Every administrator knows the survey is coming. Very few can say, with data, what they are most likely to be cited for.

We analyzed every health survey deficiency CMS recorded nationally in 2026 through August 23. That is 44,016 surveys across 14,721 facilities, producing 77,508 deficiencies. The full breakdown is in our 2026 Nursing Home Survey Risk Report. One pattern inside it deserves more attention than it gets, because it runs straight through the assessment office.

Four tags, one process

Together, 8,649 citations. About one in nine deficiencies cited in the country this year.

Most facilities read that table as four separate problems assigned to four separate people. Surveyors do not read it that way. They investigate a single chain: did the assessment match the resident, did the CAAs trigger what they should have, did the care plan follow, and were both revised when the resident changed.

That is why an assessment problem rarely stops at F641. F641 is the smallest number in the table and the one furthest upstream. When the coding is wrong or late, the finding moves downstream into care planning and quality of care, where the surveyor widens the sample and the scope and severity climb with it.

Why scope matters more than the tag

Across all 77,508 deficiencies cited this year, 65.8% were scored at D: no actual harm, isolated. Another 20.6% were scored at E, meaning the surveyor found a pattern rather than a single missed task.

That distinction is the whole game. An isolated D is a correction. A pattern E says the failure is systemic, and systemic failures escalate on the next survey. Assessment errors are particularly prone to being scored as patterns, because coding practice is rarely wrong on one chart only. If the coordinator misunderstood a Section GG item, the surveyor will find it on the second chart and the third.

There is a practical corollary. Scope and severity are assigned by the survey team, not by a formula. The difference between an isolated D and a pattern E often comes down to how many residents the team sampled and what the facility could produce while they were standing there. Documentation you can retrieve in two minutes is worth more than documentation you can retrieve in two hours.

The Risk-Based Survey raises the stakes on data accuracy

Beginning September 8, 2026, CMS is implementing the Risk-Based Survey process nationwide under memo QSO-26-14-NH. Qualifying facilities receive a more focused recertification survey, conducted in roughly half the time with fewer surveyors, and a high-performing facility icon on Care Compare.

CMS estimates roughly 12% of nursing homes initially qualify. Eligibility is built on overall star rating, staffing performance, survey history, and accurate data reporting.

Read that last criterion carefully. Accurate data reporting is MDS submission. Your star rating is largely built from quality measures, and your quality measures are calculated from MDS coding. Two of the four eligibility factors run through the assessment office.

This changes the conversation about MDS accuracy. It has always been a compliance and reimbursement issue. As of September, it is also a public-facing designation that families will see when they are deciding where to place a parent.

The same coding sets your rate

The MDS is not only a survey document. It is the source of your PDPM classification across all five components.

Under-coding costs revenue you already earned. Over-coding invites an audit you do not want. Both start in the same place: a coordinator working from an incomplete record, without time to chase down the supporting documentation before the window closes.

In the MDS audits we conduct, we typically identify $10,000 to $20,000 over a year of assessments that a facility was entitled to and did not capture. Findings vary by facility, census, and current coding practice, but the pattern is consistent enough that we now offer the audit at no charge. Most operators have never had anyone outside the building validate their coding against the chart.

What a real MDS audit looks at

If you are evaluating your own assessments, or evaluating a consultant who wants to do it for you, these are the checks that matter.

• ARD accuracy and timing. Are assessment reference dates set correctly, and are assessments completed and transmitted inside the window? Late is a finding regardless of how good the coding is.

• Certification and diagnosis validation. Does the primary diagnosis support the PDPM classification, and is the physician certification present and dated correctly?

• Section GG. Is functional coding supported by what therapy and nursing actually documented, or is it an estimate entered on the last day of the window?

• CAA logic. Did the triggered CAAs get addressed, and does the summary show clinical reasoning rather than template language?

• Care plan linkage. Does the care plan reflect what the assessment triggered? This is where F641 becomes F656.

• Revision on change of condition. When a resident declined, was the assessment updated and the care plan revised? This is F657 and it is one of the easiest findings for a surveyor to make.

• Interview items. Were resident interviews conducted, and by someone who knows how to conduct them?

• Quality measure impact. Which coded items are driving your worst-performing QMs, and are those items accurate?

If you cannot answer most of these for your last thirty assessments, that is your starting point.

Who is most exposed

Three situations account for most of the assessment problems we find.

The coordinator who learned on the job. Competent nurse, no formal MDS training, inherited the role when someone left. They are usually doing everything they were shown, and what they were shown was already wrong.

The solo coordinator. One person carrying the entire assessment schedule with no backup and no second set of eyes. Errors are not the result of carelessness. They are the result of arithmetic.

The vacancy. The position is open, and assessments are being completed by whoever has capacity that day. Every week that continues, the backlog compounds and the coding drifts.

None of these are moral failures. They are staffing realities. But CMS does not score intent.

Start with a free MDS audit

SMK Medical offers a no-cost MDS audit. We review a sample of your assessments and tell you what we find: revenue and PDPM capture, the coding driving your quality measures and star rating, and the ARD, CAA and care plan gaps that become F641, F656, F657 and F684 on a 2567.

No obligation, and no sales cycle.

Call (313) 314-8267 to schedule, or visit MDSWizard.com for our full range of MDS training, auditing and completion support.

When you need the system fixed, not just the assessment

An audit tells you what is wrong today. Holding the correction in place is a different problem.

SurveyGuard is our compliance and survey readiness platform for post-acute care, built on the same data as the report referenced above. Digital mock surveys aligned to the CMS pathways, AI-assisted Plans of Correction written for 2567 acceptance, audit tracking with named accountability, grievance and QAPI management, and a command center showing entrance readiness and active risk across every building you oversee. HIPAA compliant and EMR-integrated.

About the author

Demetrius D. Kirk, DNP, MBA, MSN, RN, LNHA, FACHCA has worked in post-acute care since 2007, from bedside nursing through nursing home administration to service as a hospital Chief Administrative Officer. He leads SMK Medical, a national MDS, mock survey and compliance risk management practice serving more than 100 facilities, and is the founder of SurveyGuard. He is an approved continuing education provider and maintains an expert witness practice in long-term care litigation. His doctoral work produced an 86.6% reduction in resident falls through structured risk assessment.

Download the full 2026 Nursing Home Survey Risk Report for every tag ranked, the complete scope and severity distribution, enforcement and penalty data by CMS region, and a ten-question self-audit framework.

Data source: CMS CASPER Citation Frequency and Deficiency Count Reports, health surveys, January 1 through August 23, 2026. Risk-Based Survey details from CMS memo QSO-26-14-NH, issued July 16, 2026.