Direct answer
CMS scope-and-severity codes do not produce a clean SNF-only national benchmark by themselves. The current CMS Health Deficiencies file is a citation-level dataset for Medicare- and Medicaid-certified nursing homes, so skilled nursing operators should use A-L patterns to judge seriousness, survey exposure, and follow-up discipline, not to claim a standalone SNF distribution. ePeople helps organize that follow-up.
Key takeaways
- CMS publishes one deficiency citation per row, and the current public Health Deficiencies file shows 419,479 rows and 23 columns in the August 26, 2026 release.
- The public file covers nursing homes in the last three survey cycles, not a clean SNF-only universe and not simply one survey equals one row.
- A-L codes combine severity and scope: A-C are minimal-harm findings, D-F are potential-for-more-than-minimal-harm findings, G-I are actual-harm findings, and J-L are immediate-jeopardy findings.
- Operators should not equate the public deficiency file with the Five-Star health inspection score because CMS excludes some displayed citations from scoring and applies separate weighting rules.
- A recent external benchmark found about 94% of cited health deficiencies were below actual harm in 2022-2024, while immediate-jeopardy findings were 2.4% to 2.7% by year.
What exactly is the CMS Health Deficiencies file measuring?
It measures deficiency citations, not facilities, not surveys, and not a SNF-only universe. CMS describes the public Health Deficiencies file as one citation per row, and the current release dated August 26, 2026 shows 419,479 rows and 23 columns on the dataset page. The CMS nursing home data dictionary adds the more precise methodology point operators need: the file covers citations from the last three survey cycles, not simply three calendar years. That distinction matters when a building had an irregular inspection cadence or several complaint-driven events. See the Health Deficiencies dataset and the CMS nursing home data dictionary.
How are the A-L codes structured in the federal scope-and-severity matrix?
CMS uses a combined seriousness grid in Chapter 7 of the State Operations Manual. A-C represent no actual harm with potential for minimal harm. D-F represent no actual harm with potential for more than minimal harm. G-I represent actual harm that is not immediate jeopardy. J-L represent immediate jeopardy. That means one letter is already doing two jobs at once: telling you how serious the outcome or risk was and how broadly the issue was spread. For operators, that makes the letter mix more useful than raw deficiency counts when prioritizing corrective work. See Chapter 7 - Survey and Enforcement Process for Skilled Nursing Facilities and Nursing Facilities.
What is the difference between scope and severity in practice?
Severity answers how serious the resident impact or risk was. Scope answers how pervasive the problem was. CMS describes scope as isolated, pattern, or widespread, based on how many residents or staff were affected and whether the issue was occasional, repeated, or systemic. That is why an H-level finding is not just “worse” than a G-level finding in the abstract; it also tells you the problem reached a broader footprint. When a skilled nursing facility reviews its own history, separate “how harmful” from “how pervasive” before assigning work to DON, unit managers, educators, or schedulers. ePeople is most useful when that handoff is explicit rather than buried in a PDF statement of deficiencies. See Chapter 7.
How CMS scope-and-severity codes translate into operator meaning
| Code band | CMS seriousness level | Operational reading for a skilled nursing facility | Five-Star health inspection points |
|---|---|---|---|
| A-C | No actual harm with potential for minimal harm | Low-severity finding, but still a process signal if the same tags repeat across units or cycles | 0 points |
| D-F | No actual harm with potential for more than minimal harm | Common risk band for survey follow-up; often where poor process control starts to affect rating pressure | D 4, E 8, F 16 points |
| G-I | Actual harm that is not immediate jeopardy | Resident-impact band that should trigger tighter executive review, proof of correction, and recurrence checks | G 20, H 35, I 45 points |
| J-L | Immediate jeopardy | Highest-response band with remedy and enforcement implications; leadership should read these as enterprise events, not isolated paperwork issues | J 50, K 100, L 150 points |
Why can’t an operator call this a clean SNF-only national distribution?
Because the public Health Deficiencies file is broader than SNFs and does not carry Provider Type as one of its fields. CMS’s Provider Information file contains Provider Type, while the deficiency file does not, so a true SNF-only cut requires joining the datasets by CCN before any percentage is quoted. That is the methodological guardrail for any operator dashboard, board packet, or consultant slide that claims to show “SNF national severity distribution.” Without that join, the safer phrasing is “CMS nursing home public deficiency data” interpreted for skilled nursing operators. See the Health Deficiencies dataset, Provider Information dataset, and CMS nursing home data dictionary.
