| |
Within-patient
studies, whether cross-over or n-of-1 trials, have an important
part to play in the study of chronic disease. Care must be taken
to ensure that they are used appropriately and that there is little
danger of carry-over affecting the results. In this context, adequate
wash-out is essential. Pre-testing for carry-over when analyzing
such trials is inappropriate and adjusting for carry-over is unlikely
to be successful. Within-patient studies are particularly useful
for measuring individual response to treatment, provided that
patients have been treated more than once with the same treatment.
In analyzing such 'n-of-1 trials', random effect models that take
appropriate account of various sources of information should be
employed. Analyzing each patient independently is a wasteful strategy.
|