Digital Therapeutics: When Software Becomes a Prescription

A doctor writes a prescription, and the patient fills it not at a pharmacy counter but on an app store. No pills, no injections — just a piece of software, cleared by a regulator, delivering a clinical intervention with an evidence base behind it. This is the premise of digital therapeutics, and it has quietly moved from a niche category pitched by a handful of startups to a recognized line item in how some health systems treat chronic disease, insomnia, substance use disorder, and pediatric ADHD.

The idea sounds almost too simple: software that works like medicine. But underneath that simplicity is a genuinely distinct category with its own regulatory pathway, its own reimbursement headaches, and its own track record of both remarkable clinical wins and high-profile commercial failures. Understanding what digital therapeutics actually are — and are not — matters for anyone building health software, evaluating vendors, or trying to make sense of where healthcare AI is actually generating value versus where it's just marketing.

What Digital Therapeutics Actually Are

Digital therapeutics, often shortened to DTx, are software-based interventions designed to prevent, manage, or treat a medical condition, and validated through clinical evidence in the same way a drug or device would be. The Digital Therapeutics Alliance, the industry's main standards body, defines them as delivering "evidence-based therapeutic interventions to patients that are driven by high quality software programs to prevent, manage, or treat a medical disorder or disease."

That definition does a lot of work, and it's worth unpacking the parts that distinguish DTx from the broader universe of health apps:

• Evidence-based: A DTx product typically has clinical trial data, often randomized controlled trials, showing it produces a measurable therapeutic outcome — not just "correlates with better habits" but "reduces HbA1c" or "reduces relapse rates."
• Regulated: Many (though not all) digital therapeutics go through a regulatory clearance or approval process, most commonly the FDA's 510(k) or De Novo pathways in the United States, or a CE mark under the EU Medical Device Regulation.
• Prescribable: Some products require a clinician to prescribe them, generating an access code the patient redeems, much like filling a prescription for a drug. Others are available over the counter but still marketed on clinical evidence.
• Outcome-oriented: The product exists to change a clinical measure, not just to inform, educate, or motivate in a general wellness sense.

This is the line that separates a digital therapeutic from a fitness tracker or a meditation app. A step counter might improve your health in a diffuse, unmeasured way. A DTx for insomnia has to show, in a trial, that it reduces time-to-sleep-onset or improves a validated insomnia severity index, and it has to keep demonstrating that to stay on the market.

The DTx Spectrum

Not all digital therapeutics look alike. Broadly, the category splits into a few functional types:

The common thread is that all of them are trying to occupy the same regulatory and reimbursement space that pharmaceuticals have occupied for decades — evidence, approval, prescription, reimbursement — just delivered through code instead of chemistry.

It's also useful to place DTx against the two categories it's most often confused with: telehealth and remote patient monitoring. Telehealth is a delivery channel — a clinician seeing a patient over video instead of in person — and the therapeutic content is whatever the clinician decides in that visit. Remote patient monitoring collects data (blood pressure, glucose, activity) and flags it for clinical review, but the software itself isn't the treatment. A digital therapeutic is different from both: the software is the intervention. There's no clinician on the other end delivering the therapy in real time; the program itself, running on a phone or a device, is what's been shown in trials to move a clinical outcome.

How the Regulatory and Evidence Pipeline Works

Building a digital therapeutic looks less like building a typical consumer app and more like running a drug development program, compressed but structurally similar.

1. Mechanism definition: The team specifies the clinical mechanism — cognitive behavioral therapy delivered through structured software modules, for instance, or a video-game-like attention task with a defined neurocognitive rationale.
2. Feasibility and pilot studies: Small studies establish that the intervention is usable and shows a signal of effect.
3. Pivotal clinical trial: A larger, often randomized and controlled trial establishes efficacy against a defined endpoint — the same endpoint a drug targeting that condition would use.
4. Regulatory submission: Depending on the risk classification, the company files for FDA clearance (510(k), showing substantial equivalence to a predicate device), De Novo classification (for genuinely novel device types with no predicate), or in some cases pursues designation as "Software as a Medical Device" (SaMD) under frameworks harmonized internationally.
5. Post-market evidence: Regulators increasingly expect continued real-world evidence generation, especially for software that can be updated after clearance — a requirement pharma doesn't usually face in the same way, since a pill's formulation doesn't change after approval.

