%201.svg)
The Centers for Medicare and Medicaid Services (CMS) uses OASIS (Outcome and Assessment Information Set) functional scoring to determine how much Medicare pays for each 30-day home health episode. Within the Patient-Driven Groupings Model (PDGM), every patient is assigned a functional impairment level, Low, Medium, or High, based on responses to eight OASIS items. That level directly determines the case-mix weight applied to the episode payment (CMS PDGM Overview). An inaccurate functional score misrepresents the patient's clinical condition, assigning the episode to the wrong payment group and the wrong LUPA (Low Utilization Payment Adjustment) threshold for the entire 30-day period.
The CY 2026 Home Health Prospective Payment System Final Rule recalibrated PDGM case-mix weights and updated functional impairment levels using CY 2024 claims data. The financial consequence of OASIS scoring errors is more precisely reflected in payment under the recalibrated model than at any prior point since PDGM launched in January 2020. This is because functional scoring is not just a clinical but also a documentation risk now. In addition, CMS has reduced Medicare home health payments by 1.3% in CY2026 compared to CY2025, placing further pressure on already-tight agency margins.
Agencies whose clinicians are completing GG items from patient self-report rather than structured observed performance are systematically undercoding functional impairment, receiving lower case-mix weights than the patient population warrants, and losing revenue without generating a denial that would surface the problem.
This guide covers how OASIS functional scoring works under PDGM, what the eight scoring items measure, how points translate into impairment levels, where scoring errors originate, and what operational quality assurance (QA) processes prevent them.
For a broader view of how OASIS documentation connects to PDGM payment, refer to the Complete Guide to Home Health Coding and OASIS Review in 2026. For OASIS-E2 instrument changes effective April 2026, refer to the OASIS-E Documentation Guide.
Key Takeaways
- OASIS functional scoring under PDGM is determined by eight OASIS items across GG0130 (Self-Care) and GG0170 (Mobility). Each item is scored on a 6-point scale. Points are summed and compared against clinical group-specific thresholds to assign Low, Medium, or High functional impairment level.
- The functional impairment level is one of five variables ( Clinical Group, Functional Impairment Level, Comorbidity Adjustment, Admission Source and Timing, LUPA Threshold) that determine the HIPPS (Health Insurance Prospective Payment System) code and the episode payment group under PDGM. Functional impairment levels and their associated case-mix weights were updated in the CY 2026 Final Rule using CY 2024 claims data.
- GG scoring requires documented observation of actual task performance. CMS guidance requires clinicians to score based on what the patient demonstrates, not what the patient reports. Patients with pain, fatigue, or cognitive impairment consistently report higher functional capability than they demonstrate under observation.
- The most consistent OASIS scoring error across home health agencies is not a misunderstanding of the rating scale. It is completing GG items from patient self-report rather than structured performance observation, which systematically produces lower impairment scores and lower case-mix weights.
- OASIS scoring errors are rarely caught by claim scrubbers. They do not generate denials. They generate silent underpayment that surfaces only in PEPPER report outlier data, MAC audit findings, or retrospective denial pattern reviews.
- LUPA thresholds are set per HIPPS code. When OASIS functional scoring errors lower the functional impairment level, they change the HIPPS code and lower the LUPA threshold, increasing the risk of LUPA payment without any change in actual visit patterns.
How OASIS Functional Scoring Works Under PDGM
Under PDGM, functional impairment scoring is a three-step process:
- First, clinicians score eight specific OASIS items using a standardized 6-point rating scale.
- Second, the responses are assigned point values and summed to produce a total functional score.
- Third, that total score is compared against the functional impairment thresholds for the patient's specific clinical group to assign Low, Medium, or High impairment level (CMS PDGM Functional Impairment).
The thresholds differ by clinical group because resource use patterns differ across patient populations.
This three-step process means the functional impairment level is the mathematical output of how eight specific items are scored, not an independent clinical impression. A clinician who scores GG0170C (Lying to sitting on side of bed)) as 03 (Partial Assistance) rather than 02 (Substantial Assistance) because the distinction feels marginal may shift the total point score across a threshold and change the impairment level for the episode. That scoring decision carries a payment consequence that the clinician completing the assessment is unlikely to see.
