Appointment Attendance Forecast
The attendance forecast estimates, for each future clinic session, how many booked patients are likely to attend, so that clinic capacity can be planned realistically.
Keywords
- Tags: Forecasting
- Tags: Capacity planning
- Tags: Health
- Stage: Production
- Type: Time-series regression
- Sector: Health and social care
- Language: English
How does our product work?
The model looks at attendance history for a clinic type, the day of the week, the time of year, how far ahead appointments were booked and local factors such as school holidays. It produces an expected attendance rate for the session as a whole.
It makes no prediction about any individual patient, and no individual is ever contacted, deprioritised or refused an appointment on the basis of its output.
Overview
Empty clinic slots waste clinician time that other patients are waiting for. Overbooking to compensate produces long waits in the department. Both are consequences of planning with an average rather than a forecast.
Session-level forecasting lets schedulers size clinics to what is likely to happen, and has reduced both wasted capacity and same-day overcrowding since it went live.
Owner and responsibility
More detailed information on the system
Here you can get acquainted with the information used by the system, the operating logic, and its governance in the areas that interest you.
System description
Forecasts are produced weekly for the following twelve weeks and are shown to schedulers as a range rather than a single number, with the historical accuracy of past forecasts displayed alongside.
The deliberate design constraint is that the model operates at session level only. Patient-level attendance prediction was considered and rejected: it would create a strong incentive to treat some patients as less worth booking.
Data sources
Aggregated historical attendance by clinic and session. Public holiday and school term calendars. No clinical records and no patient identifiers are used.
References
Twelve-month evaluation of forecast accuracy against outturn.