This is a meaningfully higher bar than most health apps clear. It's also why digital therapeutics companies have research and clinical affairs teams that resemble a pharmaceutical company's more than a typical software startup's, and why development timelines run into years rather than product sprints.

The distinction matters for a second reason: it creates a defensible moat. A wellness app can be cloned in a weekend. A digital therapeutic with a completed pivotal trial and FDA clearance cannot — the evidence and regulatory record are hard-won and slow to replicate, even if the underlying software is technically simple.

How DTx development compares to drug development

The parallels to pharma go beyond vibes — the practical workflow rhymes at nearly every stage, even though the underlying product is code rather than a molecule.

The main structural difference is speed of iteration: a drug's formulation is fixed at approval, while software can be updated constantly — which is a strength for improving the product but a regulatory headache, since updates to the therapeutic mechanism can trigger new filings.

Why It Matters Right Now

Digital therapeutics sit at an unusual intersection of several forces that are each independently reshaping healthcare delivery: chronic disease burden that outpaces clinical capacity, payer interest in interventions that scale without adding headcount, and a regulatory apparatus that has spent the last several years building out formal pathways for software-as-medicine rather than treating it as a novelty.

The category has also been a useful stress test for a broader question in health tech: can software alone, without a device or a drug attached, produce a clinical effect large enough and durable enough to justify a price tag and a prescription? The answer, based on published trial data across categories like insomnia, substance use disorder treatment, and pediatric attention disorders, has generally been yes for a defined set of conditions — particularly those where the underlying clinical intervention (cognitive behavioral therapy, structured behavioral coaching, attention training) was already delivery-agnostic and translates reasonably well into a structured software program.

At the same time, the category has had a rough commercial reckoning. Several well-funded DTx companies that won regulatory clearance struggled to get insurers to pay for their products at a price point that made the business viable, and some scaled back or shut down despite having clinical evidence that would have been the envy of most pharma pipelines. That's a genuinely important, underreported part of the story: clinical validation and commercial viability are not the same problem, and DTx has repeatedly solved the first while stumbling on the second.

Benefits of Digital Therapeutics

When a digital therapeutic works, it changes who can receive an evidence-based intervention and how consistently it is delivered.

Evidence-based therapy without a clinician in every session

Many effective treatments, such as cognitive behavioral therapy for insomnia, depend on trained clinicians who are in short supply. A DTx packages the structured parts of that intervention into a program the patient works through on their own schedule. Clinicians still prescribe and oversee, but the per-session time cost drops sharply, so more patients can start a recognized treatment instead of waiting for a specialist appointment.

Consistent delivery of the protocol

Human-delivered therapy varies with the practitioner's training, workload, and style. Software delivers the same modules, in the same order, with the same content to every patient. That consistency is part of what lets a trial result generalize: the intervention patients receive after clearance is, by design, the one that was tested, as long as the underlying mechanism hasn't been changed.

Access outside the clinic

A program on a phone is available in the evening, at weekends, and in places far from specialist services. For conditions where the main barrier is reaching a trained provider rather than the treatment itself, that can widen access, with the important caveat that it depends on the patient having a suitable device, connectivity, and confidence using it.

Measurable engagement and outcomes

Because the therapy runs in software, engagement is visible: which modules were completed, when patients dropped off, and how self-reported measures changed. Clinicians can see whether a patient is actually using the treatment, something that is much harder to know for many medications. That data also feeds the post-market evidence regulators and payers increasingly expect.

A stronger case with payers and health systems

A product with a defined endpoint and published trial results can be evaluated like other treatments, rather than as a wellness perk. That gives health systems and payers a concrete basis for comparing options, and it gives builders a defensible position that a quickly cloned wellness app cannot match. Clear evidence also makes it easier for clinicians to decide when a DTx is appropriate and when another option fits better.