The Eight OASIS Items That Feed Functional Scoring
CMS specifies eight OASIS items from Sections GG0130 (Self-Care) and GG0170 (Mobility) that contribute to the PDGM functional impairment score (CMS OASIS-E2 Manual, Section GG). These eight items are used in PDGM functional level assignment:
These eight items are not the complete GG section. GG0130 and GG0170 contain additional items that are used for quality reporting and HHVBP (Home Health Value-Based Purchasing) scoring but do not feed directly into PDGM functional impairment level assignment. Clinicians completing the full GG section should understand which items affect payment grouping and which affect quality reporting, as the scoring disciplines for each purpose differ.
The 6-Point Rating Scale
Each of the eight scoring items uses the same standardized 6-point rating scale (CMS GG Self-Care and Mobility Decision Guide). The scale measures the type and amount of assistance required for the patient to complete the activity:
Four additional codes apply when the activity was not attempted: 07 (Patient refused), 09 (Not applicable), 10 (Not attempted due to environmental limitations), and 88 (Not attempted due to medical condition or safety concerns). When one of these codes is used, no points are contributed from that item to the functional score. Agencies should ensure that activity-not-attempted codes are used only when genuinely applicable and documented, as systematic use of these codes suppresses the functional score in a way that affects payment grouping.
How Points Translate to Impairment Levels
Points are assigned based on the inverse of the score value. A score of 01 (Dependent) contributes the highest points, reflecting the highest resource use. A score of 06 (Independent) contributes zero or minimal points. The points from all eight items are summed to produce the total functional score for the period.
That total score is then compared against the functional impairment thresholds for the patient's PDGM clinical group. Because different clinical groups have different patient populations with different functional profiles, the point thresholds for Low, Medium, and High impairment differ across the 12 clinical groups. A score that qualifies as Medium impairment in the Musculoskeletal Rehabilitation group may qualify as Low in the Behavioral Health group, reflecting the different resource use patterns CMS observed in those populations.
The CY 2026 Final Rule updated these clinical group-specific thresholds using CY 2024 claims data. Agencies that have not reviewed their patient population's functional score distribution against the updated 2026 thresholds may be working from assumptions about which scores produce Medium versus High impairment that no longer reflect the current payment model.
GG0130: Self-Care Scoring in Detail
GG0130 covers self-care activities. The three items within GG0130 that feed the PDGM functional score are Eating (GG0130A), Oral Hygiene (GG0130B), and Shower/bathe self (GG0130E). Other items in GG0130, including Grooming, Dressing Upper Body, Dressing Lower Body, Bathing, and Shower or Tub Transfer, contribute to quality reporting and HHVBP scoring but do not feed directly into the PDGM functional impairment level calculation.
Assessment Requirements for GG0130
GG0130 items must be completed based on what the patient actually does during the assessment period, not what the patient reports they can do or what the clinician estimates based on the patient's general presentation. CMS guidance requires the clinician to set up the activity and observe the patient's performance before assigning a score.
The assessment period for GG0130 at Start of Care (SOC) is the first 5 calendar days of the home health episode. Clinicians are not required to observe every activity on the day of the SOC assessment. They may gather information across multiple visits during the 5-day window. However, the score must reflect the patient's actual demonstrated performance, not a clinical estimate.
The most significant scoring error in GG0130 is assigning a score based on patient self-report during the SOC visit without structuring the visit to include observation of the actual tasks. A patient who reports being able to eat independently but who, when observed, requires supervision for aspiration risk, should be scored 04 (Supervision or Touching Assistance), not 06 (Independent). The difference of two score points can shift the item's contribution to the total functional score meaningfully across a threshold.
Red Road Insight: The GG0130 self-care items that generate the most scoring inconsistency across agencies are not the ones involving physical effort. They are the ones involving safety judgment, specifically GG0130A (Eating) for patients with dysphagia risk and GG0130E (Shower/bathe self) for patients with fall risk. Clinicians consistently score independence in these activities without documenting the safety assessment that would support that score, which leaves the agency without a defensible basis if the scoring is reviewed.
Escalation Level: Immediate: Any GG0130 score assigned without documented observed performance in the clinical notes should be corrected before pre-billing QA sign-off. A score that cannot be traced to documented observation is not defensible in an ADR response.
GG0170: Mobility Scoring in Detail
GG0170 covers mobility activities. The five items within GG0170 that feed the PDGM functional score are Roll Left and Right (GG0170B), Lying to sitting on side of bed (GG0170C), Sit-to-stand (GG0170D), Chair/bed-to-chair transfer (GG0170E), and Toilet Transfer (GG0170F). Additional items in GG0170, including walking activities and wheelchair mobility, contribute to quality reporting but do not feed directly into the PDGM functional impairment calculation.
