Deciding how much help a patient needs looks simple until you have to write it down. Is that transfer "min assist" or "contact guard"? Does a verbal cue count as assistance at all? And why does the same patient get scored one way on your evaluation and a completely different way on the Medicare item set?
The levels of assistance in occupational therapy are the shared language we use to answer those questions, and getting the language right drives everything downstream: your goals, the care team's picture of progress, and whether the visit gets paid.
Here is the short version. Two systems are in play today. The traditional levels of assistance (Independent, Modified Independent, Supervision or Standby, Contact Guard, Minimal, Moderate, Maximal, and Dependent) grew out of the Functional Independence Measure (FIM) and still dominate day-to-day clinical notes. Since 2019, Medicare's post-acute assessments instead use CMS Section GG, a six-point scale with its own rules. This guide defines every level, gives you a quick-reference chart, shows how the two systems line up (and where they deliberately do not), and covers the piece most articles skip entirely: cueing and cognitive assistance.
A level of assistance answers one question: how much of a task did the patient do, and how much did you do for them? Every scale you will encounter is really tracking three things at once.
That last one trips people up. A patient can be physically independent and still need substantial help, because the help they need is cognitive. A level of assistance is not only about muscle. It is about the total support a person requires to complete a meaningful occupation safely, and that is exactly why occupational therapy leans on these scales for daily documentation and goal writing.
Rule of thumb: score what the patient actually did during the activity, not what you believe they could do on a good day. Levels of assistance describe observed performance, not potential.
Here is the traditional scale at a glance, ordered from most to least independent. The percentages describe the share of the effort the patient contributes, and they come from the FIM scoring anchors - not from CMS or AOTA, a distinction we return to below. Definitions for each level follow the chart.
| Level of Assistance | Patient Contributes | Hands-On Contact? | Typical FIM Score |
|---|---|---|---|
| Independent (I) | 100%, safely and in reasonable time | No | 7 |
| Modified Independent (Mod I) | 100%, but uses a device, extra time, or has a safety concern | No | 6 |
| Supervision / Standby Assist (SBA) | 100%, with cueing or setup only | No | 5 |
| Contact Guard Assist (CGA) | ~100%, hands-on only to prevent a loss of balance | Yes - protective, not lifting | ~4 (touching boundary) |
| Minimal Assist (Min A) | 75% or more | Yes - incidental "touching only" help | 4 |
| Moderate Assist (Mod A) | 50% to 74% | Yes | 3 |
| Maximal Assist (Max A) | 25% to 49% | Yes | 2 |
| Total Assist / Dependent (D) | Less than 25%, or unable to participate | Yes - helper does most or all | 1 |
You will see slightly different percentage boundaries depending on the source. The canonical FIM manuals put Moderate at 50 to 74%, while the American Academy of Physical Medicine and Rehabilitation's functional assessment table writes it as 50 to 75%. The variance is trivial in practice, but it is worth knowing the ranges are conventions, not a single federally fixed formula.
Levels of assistance never float free. They always attach to a specific occupation, and in practice that occupation is usually an activity of daily living. It helps to be precise about the two families of tasks, because the type of task changes the kind of help a patient needs.
Basic activities of daily living (BADLs) are the self-care tasks we learn first in life. The commonly used clinical set includes bathing, dressing, toileting, transferring, continence, feeding, and grooming. The classic instrument here is the Katz Index of Independence in Activities of Daily Living, which scores six basic self-care items. According to the National Library of Medicine's StatPearls overview of activities of daily living, the term was coined by Sidney Katz in the 1950s and remains a foundation of functional assessment.
Instrumental activities of daily living (IADLs) are the more complex tasks that let a person live independently in a community: managing finances, managing medications, meal preparation, housekeeping, transportation, shopping, and communication. The Lawton Instrumental Activities of Daily Living Scale (Lawton and Brody, 1969) is the standard measure, and the American Occupational Therapy Association's Occupational Therapy Practice Framework (OTPF-4) formally defines both ADLs and IADLs within the profession's scope of practice.
Why it matters: a patient can be Independent with feeding but require Moderate Assist for medication management. Always tie the level of assistance to the named task, never to the person as a whole.
Now the definitions. These are the labels you will write in a treatment note or use to set a goal in acute and outpatient settings.
The patient completes the entire task safely, in a reasonable amount of time, with no help, no device, and no supervision. This is full independence with no qualifiers.
