Outcome Instruments in Functional Neurosurgery

Evidence status. Thresholds vary by study, baseline severity, intervention, and anchor. This page teaches interpretation and does not reproduce copyrighted scale items. Use licensed, official instruments and current scoring manuals. Scale ranges, subscale names, and scoring structure are facts and are stated here; item wording and printed forms are the copyrighted part and are not.

1.Absolute, percent, and clinically important change

Absolute change preserves scale units. Percent change is intuitive but unstable near low baseline values. A minimal clinically important difference is an anchor- or distribution-based estimate of meaningful change, not a universal biological cutoff. Response and remission answer different questions.

Always report baseline and follow-up distributions beside a responder percentage. A patient can cross a threshold by one point; another can improve substantially and remain just outside it.

2.State and rater matter

Parkinson motor scores depend on medication and stimulation state. Dystonia ratings depend on video angle, task, duration, and whether disability or pain is included. Depression and OCD scales depend on trained interviewing and time frame. Trigeminal neuralgia scores may disagree with patient-reported pain interference.

3.MDS-UPDRS

The Movement Disorder Society revision has four parts and requires permission for use and standardized rater training. Part I covers nonmotor experiences of daily living and runs 0 to 52, Part II motor experiences of daily living 0 to 52, Part III the motor examination 0 to 132, and Part IV motor complications 0 to 24, for a total of 0 to 260. Carry those denominators: a six-point change means one thing on a 132-point motor examination and something else on a 24-point complications score. DBS studies commonly emphasize Part III but should also include Part II daily living, motor complications, cognition, gait/falls, medication dose, and adverse effects. Specify the four medication/stimulation combinations when the scientific question requires them. Anchor-based minimal-change estimates for Part III exist in the Parkinson literature; take the value from the primary source you intend to cite rather than from any secondary summary, including this one.

4.Y-BOCS and MADRS

For OCD, the international consensus combines a Y-BOCS reduction of at least 35% with CGI-I 1 or 2 for at least one week to define response. Remission means no longer meeting diagnostic criteria for at least one week; when a structured interview is unavailable, the operational alternative is Y-BOCS 12 or lower plus CGI-S 1 or 2 for at least one week. A Y-BOCS-only endpoint should be identified as such. The empirical support is close but not identical. Signal-detection work in 288 patients pooled from four randomized trials found that a reduction of at least 35% best predicted response, that a posttreatment score of 14 or lower best predicted response by global impression, and that 12 or lower best predicted wellness (remission together with good quality of life and adaptive functioning), a stricter construct than remission alone. A 2024 individual-patient meta-analysis of 25 randomized trials and 1,235 participants found the optimal empirical thresholds were lower, at 30% for response and 15 for remission, but recommended keeping the consensus definitions because contemporary trials enroll more refractory patients. Prespecify the consensus values, report continuous change beside them, and state which definition you used: moving the line from 35% to 30% moves patients across it.

For depression, trials conventionally use a 50% MADRS reduction for response and a score of 10 or lower for remission. These are common conventions with empirical validation literature, rather than universal cutoffs. Studies anchored to global severity or diagnostic remission have produced different MADRS thresholds; state whether the protocol uses less than 10 or 10 or lower, since they are not identical. These conventions facilitate comparison; they do not substitute for function, durability, or safety.

5.TWSTRS, BFMDRS, and BNI

TWSTRS runs to 85 points and separates severity, disability, and pain. Report the components, because the same total can arise from different clinical changes. Clinically important change on TWSTRS is anchored, not intrinsic. The most-cited estimate comes from CD PROBE, a 479-patient real-world registry of onabotulinumtoxinA for cervical dystonia: against the patient global impression of change, total-score improvements of 8, 9, and 11 points corresponded to minimally improved, much improved, and very much improved, and the authors settled on 8 points as the minimal clinically important change. Two cautions before carrying that number into a DBS clinic. It was derived under chemodenervation, in patients whose mean baseline total was 39, and an anchor-based threshold does not travel cleanly to a different intervention, a different baseline severity, or a different follow-up interval. CD PROBE also found the total mapped to global impression better than any single subscale, which is a reason to report the total, not a reason to stop reporting severity, disability, and pain separately, since the same total change can be all pain or all severity.

