A Binary and a Normally Distributed Endpoint

When T is a normally distributed endpoint and S is a binary endpoint, the function FixedContBinIT() can be used to evaluate the appropriateness of the candidate surrogate endpoint. Similarly, when T is a binary endpoint and S is a normally distributed endpoint, the function FixedBinContIT() can be used. The first part of these function names indicate that a Fixed- effects model is fitted to estimate the trial-specific treatment effects a and в (for details, see Chapter 9). The second part of the function names indicate whether T is a binary endpoint (function FixedBinContITO) or a normally distributed (continuous) endpoint (function FixedContBinlTQ). Similarly, the third part of the function names indicate whether S is a binary or normally distributed endpoint (functions FixedContBinlTQ and FixedBinContITO, respectively). The last part of the function names indicate that the Information-Theoretic (IT) approach is used. In addition, the arguments Model="Full" or Model="Reduced" can be used in the function call to specify whether a full or a reduced model should be fitted (to obtain estimates for a and вь for details, see Section 4.3.2). Finally, the arguments Weighted=TRUE and Weighted=FALSE can be used to specify whether measurement error should be accounted for or not (for details, see 4.3.4). The different models that can be fitted using the Surrogate package are identical to those shown in Table 13.3, where FixedBinBinIT() is replaced by FixedBinContITO or FixedContBinIT().

Main function arguments

The functions FixedContBinIT() and FixedBinContITO require the same arguments as those of the function FixedBinBinIT(). For details, see Section 13.4.

 
Source
< Prev   CONTENTS   Source   Next >