When data on exposure history (e.g., annual average intensities) are available, the code can be used for case-control data (conditional/unconditional logistic regression) or cohort data (Cox regression) in order to estimate a latency function describing the relative risk per unit exposure by time since exposure, i.e., latency, including pointwise confidence intervals. Below is an example based on lung cancer mortality and occupational exposure to radon gas from a study of Colorado Plateau Uranium Miners (Hauptmann et al., 2001).
In order to use the code, EPICURE software is required. The code runs as a series of scripts. Some editing of the scripts is required to tailor the code to the particular application. Two MATLAB scripts are also included in order to graphically present the results. The method has been used in several analyses, including
More detailed documentation is included in the README.PDF document. Download zipped scripts here. If you have any questions, please send an e-mail to . We are currently developing a SAS macro for use with the PHREG procedure.