Methodology¶

This tool is an accompaniment to the OpenPlan demand model. It has been designed to make it easy, clear and consistent to convert the predicted hospital activity produced from that model into estimates of the capacity needed to supply that activity.

What happens 'under the hood'?¶

This tool takes one financial year's worth of hospital activity data, groups it according to which resources it uses, applies a set of assumptions about how the hospital operates, then calculates how many beds, rooms and other resources might be needed to meet that demand.

The essential components needed to perform a conversion are:

• A funtional area (a resource and the context in which it is being used (e.g. BEDS_DAYCASE_RECOVERY is hospital beds that are being used for the recovery period following daycase procedures)
• Subgroups/classifications of the row-level activity data
• Relevant assumptions about how resource is used
• a conversion formula

For example:

• Functional Area: A&E beds (BEDS_AE)
• Classification: adult majors
• Assumptions: assumed length of stay per episode and hours per day the ward operates
• Conversion formula (a formula converts this into ADULT_MAJOR_AE_BEDS)

To see how all these elements relate to each other, please see the calculation traceability matrix.

How the model handles uncertainty¶

The OpenPlan capacity conversion model uses the results generated by the OpenPlan hospital demand model. The demand model results include a range of future demand forecasts, reflecting the uncertainty in assumptions about future hospital activity. The full range of demand forecasts is used to create the capacity projections, ensuring that the uncertainty in the demand forecasts is carried through to the capacity results. Importantly, the capacity conversion model does not add any additional uncertainty around the assumptions it employs; all uncertainty originates from the demand modelling stage.

For each scenario, the demand model generates at least 256 demand forecasts. Each of these forecasts is converted into a corresponding capacity requirement. The resulting distribution of capacity projections is then summarised to produce the reported capacity conversion results (the mean, p10 and p90 for each capacity type).