Spending on Home Health Care Down 21.6% Under PDGM

Spending on home health care services is down substantially under the Patient-Driven Grouping Model (PDGM), a recently unveiled analysis suggests.

While the Bipartisan Budget Act of 2018 requires PDGM to be budget neutral, spending on home health care has actually been about 21.6% lower than projected, according to a detailed analysis from Dobson DaVanzo & Associates (DDA), a health economics and policy consulting firm.

There are several reasons for the underspending, but it essentially boils down to the U.S. Centers for Medicare & Medicaid Services (CMS) being wrong about how providers would behave differently under PDGM compared to the old Prospective Payment System, which had been in place for about two decades prior to Jan. 1 of this year.

The Partnership for Quality Home Healthcare (PQHH) — a Washington, D.C.-based industry group that represents many of the largest home health providers in the nation — commissioned the DDA analysis. It submitted the findings to CMS last week in its comments on the 2021 proposed payment rule.

“The assumptions made by CMS were largely wrong,” PQHH Executive Director Joanne Cunningham told Home Health Care News. “And we’re essentially underspending in home health care right now.”

In crafting the presumably budget-neutral PDGM, CMS believed home health providers would change their documentation and coding practices to always put the highest-paying diagnosis code as the principal diagnosis.

The agency also believed providers would do everything possible to meet Low Utilization Payment Adjustment (LUPA) thresholds in order to receive a full episodic payment.

CMS additionally assumed home health providers would adjust their documentation and coding practices to receive a corresponding payment for patients’ functional limitations and comorbidities.

Combined, those three things contributed to a 4.36% “behavioral adjustment” that was baked into PDGM.

But at least two of those assumed behavior changes aren’t playing out in reality, according to the DDA analysis, which looked at the first four months of 2020 Medicare claims data and more than 2.6 million 30-day PDGM episodes.

“PDGM does not appear to be budget neutral compared to prior years measured either by average or total aggregate payments,” the DDA analysis states.

For the most part, DDA found that case-mix groups under PDGM are more similar to historical trends of primary diagnoses rather than payment-optimized groupings as projected by CMS in its behavioral assumptions. In other words, providers are coding similar to how they always have.

Looking at PDGM’s 12 clinical groups, CMS was particularly off in its projections around the MMTA-Endocrine and Neuro groups.

The fact that home health providers aren’t “upcoding” isn’t shocking, Cunningham said.

As a general rule, providers adhere to the medical information included in the patient record and what the referring physicians determine. Moreover, following best practices for coding and documentation has been a priority for many providers in light of CMS’s initiatives to reduce improper billing and fraud in Medicare.

“We made that point many, many times to CMS,” Cunningham noted.

The DDA analysis echoed Cunningham’s point about upcoding.

“This behavioral assumption would require agencies to substantially disregard international agreed coding schemas, so it is unsurprising shifts did not occur to the extent predicted in the behavioral assumption,” the analysis reads.

When it comes to LUPAs, CMS assumed that in one-third of instances when a case is one or two visits away from the LUPA threshold, providers would provide extra visits to receive a full 30-day episode payment, effectively lowering their LUPA rates in the process. In actuality, LUPA rates in 2020 have been much higher than anticipated.

In the first four months of the year when home health agencies were adjusting to PDGM, the national LUPA rate was 24.4% — with an all-time high of 28.7% in March.

The COVID-19 pandemic dramatically impacted home health activities in the spring, causing LUPAs to skyrocket due to concerned patients declining visits. But the increase in LUPA rates cannot be entirely linked to the COVID-19 virus, as DDA observed a spike in LUPA rates even prior to the onset of pandemic responses in March.

“That suggests that providers were not responding to LUPAs as CMS predicted, even early on,” Cunningham said.

Over the first four months of PDGM implementation, observed average case payments were 6% lower than projected, with an average case payment of $1,706 compared to the projected $1,815 from the FY 2020 CR file. After accounting for case-volume reductions, total aggregate payments are about 21.6% lower than projected, or $4.5 billion observed compared to $5.8 billion expected for the period.

Moving forward, CMS should eliminate the 4.36% behavioral adjustment in 2020 and consider getting rid of it for 2021 as well, Cunningham argues.

It should do so not only to ensure Medicare beneficiaries have access to fully funded home health care services during a national health crisis, but also to avoid falling deeper and deeper into a spending hole it will ultimately have to climb out of.

“If you see this level of underspending in PDGM and have a rate cut that’s built into the model, well, at some point, there has to be a reconciliation, according to the law,” Cunningham said. “So the further you get away from doing that, the bigger the hole is going to get.”