Assessment Requirements for GG0170
GG0170 mobility items require the same observed performance standard as GG0130. The clinician must structure the SOC visit or the 5-day assessment period to observe the patient attempting each activity. A clinician who scores GG0170E (Chair/bed-to-chair transfer) based on asking the patient how they get up from a chair without observing the patient attempt the transfer is not meeting the assessment standard CMS requires.
The challenge with GG0170 is that mobility assessment in the patient's home environment introduces variables that are not present in an institutional setting. A patient may perform a chair transfer more independently in a familiar chair than in a clinical demonstration environment or more dependently due to pain or fatigue on the day of assessment. CMS guidance addresses this issue by asking clinicians to score the patient's usual performance, not their best or worst performance on a single day. When functional status varies, the clinician should use clinical judgment to determine what level of assistance the patient typically requires, supported by observations across the assessment period.
The Sit-to-stand Error Pattern
GG0170D (Sit-to-stand) is the mobility item most frequently scored inaccurately across home health agencies. The error pattern is consistent: clinicians observe the patient pushing up from a chair with upper extremity support and score the activity at 05 (Setup Assistance) or 04 (Supervision), when the actual level of effort and safety risk the clinician is providing during the transfer warrants a score of 03 (Partial Assistance) or 02 (Substantial Assistance).
The distinction between 04 (Supervision or Touching Assistance) and 03 (Partial Assistance) is whether the helper is providing only guidance and safety oversight, or whether the helper is actively contributing physical effort to help the patient complete the transfer. Clinicians who are physically guiding a patient through a Sit-to-stand are contributing effort that qualifies as Partial Assistance, but they score it as Supervision because they perceive the patient as doing the majority of the work. This scoring decision consistently understates the patient's functional impairment.
Escalation Level: Short-term: Agencies should audit GG0170E scores against the narrative clinical notes for the prior 60 days for any clinician where Chair/bed-to-chair transfer is consistently scored at 04 or 05. If the notes describe active clinician contact during the transfer, the GG0170E scores require review.
OASIS Functional Scoring and HHVBP Quality Measures
OASIS functional scoring affects not only PDGM payment but also performance under the expanded Home Health Value-Based Purchasing (HHVBP) model. The Discharge Function Score quality measure, which became the primary functional outcome measure in the HHVBP model from CY 2025, calculates the percentage of home health patients who achieve a risk-adjusted expected functional score at discharge.
The Discharge Function Score uses 11 GG items: three from GG0130 (Eating, Oral Hygiene, Toilet Hygiene) and eight from GG0170 (Roll Left and Right, Lying to Sitting on Side of Bed, Sit-to-stand, Chair/Bed-to-Chair Transfer, Toilet Transfer, Walk 10 Feet, Walk 50 Feet with 2 Turns, and Wheel 50 Feet with 2 Turns). These 11 items overlap significantly with the 8 items that feed the PDGM functional impairment level, which means inaccurate scoring at SOC affects both payment grouping and the baseline from which discharge functional improvement is measured.
When SOC functional scores are lower than the patient's actual capability, the baseline appears to show greater impairment than is present. This may appear to improve the HHVBP outcome score if the discharge score reflects actual capability, but it creates a documentation inconsistency that auditors can identify. When SOC scores are higher than actual capability because the clinician scored based on self-report, the HHVBP baseline is optimistic, and the discharge improvement score may be understated.
Agencies above the HHVBP volume threshold face a payment adjustment of up to 5% based on HHVBP performance, making GG scoring accuracy a direct financial variable in two separate payment mechanisms simultaneously (CMS HHVBP Model).
Common OASIS Scoring Errors and Their Payment Consequences
OASIS scoring errors in home health agencies follow predictable patterns. Understanding the specific error types and their payment consequences allows agencies to target QA processes at the highest-impact points rather than treating all GG items with equal review effort.
Error 1: Scoring from Self-Report Without Observed Performance
This is the most prevalent OASIS scoring error across home health agencies. The clinician completes the SOC visit, interviews the patient about their functional abilities, and assigns GG scores based on what the patient reports. The clinical notes document the patient's self-report rather than the clinician's observation of actual task performance.
