PDGM pays home health in thirty-day periods, grouped by admission source, timing, clinical category, functional level and comorbidity. Therapy volume no longer drives payment, and the diagnosis coded as primary determines the clinical group — which makes coding accuracy the single largest revenue variable.
PDGM changed what home health gets paid for, and several years on there are still agencies running processes designed for the model it replaced.
What replaced what
Sixty-day episodes became thirty-day payment periods. Therapy visit thresholds, which used to drive payment directly, were removed entirely. In their place sits a grouping built from patient characteristics.
Five variables determine the group: admission source, whether the period is early or late, the clinical grouping from the primary diagnosis, functional impairment level, and comorbidity adjustment.
Coding became the revenue lever
The primary diagnosis assigns the clinical grouping. Choose a diagnosis that does not group and the claim returns for correction — it does not pay at a lower rate, it does not pay at all until fixed.
That makes the coder’s access to complete clinical documentation a direct revenue question rather than a compliance one.
Comorbidities are frequently left on the table
Secondary diagnoses can produce a comorbidity adjustment, and agencies routinely under-report them because the documentation supports conditions nobody coded.
Reviewing the comorbidity capture rate across periods is one of the higher-yield audits an agency can run on itself.
Admission source and timing
Whether the patient came from an institutional setting and whether the period is early or late in the episode both change payment. Both are factual and both are frequently recorded incorrectly at intake.
Fixing intake accuracy costs nothing and moves payment on every period.
LUPAs are now variable
Under the old model the low utilization threshold was a single number. Under PDGM it varies by grouping, which means the visit count that triggers a per-visit payment differs from patient to patient.
Scheduling without visibility of the applicable threshold produces LUPAs nobody intended.
Cash flow moved
Shorter periods mean more frequent billing and more frequent claim cycles. Agencies that did not adjust their billing cadence found their A/R behaving differently for reasons that had nothing to do with payers.
The request for anticipated payment structure also changed, which affected the front end of every episode.
What to check in your own operation
Comorbidity capture rate, the share of periods returning for non-grouping diagnoses, LUPA frequency against expectation, and whether admission source is recorded accurately at intake.
Those four explain most of the variance between agencies operating under identical rules.
Requests for anticipated payment are gone
The old split-percentage advance was phased out and replaced with a notice of admission requirement carrying its own timeliness penalty. Late submission reduces payment for the period.
That is a front-end deadline with a direct financial consequence, and it belongs to intake rather than to billing.
Common questions
- What is PDGM?
- The Patient-Driven Groupings Model, which pays home health in thirty-day periods based on patient characteristics rather than therapy visit volume.
- How long is a payment period?
- Thirty days, with the first period in an episode paid differently from subsequent ones. Timing is one of the grouping variables.
- Does therapy volume still affect payment?
- No. Removing therapy thresholds was the central change. Staffing to a therapy target no longer increases payment and can increase cost.
- Why does the primary diagnosis matter so much?
- It determines the clinical grouping, and some diagnoses do not group at all — a claim coded to one of those returns for correction rather than paying.
- What is a LUPA under PDGM?
- A low utilization payment adjustment, paid per visit rather than per period when visits fall below a threshold that varies by grouping.
Denials Piling Up?
We handle the revenue cycle end to end — coding by certified coders, claim submission, denial management and appeals, and A/R follow-up, with six reported numbers every month.