If you’re building a digital health technology and you’ve ever tried to put a number on what your product is worth, you’ve probably hit the same wall. Health economics has a robust set of methods for valuing interventions: cost-utility analysis, value of statistical life, contingent valuation, economy-wide modelling. They’re what NICE uses. Every HTA body in every developed market relies on them. Applied to a digital product, they leave a lot of the value on the table.
The methods aren’t wrong. They were built for drugs, devices, and services with bounded effects, fixed marginal costs, and clear comparators. Digital health doesn’t behave like that. It generates data, makes predictions, has optionality, and reshapes workflows in ways that ripple beyond the patient in front of you. The toolkit is the foundation, not the whole building.
This is the first of four pieces on monetising digital health value. The series builds on the World Bank’s 2023 framework for evaluating digital health interventions — specifically Table 7, where the authors offer a thin set of pointers to monetisation methods for value categories like data, prediction, and disruption potential. Three years on, those gaps remain. Pieces two through four work through them. This piece sets the foundation.
Why this matters now?
US digital health funding hit $14.2 billion in 2025, up 35% on 2024, and AI-enabled startups captured 62% of H1 2025 venture funding, the first majority share for AI-enabled companies since Rock Health began tracking the sector. Capital is concentrating in workflow tools: ambient scribes, decision support, revenue cycle automation, where the value proposition rests on outputs traditional health economics struggles to price: clinician time, data flywheels, predictive accuracy.
NICE has just confirmed the first change to its cost-effectiveness threshold in 26 years, from £20–30k per QALY to £25–35k from April 2026. Founders pitching DHTs into the NHS are operating against a valuation framework that took a quarter of a century to budge by £5,000. If your product’s value lives mostly inside that framework, fine. If it lives mostly outside it, the existing toolkit can’t price what you’ve built.
The four established methods, and what they’re good for
The Wilkinson framework cites four well-established approaches to monetising health intervention value. Each is robust and well-evidenced. None alone does justice to a digital product.
1. Value of Statistical Life (VSL)
VSL values mortality risk reduction by asking how much a population would pay to lower their risk of death by a small amount, then backing out the implicit value of a life saved. It’s codified for global health in the Reference Case Guidelines by Robinson and colleagues at Harvard, and remains the standard reference.
Where it works: a digital therapeutic for severe heart failure that demonstrably reduces 30-day mortality. Where it strains: an ambient AI scribe. The product saves nobody’s life directly — it saves clinicians 90 minutes a day. Mortality impact is several causal links downstream and the signal-to-noise kills you. VSL gives nothing useful for the products attracting most of the capital right now.
2. Efficiency analysis
Efficiency analysis values process improvements: reduced waiting times, fewer unnecessary appointments, lower administrative burden. For a DHT founder it’s often the most immediately useful framework — most digital products are sold on efficiency stories.
Where it works: a triage chatbot reducing low-acuity GP appointments. You quantify the time saved, monetise it at a defensible hourly rate, and drop it into a commissioner’s business case. Where it strains: efficiency gains that don’t cash out. The classic NHS problem — a tool saves 90 GP minutes a day, but the practice doesn’t fire 0.2 of a GP. The time gets reabsorbed. Real value is created; none shows up on the balance sheet. Commissioners discount these numbers aggressively, often by 60–80%.
3. Contingent valuation
Contingent valuation asks people directly how much they’d pay for a benefit. It’s the workhorse method when there’s no observed market — used in environmental economics, and in health for valuing things like reduced anxiety, dignity, or convenience. Bayoumi’s review in PharmacoEconomics remains the standard reference.
Where it works: consumer-facing products where willingness-to-pay is the revenue model — a wearable, a wellness app, a DTC diagnostic. Where it strains: B2B digital health, where the buyer (NHS commissioner, US payer, hospital procurement) is not the user (patient, clinician). Asking patients what they’d pay for a tool a commissioner is buying tells you nothing about what the commissioner will pay. Asking commissioners directly produces lowball anchored numbers untethered from the value being created.
4. Economy-wide effects
Economy-wide modelling captures wider productivity and labour market impacts: people back at work, reduced carer burden, second-order GDP effects. Strzepek and colleagues’ work for the Benefit-Cost Analysis Reference Case is the methodological anchor.
