Symbol Digit Modalities Test (SDMT) 

The Symbol Digit Modalities Test (SDMT) is a 90-second timed substitution task that measures processing speed, sustained attention, and visual scanning by having a patient pair symbols with digits using a printed key. It is widely used in multiple sclerosis and concussion care because it is brief, sensitive to mild processing-speed deficits, and easy to repeat serially.

What the SDMT Measures

The SDMT asks a patient to work through a row of abstract symbols, converting each one to its corresponding digit using a key at the top of the page, as quickly and accurately as possible. Unlike memory or language tasks, the SDMT isolates a fairly narrow but clinically important construct: how fast the brain can scan, match, and respond to simple visual information under time pressure. Because processing speed is one of the cognitive domains most consistently affected by mild traumatic brain injury, demyelinating disease, and a range of medical and psychiatric conditions, the SDMT functions as a general-purpose "check engine light" for cognitive slowing even when other domains (memory, language, executive function) test normally.

The test has two common administration formats:

• Written SDMT — the patient writes the corresponding digit beneath each symbol on paper.
• Oral SDMT — the patient speaks the digit aloud while an examiner records responses, which removes the motor/handwriting component and is useful when a patient has upper-extremity limitations or when isolating cognitive speed from motor speed matters.

Written and oral forms are not interchangeable — they draw on different normative datasets and can produce meaningfully different scores in the same patient, so a clinic should stay consistent with format across baseline and follow-up administrations for the same patient.

Administration Time

The timed portion of the SDMT is exactly 90 seconds per trial. Including instructions, a practice row, and recording the score, total administration time is typically around 5 minutes — one of the shortest tasks in a standard neurocognitive battery, which is part of why it is used so frequently for serial (repeat) monitoring rather than only at intake.

Scoring and Interpretation

The raw score is the number of correct symbol-digit substitutions completed within the 90-second window. That raw score is then converted to an age- and education-normed T-score or percentile using a published normative dataset (commonly the original Smith 1982 manual norms, or newer MS-specific normative datasets, depending on which reference the clinic has adopted).

There is no single universal cutoff that marks a raw SDMT score as "normal" or "impaired" in isolation — a raw score only becomes clinically meaningful once compared against age- and education-matched norms, and interpretation should always specify which normative dataset was used, since datasets differ and are not directly interchangeable. In practice, most clinics use the SDMT in one of two ways: (1) a single administration compared to normative percentiles to flag possible processing-speed impairment, or (2) serial administrations compared to the patient's own prior scores to track recovery trajectory after concussion or track decline/stability in a chronic condition. Because the test has a modest practice effect, clinicians should be cautious about attributing a small improvement on repeat testing entirely to genuine recovery rather than familiarity with the task.

Clinical Use Case

The SDMT is used across several patient populations relevant to Kavera's modules:

• Concussion / post-concussion syndrome — screening for processing-speed slowing that patients often describe as "brain fog," and tracking whether that slowing resolves along an expected recovery timeline.
• Return-to-play and return-to-learn decisions — as one data point (not a standalone gatekeeper) supporting return-to-play criteria and return-to-learn planning, since processing-speed deficits can affect classroom and sideline readiness even when symptoms have largely settled.
• Multiple sclerosis and other conditions affecting processing speed — the SDMT's strongest historical evidence base is in MS, where it is a standard outcome measure; concussion-care and cognitive-health practices have adopted it because the same processing-speed sensitivity generalizes well to mild TBI and other causes of cognitive change.
• General cognitive-health monitoring — as a brief repeat-testable component within a broader battery for patients presenting with subjective cognitive complaints of unclear cause.

The SDMT is a screening and monitoring tool, not a standalone diagnostic instrument — an abnormal result should prompt either more comprehensive neuropsychological testing or clinical correlation, not a diagnosis on its own.

Billing: Which CPT Code Covers the SDMT

The SDMT is not billed as a standalone code. It is billed as part of a neurocognitive test administration under the standard psychological/neuropsychological testing CPT family, based on who administers it and how it is delivered:

• 96136 / 96137 — administration and scoring by a physician or other qualified health care professional (QHP), first 30 minutes / each additional 30 minutes, when the SDMT is administered as part of a battery of two or more tests.
• 96138 / 96139 — the same administration and scoring performed by a trained technician, first 30 minutes / each additional 30 minutes.
• 96146 — automated administration and scoring, applicable when the SDMT is delivered and scored through a computerized/digital platform rather than by a live administrator.

Which code applies depends on the delivery method and who performed the administration, and is governed by the applicable CMS billing and coding article (opens in a new tab) and current AMA CPT guidance (opens in a new tab) — verify current payer policy before billing, since coverage and documentation requirements vary by payer and can change annually. For practices building a broader RTM workflow around between-visit cognitive monitoring, see RTM 98978 and the RTM for cognitive-behavioral monitoring guide.

Where SDMT Sits in the Kavera Protocol

Within Kavera's assessment battery, the SDMT contributes the processing-speed data point inside the Cognitive Health module and is also commonly used within the Concussion module to track recovery of cognitive speed between in-office visits. Because it is short and repeatable, it fits naturally into between-visit monitoring workflows — generating the kind of serial, time-stamped cognitive data that supports both clinical decision-making and RTM/neurocognitive testing documentation.

FAQ

How long does the SDMT take to administer?

The timed trial itself is 90 seconds. With instructions and a practice item, total administration time is typically around 5 minutes.

Is there a pass/fail score on the SDMT?

No. The SDMT produces a raw substitution count that must be converted to an age- and education-normed score before it has clinical meaning — a raw number alone cannot be labeled "normal" or "impaired."

Can the SDMT be used to track recovery after concussion?

Yes, serial administration comparing a patient's score over time to their own prior scores is a common use, though clinicians should account for practice effects when interpreting improvement on repeat testing.

What CPT code do I bill for the SDMT?

There is no standalone SDMT code. It is billed under 96136/96137 (physician/QHP administration), 96138/96139 (technician administration), or 96146 (automated/computerized administration), depending on how it was delivered.

Is the written or oral version better?

Neither is universally "better" — they draw on different norms and isolate slightly different skills (oral removes the handwriting/motor component). The key is administering the same format consistently for a given patient across repeat testing.

How Kavera Handles This

SDMT is a core between-visit cognitive test in the Concussion and Cognitive Health modules. Kavera delivers it on your cadence, scores it, and plots processing speed against baseline. Serial SDMT is one of the clearest recovery signals in concussion; the record supports 96136–96139 and feeds RTM documentation. Self-Serve practices run this with their own staff. On Managed, Juliet Mott's team runs it and bills it under your credentials.