Reading MDS 3.0 Like a Plaintiff Attorney

MedLegal AI Editorial · April 18, 2026 · 12-minute read

The Minimum Data Set 3.0 is a facility's federally-mandated, sworn assessment of a resident's needs. It is also the single most concentrated document for detecting the gap between what the facility claimed was true and what the contemporaneous care records actually show. Attorneys who learn to read it forensically — section by section, against the nursing notes and the Braden scale — win cases their peers settle.

What the MDS 3.0 is, and what it claims

Every Medicare or Medicaid-certified skilled nursing facility is required to complete an MDS 3.0 assessment for every resident on admission, at scheduled intervals (5-day, 14-day, 30-day, 60-day, 90-day), quarterly, annually, and on significant change. The assessment is electronically submitted to CMS. It drives reimbursement through the PDPM case-mix model. It triggers Care Area Assessments that drive care plans. It is signed under penalty of perjury by the MDS coordinator and the facility's administrator.

The MDS is, legally and practically, a statement the facility is making about what the resident needed. That is what makes it useful to a plaintiff attorney. If the MDS says the resident was at low risk for pressure ulcers and two weeks later the resident had a stage III heel ulcer, one of three things is true: the MDS was wrong, the contemporaneous records are wrong, or the resident's condition changed catastrophically in a window the facility should have caught with more frequent reassessment.

All three of those are litigable. Reading the MDS well is what lets you identify which one is in front of you.

The sections that matter most

A full MDS 3.0 has twenty sections (A through Z with omissions). Most of them are administrative or demographic. The forensic work concentrates in five sections.

Section G — Functional Status / ADL

Records how much assistance the resident needs with bed mobility, transfers, locomotion, dressing, eating, toilet use, personal hygiene, and bathing. Scored 0 (independent) through 4 (total dependence), with separate columns for "self-performance" and "support provided."

Plaintiff-relevant gaming patterns:

• Spontaneous improvement before MDS cutoff. ADL scores that drop (improve) by 1 point across multiple categories in the 2-3 day window before a scheduled MDS. This inflates PDPM reimbursement by classifying the resident as needing less care than the clinical trajectory supports. Cross-check against nursing notes from the same days. If the notes don't describe any intervention that would have produced the improvement, the improvement is suspect.
• Static scores across catastrophic changes. The opposite pattern: the resident has a documented stroke, fall, or hospitalization, and the post-event MDS shows ADL scores unchanged from pre-event. This is the facility failing to trigger a significant-change assessment — independently actionable under F-Tag 637.
• Bed mobility 0-1 with stage III+ pressure ulcer. Bed mobility scored as "independent" or "supervision only" on a resident who developed a sacral or heel pressure ulcer requiring pressure-redistribution surfaces is internally contradictory. Either the bed mobility scoring is wrong or the pressure-ulcer care plan was inadequate.

Section I — Active Diagnoses

Checkbox list of active medical conditions. Simple on its face; devastating when the omissions are read in context. Plaintiff-relevant questions: Was diabetes mellitus checked on admission and then unchecked later, conveniently before a lower-extremity amputation? Was a dementia diagnosis omitted on a resident who was chemically restrained? Was CHF omitted on a resident who was allowed to over-hydrate to the point of pulmonary edema?

What makes Section I powerful is that omissions propagate into care planning. A diagnosis that isn't checked doesn't generate the CAA that would have triggered the care plan intervention. "We didn't know" becomes a harder defense when the diagnosis is in the hospital transfer summary, the H&P, and the physician orders — but not in Section I.

Section M — Skin Conditions

The section plaintiff attorneys doing pressure-ulcer cases read first. Records presence, stage (I-IV, unstageable, suspected deep-tissue injury), site, size (length × width × depth in cm), tissue type in the wound bed, exudate, and whether the ulcer was present on admission or facility-acquired.

