In this work, we consider the problem of using routinely collected primary care data to monitor the progression of undiagnosed incipient renal failure in a large sample of people in primary care. Our proposed strategy, is to build a dynamic regression model in which a subject?s rate of change in eGFR, relative to the expected profile of eGFR for all subjects with the same values for age, sex and other explanatory variables, is modelled as a stochastic process B(t), which is realised independently for each subject.
Key-Words: longitudinal studies; random effects; Kalman-Filter algorithm |