The Crossover Design 2/22/13 2 pm

Disadvantages:

From:

The Two-Period Crossover Design in Medical Research. John R. Woods, PhD; James G. Williams, MS; and Morton Tavel, MD, Annals of Internal Medicine, 1989; 110; 560-566

1. Must insure no “carryover effect between periods. There could exist a failure to return subject to baseline state.

- There could exist non-uniform pharmacologic and psychologic carry-over effects as well.

- Non-uniform changes in a patient over time.

- Needless to say, if one treatment provides a cure, then there is no way to test the 2nd treatment for that particular patient.

2. Impractical if len...

-566 1 110 1989 1998 2 2/22/13 221 2nd 3 3rd 4 5 50 560 6 ab advers advisori also analysi analyz annal arkansa armitag attrit b ba balanc base baselin bemac bias biometr brown carri carry-ov carryov caus chang choic circumst clinic committe compar comparison consequ consider could critic cross cross-ov crossov cure data demet design detect determin differ difficult disadvantag discard discern discourag distort dropout due edit effect effici enough epidemiolog especi estim etc evid exceed exist experi extrem failur fda first follow found friedman fundament furberg g general great greater group hamer heavyweight hill idea impract independ insur intern investig jame john johnson lengthi lens liabil list littl longer loss low m make may md medic medicin methodolog might model mortal morton ms must near necessari needless nj non non-uniform number observ occur one order outlier paper paradox parallel parallel-group particular patient period pharmacolog phd pippa piscataway pm power profound protocol provid psycholog r receiv relat report requir research respons result return robert rock sas say school scienc seen sequenc shelli simpson sometim springer state studi subject sugi summari tavel test thee theoret time total treatment trial two two-period umdnj underpow unequivoc uniform univers use useless violat washout way well whether william wood world would