How do the TPMAs work in the model?

Why do some TPMAs have sample rates?

Some TPMAs, e.g. Obesity related admissions IP-AA-026, Alcohol related admissions IP-AA-001, IP-AA-002, IP-AA-003, and Smoking related admissions IP-AA-029, have a sample rate. Whilst most activity mitigation strategies identify all spells based, in this case the model only selects a proportion of spells based on the sample rate for that particular row of activity.

If a specific inpatient admission has a sample rate of 0.36 for the Obesity related admissions TPMA, it will get selected to be impacted by this mitigator in roughly 36% of model runs.

Are there overlaps between TPMAs? What happens if a spell has more than one TPMA that applies?

Yes. All activity could have more than one TPMA applied to it. There is ongoing work to show the TPMA overlaps at a national level - this analysis could also potentially be reproduced at other geographies?

If an inpatient spell has more than one TPMA that applies:

• if there is a sample rate, the TPMA may not apply in every model run.
• activity avoidance TPMAs: the factors for each activity avoidance TPMA applying to the spell are multiplied together to produce an activity avoidance summary parameter (step 7 in the worked example of the model methodology)
• efficiencies TPMAs: Only one will get selected in each of the model runs.

How do the activity avoidance TPMAs work?

See step 7/8 in the worked example of the model methodology)

How do the efficiency TPMAs work?

See step 9/10 in the worked example of the model methodology)