1 Steps
As a primary outcome of this tutorial is to understand the bayesian requirements for modeling vancomycin in neonates for bayesian forecasting dose recommendations, the data and parameter values will use existing vancomycin values to get in the ballpark of outcomes to be expected from real data.
At a very high level, the following steps must be done:
- Data Generation
- Generate covariate distributions
- Simulate profiles for population of neonates
- Scenario ‘filters’
- sampling times
- balance of peak/trough
- dose modifications?
- Model fitting via FOCE
- model building
- covariate model building
- proper diagnostic analyses for evaluating model adequacy/selection
- Model fitting via Bayes
- model building
- covariate model building
- prior selection
- proper diagnostic analyses for evaluating model adequacy/selection
- Model forecasting
- how to handle propogation of uncertainity for probability of target attainment (PTA)
- influence of PTA output on recommendations (dose and future sampling)