CMS guidance is explicit that GG scores must reflect what the patient actually does, not what the patient says they can do. Patients with pain, fatigue, cognitive impairment, or anxiety about their clinical situation consistently over-report their functional capability. A patient who reports being able to transfer independently but who, when observed, requires substantial assistance is a patient whose GG score based on self-report will understate impairment by two to three scoring levels.
Error 2: Inconsistency Between GG Score and Narrative Notes
The second consistent error is a mismatch between the GG score recorded in the OASIS and the description of the assessment in the clinical narrative. A clinician scores GG0170F ( Toilet Transfer) as 04 (Supervision) while the note describes standing next to the patient, guiding arm placement, and providing physical support during the transfer. The note documents Partial Assistance but the score reflects Supervision.
This inconsistency is one of the most frequently cited findings in Medicare Administrative Contractor (MAC) ADR (Additional Documentation Request) reviews because it is visible in the record without requiring clinical judgment to identify. When the coded GG score and the narrative description do not match, the reviewer treats the mismatch as a documentation integrity problem.
Error 3: Using Activity-Not-Attempted Codes as Default Responses
When a clinician cannot complete the GG assessment for a specific item, activity-not-attempted codes are available: 07 (Patient refused), 09 (Not applicable), 10 (Environmental limitations), and 88 (Medical condition or safety concerns). These codes contribute zero points to the functional score. When used appropriately, they represent genuine assessment limitations. When used as defaults for difficult or time-pressured assessments, they suppress the functional score and reduce the PDGM impairment level below what the patient's clinical status warrants.
Error 4: Scoring Usual Performance Without Documenting the Basis
CMS guidance asks clinicians to score usual performance when functional status varies across the assessment period. Scoring usual performance requires documentation of the observations that support the determination. A clinician who scores GG0170D ((Sit-to-stand)) as 03 because the patient performed better on day 2 of the assessment period than on day 1 must document both observations and the basis for determining that the day 2 performance reflects usual status. Without that documentation, the score is not auditably defensible.
Red Road Insight: The audit risk from OASIS scoring errors is rarely triggered by the score itself. It is triggered by the gap between the score and what the clinical record supports. Agencies that score accurately but document the basis for the score consistently resolve MAC reviews faster and with lower denial rates than agencies where the score and the notes tell different stories.
Escalation Level: Monthly: Agencies should run a monthly comparison of GG score distributions against the clinical narrative documentation across all SOC assessments. When the percentage of 04, 05, or 06 scores across GG0170 items exceeds the expected distribution for the agency's clinical population, it indicates a systematic scoring pattern that warrants structured review.
Building a QA Process for OASIS Functional Scoring Accuracy
Effective QA for OASIS functional scoring in home health requires a structured process that operates at three points in the episode timeline. Retrospective sampling after claims are submitted identifies patterns after the payment consequence has already occurred. Concurrent pre-submission review is the mechanism that prevents incorrect scores from reaching the claim.
At Start of Care
SOC QA for OASIS functional scoring should include a comparison of each GG score against the corresponding narrative documentation. For each of the eight scoring items, the reviewer should confirm that the note describes an observed assessment, not a self-report interview, and that the level of assistance described in the note is consistent with the score assigned. A GG0170E (Chair/Bed-to-Chair Transfer) score of 05 (Setup/Cleanup Assistance) paired with a note documenting that the clinician physically supported the patient's weight while transferring from the bed to the chair is a scoring error. The note describes Substantial Assistance (02), not Setup/Cleanup Assistance, and should be corrected before the SOC OASIS is submitted.
SOC QA should also confirm that activity-not-attempted codes are accompanied by the clinical basis documented in the note. Code 88 (Not attempted due to medical condition or safety concerns) requires documentation of the specific medical condition or safety concern that prevented the assessment. Code 07 (Patient refused) requires documentation of the refusal and any attempt to obtain the patient's participation.
At Recertification
Recertification functional scoring should be compared to the SOC baseline and to the clinical documentation from the intervening visits. Functional scores that improve significantly between SOC and recertification without corresponding documentation of functional progress in visit notes are a pattern that reviewers identify as inconsistent documentation. Functional scores that remain unchanged when visit notes document significant functional improvement are equally problematic from a quality reporting perspective.