What does the current public file tell operators right now?
It tells them the file is large, current, and useful for pattern detection, but not sufficient for every benchmarking claim. As of September 17, 2026, CMS shows the current public Health Deficiencies release as issued August 26, 2026, last modified August 1, 2026, and scheduled for another update on September 30, 2026. That timing matters for operators comparing internal survey prep work with public reporting lag. A building may have corrected a practice operationally while its public citation history still influences outside perception. See the Health Deficiencies dataset.
How is the public deficiency file different from the Five-Star health inspection score?
They overlap, but they are not the same input set. CMS states in the nursing home data dictionary that the public file includes citations not used for the health inspection rating, including citations under IDR or IIDR and citations from cycle 3 standard surveys. The Five-Star technical guide then applies its own weighting rules, revisit logic, exclusions, and scoring structure. So if an operator sees a sharp rise in public-file deficiencies, the right next question is not “Did our star rating automatically drop?” but “Which displayed citations actually feed the scoring method, with what weight, and for how long?” See the CMS data dictionary and the Five-Star Technical Users’ Guide.
Which quantitative benchmarks are safe to use without inventing a current national total?
Two are safe from the available dossier. First, CMS’s current public file size itself is a real benchmark: 419,479 citation rows in the August 26, 2026 release. Second, a September 2026 external benchmark from the Long Term Care Community Coalition found that about 94% of health deficiencies in its 2022-2024 QCOR scope-and-severity database were below actual harm, while immediate-jeopardy findings were 2.4% in 2022, 2.6% in 2023, and 2.7% in 2024. Those figures are useful context, but they are not a substitute for a fresh August 2026 CMS aggregation. See Health Deficiencies and The Accountability Gap.
Why do scope-and-severity bands matter more than raw tag counts for survey readiness?
Because CMS ties seriousness to enforcement and return-to-compliance logic. The CMS enforcement overview says scope and severity are part of remedy selection, and Chapter 7 states substantial compliance exists when deficiencies are cited at no more than the minimal-harm level. In practice, that means one G-level or J-level event can matter more operationally than several low-level repeat citations. A serious distribution shift is a staffing, training, auditing, and leadership problem before it is a spreadsheet problem. That is also why the archive’s survey readiness checklist and risk-based survey post are better companions to this article than a simple tag-count list. See Nursing Home Enforcement and Chapter 7.
How should a skilled nursing facility read complaint and infection-control rows?
Carefully, because the denominator is uneven by design. CMS includes standard, complaint, and focused infection-control deficiency indicators in the public file, but the data dictionary notes that complaint inspection dates are included only when the complaint inspection produced one or more citations. That means complaint-derived rows can overrepresent “cited complaint events” relative to total complaint survey activity. Operators comparing survey origin should say exactly what denominator they are using: all citation rows, only complaint-citation rows, or all inspection events. See the Health Deficiencies dataset and CMS nursing home data dictionary.
What is the practical workflow for turning severity distribution into action?
Start with bands, not anecdotes. Group your own citations into A-C, D-F, G-I, and J-L. Then break each band by tag family, survey origin, unit, and recurrence across cycles. After that, connect the work to who actually closes the risk: DON for practice control, educator or DSD-equivalent training lead for remediation plans, administrator for proof and escalation, and scheduler for staffing-related recurrence signals. If you already use the archive’s F-tag frequency post and care plan workflow, this article gives you the seriousness lens that should sit on top of those tag-level reviews. ePeople can support that by surfacing the work queue and evidence trail, but the underlying CMS categories still need to be interpreted correctly. For product context, see Staffing Perfection.
- State the denominator first: citation rows, surveys, or facilities.
- Separate public-file counts from Five-Star scoring inputs.
- Treat G+ as a leadership review threshold and J+ as an enterprise event.
- Do not publish a 'SNF-only national distribution' unless you joined CCN to Provider Information.
- Track repeated D-F findings as early-warning process failures, not harmless noise.
- Keep evidence of correction aligned to the seriousness band, not just to the tag text.
What should operators avoid saying when they present these numbers internally?
Avoid three shortcuts. First, do not say the public CMS file equals the Five-Star score. Second, do not say a nursing-home-wide file proves a SNF-only national benchmark unless you performed the provider-type join. Third, do not say a low share of G+ findings means the building is safe from survey pressure; repeated D-F patterns still create rating drag and operational rework. The more careful framing is that scope-and-severity distribution helps skilled nursing operators rank survey risk and corrective effort. That framing survives scrutiny and still gives executives something usable.
