Direct answer
Public CMS deficiency files show that a short list of F-tags carries a large share of survey volume in skilled nursing, with infection control, accident prevention, food sanitation, care planning, and records repeatedly near the top. ePeople uses that pattern to help SNF operators prioritize daily controls, not just annual policy review.
Key takeaways
- Public CMS health deficiency data currently spans 419,479 citation rows in the three-year file, which is enough volume to show repeat operational patterns rather than one-off anecdotes.
- Frequency and seriousness are not the same. High-volume tags like F880 and F812 are common, but F689 stands out because harm-level findings appear far more often within that tag.
- A focused watchlist beats a generic binder review. In California survey data, the top 10 tags accounted for 37% of all citations and the top 20 for 54%, showing why concentrated prep is practical.
- KFF found that 5% of deficiencies in the last survey cycle rose to actual harm or immediate jeopardy, and about 28% of facilities had at least one such finding in 2026.
- ePeople is most useful when operators turn public tag frequency into assigned work by shift, department, and proof standard rather than treating every F-tag as equally likely.
Last updated: September 10, 2026.
Most SNF operators do not need a longer F-tag list. They need a shorter priority list. The public CMS file for health deficiencies currently displays 419,479 citation rows, and that volume makes one point obvious: survey exposure is concentrated. Some tags show up again and again because the underlying work is repetitive, cross-departmental, and easy to drift on between surveys.
This post looks at public CMS deficiency data as an operations problem, not a trivia exercise. It combines the current CMS Health Deficiencies file, the CMS Citation Code Look-up, CMS survey context, and published analyses built from the same public file. The goal is simple: decide which tags deserve daily controls, which ones deserve deeper root-cause review, and where ePeople fits when a facility wants proof instead of a binder reminder.
What does the public CMS deficiency file actually include?
The CMS health deficiency dataset is the national public file behind this discussion. CMS describes it as a list of health citations from the last three years, with one citation per row, and the current catalog page shows 419,479 rows. Each row includes the F-tag, survey date, scope and severity code, survey type, and correction status. That matters operationally because frequency, complaint origin, and severity can all be separated instead of blurred into one total deficiency count.
How concentrated is F-tag volume in skilled nursing?
Very concentrated. A California analysis built from state and federal survey files found that surveyors cited 205 distinct F-tags, but the top 10 tags accounted for 37% of all citations, the top 20 for 54%, and the top 50 for 80%. That is a useful operating lens even outside California: frequency is usually driven by recurring workflows such as infection control, accident prevention, care planning, medication handling, food service, and records discipline.
High-frequency F-tags that operators should separate into different workstreams
| F-tag | What it covers | Why it repeats | Better owner than 'everyone' |
|---|---|---|---|
| F880 | Infection prevention and control | Requires consistent shift-level behavior, audits, and isolation practice | IP, DON, unit managers |
| F689 | Accident hazards and supervision | Combines assessment, environment, observation, and follow-through | DON, rehab, nurse managers |
| F812 | Food procurement, storage, preparation, service | Touches kitchen process, temperatures, dating, sanitation | Dietary manager, administrator |
| F656 | Comprehensive care plan | Breaks when assessment findings do not reach the plan or revision control slips | MDS, DON, IDT |
| F842 | Resident records | Late entries, missing signatures, and inconsistent documentation standards recur | HIM, DON, department leads |
| F761/F756 | Medication storage and regimen review | Daily med-pass habits drift faster than policy updates | Pharmacy, DON, med room leads |
Which tags appear most often in national CMS-based tallies?
A May 2026 national aggregation built from CMS survey data reported 628,761 total citations across 14,696 facilities. In that tally, F880 infection control led with 24,296 citations across 11,916 facilities, or 81.1% of homes. F689 followed with 21,911 citations across 10,134 facilities, or 69% of homes. F812 food safety had 20,409 citations across 11,115 facilities, or 75.6% of homes. F656 care planning had 15,168 citations across 8,526 facilities, and F761 pharmacy-related storage and labeling issues had 13,856 citations across 8,798 facilities.
- F880 infection control: 24,296 citations; 11,916 facilities; 81.1% of homes.
- F689 accident hazards/supervision: 21,911 citations; 10,134 facilities; 69% of homes.
- F812 food safety and sanitation: 20,409 citations; 11,115 facilities; 75.6% of homes.
- F656 care planning: 15,168 citations; 8,526 facilities; 58% of homes.
- F761 pharmacy services and medication storage issues: 13,856 citations; 8,798 facilities; 59.9% of homes.
Treat those numbers carefully. They do not mean every F880 is equivalent to every F689, or that a facility with one common citation is automatically unsafe. They do mean a facility that wants fewer surprises should build recurring proof around the short list that shows up most often in the public record.
Which high-frequency tags more often rise to actual harm or immediate jeopardy?
Frequency alone can mislead. KFF found that 5% of deficiencies in the last survey cycle rose to the level of actual harm or immediate jeopardy, totaling about 6,700 deficiencies, and about 28% of facilities had at least one such finding in 2026. Within the CMS-based national tally above, F689 stands out: 7,221 of its 21,911 citations were classified as actual harm or jeopardy, far higher than the harm counts attached to F880 or F812 in the same analysis.