Digital Therapeutics Use Cases

The category has found its footing in conditions where the underlying intervention was already structured and didn't depend on physical, in-person care.

Chronic insomnia

Cognitive behavioral therapy for insomnia is a well-established first-line approach, but access to trained therapists is limited. DTx products translate its components, such as sleep scheduling, stimulus control, and cognitive techniques, into guided software modules. Trials in this area measure outcomes like sleep-onset time and validated insomnia severity scores, and it remains one of the clearest examples of software delivering a recognized therapy.

Substance use disorder

Prescription DTx have been used alongside standard treatment for substance use disorders, delivering structured behavioral therapy content and reinforcing skills between clinic visits. The software doesn't replace medication or counseling; it extends the therapeutic contact patients receive between appointments, when relapse risk can be highest. Outcomes in this category are typically tied to measures such as abstinence or retention in treatment.

Pediatric ADHD

Attention-training programs designed like video games have been developed for children with ADHD, built around a defined neurocognitive rationale and tested against attention measures. These products show how a DTx can look nothing like a traditional health app while still following the evidence and clearance path described above.

Chronic pain

Programs for chronic pain typically combine education, behavioral techniques, and guided exercises drawn from established psychological approaches. They are aimed at helping patients manage symptoms and function day to day, often as part of a wider treatment plan rather than a standalone solution. The appeal for health systems is giving patients structured support between appointments, when pain management decisions are actually made, without adding clinician hours.

Diabetes management as an adjunct

Companion therapeutics paired with existing drug therapy help patients adjust behavior, track relevant measures, and stay on their treatment plan. Here the software supports the effect of a medication rather than acting alone, and outcomes are linked to clinical measures such as HbA1c. This adjunct model is also where integration into existing care pathways tends to be most natural, because the clinician managing the medication can see how the patient is engaging with the companion program.

Digital Therapeutics Best Practices for Builders and Health Systems

For teams building or evaluating digital therapeutics, a few practical realities shape what's actually achievable.

For companies building DTx products

• Budget for trial infrastructure, not just engineering. The clinical trial, biostatistics, and regulatory affairs cost of a DTx program frequently exceeds the software development cost. Teams that treat this as an app-development project first and a clinical-evidence project second tend to underbuild the parts that actually create defensibility.
• Design for a reimbursement pathway from day one. A product with strong clinical data but no CPT code, no payer contract, and no clear prescribing workflow will struggle regardless of efficacy. Reimbursement strategy needs to be part of the initial product design, not an afterthought bolted on after clearance.
• Plan for post-market evidence generation. Regulators and payers alike increasingly want to see that the product keeps performing once it's in the real world, not just in a controlled trial population.
• Treat software updates as regulatory events. Unlike a typical SaaS product, a materially changed algorithm or user flow in a cleared DTx can trigger a new regulatory filing. Release cadence has to account for this.

For health systems and payers evaluating DTx

• Ask for the specific endpoint, not just "clinically validated." A product that improved a self-reported satisfaction score is a different claim than one that reduced hospitalization rates. The rigor varies widely across products carrying the same "DTx" label.
• Check what happens after the prescription. Adherence for digital interventions tends to drop off sharply after the first few weeks, similar to patterns seen in medication adherence — but with less institutional infrastructure built around monitoring it.
• Understand the reimbursement mechanism before piloting. Some products bill through medical benefit, some through pharmacy benefit, and some have no established coding pathway at all, which shifts cost onto the health system or the patient.

Common Digital Therapeutics Mistakes

The category's commercial setbacks trace back to a handful of avoidable decisions, made by builders and buyers alike.

Running a DTx program like an app project

Teams from consumer software often plan around engineering sprints and treat clinical work as a later phase. In practice, trial design, biostatistics, and regulatory affairs usually cost more and take longer than the software. Underinvesting in them produces a polished product without the evidence that makes it defensible or prescribable.

