Whether the outcomes are periodic patient reports, repeatedly-observed disease signs, or serial biomarkers, to study chronic disease requires dealing with time-varying data. this Phase II SBIR grant application concerns the continued development of TI (pronounced """"""""T-sub-I""""""""), a PC program to analyze such data. With the advent of quasi-likelihood, generalized estimating equations, and new ways to handle missing data, has come a striking increase in applicable models, perhaps more than after the 1972 introduction of the cox model for survival data. TI will provide graphics for data exploration and model validation. TI will run under most current version of Windows, use Civilized Software's MLAB as its programming language, and will offer context-sensitive help, a dictionary of relevant terms, identification of file entries from their screen position, and simple animation. TI will permit the user to analyze time-varying data from the two perspectives of repeated measured using a variety of models and time series where both time and frequency domain tools will be provided.
