Announced last summer, CMS has now instituted the WISeR program to force prior authorization into Medicare Part B using outside venders and their software.
There's a new piece about it in JAMA Internal Medicine currently. Kannarkat et al. They note,
- CMS puts forth a thoughtful effort through WISeR to reduce health care waste....With WISeR, AI tools prescreen PA documentation for completeness and match against Medicare coverage rules before human review,
It's a mess. The CMS NCD is strikingly frozen in 2003 language. It still covers the basic durable points — idiopathic PD, levodopa responsiveness, STN/GPi targets, movement-disorders involvement, and exclusion of atypical parkinsonism — but a modern DBS center would think about DBS much differently than the NCD reads. CMS itself still lists NCD 160.24 as Version 1, effective April 1, 2003, with no ending effective date. (Centers for Medicare & Medicaid Services)
1. Modern DBS is not just a “late advanced PD” last resort
The 2003 NCD frames STN/GPi DBS around “advanced idiopathic PD” using Hoehn & Yahr or UPDRS III, with persistent disabling symptoms despite optimal medical therapy. Modern expert thinking is more functional: consider DBS when troublesome motor fluctuations, dyskinesia, or medication-refractory tremor interfere with functioning or quality of life, not only after the patient reaches some rigid “advanced” stage. A 2025/2026 international expert consensus specifically says candidacy should not rest solely on UPDRS/MDS-UPDRS score, Hoehn & Yahr stage, age, or disease duration, and warns that “advanced PD” language can delay referrals until patients are worse candidates. (Nature)
2. “Optimal medical therapy” is now a more nuanced concept
The NCD requires disabling symptoms or medication side effects despite optimal medical therapy, but does not define it. Modern care would document the medication history more granularly: levodopa schedule, OFF time, dyskinesia burden, adjunctive agents, intolerances, dopamine agonist complications, rescue therapies, and whether advanced alternatives such as levodopa infusion, apomorphine infusion, or other device-aided therapies have been considered. The expert consensus notes that optimal medical management for motor fluctuations may reasonably include at least four levodopa doses per day and trials of two or more adjunctive oral medications, but also says patient-specific issues prevent a rigid one-size-fits-all definition. (Nature)
3. Levodopa responsiveness still matters, but it is not a crude gate
The 2003 rule requires patients to be L-dopa responsive with clearly defined ON periods. That remains broadly right, because DBS best treats levodopa-responsive motor features. But modern care is less mechanical. A levodopa challenge is still important, yet contemporary experts caution that levodopa responsiveness should be assessed and considered, not used as the sole criterion. Tremor, for example, can be relatively medication-refractory yet respond well to DBS; conversely, gait, balance, speech, swallowing, and cognitive symptoms are less predictably helped. (Nature)
4. Cognitive and psychiatric exclusions are less binary
The NCD excludes patients with cognitive impairment, dementia, or depression if these would be worsened by or interfere with DBS benefit, and excludes current psychosis or substance abuse. Modern DBS programs still take cognition and psychiatric stability very seriously, but they usually approach this as risk stratification, not a simple yes/no exclusion. The newer consensus says there is no data-driven cognitive “cutoff,” that some patients with cognitive impairment may still benefit for motor symptoms, and that mood/anxiety/psychiatric symptoms do not necessarily exclude candidacy if addressed before surgery. (Nature) In practice, this means formal neuropsychological testing, psychiatric review when needed, caregiver assessment, and a target-selection discussion — for example, being cautious about STN in patients with cognitive or mood vulnerability.
5. Target choice is more individualized: STN vs GPi is a strategic decision
The NCD lists STN or GPi as covered targets, but it does not reflect the modern “which target for which patient?” conversation. Contemporary reviews summarize a practical distinction: STN and GPi both improve motor symptoms, but STN DBS more often allows medication reduction, while GPi has a direct anti-dyskinetic effect and may be favored in some patients where cognitive, mood, or medication-reduction risks are concerns. (NCBI) VIM DBS for Parkinsonian tremor is now usually understood as a tremor-focused option that does little for bradykinesia, rigidity, fluctuations, or dyskinesia. (NCBI)
6. The “conscious operative procedure” requirement is dated
The NCD repeatedly requires willingness and ability to cooperate during a conscious operative procedure. That reflects the older awake-DBS paradigm with intraoperative testing. Many centers still do awake DBS, but modern practice also includes asleep DBS under general anesthesia using high-resolution imaging, intraoperative CT/MRI, stereotactic planning, and sometimes robotic assistance. Recent literature reports broadly comparable outcomes between awake and asleep DBS in Parkinson’s, so a blanket “must cooperate while awake” requirement is no longer aligned with common practice. (PubMed)
7. The MRI prohibition is obsolete as written
The 2003 NCD says DBS patients should not be exposed to “any type of MRI” because it may adversely affect the DBS system or brain tissue. That is probably the most visibly dated sentence. Modern DBS systems are often MR Conditional under specified manufacturer conditions. For example, Medtronic describes a portfolio of full-body MR Conditional DBS systems for 1.5T and 3T scanning under specified limits, and FDA materials in 2025 describe newer DBS systems with adaptive DBS features. (Medtronic) The modern rule is not “no MRI”; it is “MRI only under the exact device-specific conditions, with the system identified and programmed appropriately.”
8. Device technology has moved far beyond 2003 stimulators
The NCD basically imagines DBS as fixed implanted electrical stimulation. Modern devices include directional leads, more flexible programming, rechargeable IPGs, expanded MRI labeling, sensing-capable systems, remote/programming refinements, and now adaptive DBS. FDA’s 2025 SSED for Medtronic’s Activa/Percept/SenSight system describes a supplement introducing an optional adaptive deep brain stimulation feature for Parkinson’s disease. (FDA Access Data) That does not mean adaptive DBS is yet the universal standard, but it shows how far the technology has moved from the 2003 coverage model.
9. Modern evaluation is explicitly multidisciplinary
The NCD does require neurosurgical expertise and movement-disorder physician involvement, which remains appropriate. But modern DBS programs are more explicit about a team evaluation: movement-disorders neurology, functional neurosurgery, neuropsychology, psychiatry when needed, speech/swallow evaluation, physical therapy/gait assessment, social support, goals-of-care discussion, and careful expectation-setting. The current consensus recommends at minimum evaluation by a neurologist, neurosurgeon, and neuropsychologist, with rehabilitation and other disciplines when available. (Nature)
10. Non-motor symptoms are now part of counseling, not the primary indication
The 2003 NCD is almost entirely motor-symptom oriented. Modern care still uses DBS primarily for motor fluctuations, dyskinesia, tremor, rigidity, and bradykinesia, but clinicians now counsel patients more explicitly about sleep, autonomic symptoms, mood, fatigue, cognition, speech, swallowing, gait, and freezing of gait. The consensus notes that DBS may improve sleep and some non-motor fluctuations, but these are not currently primary indications in the absence of motor fluctuations, dyskinesia, or tremor. (Nature)
How Modern DBS Care Has Evolved Beyond CMS's 2003 NCD
The 2003 NCD reflects the standard of care from over two decades ago. Modern practice has shifted substantially in several clinically important ways. Below are the major areas where current evidence and practice diverge from the 2003 rules.
###
The 2003 NCD is not a guideline that has drifted from practice — it is the operative, unrevised national coverage rule that any prior-authorization model demonstration must apply. MACs cannot fix the gaps below through LCDs because LCDs must be consistent with NCDs.
Only CMS, through the NCD reconsideration process, can update the criteria. The fact that this hasn't happened in 23 years means the prior-auth demo would be enforcing rules written:
- Before the FDA expanded DBS labeling for earlier-stage PD (2016, post-EARLYSTIM)
- Before asleep / image-guided DBS became a standard technique (circa 2013 onward)
- Before FDA approval of DBS for dystonia (HDE 2003, same year as NCD but not in it), OCD (2009), and epilepsy (2018)
- Before MRI-conditional DBS systems existed (the NCD flatly prohibits "any type of MRI")
- Before rechargeable IPGs, directional leads, and sensing/adaptive DBS (Percept BrainSense, 2020) existed
- Before MDS-UPDRS replaced the original UPDRS (2008)
- Before MR-guided focused ultrasound became an FDA-approved alternative (2016 ET, 2018 PD)
Six specific binding-rule problems
A few of these are worth highlighting because they show the NCD doesn't just lag practice — it would, if applied literally in a prior-auth check, deny coverage for care that meets the current FDA label:
1. "Advanced idiopathic PD" is a hard NCD criterion. The FDA-approved label since 2016 includes PD patients with at least 4 years of disease and recent-onset motor complications — not "advanced." A prior-auth contractor reading the NCD strictly would deny a patient whose surgery is fully on-label.
2. "Willingness and ability to cooperate during conscious operative procedure" is a hard NCD criterion. Asleep DBS under general anesthesia with intraoperative imaging is now standard at many high-volume centers and has comparable or better outcomes in published series. A literal NCD read denies the modern technique.
3. The MRI prohibition is absolute in the NCD text. It states patients "should not be exposed to … any type of MRI." Current FDA-labeled MRI-conditional DBS systems are explicitly designed for safe MRI under defined parameters. The NCD text and the FDA label are in direct conflict, and the NCD cannot be locally corrected.
4. The indication list is closed. The NCD covers only ET and PD at VIM/STN/GPi. Dystonia (FDA HDE 2003), OCD (FDA HDE 2009), and epilepsy (FDA PMA 2018) all fall outside it. Coverage for these has had to be cobbled together through other NCDs (e.g., 160.2 for "motor function disorders"), case-by-case review, or non-coverage — an artifact of the NCD never being reopened. MAC policies route non-ET/PD indications to NCD 160.2 (Treatment of Motor Function Disorders with Electric Nerve Stimulation) rather than 160.24, which is itself a workaround for the closed list. Getbynder
5. The "conscious" requirement also functions as an indirect exclusion of pediatric and severely symptomatic patients who can't tolerate awake surgery — even when an asleep approach is medically appropriate and FDA-cleared.
6. Categorical exclusions for "cognitive impairment, dementia or depression" are written as bright lines. Modern practice uses neuropsych evaluation to stratify risk rather than exclude categorically. A prior-auth checklist applying the NCD literally would screen out patients current movement-disorder centers would accept.
Your Argument is Worthwhile
The point isn't just "the science has moved on." It's structural:
- The NCD is binding nationally.
- LCDs cannot fix it (42 CFR 405.1060; Medicare Program Integrity Manual Ch. 13).
- Only CMS reconsideration can update an NCD, and that hasn't happened since 2003.
- A prior-authorization model demonstration that applies the NCD as a checklist therefore enforces 2003-era medicine as a condition of payment in 2026 — including against care that is on-label with FDA and consistent with specialty society guidelines.
That's the cleanest framing: the rule isn't outdated by accident of practice drift; it's outdated by 23 years of CMS inaction on a binding national rule that no downstream contractor has authority to modernize. A prior-auth demo that operationalizes it accelerates and hard-codes that gap rather than narrowing it.