That is why operators should separate 'common' from 'consequential.' Infection control and food sanitation deserve disciplined routine controls because they recur constantly. Accident prevention deserves that same routine discipline plus a more serious escalation path, because a break in supervision, transfer technique, environmental control, or care-plan follow-through can move from paperwork to resident harm quickly.
Why does F880 stay near the top year after year?
Because it depends on repeated behavior, not a one-time policy. CMS has repeatedly highlighted infection prevention as a major survey issue, and a CMS congressional justification noted that infection control deficiencies were the most common type of deficiency cited, with 82.5% of homes, or 14,742, receiving such a deficiency from 2019 through 2021. Earlier GAO work found a similar pattern, reporting infection prevention and control deficiencies in 13,299 homes, or 82% of all surveyed homes, from 2013 through 2017.
Operationally, F880 repeats because it lives in hand hygiene, PPE practice, isolation setup, cleaning process, competency reinforcement, and leadership follow-up. Those are shift behaviors. A policy can be current and the practice can still drift by Tuesday afternoon. That is one reason ePeople is more useful when it surfaces missing follow-through in daily work than when it acts like a static policy library.
Why does F689 deserve more attention than its rank alone suggests?
F689 is not just a falls tag. The CMS look-up describes it as a requirement to keep the facility free from accident hazards and provide adequate supervision to prevent accidents. In practice, that pulls together resident-specific risk assessment, device use, transfer technique, environmental checks, rounding, and response after a change in condition. The tag is broad, but that is exactly why it creates recurring exposure across departments.
Older oversight work also shows how long this problem has persisted. A GAO review identified accident prevention as one of the most frequently cited quality requirements nationwide and used it as a case example of real resident harm tied to missed prevention planning and follow-through. When a tag remains common over time and also shows a heavier share of harm-level findings, it belongs on an operator short list even if another tag appears more often overall.
What do F812, F656, and F842 tell an operator about ordinary process reliability?
They usually point to routine process failure, not exotic clinical complexity. F812 food procurement, storage, preparation, and service is a kitchen discipline signal. F656 care planning is usually a handoff and revision-control signal. F842 records is usually a documentation-timeliness and completeness signal. In California, those three all landed in the statewide top 10, with 2,477 F812 citations, 2,475 F656 citations, and 1,354 F842 citations.
That matters because these tags are usually fixable through ownership clarity. Dietary owns temperatures, dating, sanitation rounds, and proof. MDS and nursing leadership own whether assessment changes reach the comprehensive plan. HIM and department leads own standards for signatures, late entries, scanned records, and retrieval. When the owner is vague, the citation tends to repeat.
How should a SNF use frequency data without overreacting to every citation?
Start with three buckets instead of one master list. First, high-frequency routine tags such as F880 or F812 need recurring audit loops. Second, high-harm tags such as F689 need faster escalation and leadership review. Third, documentation-control tags such as F656 and F842 need revision tracking and proof retrieval. This structure is more useful than a generic survey board because it matches how work actually fails.
- Build a top-10 tag watchlist from public data, not just memory.
- Assign one executive owner and one day-to-day owner to each recurring tag.
- Tie each tag to a proof source: audit form, care-plan revision, training log, staffing record, or exception report.
- Review repeat tags monthly in QAPI instead of waiting for the annual mock survey.
- Use complaint-origin findings to test whether your escalation process works after hours and on weekends.
What does the survey context say about where these findings are discovered?
Routine surveys still matter more than many operators assume. KFF found that 87% of facilities received at least one standard survey in the prior 15 months, and about three in four deficiencies in the most recent survey cycle were identified during routine inspections rather than complaint surveys. That means a facility cannot rely on fixing only the issues that residents or families are most likely to report. Survey exposure is still largely built in ordinary daily operations.
For leaders, the implication is practical. If your prep model focuses only on response after an allegation, you will miss the tags that surveyors see during standard rounds: dated food, unlabeled items, inconsistent care-plan implementation, incomplete records, missed infection controls, and supervision gaps that are visible in ordinary resident care.
Where does ePeople fit if the public data already tells you the top tags?
The public file tells you where the industry gets cited. It does not tell you whether your building caught today's missed break in supervision, yesterday's undocumented care-plan revision, or this week's repeated hand-hygiene drift on one hall. That is the gap ePeople is meant to close. The value is not that it knows F-tags exist. The value is turning known repeat-risk areas into assigned work, reminders, escalations, and proof that supervisors can actually retrieve when surveyors ask.
A useful F-tag dashboard does not just rank risk. It shows who owns the next proof step, by shift and by department.
What should operators do with this data in the next 30 days?
Pick your five. For most SNFs, that means a local watchlist built around F880, F689, F812, F656, and F842 or the nearest equivalents in your recent survey history. Then test whether each tag has four things: a named owner, a current audit method, a retrieval path for proof, and a trigger for escalation when drift repeats. Public frequency data is most useful when it narrows attention. If it creates a longer to-do list without ownership, it has not helped operations.
