I:  THE VALIDITY OF CLAIMED OUTCOMES OR SAVINGS

A typical RFP will ask: “What is your ROI and how do you measure it?” And perhaps: “Who has validated it?” And “What is your guarantee?”

Those questions are not helpful.

You don’t ask them how they measure their savings. You tell them how they are to measure their savings.

(a) Selecting the codes to measure

You start by determining which diagnosis or procedure codes you want to reduce in ER and inpatient. We will use a diabetes example here. Diabetes is a non-elective common chronic admission. The “control” would be other common chronic admissions. Here are the codes for diabetes and the other common chronics.

You expect separation – admission rates and ER rates should fall for diabetes relative to the other common chronics.

Critically, this separation is for entire group. Not people who were already identified. They will always decline, as previously unidentified people take their place. That is regression to the mean, described at length in Part One of the accompanying series.

It is also not for participants-vs-non-participants. That creates participation bias, many examples of which are in Part Two of the accompanying series.

As to the last point, the vendor will complain that they can’t be responsible for people they can’t enroll. That is precisely the point. They must commit to enrolling as many people as possible and then do well enough on them to move the needle for the entire group.

(b) Picking the baseline

A vendor will justifiably say they can’t promise outcomes without knowing the baseline rates of admissions and ER visits coded to each. There are two ways to determine the baseline. The first is to get your de-identified codes from the carrier or TPA. This might take a lot of time, of course. The second is to estimate the baseline in order to get the RFP out, with an algorithm to reconcile the actual baseline to the estimated baseline.

These figures will suffice as baseline estimates:

Those numbers, summing to 10.4 events/1000, might seem low to you, given that your total ER visit and inpatient admission rate/1000 is roughly 200-250 in total. However, this is directionally accurate for most large groups.

(c) Creating the template for the response

Current RFPs not only fail to get valid savings estimates. They also fail to get comparable savings estimates. For the latter, you set out the event reduction or savings estimate you expect to see, along with the amount at risk.

Then the respondent fills in the grid, like so, supposing for this example that their base price is $1 PMPM. The more risk they take and the higher they set the bar, presumably the higher their price would be, incorporating a “risk premium.” They may also be shy about overcommitting, as in three of the cells here, and not bid at all.

There is also no magic in the axes here. You can alter them, of course. The important takeaway is the concept itself.

There are a few second-order effects that you need to determine for yourselves. Like if they miss the first year and must return fees and overperform in the “out” years and get their money back, how is the time value of money handled? And in this particular case, what if the diabetes program is so successful that it reduces cardiac events as well? (There are several easy workarounds for that. Like tracking cardiac events with diabetes as a secondary diagnosis vs. other cardiac events.)

(d) Reconciling estimated and actual baselines

This is easy. Just set a formula that adjusts the expected reduction if the baseline is lower. For example, if the diabetes baseline is 10% lower than 3.4 events per 1000, the contracted event reduction percentage might be reduced by 20%. (The lower the baseline, the harder it is to reduce it further.) Just lay out the algorithm in the RFP.

(e) Other categories this methodology works for

Some variation of the codes above can be used for wellness/disease management as a whole. Since those programs are almost all focused on cardiometabolic and not asthma or COPD, you can compare the three cardiometabolic codes to the two respiratory codes and look for separation.

As part of any subscription, you can request our sets of codes for:

• • MSK
  • Neonatal
  • Diabetes drug utilization
  • Health literacy (which can also be procured In conjunction with wellness)
  • Offsets to weight loss drugs
  • Mental health

Further, if you are Gold or Platinum, you can call our experts to design an RFP component for you using this concept for other point solutions as well.

(f) Validation

This RFP language justifiably gives vendors validation by the Validation Institute an edge. However, it is possible that a vendor gets validated by a different entity…and the validation is real. “Are these outcomes claims validated by the Validation Institute?”