Trainees and Professionals Model and Virtual Experiments

From SystemsWiki

Jump to: navigation, search

Consider a group of health professionals. We are about to face the problem of a sudden increase in the leaving rate of professional assistants . We need to replace these professionals with trainees who take 6 months of training and supervision to become registered professionals. Our simple rule for hiring is to link the trainee intake with the rate at which professionals leave.


Fact: The system is in steady state. A stock of 60 professionals is being maintained (and has been for 10 months). 10 professionals leave and 10 trainees start per month Training takes 6 months

What happens to the lines on the graph, if at 10 months the leaving rate is increased to 15 per month?


Here is a model of Trainees and Professionals nearly identical to the Rookies & Pros Model


In our workshops this blackboard exercise produces many results and many interpretations. Describing it explicitly in a diagram makes what we mean precise and rigorous. Based on the description and the model logic, we can provide parameter and initial values for the software. Then the explicit behaviour over time becomes the output of the computer model:

Fig. 1 - Output Graph of Rookies and Pros Model [Source]
Fig. 2 - Trainees Professionals model for Virtual Experiments [Source]

Model 2 Experiment to get to target

Set a target for pros say 100 and hire according to the gap Note sliders for virtual experiments The following behaviour over time graph shows the results of this policy rule

The oscillations are due to the delay due to the time spent in training. So we need to adjust or hiring to take account of the trainees already in the pipeline.


Model 3 shows how this feedback from trainees and professionals changes the structure of the model The model output below shows the oscillations, swings (or hunting) have disappeared and there is no overshoot, but there is still a significant delay in replacing the professionals. The above policy requires spikes in training intake which may not be feasible due to shortages in training staff and places or inability to recruit trainees in surges.

Compare this discussion using simulation runs and stock flow model structure with causal loop diagrams representing the problem

Fig. 3 - Causal Loop Diagram of Trainees and Professionals [Source]

Use of technology to reduce training time, but there may be other unintended effects due to delays and feedbacks and contextual factors, such as the trainers quitting due to the rapid, forced introduction of the new technology

This simple model is relevant to current health workforce issues, particularly responses due to the expected retirement of baby boomers.

SO far we have illustrated the use of concept maps including causal loop diagrams, stock-flow diagrams, and virtual experiments running simple models and successively showing the effects of adding more items to the model. IN abstracting the essentials of real world problems we need to build models that are simple enough, but no simpler. And we perhaps need to remember when to stop. As a colleague put it

“There is no perfection in this life, therefore we must settle for success.”

Questions & Comments to Geoff McDonnell
Personal tools