8204 5907
Flinders Office of Clinical Nursing Research
Flinders Medical Center
Bedford Park SA 5042
email:[log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|