A cognitive screen is a small arithmetic problem with clinical stakes. You administer the SLUMS, add up the component points, then read the total against a cutoff that shifts depending on the patient's education. Get the addition right, pick the correct cutoff, and the same number means normal for one patient and mild cognitive impairment for another. It is exactly the kind of step that is easy to do and easy to do slightly wrong at the end of a long visit.
In most charts the screen lands as a sentence. "SLUMS 24, likely mild NCD." The number is there, but the parts that produced it are not. Which components did the patient miss? Was the cutoff the high-school band or the lower one? And when you screen the same patient again in six months, the only way to compare is to find the old note and read the two sentences side by side.
Where the cognitive screen score usually ends up
The trouble is that a SLUMS result is structured data trapped in prose. Eight components, each with its own maximum, sum to a total out of 30, and that total maps to one of three interpretation bands through a cutoff that depends on education level. Written as free text, all of that collapses into a single number and a guess at the band.
That makes two ordinary tasks harder than they should be. The first is getting the interpretation right in the moment, because the education-adjusted cutoff is a detail that is simple to forget. The second is tracking change over time, which is the whole point of screening a patient more than once. A two-point drop matters, but only if you can line up this visit's components against the last one, and free text does not line up.
How do you record a SLUMS cognitive screening score in an EHR?
In Oli Health you record it with the Cognitive Screen block inside a progress note. You enter the eight SLUMS component scores and select the patient's education level, and the block sums the total out of 30 and assigns the interpretation band for you. The result is stored as structured data, not as a sentence, so every component, the total, and the band are kept for each visit.
What the Cognitive Screen block captures
The block lives inside a progress note. You do not open a separate assessment tool. Type /cognitive-screen (or just /cog) and pick Cognitive Screen from the Common group.
What drops in is a structured form, not a paragraph. The instrument is SLUMS, shown as a fixed field. You set one selector, Education Level, with two options: High School or More, and Less Than High School. That single choice is what decides the interpretation cutoff, so it is the one thing you confirm before you read the result.
The exam itself is eight numeric fields, each capped at its real SLUMS maximum:
- Orientation, 0 to 3
- Calculation, 0 to 3
- Animal Naming, 0 to 3
- Delayed Recall, 0 to 5
- Digit Span, 0 to 2
- Clock Drawing, 0 to 4
- Shape Recognition, 0 to 2
- Story Recall, 0 to 8
Each field holds whole numbers only and will not accept a value above its cap, so a slip of the keyboard cannot push Clock Drawing to a 7. There is a Notes field for anything the numbers do not carry, and the block stamps its own date and time so the screen is pinned to when it was done.
The total and interpretation are computed, not typed
You never add up the SLUMS yourself, and you never type the band. As you enter the components, a summary strip in the header keeps a running total out of 30, a count of how many of the eight components you have filled in, and the interpretation badge for the current score.
The interpretation follows the education-adjusted SLUMS cutoffs. For a patient with high school or more, 27 to 30 reads as normal, 21 to 26 as mild neurocognitive disorder, and 20 or below as dementia. For a patient with less than high school, the bands shift down: 25 to 30 normal, 20 to 24 mild NCD, and 19 or below dementia. Change the Education Level selector and the band updates against the right cutoff on its own.
Because the total and the band are derived from the components, they cannot drift away from them. The block computes the total, the maximum, and the interpretation from the eight scores and the education level, and a hand-edited total that does not match the parts is rejected rather than saved. The number in the chart is always the number the components produce.
Does the SLUMS cognitive score adjust for a patient's education level?
Yes. The Cognitive Screen block applies the education-adjusted SLUMS cutoffs. For a patient with high school or more, 27 to 30 is normal, 21 to 26 is mild neurocognitive disorder, and 20 or below is dementia. For a patient with less than high school, the bands are 25 to 30, 20 to 24, and 19 or below. Selecting the education level changes which cutoff the interpretation uses, automatically.
Less typing with AI charting
The block is also wired into Oli's AI charting. When the Cognitive Screen block is part of your note template and you record the encounter, the AI reads the transcript and fills the block in as it drafts the rest of the note. If you walk through the SLUMS aloud and call out the points, the component scores land in the right fields and the education level and administration mode are captured alongside them. The total, the maximum, and the interpretation band are then derived on the server from those scores, the same way they are when you type them by hand.
As with everything in Oli, this is a draft you check, not an entry that posts itself. You open a note that is already filled in, read it against what you remember of the exam, fix anything the model misheard, and sign off. Nothing reaches the chart without your review.
Can AI charting fill in a cognitive screen from a visit recording?
Yes. If the Cognitive Screen block is in your note template and you record the visit, Oli's AI charting reads the transcript and populates the eight SLUMS component scores, the education level, and the administration mode as it drafts the note. The total and interpretation band are then computed from those scores. The clinician reviews the draft, corrects anything inaccurate, and signs off before it becomes part of the chart.
Every screen kept as structured clinical data
Capturing one clean screen is useful on its own. Capturing every screen the same way is what makes a second and third screen worth something. Because each result is stored as structured data, the components, the total, the band, the education level, and the date, a repeat screen is directly comparable to the one before it rather than a sentence you have to find and reread.
The block also records an administration mode, marking whether the SLUMS was clinician administered or entered from a score obtained elsewhere, so a result imported into the chart is never confused with one performed in the room. When the block is used inside a neurodevelopmental or functional workflow, its severity feeds the ND Matrix and Timeline header, so a cognitive result sits next to the rest of the picture for that patient instead of off on its own.
There is no separate trend card for cognitive screens today, so this is data retention and comparability rather than a charted line. What you get is a clean, structured record of each screen, captured the same way every time, ready to read against the next one.
Where the cognitive screen block fits
Anywhere a SLUMS gets done and the result needs to last. Primary care and geriatrics running baseline and follow-up screens across years. Memory and cognitive clinics where the components, not just the headline number, are the clinical detail. Functional and naturopathic practices tracking cognition alongside the rest of a patient's chart. In each case the value is the same: the exam becomes structured data the moment you enter it, and the interpretation is computed against the right cutoff every time.
It pairs naturally with the rest of the structured chart. A cognitive result reads differently next to an active problem list that carries dementia or mild cognitive impairment across visits, and it sits in the same kind of structured field as the objective vitals you trend over time. It is the same idea behind capturing a validated score the same way each visit so the change is provable, applied to cognition.
None of this diagnoses anything. The block is a record you complete and review each visit, the interpretation follows the published SLUMS cutoffs rather than any judgment of its own, and nothing the AI drafts reaches the chart until you confirm it. Oli keeps the screen structured, scored, and comparable. The clinical call stays yours.
Try it on one patient's chart
The fastest way to see it is to run a SLUMS on one patient, enter the eight components, and watch the total and the interpretation band settle as you go. Flip the education level and see the band move against the right cutoff. Next visit, screen again and read the two results as structured records instead of two sentences in two notes.
You can do that on Oli Health for free. Practitioners who create an account before August 31, 2026 get the full platform free for the life of the account, with the clinical blocks included rather than locked behind a tier.
Create your free-forever account and try it on a real patient, or book a live demo and we'll walk you through it.