The patient still does 100% of the task without another person, but one or more qualifiers apply: they use an assistive device or adaptive equipment, they need more than a reasonable amount of time, or there is a safety consideration. A patient who dresses independently using a reacher and a sock aid is Modified Independent.
The patient performs the whole activity themselves, but you need to be right there for cueing, coaxing, or setup, without any physical contact. Standby assist (SBA) and supervision live here. Think of a patient who can complete a tub transfer but needs verbal reminders to lock the wheelchair and reach for the grab bar first.
The patient still does essentially 100% of the work, but you keep a hand on them or on a gait belt to catch a loss of balance if it happens. The contact is protective, not assistive - you are not lifting or moving them. This is the single most confused pair on the whole scale, because contact guard (CGA) and standby (SBA) look almost identical except for one thing: whether you are touching the patient. We break the distinction down in depth in our guide to contact guard assist vs standby assist.
Heads up: a gait belt is a guarding and steadying tool, not a lifting device. If you are using it to hoist a patient's body weight, you have moved well past contact guard into hands-on assistance and your documentation should say so.
The patient performs 75% or more of the task, and you provide incidental, "touching only" help to finish it. A steadying hand at the hips during the last few degrees of a sit-to-stand is a classic minimal assist.
The patient does 50% to 74% of the work, and you provide the rest. You are actively helping to lift, hold, or support a limb or the trunk, but the patient is still contributing more than half of the effort.
The patient contributes 25% to 49% of the effort. You are doing more than half of the work. The patient participates meaningfully but cannot carry the task.
The patient contributes less than 25% of the effort, or cannot participate at all, and the helper performs most or all of the task. In the Medicare item set, this level also captures activities that require two or more helpers.
The Functional Independence Measure (FIM) is a seven-point ordinal scale built from 18 items (13 motor and 5 cognitive), with a total score that ranges from 18 to 126. It is the source of the percentage anchors in the chart above, and it is where the familiar "min, mod, max" vocabulary comes from. The FIM is a proprietary instrument owned and licensed by the Uniform Data System for Medical Rehabilitation (UDSMR), so it still exists and is still used under license - it is not a free public-domain tool.
Here is the part that surprises clinicians who trained a decade ago. Medicare no longer uses the FIM in the inpatient rehabilitation setting. In the FY2019 Inpatient Rehabilitation Facility Prospective Payment System Final Rule, CMS removed the FIM instrument and its function modifiers from the IRF Patient Assessment Instrument. Per the Federal Register (CMS-1688-F), that removal applies to IRF discharges occurring on or after October 1, 2019. The FIM did not disappear from clinical life, but it stopped being the federal standard for post-acute payment.
It is also worth a note of humility about the FIM's precision. Research on the measure has found that agreement on individual item scores between raters can be modest. A study of FIM item scores across rehabilitation settings reported only fair agreement for individual items, a reminder that these levels are structured clinical judgments, not laboratory measurements. Clear, consistent documentation matters precisely because the scoring has some built-in subjectivity.
Section GG is the standardized functional assessment CMS now uses across all post-acute care settings, a change mandated by the IMPACT Act of 2014. The same Section GG items appear on the Minimum Data Set in skilled nursing facilities, the IRF-PAI in inpatient rehab, the LTCH CARE Data Set in long-term care hospitals, and OASIS in home health, so a patient's function is measured the same way as they move between settings.
Section GG uses a six-point scale for self-care and mobility. The verbatim CMS code definitions, drawn from the CMS Section GG coding guidance and reproduced in AOTA's Section GG functional assessment form, are:
| Code | Label | What It Means |
|---|---|---|
| 06 | Independent | Patient completes the activity safely with no help from a helper. |
| 05 | Setup or Clean-Up Assistance | Helper only sets up or cleans up before or after; patient does the activity. |
| 04 | Supervision or Touching Assistance | Helper provides verbal cues and/or touching or steadying and/or contact guard assistance. |
| 03 | Partial / Moderate Assistance | Helper does less than half the effort (lifts, holds, or supports). |
| 02 | Substantial / Maximal Assistance | Helper does more than half the effort. |
| 01 | Dependent | Helper does all the effort, or two or more helpers are required. |
When an activity does not happen, Section GG uses four "activity not attempted" codes instead: 07 (patient refused), 09 (not applicable), 10 (not attempted due to environmental limitations), and 88 (not attempted due to medical condition or safety concerns).
