Overview

PMPM - Per Member Per Month - is a significant metric for evaluating your network or comparing two networks. To create these metrics we need to consider costs, patient attribution and eligibility, and then apply a risk factor conversion in order to provide a level playing field for comparison. CMS provides extensive details that we follow to calculate this metric.

 

Where you will find PMPM Metrics
DCEs on the Home page	MSSP ACOs on the Analyze page

 

 

How we calculate PMPM

PMPM is the average costs per beneficiary per month. The name itself, Per Member Per Month, pretty much says it all.

So, here are the steps we use. And yes, this is an oversimplification at an almost epic scale.

1. Add up all costs reimbursed to a given network for a chosen year. 
2. Identify "eligible" patients
3. Count the months in which a patient was eligible as a beneficiary.
4. Use the following formula to calculate PMPM

 

Determining Eligibility Months

To calculate on a monthly basis we need to identify eligibility for each beneficiary to assure that we only use costs and count months for which each beneficiary meets certain criteria. Think of it as making sure that the average isn't skewed through the inclusion of wrong data. Some examples will show how we limit eligibility months:

• Any patient who dies six months into the year is only counted for six months, not for the whole period.
• The beneficiary must be enrolled in Part A and Part B for the entire 2 year alignment period
• The beneficiary can't be enrolled in Medicare Advantage.
• The beneficiary can't be using Medicare as a secondary payer.

To simplify, for each month we check to see if the beneficiary is eligible and if so, we add that month to the overall count that we use in the denominator of the equation above.

 

 

How does Risk Score impact PMPM

In order to have a level playing field for comparison, we create a risk score adjustment to the final cost. We also use a normalization factor to "center" risk scores around 1.0 for the population being evaluated. Doing so provides an accurate basis for comparison:

• A normalized risk score of 1.0 means that beneficiary/physician is at the average risk level for the population.
• A normalized risk score of 2.0 means that beneficiary/physician is at twice the average risk level for the population.

The normalization factors that we have used are in the following table:

Performance Year	Factor applied
2019	​1.1781502149
2020	1.1933586994
2021	0.9472285591
2022	1.1127651982
2023	1.1343580715

 

Example of 2021 Risk Score Normalization

For MSSP ACOs and DCEs, Trella has normalized patient risk scores based on the national average FFS risk scores (note this is not limited to just DCE assigned lives) and now performs the adjustment at the aggregate cost level, rather than at the beneficiary level. Normalization creates an index that allows for a more accurate comparison of risk across providers and better aligns our methodology to what CMS uses.

The normalization factor centers risk scores around 1.0 for the patient population being evaluated. A beneficiary with a normalized risk score of 1.0 carries the same level of risk as the population average. For providers, a normalized risk score of 1.0 indicates a patient population with the same level of risk as the overall FFS population. Along the same lines, a normalized risk score of 2.0 indicates a beneficiary carries twice the level of risk as the population average. Providers with a normalized risk score of 2.0 treat a patient population that carries twice as much risk as the FFS average.

Some equations:

Keep these ready on the blackboard, we will come back and use them:

 

Table of samples

Physician	Beneficiary	Unadjusted Risk Score	Unadjusted Cost ($)
Dr. I. Jones	Patient Doe	1.123	132,554
Dr. I. Jones	Patient Buck	1.023	213,243
Dr. I. Jones	Patient Ram	0.987	218,612
Dr. I. Jones	Patient Lamb	0.957	123,432
Dr. E. Brown	Patient Sturgeon	0.998	174,633
Dr. E. Brown	Patient Palver	1.342	194,432
Dr. E. Brown	Patient Parker	1.256	212,283

 

Example 1 - Dr. I. Jones

Four patients, Doe, Buck, Ram, and Lamb.

Use Equation to calculate the Unadjusted Physician Risk Score

Thus, the Unadjusted Physician Risk Score is 1.0225

 

Use Equation to calculate the Normalized Risk Score

The Normalized Risk Score for the physician is 1.0743

 

Next we add up the Unadjusted Physician Cost:

132,554 + 213,243 + 218,612 + 123,432 = $687,481

 

Use equation to put all the pieces together:

 

The Adjusted Physician Cost is $640,269

 

 

Example 2 - Dr. E. Brown

Three patients, Sturgeon, Palver, Parker.

Use Equation to calculate the Unadjusted Physician Risk Score

Thus, the Unadjusted Physician Risk Score is 1.1987

 

Use Equation to calculate the Normalized Risk Score

The Normalized Risk Score for the physician is 1.3098

 

Next we add up the Unadjusted Physician Cost:

174,633 + 194,432 + 212,283 = $581,348

 

Use equation to put all the pieces together:

 

The Adjusted Physician Cost is $443,845