TWSTRS itself has been revised. The Comprehensive Cervical Dystonia Rating Scale packages a modified TWSTRS-2, with subscales for motor severity, disability, and pain, alongside a psychiatric screening module and a disease-specific quality-of-life measure, and it has been clinimetrically tested. If you are designing a cervical dystonia protocol now, choose deliberately between classic TWSTRS, which is what the historical DBS literature is expressed in, and the comprehensive scale, which measures more of what the patient notices.

For generalized and segmental dystonia the instrument is the Burke-Fahn-Marsden Dystonia Rating Scale, and nearly every published DBS number you will meet is a BFMDRS movement percentage. It carries a movement subscale and a separate disability subscale. Record both at baseline and report both: they diverge as fixed deformity accumulates, and a cohort whose movement score falls by two-thirds while disability barely moves has told you something about those patients that the movement score alone conceals.

The BNI pain intensity score is efficient for trigeminal neuralgia because it folds medication into the category: I, no pain and no medication; II, occasional pain not requiring medication; IIIa, no pain on medication; IIIb, pain adequately controlled on medication; IV, pain not adequately controlled; V, no relief. Two success definitions circulate and they are not the same number. Grade I is pain-free off medication. Grades I through IIIb mean adequate control with or without medication, and that is the band most radiosurgical and percutaneous series report. Read the definition before you read the percentage. The scale remains ordinal, so add attack frequency, triggers, complications, and patient-reported quality of life; facial numbness has its own separate BNI score and belongs in its own column rather than folded into the pain grade.

6.Pain, seizures, tremor, and spasticity

Pain. Intensity on a 0 to 10 numeric rating scale is the near-universal primary, and the near-universal responder definition is a 50% reduction, with 30% widely treated as a smaller but still meaningful tier. Both are conventions carried over from analgesic drug trials rather than thresholds derived in implanted patients, and a single 50% line can make a device therapy's success rate look cleaner than the underlying distribution. Intensity alone is insufficient: add an interference or disability measure, a global impression of change, opioid dose, and the number most often missing, which is explantation and device revision reported on the same denominator as the responder rate. A cohort with 70% responders and a quarter of devices explanted at five years has not been described by its responder rate.

Seizures. The instrument behind essentially every neuromodulation figure in epilepsy is a patient-kept diary, and the convention is a 50% or greater reduction in frequency from a prospectively recorded baseline. Seizure diaries have important ascertainment limitations. In video-EEG monitoring of 91 adults with focal epilepsy, patients failed to document 55.5% of recorded seizures, and the omissions largely reflected seizure unawareness (including seizures arising from sleep) rather than poor compliance, so reminders alone do not fix the problem. A diary-based reduction estimates reported seizures; its relationship to total clinical and electrographic seizure burden depends on stable ascertainment and the seizure types being counted. Require an adequate prospective baseline, define a countable event before the study starts, and corroborate the diary electrographically where the question will bear it.

Tremor and spasticity. Tremor outcomes commonly use the Fahn-Tolosa-Marin Clinical Rating Scale for Tremor (FTM-CRST) or the Essential Tremor Rating Assessment Scale (TETRAS), with accelerometry and spiral ratings as adjuncts; specify which limb, which task, and whether the rater was blinded. Spasticity and rhizotomy outcomes need a function measure and a tone measure together: GMFCS stratification, GMFM for gross motor function, and modified Ashworth or Tardieu for tone, with instrumented gait when it will change a decision. In both, accompany group estimates with uncertainty, individual trajectories, function, and adverse outcomes.

7.Core, disease, and mechanism layers

Use a core layer across procedures: serious adverse events, reoperation, infection, cognition, mood, quality of life, and patient goal attainment. Add a disease layer such as MDS-UPDRS, Y-BOCS, MADRS, TWSTRS, BFMDRS, a pain numeric rating scale, a seizure diary, or the BNI grade. Add a mechanism layer matched to the intervention: tremor accelerometry, LFP burden, pain unpleasantness, gait sensors, or stimulation energy.

8.Responder analyses without threshold theater

Prespecify one primary threshold and provide sensitivity analyses at adjacent thresholds. Report continuous change and confidence intervals. Show deterioration as well as improvement. When several scales are tested, identify the hierarchy and correct or contextualize multiplicity.

9.Time is an outcome

Early microlesion effects, programming optimization, placebo, disease progression, tolerance, and hardware changes alter trajectories. Choose time points that match the intervention and include durability. A useful report shows each patient's course, not just group means at a convenient visit.