Cohort Markov and Partitioned Survival Models in Excel
We are accomplished in VBA and Excel manipulation. We often program in the most complicated analysis elements needed in an Excel models aimed at HTA’s. A recent example was programming in various Cure models with PSA capabilities into an existing Excel model produced by a much larger agency that could not do it. Other examples include programming latent class models, Royston-Parmar flexible survival, four parameter generalised F distributions.
We understand the theory behind complicated stats models enabling such programming – acquiring such knowledge is what we enjoy most – unashamedly nerdy!
Larger agencies are better suited to building the majority of any Excel model – employing staff that solely focus on doing such – we hate the many hours spent on improving aesthetic features.
Discrete Event Simulation Models Built in R / C++ or Julia
DES models are theoretically superior to any form of Markov model, focusing on the individual and not the cohort. We believe such models will dominate in the years to come but they cannot be efficiently programmed in Excel if PSA is required.
We build DES models that are easy for the user to navigate – inputs are stored and manipulated in Excel, whilst the analysis is efficiently conducted in R/C++ or Julia.
We prefer Julia to R as it is more efficient and have built a DES Julia template that is much more comprehensive than the R template in NICE TSD 15.