Leaving reimbursement until after clearance

Several clinically strong products struggled because nobody had worked out who would pay, under which code, and through which benefit. Clearance alone doesn't generate revenue. Treating payer strategy and the prescribing workflow as day-one design inputs avoids building a product that clinicians can't easily prescribe and insurers won't readily cover.

Accepting "clinically validated" at face value

Buyers sometimes treat the DTx label as a quality guarantee. The strength of evidence varies widely, from rigorous randomized trials on hard endpoints to small studies on satisfaction scores. Health systems that skip reading the actual study design risk paying for an intervention that doesn't deliver the outcome they care about.

Ignoring adherence after the prescription

A therapeutic only works if patients use it. Programs that launch without a plan for monitoring engagement, re-engaging patients who drop off, and involving clinicians when adherence falls tend to see real-world results far below trial results. Adherence needs owners and processes, not just better app design.

Shipping algorithm changes like ordinary updates

In conventional software, continuous releases are normal. In a cleared DTx, a material change to the therapeutic mechanism can require documentation, validation, and sometimes a new regulatory submission. Teams that don't separate routine fixes from changes to the clinical core can end up with a product that no longer matches what was cleared.

Real Limitations and Open Questions

Digital therapeutics are not a universal substitute for pharmacological or in-person clinical care, and the category has real, unresolved friction points.

Reimbursement remains fragmented. Unlike drugs, which move through well-established pharmacy benefit and formulary structures, digital therapeutics don't have a single, consistent path to get paid for — a friction point closely related to the prior authorization bottlenecks that already frustrate clinicians prescribing conventional treatments. Some are billed as durable medical equipment, some through novel CPT codes designed for remote monitoring or digital interventions, and some simply aren't covered at all, leaving the cost with employers, patients, or pilot programs that eventually need to prove ROI to continue.

Adherence is a genuine clinical problem, not just a UX problem. A digital therapeutic only works if the patient actually engages with it over the prescribed course. Attrition curves for many digital interventions look similar to attrition curves for other self-directed treatments — steep dropoff after the initial weeks — which undermines the real-world effectiveness even when trial efficacy was solid.

The evidence bar varies by product, and the category label doesn't guarantee rigor. "Digital therapeutic" is not a single legally protected term with one fixed evidence threshold everywhere in the world; the strength of the underlying clinical trials varies considerably between products, and health systems evaluating vendors need to look at the actual study design rather than the marketing category.

Software maintenance creates ongoing regulatory obligations that most software companies aren't built for. Bug fixes are routine in software; in a regulated DTx, a fix to the clinical algorithm may require documentation, validation, and in some cases a new regulatory submission. This slows iteration speed relative to conventional software development in a way that surprises teams coming from a typical product background.

Equity and access questions are unresolved. A prescription that requires a smartphone, reliable data connectivity, and a baseline level of digital literacy is not equally accessible to every patient population, and there's limited long-term data on whether DTx narrows or widens existing gaps in care access.

What to Watch Next

A few trends are likely to shape where digital therapeutics go from here:

• Consolidation of reimbursement pathways. Payers and standards bodies are gradually working toward more consistent coding and coverage frameworks for software-based interventions, which would reduce the friction that has hurt commercial viability for several clinically strong products.
• AI-native DTx. Newer entrants are building adaptive, AI-driven personalization directly into the therapeutic mechanism rather than bolting AI onto a static program, which raises new regulatory questions about how adaptive algorithms get validated and re-validated over time.
• Integration into standard care pathways rather than standalone prescribing. The products gaining traction increasingly sit inside existing clinical workflows — prescribed alongside a drug, monitored through the same portal a clinician already uses — rather than existing as a separate app a patient has to seek out independently.
• Sharper scrutiny of real-world effectiveness versus trial efficacy. As more products accumulate years of post-market data, expect more public reckoning with the gap between what happened in a controlled trial and what happens when a product is prescribed at scale.
• Consolidation among vendors. Given the commercial difficulty several standalone DTx companies have faced, expect more of the category to get absorbed into larger health tech, pharma, or payer organizations that can subsidize the go-to-market cost with an existing distribution channel.

