THE ANALYSIS OF COMPLEX PATTERNS OF LONGITUDINAL BINARY RESPONSE - ANEXAMPLE OF TRANSIENT DYSPHAGIA FOLLOWING RADIOTHERAPY

In many clinical trials, treatment is given in phases and the prevalen ce of symptoms is recorded longitudinally. As a result, complex non-li near response patterns may be observed as the prevalence of symptoms c hanges as a consequence of treatment. In such cases, although profiles of the marginal response over time give an informative description of the data, they do not allow a formal treatment comparison or adjustme nt for covariates of interest. Here we analyse previously reported dat a on transient dysphagia in patients with non-small-cell lung cancer t reated with radiotherapy. We use a generalized estimating equation app roach for repeated measurement binary response to give robust standard errors taking account of the dependence of observations taken on the same subject, alongside a natural cubic spline to represent the comple x shape of the marginal response. This provides a reasonable model for the marginal response and allows unbiased estimation of an apparent t reatment difference. The effect of different choices for the working c orrelation matrix is discussed, as is the modelling of treatment group differences that vary over time. We conclude that these models provid e a powerful tool for the analysis of such data that can now be applie d using generally accessible software. (C) 1998 John Wiley & Sons, Ltd .