We do a number of things in our processes for developing/releasing the model and onboarding trusts.
Where should we document this? It doesn't feel like it fits into assumptions log, QA log or plan? Where should it live? Just a page on the project information?
Code development
The model is expected to evolve and as a consequence QA review will be a continuous process. Code testing and coverage will be kept under constant review, and user acceptance of outputs will be carried out whenever any changes are made to the outputs. Some QA processes will be repeated after the model has undergone significant changes, namely:
Checking of outputs and sensitivity
External QA of code
GitHub Actions
PR review
Release checklists
Onboarding trusts
The following parameters are set by individual trusts
Baseline adjustments
Population changes
Waiting list imbalances
Expatriation/ repatriation
Inequalities
Activity mitigators
Bed occupancy
In all cases these parameters are set after an extensive support process from a team allied to the team building the model, and the parameters are set with appropriate involvement of stakeholders and with governance processes as determined locally.
Data is further verified by consistency checks carried out by hospital trusts using the model. Information about their baseline activity is provided to trusts and they are able to check their own records of activity against the data used in the model. Any inconsistencies are explored and resolved with each trust before the model is run.
Where to add general processes?
We do a number of things in our processes for developing/releasing the model and onboarding trusts. Where should we document this? It doesn't feel like it fits into assumptions log, QA log or plan? Where should it live? Just a page on the project information?
Code development
The model is expected to evolve and as a consequence QA review will be a continuous process. Code testing and coverage will be kept under constant review, and user acceptance of outputs will be carried out whenever any changes are made to the outputs. Some QA processes will be repeated after the model has undergone significant changes, namely:
Onboarding trusts
The following parameters are set by individual trusts
In all cases these parameters are set after an extensive support process from a team allied to the team building the model, and the parameters are set with appropriate involvement of stakeholders and with governance processes as determined locally.
Data is further verified by consistency checks carried out by hospital trusts using the model. Information about their baseline activity is provided to trusts and they are able to check their own records of activity against the data used in the model. Any inconsistencies are explored and resolved with each trust before the model is run.