Teams evaluating or building in this space, from regulatory strategy through clinical software architecture, can find hands-on support through Woyce Technologies.

FAQ

What's the difference between a digital therapeutic and a regular health app?

A digital therapeutic has to demonstrate a measurable clinical outcome through structured evidence, typically clinical trials, and often goes through formal regulatory clearance. A general health or wellness app usually makes no such regulated clinical claim and isn't held to the same evidence standard. In practice, look for three things: a named clinical endpoint, published trial results, and a regulatory status you can verify in the relevant public database. A product missing all three is a wellness app, however clinical its marketing sounds.

Do you need a prescription for a digital therapeutic?

Some do, known as prescription digital therapeutics (PDTx), which require a clinician to authorize access, similar to filling a drug prescription. Others are cleared for over-the-counter use and don't require a prescription, even though they carry the same underlying clinical evidence base. Whether a product is prescription-only depends on its regulatory classification and intended use, and the same underlying software can occasionally be offered under different access models in different countries.

Are digital therapeutics covered by insurance?

Coverage is inconsistent. Some are billed through pharmacy or medical benefits using specific procedure codes, some are covered only through employer pilot programs, and some have no established reimbursement pathway at all, which has been a major obstacle to commercial adoption despite clinical evidence of effectiveness. Before prescribing or piloting a product, health systems should confirm exactly which benefit it bills through, which codes apply, and whether coverage depends on the patient's specific plan, because the answer can differ between payers.

What conditions do digital therapeutics currently treat?

Common categories include chronic insomnia, substance use disorder, pediatric ADHD, chronic pain management, and diabetes management as an adjunct to existing drug therapy. The category tends to work best for conditions where the underlying clinical intervention, such as cognitive behavioral therapy or structured behavioral coaching, doesn't inherently require in-person delivery.

How are digital therapeutics regulated?

In the United States, most go through the FDA's 510(k) or De Novo pathways as medical devices, sometimes as Software as a Medical Device (SaMD). In the EU, they typically require a CE mark under the Medical Device Regulation. Both frameworks generally require evidence of safety and efficacy before market clearance. International groups such as the International Medical Device Regulators Forum have published shared terminology for software as a medical device, which helps companies plan submissions across several markets at once.

Why have some digital therapeutics companies failed despite FDA clearance?

Regulatory clearance proves clinical efficacy, not commercial viability. Several well-evidenced DTx products struggled because payers were slow to reimburse at a sustainable price, adoption workflows for prescribing clinicians were unclear, and patient adherence outside the trial setting was lower than expected — a reminder that clinical validation and market fit are separate problems.

Can AI make digital therapeutics more effective?

AI can enable more adaptive, personalized interventions that adjust in real time to a patient's behavior or symptoms, which is an active area of development. It also raises new regulatory complexity, since an algorithm that changes its own behavior over time is harder to validate with a single static clinical trial than a fixed software program. Regulators have been developing approaches such as predetermined change control plans, where a company describes in advance how an algorithm is allowed to change, but builders should expect adaptive AI features to add validation work rather than remove it.

Conclusion

Digital therapeutics take a simple idea, software that treats disease, and hold it to the standards of medicine: a defined clinical mechanism, controlled trials against real endpoints, regulatory clearance, and ongoing post-market evidence. That bar is what separates a DTx from the wider world of health apps, and it is also what makes a cleared product hard to copy.

The harder lesson from the category is that clinical success does not guarantee commercial survival. Reimbursement is still fragmented, adherence drops off outside trial conditions, and every meaningful software update can carry regulatory weight. Access depends on smartphones, connectivity, and digital literacy that not every patient has. Health systems evaluating products should look past the "clinically validated" label to the specific endpoint, the adherence data, and the billing route. Builders should plan the evidence program and reimbursement strategy as seriously as the software itself.

If you are planning a regulated health product, start by writing down the clinical endpoint you intend to move and how a payer would pay for moving it. When you need engineering help with the software, data, and validation infrastructure behind that plan, our healthcare AI development team can help you scope it.