Pre-Billing
The pre-billing QA checkpoint for functional scoring should include a final confirmation that the GG scores on the claim are consistent with the OASIS submission and that the OASIS submission is consistent with the clinical notes. It should also confirm that the HIPPS code produced by the submitted scores reflects the correct functional impairment level for the patient's clinical group. The PDGM Grouper Tool (CMS PDGM Grouper Software), updated to v07.1.26 for April 2026, should be used to validate the HIPPS code before submission. A claim submitted with an incorrect HIPPS code due to functional scoring errors will be paid at the wrong rate without generating a denial.
QA Cadence
How External OASIS Review Addresses Functional Scoring Gaps
For home health agencies managing high SOC volumes across multiple clinicians and branches, concurrent pre-submission review of every GG score against the clinical documentation is operationally demanding. The assessment burden at SOC is significant, and the time pressure on clinicians completing assessments in the patient's home environment creates conditions where self-report scoring becomes the default rather than the exception.
An external OASIS review function operating as a concurrent pre-submission checkpoint provides the review capacity that agency-level QA at volume cannot consistently maintain. For OASIS functional scoring specifically, the review function compares each GG score against the narrative documentation, flags narrative-score mismatches for correction before submission, tracks scoring patterns at the clinician level to identify systematic undercoding, and confirms that the functional impairment level produced by the scores reflects the patient's documented clinical status.
Agencies that have implemented concurrent OASIS review also have access to scoring accuracy data at the clinician, branch, and clinical group level that identifies where GG scoring drift is concentrated. This data supports targeted re-training on the specific items and scoring distinctions where errors are most frequent rather than broad-spectrum training that addresses everything equally. More detail on how OASIS assessment accuracy connects to coding and payment is in the 7 Home Health Coding Errors guide and the OASIS-E2 vs. OASIS-E1 guide.
The Bottom Line
OASIS scoring home health agencies rely on is the foundation of PDGM payment. The eight GG items that determine functional impairment level are the clinical data points that set the case-mix weight and LUPA threshold for every 30-day episode. Under the CY 2026 recalibrated model, the payment consequence of scoring those items from self-report rather than observed performance is more precisely reflected in the episode payment than at any prior point since PDGM launched.
Agencies that treat OASIS functional scoring accuracy as a daily operational discipline, with concurrent QA that confirms observed performance documentation before submission, accumulate less scoring error and less payment exposure than agencies that treat GG scoring as a clinical judgment left entirely to the assessing clinician without structured review.
How Red Road Supports OASIS Functional Scoring Accuracy
Sit-to-stand scoring is where the largest gap consistently shows up between what the clinical notes describe and what the GG0170D score reflects. A nurse documents that the patient required a two-step verbal cue before initiating the movement and paused mid-transfer for balance, then scores GG0170D as 05 (Setup/Cleanup Assistance). Verbal cueing and safety monitoring during an unsteady transfer reflects Supervision (04) at minimum, not Setup/Cleanup Assistance. The distinction matters because GG0170D carries significant weight in the total functional score for several clinical groups.
The same pattern shows up in GG0130. The self-care items that generate the most scoring inconsistency are the ones involving safety judgment, specifically GG0130A (Eating) for patients with dysphagia risk and GG0130E (Shower/bathe self) for patients with fall risk, rather than the ones involving physical effort. And the audit risk from scoring errors overall comes less from the score itself and more from the gap between the score and what the clinical record supports. Agencies that document the basis for their scores consistently resolve MAC reviews faster than agencies where the score and the notes tell different stories.
Red Road's OASIS review service provides Registered Nurse clinical reviewers who validate every GG score against the corresponding clinical documentation before billing. Red Road has reviewed more than 2 million charts and coded more than 2.5 million charts at a consistent 98% accuracy rate. For OASIS functional scoring specifically, each review confirms that the 8 scoring items reflect documented observed performance, that narrative descriptions are consistent with the scores assigned, that activity-not-attempted codes are accompanied by the documented clinical basis, and that the functional impairment level produced by the scores is consistent with the patient's documented clinical status.
Scoring accuracy data is tracked at the clinician level and the item level, allowing agencies to identify whether GG scoring errors are concentrated in specific items (such as GG0170E Chair/bed-to-chair transfer or GG0170F (Toilet Transfer), specific clinicians, or specific branches. This item-level tracking makes corrective re-training more targeted and more effective than aggregate accuracy reporting alone supports.
This content reflects CMS guidance and regulatory standards as of 31 July 2026. Agencies should verify current requirements against the most recent CMS OASIS and PDGM guidance and MAC bulletins, and consult their compliance advisors for guidance specific to their operations.
.svg){kind=icon}