Plaintiff-relevant gaming patterns:

• Systematic under-staging. Reporting a stage III ulcer as stage II, or a stage IV as "unstageable" to avoid F-Tag 686 survey deficiencies. The contemporaneous wound-care nursing notes, which describe the same ulcer on the same date, usually reveal the mismatch.
• Present-on-admission fabrication. Ulcers that appear on Section M as "present on admission" when the hospital transfer summary and admission skin assessment show no such finding. This is one of the easiest-to-prove misrepresentations in the MDS if you have the transfer records.
• Omission of deep-tissue injury (DTI). Purple discoloration under intact skin, characteristic of forming pressure injury, often omitted from Section M because DTIs are more difficult to explain away than an open stage II ulcer.

Section K — Swallowing/Nutritional Status

Weight, weight loss, nutritional approaches (feeding tube, IV, mechanically altered diet), swallowing disorders. Matters disproportionately in dehydration, malnutrition, and aspiration cases. Plaintiff attorneys should check:

• Weight trajectory over 30, 90, 180 days — documented weight loss of 5% in 30 days or 10% in 180 days should have triggered a significant-change MDS and a nutrition-focused CAA. Weight loss documented in the chart but not in Section K is both evidentiary and independently a deficiency.
• Swallowing problems flagged in speech therapy notes but not reflected in Section K, on a resident who died of aspiration pneumonia.
• "Mechanically altered diet" coded as No while the diet order in the chart says pureed — small detail, but these contradictions accumulate and give the jury a pattern.

Section J — Health Conditions

Pain frequency and intensity, falls and fall-related injuries, prognosis conditions. For falls cases, Section J is the trigger for the falls CAA. A resident with multiple falls in a quarter whose Section J reports "no falls" is either a charting error or a concealment. For pain cases, Section J is where under-assessment of pain in non-verbal dementia patients shows up — and where the facility's failure to use behavioral pain-assessment tools (PAINAD, for example) becomes visible.

Cross-checking — where the evidence actually lives

No single MDS tells you the truth. The forensic work is the comparison.

Practical workflow for a records review

1. Extract every MDS in the chart. Admission, 5-day, 14-day, each quarterly, each significant change, annual, and discharge if applicable. Build a side-by-side table with one column per MDS and one row per key section field.
2. Flag any score change without a clinical event. Anywhere a Section G, K, or M field changes between two adjacent MDSes, look for the clinical event that would explain the change. If there isn't one, circle it.
3. Pull the 7 days of nursing notes before each MDS. Paralegal-level task. The assessment is supposed to reflect the 7-day look-back window immediately before the assessment reference date (ARD). If the notes contradict the assessment, that's the discrepancy.
4. Cross-check Braden specifically against Section M. Any resident with a Braden >18 who has a Section M pressure ulcer entry is a candidate for the chart-gaming argument. Either the Braden is over-rated or the care response was inadequate.
5. Check for missed significant-change MDSes. Every time the chart shows a catastrophic event — fall with injury, hospitalization, new stage III+ ulcer, major diagnosis — the rule requires a significant-change assessment within 14 days. If one is missing, that's a federally-required document the facility didn't produce.

Steps 1-5 are mechanical. They can be done by hand in 4-6 hours for a 10,000-page chart, or in 30-60 minutes of compute with a tool that's been specifically tuned on MDS parsing. Either way they produce the same output: a structured list of discrepancies each of which is either evidence or requires explanation.

What this buys you at trial

Jurors do not know what the MDS is. They do understand "the facility signed a sworn document saying the resident needed X care, and the care records from the same week show the facility was providing less than X." The MDS discrepancy is the piece of evidence that turns an abstract duty-of-care argument into a concrete documentation contradiction.

Defense experts will explain that MDS completion is imperfect, that the 7-day look-back is squishy, that coding rules are confusing. All true. None of it explains a 4-point ADL improvement that coincides exactly with a reimbursement window and doesn't correlate with any documented clinical intervention. Jurors can do arithmetic. They cannot be talked out of it by an expert witness.

What the MDS is not

Not a substitute for the full medical record. Not a standalone basis for a case theory. Not a substitute for standard-of-care expert testimony. The MDS is a concentrated evidentiary vein — but it needs to be connected to the complete record, the depositions, and the expert opinions to do its work at trial.

A firm that reads MDS well catches cases the firm down the street takes and loses, and declines to take cases the firm down the street takes and settles for nuisance value. Either outcome is a compounding edge.