The long-stay antipsychotic measure: N047.01 and claims data

The long-stay antipsychotic measure used to be an MDS measure: N0415A1 on the target assessment. Since the January 2026 Care Compare refresh, Five-Star uses a new version, N047.01, that also reads Medicare and Medicaid claims and Medicare Advantage encounter data. What your MDS says is now only half the answer.

What changed, and when

The QM User’s Manual v18.0, effective January 1, 2026, re-specified the measure as N047.01. The Five-Star guide records the switch: effective with the January 2026 refresh, CMS replaced the old long-stay antipsychotic measure with the re-specified one and set new cut points in ten equal deciles. With the April 2026 refresh, CMS corrected the calculation so that every resident with a history of schizophrenia is excluded, as the specifications require, and reset the cut points using data from the fourth quarter of 2024 through the third quarter of 2025.

CMS explained the reason in its memo to the states. A 2021 OIG report found the MDS may not reflect every resident prescribed an antipsychotic, so claims now fill that gap, and diagnoses that exclude a resident are checked against claims too. CMS put the national rate at 14.64% under the old measure and 16.98% under the new one.

What puts a resident in the numerator

Either of two things, in the quarter:

• The MDS. The target assessment has N0415A1 checked, meaning an antipsychotic was taken at any time in the 7-day look-back. Under N047.01 this half counts only when the target assessment’s date falls inside the quarter.
• A claim. The resident has a Medicare Part D or Medicaid pharmacy record with a fill date in the quarter during the stay, or a claim for an antipsychotic given by a clinician (an injection, for example) with a service date during the stay.

A claim dated while the resident was discharged from the facility doesn’t count. The drugs are identified by CMS’s published lists of drug codes (NDCs) and procedure codes (HCPCS), in the supplemental files on the CMS quality measures page.

Why stopping the drug no longer clears the quarter

Under the old measure, a resident whose last dose came before the assessment’s 7-day look-back began came out of the numerator on that assessment. The MDS half still works that way. The claims half doesn’t look at a 7-day window; it looks at the whole quarter.

This is the right result, not a penalty: the resident did receive the drug that quarter. It does mean a dose reduction shows up in the measure one quarter later than it used to.

Who is excluded

Coverage. CMS has to be able to see the resident in claims. A resident is excluded unless they were continuously enrolled, month by month, in Medicare with drug coverage (Parts A, B and D, or Medicare Advantage with Part D) or in Medicaid only, from the start of the quarter to the end of the episode. There are similar coverage checks for the year before the target date and, for residents 65 or older admitted within the past year, the year before admission.

Diagnoses. Schizophrenia, Tourette’s syndrome and Huntington’s disease still exclude a resident, but now the diagnosis has to be in two places:

On the MDS, that means I6000 (schizophrenia), I5350 (Tourette’s) or I5250 (Huntington’s) checked on the target assessment or the prior one. On claims, it means a matching diagnosis code on a Medicare or Medicaid claim within a one-year window. For schizophrenia in a resident 65 or older who was admitted within the past year, that window is the year before admission.

A diagnosis checked on the MDS with no matching claim no longer excludes anyone. If that resident is on an antipsychotic, they are in the numerator.

Hospice. A resident is excluded if they received Medicare Part A or Medicaid hospice services, or were enrolled in hospice, in any month from the start of the quarter to the end of the episode. CMS finds this from hospice claims and Medicaid eligibility data, not from the MDS.

Common coding mistakes

• Coding by use, not by class. N0415 is coded by the drug’s pharmacological classification. The RAI Manual gives this example:

• Missing a dose given elsewhere. N0415 includes these drugs given by any route in any setting while the person is your resident, such as in an emergency room.
• Long-acting injections. The MDS counts a monthly injection only if it was given in the 7-day look-back. The claims half can still pick it up from the pharmacy fill or the claim for the injection.
• A diagnosis without support. I6000 needs a physician-documented diagnosis. The RAI Manual says a possible misdiagnosis, such as a mental disorder without the diagnostic information to support it, should not be coded. CMS audits schizophrenia coding, and a building found coding it inaccurately has its long-stay and overall QM ratings dropped to one star for six months, then gets the minimum points on this measure for months 7 to 12.

How Five-Star scores it

The long-stay antipsychotic measure is worth up to 150 points, in ten bands on the national distribution. It is not risk adjusted. The cut points were set on the new measure using data from the fourth quarter of 2024 through the third quarter of 2025:

Your iQIES Facility-Level and Resident-Level QM Reports carry N047.01, so you can see which residents CMS counted. A tracker that reads only the MDS misses drug use that shows up only in claims and can’t see the coverage and hospice exclusions, so its rate can differ from CMS’s in either direction.

SuperQM tracks the MDS half of this measure resident by resident and marks where the claims half can still change the answer.