- The functionality for hospital culture (i.e. hospital culture influencing time to admit) is currently disabled. Choosing hospital culture W/S will give the same result (= the parameters set in the config.json)
- The functionality for bed occupancy influencing time to get a bed is currently disabled. Choosing any option for bed occupancy will give the same result ( = the parameters set in the config.json)
[ ] Explore variability of simulations. This would require Monte-Carlo modelling.
[ ] Include stdevs in time calculations (e.g. time_assess, time to diagnostics) - need to get distributions from data to inform distributions.
[ ] Transfer the model to different hospitals. To prevent overfitting, we suggest a) looking for process map similarities across multiple hospitals, b) treating the model more like a machine learning problem where there is clear training, validation, test sets/metrics and avoiding data leakage.
[ ] Introduce the role of the human efficiency factor into the model.
[ ] Further QA of model, including potentially through unit tests, validation of all parameters, validation of final modelled process map, additional metric validation, etc.
[ ] Get list of occupancy of rooms (i.e. average occupancy of waiting room, majors etc) and resource use (doctors) as an easy to read output.
[ X ] Turn off SDEC completely.
[ ] Better parametrise acuity 3 MH delay time in Majors - this could be informed by PAH data, if provided.
[ ] Decide if health economics modelling is to be incorporated.
[ ] XRay and CTScan should potentially be in rooms where you wait for a resource (machine) to be free (i.e. similar to doctors and waiting/treatment rooms).
[ ] Open animations from past runs in the UI.
[ ] Increase scenario functionality in the UI.
[ ] Reload model environment and variables.
[ ] Way of preloading simulations.
[ ] Separate XRay and CTScan; and add in Bloods.
[ ] Change logos and title.