DESPatientPathway

This is a 'Discrete Event Simulation' of the flow of patients through a typical acute trust hopsital service.

Patients are referred in to be seen at an outpatient (OP) clinic. The patients either attend, or do not attend (DNA). DNAs are re-booked and these patients re-join the clinic waiting list. Patients that attend the clinic are either admitted to the treatment waiting list, re-booked for a followup appointment, or are discharged completely. If admitted the patient visits a pre-operative bed, the operating theatre, then a post-operative bed before being discharged home. The pre and post-operative beds are a shared resource on the same ward.

This model is in development. It is not yet ready to be used for planning. See the 'Notes & Assumptions' tab for more details.

Demand:

Number of new patient referrals (monthly)

Number of existing patients in the OP clinic backlog

Capacity:

Number of OP clinic slots (patients per week)

Number of theatre slots (patients per week)

Total beds (pre & post-operative combined)

Service performance:

OP DNA rate (%)

OP outcome: Book followup (%)

OP outcome: Admit (%)

Average pre-operative length of stay (days)

Average post-operative length of stay (days)

Model options:

Future period to forecast (weeks)

Number of simulations to run

Simulation results:

This table shows the simulation results patient by patient. Use the search box to search for an event, or a patient.

Patient pathway:

Source Code

The code for this project is open-source, and is on GitHub.

Notes

This model is in development. Potential future improvements are detailed below.

Need to add an additional class of 'priority' patients (eg. cancer) in parallel to the standard patients. Define the demand and queuing behaviours separately for each class of patient.

Assumptions

Working patterns

The simulation runs on a weekly pattern. It does not account for holidays, cancelled clinics, or cancelled theatre procedures. These inefficiencies should be included in the numbers selected by the users. For example:

If the OP schedule allows for 30 patients weekly, but historically 10% of clinic slots are cancelled due to consultant holiday, the number of slots input to this model should be 27.
Similarly, if there are nominally 10 operating theatre patient slots, but historic theatre session utilisation is 80%, only 8 slots should be specificed in this model.

Probability distributions and randomness

Exponential distribution are used to model the patient arrivals (referral times). This intentionally introduces randomness into the model, and is the reason that repeat runs do not give identical results.

The decision points (ie. the outcome of the clinic appointment), are also randomly sampled, introducing further randomness. The overall proportion of decisions match those specificd by the user in the user interface.

Followups are rebooked (re-join the queue) in 4 weeks. Randomness could be added here if desired.