run results processing throws numpy runtime warning on extreme values for patient proportions

[TomMonks]

Steps to recreate:

scenario = Scenario(prob_trauma=0.0)

# run the model in single run model
run_results = single_run(scenario, random_no_set=42)

This results in a numpy runtime warning as it is attempting to take a mean of an empty slice for Trauma patients. The single_run method creates an instance of SimulationSummary that attempts to run the code. The code still runs, but a warning is produced. Testing currently suppresses the runtime error.

The reverse is true if prob_trauma=1.0