OASIS Manual · For clinicians

OASIS Skip Patterns and Branching Logic

When to skip an item, when to score it, and how branching logic works in OASIS-E2.

By Reza Djangi, OTR/L·7 min read·Aligned with OASIS-E2 v1.00.0

CMS source

OASIS-E2 Guidance Manual · Verified 2026-04-26

TL;DR

A big chunk of OASIS-E2 is conditional. Whether you score an item, skip it, or enter a dash depends entirely on how an earlier "gateway" item was answered.
Three categories of skip patterns dominate: timepoint-specific items (only at SOC/ROC, only at Discharge), gateway-driven branches (score the parent, then maybe its children), and "activity not attempted" codes on the GG section.
The single most damaging skip-pattern mistake is filling in a dash on an item that's actually required at this timepoint. Dashes are tracked. Too many dashes flag your record on CASPER.

What are skip patterns?

Skip patterns are the conditional rules that determine which OASIS items must be completed at a given timepoint based on the response to a prior item. Items that are skipped per the rules should be left blank rather than scored or dashed.

Source: OASIS-E2 Guidance Manual — Ch. 3, p. 3.1 · Verified 2026-04-26

Timepoint-specific items

Each OASIS item is tagged with the timepoints at which it is collected. Items collected only at SOC and ROC (such as several care preferences and risk-screen items) must be left blank on Recert, Transfer, and Discharge records. Items collected only at Discharge (outcome-oriented items) are not collected at SOC or ROC.

Source: OASIS-E2 Data Specifications — OASIS-E2 Item Set, Sec. 1 · Verified 2026-04-26

Response-triggered branching (gateway items)

Several OASIS items act as gateway items: a specific response to the gateway item determines whether subsequent items are required, optional, or skipped. For example, if the pain assessment indicates no pain, the frequency-of-pain follow-up item is not required.

Source: OASIS-E2 Guidance Manual — Ch. 3, p. 3.4 · Verified 2026-04-26

From the field

The classic example is M1240 (Pain Assessment) gating M1242 (Frequency of Pain Interfering with Activity). If M1240 indicates no pain, M1242 doesn't apply — leave it blank, don't score it 0. Scoring M1242 = 0 when M1240 says "no pain" creates a contradiction the data validation engine will flag.

Another familiar gateway: the pressure ulcer cascade rooted at M1311. If the patient has no Stage 2 or higher pressure ulcers, the entire chain of follow-up items about ulcer characteristics doesn't apply. Trying to "complete the form" by scoring those children with 0 or NA tells CMS something different from "the patient has no qualifying pressure ulcer."

In Section D, the depression screen gates the rest of the depression items. If the screen is negative or the patient declines, you don't follow through with the longer instrument.

NA vs. not assessed vs. dash (—)

The dash response (—) indicates that the item was not assessed and that the agency was unable to obtain the necessary information. The dash is not equivalent to "not applicable" (NA) and should be used only as a last resort, since records with excessive dash use are flagged in agency-level data quality reports.

Source: OASIS-E2 Guidance Manual — Ch. 3, p. 3.5 · Verified 2026-04-26

From the field

This trips up new clinicians more than anything else on the OASIS. Three different things look similar but mean different things:

NA (Not Applicable) — the item doesn't apply because of the patient's situation. Patient has no caregiver, so caregiver-status questions are NA.
Skipped per branching — the item was bypassed because a gateway answer made it irrelevant. Leave blank, don't dash.
Dash (—) — you tried to assess and couldn't. Patient was nonverbal, no proxy available, item was unscoreable. This is the "last resort" answer.

The mistake I see most: clinician hits a question they're not sure how to answer and dashes through it. Dashes are tracked at the agency level on CASPER, and a clinician with a high dash rate stands out fast. Either the answer is NA, or the question was supposed to be skipped, or you actually need to go back and assess.

GG items and skip logic

GG self-care and mobility items use a 6-point scoring scale (06 Independent through 01 Dependent) plus four "activity not attempted" codes (07, 09, 10, 88). Each "activity not attempted" code reflects a different reason — refused, environmental limitation, medical condition, or activity not applicable — and each is a valid response, not a skip.

Source: OASIS-E2 Guidance Manual — Ch. 3, p. 3.20 · Verified 2026-04-26

From the field

GG items are different from M-items in one important way: there's almost no skipping. Each of the GG self-care items (eating, oral hygiene, toileting hygiene, etc.) and mobility items (sit-to-stand, walking 10 feet, etc.) is collected at SOC/ROC and at Discharge. If the patient didn't perform the activity, you use one of the "activity not attempted" codes — you don't leave it blank, and you don't dash.

The most frequent GG error I see: clinician marks a 1 (Dependent) when the patient didn't actually attempt the activity. The right answer was 09 (not attempted due to medical condition) or 88 (not applicable). Marking it 1 inflates dependency and skews the GG quality measures.

And don't confuse GG items with the M1830/M1840/M1850/M1860 ADL items — they are scored on a different scale and they are not interchangeable. Both sets are still on OASIS-E2.

What software does for you and what it doesn't

Common errors

OASIS Assessment Schedule — assessment-specific items depend on which RFA is in use
QA Workflow, Locking, and Corrections — catching skip pattern errors before transmission
OASIS Glossary — definitions for NA, dash, and branching terminology