Wu M C, Lan K K, Connett J E
Biostatistics Research Branch, National Heart, Lung and Blood Institute, Bethesda, MD 20892.
Stat Med. 1994 May 15;13(9):945-53. doi: 10.1002/sim.4780130905.
We discuss the monitoring of clinical trials data and propose two surrogates for total information for use with the spending function approach where there are repeated measurements and we wish to compare the rates of change in a response variable under different treatments. These surrogates are applied to a setting similar to the Lung Health Study. Although the surrogates do not require estimation of the variance parameters, they do require some knowledge of the ratio R of the within- to the between-individual variances. The effects of overestimating and underestimating R are illustrated. In situations of uncertainty we recommend overestimating R, to provide a conservative estimate of information time at each interim analysis.
