- Physical therapy
- Module 6
- Section 9 of 17
OASIS Item Codes and PDGM
Every concept in this module has been taught in plain clinical English on purpose. The field labels in a home health EMR are usually not plain English, though. A screen is more likely to read M1033 and GG0130 than "risk for hospitalization" or "functional self-care," because OASIS is a federally standardized instrument, every item carries a code, and the code is what most systems print on the field. The clinical judgment is exactly what you have already learned. The table below is the bridge between what you know and what the label on the field is likely to say.
| Plain-English Concept | OASIS Item Code | Where You Learned It |
|---|---|---|
| Risk for hospitalization | M1033 | Section 3, Risk Assessment |
| Medication management / drug regimen review | M2001-M2030 | Section 3, Medications |
| High-risk drug classes | N0415 | Section 3, Medications |
| Skin and wound status | M1306-M1342 | Section 3, Skin Conditions |
| Functional self-care | GG0130 | Section 3, Functional Status |
| Functional mobility | GG0170 | Section 3, Functional Status |
| Cognitive function (Brief Interview for Mental Status) | BIMS, C0100-C0500 | Section 2, BIMS |
| Delirium screening (Confusion Assessment Method) | CAM, C1310 | Section 2, CAM |
| Depression screening | PHQ-2 / PHQ-9, D0150 | Section 2, PHQ-2 and PHQ-9 |
| Transportation access (social determinant of health) | A1255 | Section 2, Transportation |
Nothing in this table is a new clinical concept. It is the same judgment you already know how to make, wearing the label an EMR is likely to print on the field. The first time a preceptor or a chart note says "score the M1033" instead of "what's her hospitalization risk," you will already know exactly what they mean.
How Accurate Scoring Connects to Payment: PDGM in Plain English
You do not need to become a billing expert to work in home health, but you should understand, at a structural level, why accurate OASIS scoring is not just a documentation exercise. It is a direct input into how an episode is paid under the Patient-Driven Groupings Model (PDGM), the payment system Medicare uses for home health.
Here is the mechanism, without the numbers, because the numbers change over time and the structure does not:
The reason to score accurately is not "score high to get paid more." The payment system is built to match resources to what the patient actually needs, and over-scoring and under-scoring both break that match. Score the transfer you watched and the distance you measured, the same way Documentation Efficiency and Common Mistakes already taught you.
None of the specific numbers behind case-mix weights, functional impairment thresholds, or LUPA visit counts are covered here on purpose. Those figures are set and periodically revised by CMS, and the current values sit with whoever handles billing: a biller on staff, a contracted billing service, or a screen in the EMR, depending on the agency. What has held since PDGM began in 2020, and still holds in 2027, is the structure above, and understanding that structure is what lets you follow a conversation about why a period fell below the LUPA threshold, or why documentation accuracy matters to a biller and a coder who never saw the patient.
Your Words Decide What Can Be Coded
The list above named the primary diagnosis as one of the inputs that classifies a payment period. That input deserves its own treatment, because it is the one your note controls.
Assigning the ICD-10 code is not usually the visiting clinician's job in home health. A coder does that, sometimes on staff and sometimes contracted, working from the referral, the hospital paperwork, and your evaluation note. Here is the part that surprises new clinicians: a coder can only code what the documentation supports. They are not permitted to infer, and they cannot be more specific than the record in front of them. The ceiling on how accurately an episode gets coded is set in the field, by the clinician who wrote the note.
That cuts both ways. Referrals routinely arrive saying something like "generalized weakness," which describes what the patient looks like rather than the condition producing it. If your evaluation note repeats that phrase and stops there, the coder has nothing more specific to work with, the primary diagnosis stays exactly as vague as the referral was, and the clinical grouping follows the code. None of that is the coder being careless. It is the record being vague, and the record is yours.
The boundary is worth stating plainly: you are not the one choosing the code, and you are not the one adding a new medical diagnosis to the chart. Your job is narrower and more useful than either. Document the conditions you observed, what each one is doing to this patient's function, and what you did differently because of it. That is the raw material a coder actually works from, and nobody else in the organization can produce it.
Comorbidity Capture: The Part Only You Can See
The comorbidity adjustment named in that same list works the same way as the primary. A secondary condition can only be counted if it sits in the record with enough clinical substance behind it to be coded, and you are frequently the only clinician in the house watching how those conditions actually behave. Write them the way you would describe them to a colleague, in plain clinical language, and let the coder handle the translation.
Three things earn a condition its place in the record: the condition, its functional consequence, and the intervention it shaped. A row of abbreviations at the top of the note delivers none of the three, and a condition captured that way can sit in the chart for the whole episode without ever reaching the person assigning the codes.
Write this: Referred following hospitalization for a heart failure exacerbation, now with generalized weakness and reduced activity tolerance. Type 2 diabetes with a healing surgical incision on the right lower leg: dressing dry and intact today, protective sensation reduced at the great toe and first metatarsal head, patient and daughter educated on daily foot inspection and on checking inside the shoe before walking. Heart failure limiting activity tolerance: ambulated 60 feet with a rolling walker before requesting to sit, dyspnea recovered within two minutes of seated rest, session paced with rest breaks and vitals taken before and after. Parkinson's disease with freezing at the bedroom doorway affecting transfer and turning safety: cueing sequence trained with the daughter present, device plan held at the rolling walker rather than progressing toward a cane for that reason.
Not this: Generalized weakness. PMH: DM, CHF, PD. PT for strengthening and gait training.
When the Coder Sends You a Query
A coder query is a short message, usually arriving through an EMR or relayed by a clinical manager, asking you to clarify something in your note before a code is assigned. It sounds like this. "You documented a healing wound on the right lower leg. Is it surgical or a pressure injury?" "You noted the patient is diabetic. Are you addressing anything related to it?" "You wrote generalized weakness. Weakness following what?"
A query is not a complaint about your note and it is not a request for a particular answer. It is the coder telling you the exact spot where the record ran out of specificity, which makes it some of the most specific documentation feedback you will ever get. Answer it the same day, in the record, from what you actually observed at the visit. If you know, say it plainly. If you do not know, say you do not know and say who would.
What you never do is guess toward the answer that sounds better, and you never revise a prior note to produce a cleaner code. Documentation follows the visit, not the payment. That is the same line this module already drew about scoring what you observe, and it does not move just because a different department is the one asking.
You are accountable for how accurately your patients get coded even when someone else assigns the code. What you control is the description: every condition you observed, what it is doing to this patient's function, and what you changed because of it. Write that in plain clinical language and a coder can build an accurate picture. Leave it out and no amount of coding skill downstream puts it back.