Where it works: large-scale interventions with measurable workforce impact — a national diabetes prevention programme, a mental health platform across a working-age population. Where it strains: most early-stage DHTs. The required data infrastructure — input-output tables, labour market elasticities, regional productivity coefficients — is far heavier than a Series A can produce. The analysis lives in academic papers and consultancy reports, not in business cases.
Table 1. The four established methods at a glance.
Method | What it monetises | Works well for | Where it strains for DHIs |
Value of Statistical Life | Mortality risk reduction | A DTx for severe heart failure that reduces 30-day mortality | Workflow tools — any mortality effect is several causal links downstream, signal-to-noise kills you |
Efficiency analysis | Process improvements, time savings | A triage chatbot that demonstrably reduces low-acuity GP appointments | Time gains that don’t cash out — commissioners discount these 60-80% |
Contingent valuation | Willingness-to-pay for unobservable benefits | Consumer-facing apps and wearables with direct revenue from users | B2B sales where buyer (commissioner) isn’t user (patient) — WTP signals disconnect |
Economy-wide effects | Productivity, labour market, GDP impact | National-scale prevention programmes with measurable workforce effect | Data infrastructure required is far heavier than a Series A company can produce |
The pattern: where the toolkit ends
The four methods share a model. The intervention is a discrete thing that does a discrete job. Take this drug, get this outcome. Cost the inputs, value the outputs, compare to comparator, divide. The methods are sophisticated; the underlying model is simple.
Digital health doesn’t fit. Three things sit outside it.
The data the product generates. A drug doesn’t produce data as a byproduct of being taken. A DHT does — every interaction is a data point with economic value to the operator (training), the health system (population insight), and third parties (research). None of the four methods price this. Wilkinson’s framework flags “Value of Data” as a category but cites only a Data Shapley paper from ML research and a consultancy blog post on data as enterprise asset. That’s not a method. That’s a research agenda. Piece two works through it.
The value of better prediction. Most AI-enabled DHTs are, fundamentally, prediction products — diagnoses, deteriorations, no-shows, readmissions. The economic value of a marginally better prediction is real and modellable, but ROC curves and sensitivity-specificity statistics don’t convert to monetary terms. Adjacent methods exist (decision curve analysis, value of information analysis) but sit outside the standard toolkit and are rarely applied to DHTs in commercial business cases. Piece three pulls these together.
The optionality of platforms. A digital platform isn’t valued like a drug — it’s valued like a venture. Its worth depends not on what it does today but on what it could become: new indications, new integrations, new capability layered on the same data and infrastructure. DCF systematically underprices this. Pharma R&D has used real options theory for two decades to handle exactly this problem. Almost no one applies it to digital health. Piece four makes the case.
What this means if you’re building a DHT
Know what your product’s value model actually is. If your value lives inside the established toolkit — clear health outcome, defensible efficiency saving, measurable willingness-to-pay — use the established methods. They’re well understood, well accepted, and they’ll get you through HTA. Don’t reinvent what you don’t need to.
Know what the toolkit doesn’t see. If meaningful value lives in the data your product generates, the predictions it makes, or the optionality it creates, the established methods will underprice you. Commissioners and investors using them will arrive at numbers that don’t reflect what you’ve built. The answer isn’t to abandon the toolkit — it’s to layer additional value categories on top of it.
Get ahead of the gap. The NICE Evidence Standards Framework is being criticised in peer-reviewed literature for lacking the adaptability rapidly evolving DHTs require. The methods are catching up. Founders who’ll do best over the next 3–5 years are those who can articulate value in the language of the established framework and demonstrate the value categories it doesn’t yet handle. That’s a defensible position. “Trust me, our data is valuable” is not.
What’s next in this series
Piece two tackles data value — how to monetise the data asset a DHT generates, drawing on Data Shapley from ML research and enterprise data valuation from corporate finance. Piece three handles prediction value — converting accuracy improvements into business case numbers using decision curve analysis and value of information. Piece four makes the case for real options theory as the right valuation framework for DHT platforms.
If you’re building a business case right now, the established toolkit is necessary and not sufficient. The series works through what to add.
Healthonomix advises digital health founders, NHS commissioners, and investors on health economics, market access, and reimbursement strategy. If you’re working on a business case and the established methods aren’t telling the full story, get in touch.
