Clearpol analyzed every state and federal nursing-home survey in California from mid-2017 through July 2026 — 13,352 inspections and 53,837 F-tag citations across 1,030 facilities and all 53 surveyed counties. Then we did the part nobody else does: we read every citation and sorted it into the specific way care broke down. This is what the numbers say.
Most published survey statistics stop at the tag: how many F880s, how many F689s. That tells an operator what was cited, but not why, and never what to actually fix. Our dataset parses the narrative behind each citation — isolation precautions versus surveillance versus linen handling inside F880, hazard identification versus monitoring versus supervision inside F689 — so a tag count becomes a work assignment. Because the coverage runs statewide, the same lens applies to a single building, a county, or the whole state.
The headline trend
More citations, less harm
California surveys have gotten heavier on paper and lighter on blood. Deficiencies per survey climbed four years running — 3.11 in 2023, 3.63 in 2024, 3.81 in 2025, and 3.94 year-to-date in 2026, a 27% rise. Over the same window the share of surveys reaching Immediate Jeopardy fell from 2.5% to 1.3%, and the harm-level rate roughly halved.
The modern California survey is closer to a thorough documentation exam than a hunt for catastrophe. Surveyors are writing more findings and touching more tags per visit, but a smaller share of those findings rise to actual harm. That is good news for residents and a planning signal for operators: the volume you must manage is growing even as the odds of a single visit turning into a jeopardy citation shrink.
Who knocks on the door
Two very different kinds of visit
Three out of four California surveys are complaint investigations — 9,984 of them — and they are light: an average of 1.7 deficiencies, a median of one. The annual recertification survey is the opposite animal. Standard and infection-control surveys are only about a quarter of all visits but average 11 deficiencies each, and they generate roughly 69% of every citation in the state. Only 2.7% of recerts come back clean.
The complaint visit is where jeopardy risk lives. The recert is where the citation load lives. They demand two different playbooks.
The severity ladder
Where the citations actually land
Of the 46,855 severity-rated citations, D and E — no-harm-but-more-than-minimal — account for 82%. Actual-harm citations (G and above) are 2.2% of the total, and Immediate Jeopardy (J, K, L) is 0.8%. The practical read: the volume game is won on systems and documentation at the D/E level, while the survival game is the roughly 2% of findings that reach harm.
One detail matters for anyone who reaches the top of the ladder: when jeopardy is found, it is often not isolated. Of 376 IJ citations, 30% were scoped K or L — pattern or widespread — rather than a single isolated event.
Immediate Jeopardy anatomy
Where jeopardy actually comes from
Accidents and inadequate supervision (F689) drive 85 of the 376 IJ citations — nearly a quarter. Quality of care (F684) and abuse/neglect (F600) follow at 31 each, and infection control (F880) at 26. But two findings deserve more attention than their rank suggests. The medication-safety cluster — significant med errors, 5%+ error rates, and pharmacy services combined — totals 39 IJ citations, which would rank second on its own. And CPR (F678) is the sleeper: 25 jeopardy citations from failure to initiate resuscitation, training gaps, and advance-directive confusion, despite barely registering on any top-tag list.
Jeopardy is also a repeat-offender phenomenon, not random lightning. Nineteen percent of facilities have logged at least one IJ — but the 91 buildings with two or more account for 72% of every IJ citation in the state.
Screen your portfolio for repeat-IJ buildings and manage them individually — averages hide the risk. Then drill the specific jeopardy pathways that don't show up in tag counts: supervision and elopement, medication administration, and CPR/advance-directive readiness. Those three account for far more jeopardy than their citation volume implies.
The system view
The same failures, different tag numbers
Because we categorize the narrative inside each citation, a pattern emerges that pure tag counts hide: a handful of operational failure modes generate most of the volume across completely different F-tags. Policy and procedure gaps drive 3,272 citations spanning infection control and pharmacy. "Monitoring absent" drives 2,594 across accident and nutrition tags. Risk assessment, intervention failure, and storage sanitation each cut across multiple tags the same way.
Fix the system, not the tag. A single working monitoring routine, risk-assessment process, or policy-adherence check prevents citations under five or six different F-tags at once. This is the strongest argument in the data for building a prevention loop rather than chasing individual deficiencies after the fact.
The playbook
A 20-tag playbook covers half of everything
California surveyors cited 205 distinct F-tags over nine years, but the distribution is steeply concentrated: the top 10 tags account for 37% of all citations, the top 20 for 54%, and the top 50 for 80%. A focused 20-tag playbook covers half of everything you're likely to face.
Here are the ten tags that anchor that playbook statewide:
| F-Tag | What it covers | Citations | % of all |
|---|---|---|---|
| F880 | Infection prevention & control | 3,171 | 5.9% |
| F812 | Food storage / prep sanitation | 2,477 | 4.6% |
| F656 | Comprehensive care plan | 2,475 | 4.6% |
| F689 | Accident hazards / supervision | 2,428 | 4.5% |
| F684 | Quality of care | 1,941 | 3.6% |
| F755 | Pharmacy services | 1,923 | 3.6% |
| F761 | Label / store drugs | 1,814 | 3.4% |
| F842 | Resident records | 1,354 | 2.5% |
| F550 | Resident rights | 1,262 | 2.3% |
| F658 | Professional standards of care | 1,197 | 2.2% |
But not all tags carry the same stakes. When abuse/neglect (F600) is cited it reaches harm level one time in five; F689 nearly one in six. Meanwhile tags like resident records (F842), room size (F912), and environment (F814) have never been cited above F level in nine years. Same word — "deficiency" — completely different consequences.
Geography
Your county is not the state
The same federal rulebook produces very different results depending on where a building sits. Citation intensity ranges from about 2.8 deficiencies per survey in San Luis Obispo and Kern to 5.96 in Orange — a 2.2× spread between markets under identical regulations. District and surveyor culture is a real variable, and it is knowable.
We pulled the three largest county markets apart in their own reports — where the risk concentrates, which tags over-index locally, and what buildings there should drill: