CPT code simulator — how a prescription digital therapeutic earns FDA clearance, a billing code, payer coverage, and real-world reimbursement
Prescription digital therapeutics (PDTs) are regulated by FDA as Software as a Medical Device (SaMD). Before a single dollar of reimbursement is possible, the sponsor must clear FDA review — establishing safety and efficacy through clinical evidence, almost always via the De Novo or 510(k) pathway, at Class II risk. No billing code, no payer conversation, and no claim exists until this gate opens.
FDA has no dedicated regulatory category for "digital therapeutics" — it fits them into the existing medical device framework as SaMD, software intended to treat, diagnose, or mitigate disease independent of hardware.
Two submission pathways dominate:
• De Novo classification: used when no legally marketed predicate device exists. This is how the first-in-category products entered the market — Pear Therapeutics' reSET (2017) had no predicate, so it went through De Novo, simultaneously creating a new regulatory classification that later products could cite as their own predicate.
• 510(k) premarket notification: used once a predicate exists. A sponsor demonstrates "substantial equivalence" to an already-cleared device. Once reSET existed as a predicate, subsequent substance-use and insomnia apps could pursue 510(k), which is typically faster (roughly 3–6 months of FDA review versus longer for De Novo).
EndeavorRx (Akili Interactive), cleared June 2020 for pediatric ADHD, took the De Novo route as the first game-based digital therapeutic, since no video-game-as-medicine predicate existed. Its pivotal trial was a randomized controlled trial (STARS-ADHD) showing improvement on the TOVA attention measure.
FDA clearance for a PDT typically requires at least one adequately powered randomized controlled trial (RCT) demonstrating a clinically meaningful effect on the labeled indication — not merely that the app is safe or that patients like using it.
• reSET was cleared on a 12-week RCT (n=399) showing improved abstinence versus treatment-as-usual in outpatient substance use disorder. • Somryst was cleared on two RCTs totaling >1,300 patients with chronic insomnia, showing reductions in Insomnia Severity Index scores comparable to in-person CBT-I. • EndeavorRx's STARS-ADHD trial (n=348) was a randomized, controlled, blinded study — one of the more rigorous designs FDA has accepted for a Class II SaMD.
FDA's Digital Health Center of Excellence (established 2020) coordinates SaMD review across the agency, issues guidance on clinical decision support software, AI/ML-based devices, and Breakthrough Device Designation — an expedited-review status a sponsor can request to accelerate development and review timelines for devices addressing serious or life-threatening conditions with no adequate alternative.
Clearance is not approval in the strict FDA sense — 510(k) and De Novo pathways produce "clearance," while the more rigorous Premarket Approval (PMA) pathway produces "approval." This distinction becomes commercially important later: some Category I CPT code transitions and CMS coverage policies explicitly ask for full approval, not mere clearance.
FDA clearance only authorizes marketing; it does not create a way to bill for the product. The sponsor must separately petition the American Medical Association's CPT Editorial Panel for a Current Procedural Terminology code. New, unproven technologies are assigned temporary Category III codes — trackers that let claims be submitted and counted, with zero guarantee that any payer will actually reimburse them.
CPT codes are not a monolithic list; the AMA splits them into three functionally distinct categories:
• Category I: the "real" permanent codes used for established, widely performed procedures and services. Getting here requires proof of FDA approval (in many cases) and evidence the service is performed by many physicians in multiple locations across the United States — a "widespread use" bar.
• Category II: optional supplemental tracking codes for performance measurement and quality reporting. They carry no payment value at all; they exist purely for data collection.
• Category III: temporary tracking codes for emerging technology, services, and procedures. They exist so payers and researchers can observe utilization and outcomes data before a service is mature enough to justify a permanent code — but a Category III code carries no inherent payment obligation. A payer can process a Category III claim and pay $0, pay a negotiated rate, or deny it as not medically necessary.
A DTx sponsor typically applies for a Category III code shortly after FDA clearance, since without any code, claims cannot even be submitted to a payer's adjudication system in a structured way.
Rather than creating device-specific codes for every digital therapeutic, CMS activated a Remote Therapeutic Monitoring (RTM) code family effective January 1, 2022 that many DTx and MSK/behavioral apps now bill under:
• 98975: initial set-up and patient education on use of the monitoring device • 98976 / 98977: device supply with scheduled recording/alert transmission for respiratory or musculoskeletal system programs, per 30-day period • 98980: remote therapeutic monitoring treatment management, first 20 minutes of clinical staff/physician time per month • 98981: each additional 20 minutes beyond the first
Critically, RTM codes reimburse for the clinician's monitoring and management time associated with the data a DTx generates — not for the software license itself. This is an imperfect fit: a sponsor's app can enable RTM billing by a treating physician, but the app maker often does not directly collect the RTM payment. Many DTx business models instead pursue a dedicated Category III device code (billed by the prescriber or pharmacy) running in parallel with, not as a replacement for, RTM billing.
A recurring point of confusion — deliberately clarified here — is that a CPT code answers only "how do we describe this service on a claim," not "will anyone pay for it."
Three independent decisions determine whether money ever changes hands:
• Coding: does a standardized code exist to describe the service? (AMA CPT Editorial Panel) • Coverage: has a specific payer decided this service is a covered benefit under a given plan? (individual payer medical policy) • Payment: at what rate, under what conditions, and with what prior authorization does the payer actually reimburse a covered, coded claim? (payer contracting and fee schedules)
A Category III code satisfies only the first pillar. Sponsors must separately win coverage determinations from each payer, one by one, and then negotiate payment rates — a process that can take years after the code itself exists.
With a Category III code in hand, the sponsor now faces the hardest and least standardized part of the pathway: convincing individual commercial payers and self-insured employers, one at a time, that the evidence justifies coverage. Health technology assessment (HTA) teams weigh clinical trial quality, real-world evidence, and cost-effectiveness against a $100,000–$150,000 per QALY willingness-to-pay bar inherited from oncology and pharmacy benefit review.
Unlike FDA, which asks "is this safe and effective for its labeled indication," a payer medical policy committee asks a different, harder question: "is this worth paying for, relative to alternatives, at this price, for our specific member population?"
HTA reviews typically weigh:
• Strength and quantity of RCT evidence — a single small pivotal trial is rarely enough; payers look for replication, larger real-world evidence datasets, and durability of effect beyond the trial window • Comparative effectiveness versus standard of care — does the DTx outperform, or merely match, existing low-cost interventions like generic medication or a workbook-based CBT program? • Cost-effectiveness modeling — the Institute for Clinical and Economic Review (ICER) and internal payer actuarial teams often apply a $100,000–$150,000 per quality-adjusted life year (QALY) threshold, the same range used to judge specialty pharmaceuticals • Budget impact — even a cost-effective therapy can be rejected or capped if population-level adoption would blow up near-term plan spending
Most DTx clear the safety and mechanism bar easily; the majority fail here, at the value bar, not at the clinical bar.
Prescription digital therapeutics are prescribed like drugs but delivered like software, and U.S. benefit design was never built for that combination:
• Pharmacy benefit routing treats the DTx like a drug: dispensed through a pharmacy benefit manager (PBM), assigned an NDC-like identifier, subject to formulary tiering and copay structures familiar to patients. Some digital formularies (e.g., certain CVS Health and Included Health programs) pioneered this model specifically to make DTx adoption feel identical to filling a prescription.
• Medical benefit routing treats the DTx like a device or procedure: billed under the CPT Category III or RTM code by a prescribing clinician, adjudicated like durable medical equipment or a monitoring service, and often subject to separate prior authorization and site-of-care rules.
The two routes have different utilization management, different data infrastructure, and different stakeholders inside a payer organization — meaning a single DTx sponsor frequently must run two entirely separate coverage campaigns to reach the same patient population.
Self-insured employers, who bear their own claims risk under ERISA and are not bound by state-mandated benefit rules, have historically been faster and more flexible adopters of digital therapeutics than fully-insured commercial plans — many early DTx go-to-market strategies targeted large self-insured employers before pursuing broad payer contracts.
A Category III code is built to expire. Converting it into a permanent Category I code requires proof that the service has become genuinely widespread — performed by many physicians, in many practice settings, across the country — usually alongside full FDA approval rather than mere clearance. In parallel, Medicare's decentralized coverage machinery and increasingly common value-based contracts shape whether the code, once permanent, is actually worth having.
The AMA CPT Editorial Panel will not convert a Category III code to Category I on evidence alone — it specifically requires demonstration that the underlying service or technology is:
• Performed by many physicians or other qualified health professionals, not a single sponsor's pilot network • Performed in multiple locations across the United States, not concentrated in a handful of academic centers • Supported by an evidence base sufficient for the AMA/Specialty Society RVS Update Committee (RUC) to recommend a relative value unit (RVU) and, in turn, a payment rate
For many DTx categories, this bar is where the pathway currently stalls: temporary codes get renewed rather than converted, because national-scale, multi-physician utilization is hard to demonstrate for a still-young product category. The 2023 introduction of CPT 989X6, covering digital cognitive behavioral therapy device supply, was a notable step of the AMA carving out DTx-specific code language rather than forcing sponsors to stretch existing RTM or durable-medical-equipment codes to fit.
Even a permanent Category I code does not guarantee Medicare will pay for it everywhere. CMS coverage decisions happen at two different levels:
• National Coverage Determination (NCD): a formal, nationwide CMS policy stating whether and how a service is covered under Medicare for all beneficiaries. NCDs are relatively rare, slow to issue, and CMS has generally not issued a broad NCD specifically for prescription digital therapeutics as a class.
• Local Coverage Determination (LCD): in the absence of an NCD, each Medicare Administrative Contractor (MAC) — CMS contracts with roughly a dozen regional MACs to process claims — can issue its own coverage policy for its jurisdiction. This means a DTx can be reimbursed by Medicare in one region and denied in another, simply because different MACs have reached different conclusions on the same code.
This patchwork is a major driver of why commercial payer strategy (state by state, plan by plan) often matters more to a DTx sponsor's near-term revenue than any single federal decision.
Facing payer skepticism about durability of effect and real-world adherence (which is typically lower than trial adherence), DTx sponsors increasingly negotiate value-based or risk-sharing arrangements rather than simple fee-for-service rates:
• Outcomes-based rebates: the payer pays a base rate, but the sponsor rebates a portion if a defined clinical outcome is not met at a follow-up interval • Risk corridors: sponsor and payer share financial risk if total population spend or utilization deviates from a pre-agreed forecast • Pay-for-engagement models: partial payment tied to a minimum threshold of patient engagement with the app, as a proxy for likely clinical benefit
These contract structures let a skeptical payer say yes to coverage without fully underwriting the sponsor's efficacy claims, shifting some of that risk back onto the sponsor — a trade many venture-backed DTx companies accept in order to get any commercial traction at all.
At full maturity, a DTx processes tens of thousands of claims under a mix of fee-for-service and outcomes-based contracts, competing for renewal against a five-year Category III sunset clock and a market that has already produced cautionary tales. Regulatory clearance and even a working billing code are necessary, but not sufficient, conditions for commercial survival.
By this stage, a DTx sponsor typically operates several parallel billing arrangements simultaneously: Category III or Category I device-supply claims submitted by pharmacies or prescribers, RTM claims submitted by treating clinicians for monitoring time, and direct-to-employer or direct-to-payer subscription contracts that bypass fee-for-service billing entirely.
Ongoing evidence generation does not stop at launch — payers increasingly require post-market real-world evidence (adherence rates, symptom-scale improvement such as PHQ-9 for depression or HbA1c for diabetes programs, and hospitalization or ER-utilization offsets) as a condition of contract renewal, not just initial coverage. A product that cannot demonstrate real-world effectiveness comparable to its pivotal trial risks losing coverage even after winning it once.
Category III status is explicitly temporary. If a code is not converted to Category I or formally renewed by the AMA CPT Editorial Panel within its review cycle (codes are revisited on a rolling basis, with a five-year outer horizon commonly cited for archival if no action is taken), it can be deleted from the code set entirely — eliminating the sponsor's ability to bill under that identifier and forcing a scramble to re-code claims under an adjacent code family such as RTM.
This creates a structural business-model risk unique to reimbursement-dependent digital health: the entire revenue mechanism can lapse for administrative reasons unrelated to clinical performance, on a timeline the sponsor does not fully control.
The 2023–2025 period delivered a hard reset on early DTx optimism:
• Omada Health and Big Health (maker of Sleepio) remained operating with substantial employer and payer contract books, with Omada completing a public listing in 2024 — evidence that durable reimbursement relationships, not regulatory novelty, are what sustain a DTx business • Akili Interactive sold its prescription EndeavorRx business line to Virpax Pharmaceuticals in 2024 and pivoted its core strategy toward an over-the-counter consumer product (EndeavorOTC), sidestepping the reimbursement pathway altogether after concluding it could not scale profitably through payers alone • Several smaller DTx ventures quietly wound down or were acquired for parts as investors reassessed the multi-year, multi-payer slog required to reach durable reimbursement
The throughline across these outcomes is that the reimbursement pathway — FDA clearance, then a code, then payer coverage, then Category I conversion, then durable contracts — is long, expensive, and each gate filters out most entrants, regardless of how strong the underlying clinical evidence was at the start.
Pear Therapeutics filed for Chapter 11 bankruptcy in April 2023 despite holding three FDA-cleared prescription digital therapeutics (reSET, reSET-O, and Somryst) and having pioneered the very De Novo pathway other DTx sponsors now rely on. The company could not convert regulatory and clinical success into a sustainable reimbursement rate at scale — direct proof that FDA clearance alone does not guarantee commercial viability, and that the payer-coverage and coding gates downstream of clearance are where most DTx economics actually get decided.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| Commercial (fully insured) | Formulary + HTA review, prior authorization common | ||
| Self-insured employer (ERISA) | Faster adoption, flexible benefit design, direct contracting | ||
| Medicare (via MAC/LCD) | Regional Local Coverage Determinations, no broad NCD | ||
| Medicaid (state-by-state) | State formulary variation, slowest broad adoption |