Two features of Section GG catch clinicians off guard, and both are the direct opposite of how the FIM works:
The current Section GG item definitions live in the IRF-PAI manual, Version 4.4, effective October 1, 2026. The six-point wording has been stable across recent versions, so notes written to the code definitions above remain accurate.
This is the piece most levels-of-assistance articles leave out, and it is the question OTs ask most often: how do you score a patient whose body works fine but whose brain needs help? A patient with a traumatic brain injury may be physically capable of dressing yet unable to sequence the steps without prompting. That prompting is assistance.
The clearest evidence that cueing counts comes straight from CMS. Section GG code 04 names "verbal cues" in the same breath as physical touching and contact guard, which means the federal scoring system explicitly treats a spoken prompt as a form of assistance, not as independence. If a patient needs you to talk them through a task, they are not independent at it.
Beyond that anchor, the profession documents cueing along three dimensions rather than a single fixed formula:
A caution: unlike the FIM percentages, there is no official CMS or AOTA percentage scale for cueing. If you see "minimal cue equals 1 to 25%" charts online, treat those as one clinic's documentation convention, not a national standard. Describe the cueing you gave in plain, specific language and you will never be wrong.
The honest answer is that it depends on your setting and your payer, and many clinicians end up fluent in both systems.
What you should not do is invent a one-to-one crosswalk between the FIM and Section GG. CMS has never published an official mapping between the two, and for good reason: they measure effort differently (percentages versus "more or less than half") and they group levels differently (Section GG merges standby and contact guard into 04). Any alignment you see, including the FIM column in the chart above, is an approximate conceptual guide to help you think in both languages - not a billing formula. Code each system on its own terms.
A level of assistance on its own is a conclusion. Skilled documentation shows the work behind it. The weak versions below state a level; the strong versions justify it, which is what defends the note in an audit and what our guide to writing an OT SOAP note walks through in full.
| Level | Weak (States the Level) | Strong (Justifies the Level) |
|---|---|---|
| Standby Assist | "Patient required SBA for grooming." | "Patient completed grooming at the sink independently with SBA for standby cueing to initiate each step; no physical contact required." |
| Contact Guard | "CGA for transfers." | "Patient performed stand-pivot transfer wheelchair to bed with CGA at the gait belt for balance loss at initiation; patient completed the movement without physical assistance." |
| Moderate Assist | "Mod A for lower body dressing." | "Patient required Mod A (patient contributing an estimated 60% of effort) to don pants while seated, with therapist supporting the right lower extremity through hip flexion secondary to weakness." |
| Max Assist (cognitive) | "Max verbal cues for meal prep." | "Patient required maximal verbal and gestural cues to sequence a 3-step meal-prep task, needing a prompt at each transition to prevent skipped steps; physically independent throughout." |
The level is the headline. The evidence - what the patient did, what you did, and why - is the story that makes the headline defensible.
Contact guard assist means the patient performs essentially all of a task while the therapist keeps a hand on them or on a gait belt to prevent a loss of balance. The contact is protective, not assistive - if you are lifting or supporting body weight, it is no longer CGA.
Standby assistance, also called supervision, means the patient completes the task independently but you remain close for safety, cueing, or setup, without any physical contact. The absence of touch is what separates SBA from contact guard.
Maximal assistance (Max A) means the patient contributes roughly 25% to 49% of the effort and the therapist provides more than half. The patient participates but cannot carry the task on their own. On the FIM this is a score of 2; on Section GG it corresponds to code 02.
FIM scoring is the Functional Independence Measure, a seven-point scale (1 for total assistance up to 7 for complete independence) applied to 18 self-care, mobility, and cognitive items, for a total score between 18 and 126. It is the origin of the min, mod, max percentage bands, though Medicare no longer uses it on the IRF-PAI.
There is no single fixed count - you will see anywhere from four to eight levels named depending on the source. A common short list is Independent, Supervision or Standby, Minimal, Moderate, and Maximal (with Dependent as the floor). The fuller traditional scale in this guide adds Modified Independent and Contact Guard, and Section GG uses six numbered codes. The concepts are consistent even when the count is not.
Not on the IRF-PAI. CMS removed the FIM from the inpatient rehabilitation assessment for discharges on or after October 1, 2019, and replaced it with Section GG. Some facilities still use the FIM under a UDSMR license for their own outcomes tracking.