22,000 Members Treated: How a Digital Therapy Benefit Cut Depression and Hospital Costs

22,000 People, One App, Real Results

Can a health insurer deliver effective mental healthcare through a smartphone app — and prove it works? A new case study published on October 14, 2026, by Amwell suggests the answer is yes. The study examines Discovery Health’s use of SilverCloud by Amwell to provide a fully funded digital mental health benefit in South Africa, and the numbers are striking: 22,000 member activations in the first 12 months, clinically significant improvement rates of 60 to 78 percent, 90 percent user satisfaction, and 6 percent lower adjusted mental-health-related hospital costs in the year after use.

The case study, announced from Boston and Johannesburg, tackles the three questions every health plan asks about digital care: will members actually use it, does it improve their health, and what does it do to costs? On all three, Discovery Health’s programme delivers affirmative, quantified answers.

Discovery Health introduced the benefit in late 2023 as South Africa’s first funded digital therapeutic benefit for mental health. Rather than launching a standalone digital service and hoping members would find it, the insurer embedded SilverCloud directly into its existing mental healthcare benefit, clinical referral pathway and member experience — a design choice the case study credits for the high uptake.

How the Programme Works

Eligible adult members with mild-to-moderate depression or anxiety can be referred by a GP or psychologist, activate the benefit through the Discovery Health App, and begin structured internet-based cognitive behavioural therapy — iCBT, one of the most evidence-backed psychological treatments in existence.

Members work through evidence-based digital CBT content while receiving personalised written guidance and feedback from trained human supporters, who monitor progress and help members stay engaged. Referring healthcare professionals remain part of the care pathway throughout, preserving the link between digital treatment and the traditional healthcare system.

That hybrid design — digital content plus human support plus clinical integration — distinguishes the programme from the crowded field of direct-to-consumer wellness apps. It is treatment, not content: structured, supervised and connected to the member’s broader care.

The Outcomes in Detail

The clinical results are the study’s centrepiece. Members who started with moderate, moderately severe or severe depression, and members with moderate or severe anxiety, showed clinically significant improvement rates of 60 to 78 percent across depression and anxiety measures, with the rate varying by baseline severity.

“Clinically significant” is doing important work in that sentence. It means the improvements were large enough to matter in patients’ lives — not just statistically detectable blips on a questionnaire, but changes that cross the threshold of meaningful recovery. Improvement rates of 60 to 78 percent compare favourably with outcomes reported for face-to-face therapy in similar populations.

Satisfaction was equally strong: 90 percent of surveyed users reported being satisfied with the benefit. And the cost analysis found 6 percent lower adjusted mental-health-related in-hospital costs in the year after use, after adjusting for baseline experience and confounding factors — suggesting the programme may reduce the need for expensive crisis care.

Why Activation Is the Hard Part

The case study’s emphasis on “activation first” reflects a hard-won lesson from digital health. The graveyard of mental health apps is filled with products that demonstrated efficacy in trials but failed in the real world because nobody used them. Engagement, not clinical content, is usually the binding constraint.

Discovery Health’s answer was structural: build the digital benefit into the benefit members already use. Eligible members encounter it through their GP, their psychologist or the app they already open — not through an advertising campaign for a separate service. The 22,000 activations in year one suggest the strategy worked.

The human supporters are the second half of the engagement equation. Pure self-guided digital therapy suffers from high dropout; the addition of trained supporters who provide personalised feedback is among the best-evidenced methods for keeping people engaged — and the outcomes here are consistent with that literature.

What It Means for Mental Healthcare Access

The broader significance lies in access. South Africa, like most countries, faces a severe shortage of mental health professionals relative to need. Traditional models — one therapist, one patient, one hour — cannot scale to meet population demand. Digital therapeutics, delivered through existing insurance infrastructure, offer a path to scale that does not require training tens of thousands of new clinicians.

The cost finding matters for the same reason. Health systems and insurers will fund what demonstrably works and demonstrably saves money; a 6 percent reduction in hospital costs, alongside strong clinical outcomes, makes the economic case as well as the medical one. Prevention and early intervention are cheaper than crisis care — a principle the data now quantifies.

The model is also exportable. Nothing in the programme’s design is specific to South Africa: funded digital CBT, embedded in existing benefits, with human support and clinical integration, could be replicated by insurers and health systems worldwide. The case study functions, in effect, as a blueprint.

Caveats and Open Questions

As a company-published case study rather than an independent randomised trial, the findings should be read with appropriate caution. Amwell has a commercial interest in demonstrating SilverCloud’s effectiveness, and case studies do not carry the evidentiary weight of peer-reviewed research with control groups.

Questions remain about longer-term durability: do the improvements persist beyond the first year? About selection effects: are the members who activate the benefit systematically different from those who don’t? And about the 6 percent cost figure, which depends on statistical adjustments that independent researchers would want to scrutinise.

Still, the combination of large-scale real-world uptake, strong satisfaction, clinically significant improvement rates and directionally favourable costs is more than most digital health programmes can show. As health systems worldwide grapple with the gap between mental health need and available care, Discovery Health’s experiment offers something the field desperately needs: evidence, at scale, that a different model can work.

The Global Treatment Gap Digital Care Could Fill

The Discovery Health experiment matters most in the context of the global mental health treatment gap. Worldwide, the large majority of people with depression and anxiety receive no treatment at all — a gap driven by clinician shortages, cost, distance and stigma. Traditional service models, requiring one trained professional per patient per hour, cannot scale to close it in any country, let alone in low- and middle-income settings.

Digital therapeutics attack the scaling problem directly. A well-designed iCBT programme can serve thousands of patients simultaneously, with human supporters handling caseloads far larger than face-to-face therapy allows. The marginal cost of each additional user is a fraction of traditional care — which is precisely why insurers, always attentive to unit economics, are among the earliest adopters.

The South African setting is instructive. A middle-income country with significant mental health need and constrained specialist capacity, it represents exactly the environment where scalable models are most valuable. That a commercial insurer funded the benefit — rather than a research grant or philanthropy — suggests the model has crossed from pilot to product.

What Comes Next for Digital Therapeutics

The field’s next questions are about depth and durability. Can digital programmes handle more severe and complex presentations, or will they remain best suited to mild-to-moderate cases? Do the gains persist for years, or do users need booster engagement? How do outcomes compare head-to-head with face-to-face therapy in randomised trials, rather than case studies?

Regulation is evolving in parallel. Digital therapeutics occupy a blurred space between wellness apps and medical devices, and regulators worldwide are still defining the evidentiary bar for claims of clinical effectiveness. Programmes that publish real-world outcomes data — as Discovery Health and Amwell have done here — will be better positioned as standards tighten.

Integration is the other frontier. The most promising models, including this one, do not position digital care as a replacement for clinicians but as an extension of them — triaging mild cases digitally, escalating complex ones to humans, and keeping referring professionals in the loop. In that architecture, the app is not the therapist’s rival but the therapist’s force multiplier.

If the 60-to-78 percent improvement rates hold up under independent scrutiny, the implications extend well beyond one insurer in one country. They suggest that the treatment gap — among the most stubborn failures in global health — may finally have a scalable answer. Twenty-two thousand South Africans have already tested that proposition. The rest of the world is watching.

Evidence at Scale

Digital mental health has long promised scale; Discovery Health’s programme is among the first to demonstrate it with hard numbers. Twenty-two thousand activations, majorities improving, costs falling — the case study sets a benchmark the industry will now be measured against.