Early Health Technology Assessment Explained

A promising prototype can still solve the wrong problem. It can require a workflow that no clinic can support, compare itself with outdated care, or reach pivotal testing without measuring the outcome that a health system needs. Early health technology assessment, usually shortened to early HTA, brings those questions forward while development choices can still change.

In 2025, an HTAi working group reported a consensus definition: early HTA evaluates the potential value of a conceptual or actual health technology to inform subsequent development, research, or investment. The timing matters, but the decision matters more. The task is not to produce a premature reimbursement verdict. It is to make uncertainty useful.

Start with the decision, not the device#

Conventional HTA is often associated with decisions about adoption, coverage, price, or use after a technology has accumulated substantial evidence. Early HTA moves assessment upstream. The object might be an idea, a prototype, an early diagnostic pathway, a digital tool, a service change, or a medicine whose clinical program is still being designed.

The first question is therefore not, “Is this cost-effective?” It is, “Which decision needs help now?” A public funder may be deciding whether to support another study. A development team may be choosing between two designs. A hospital may be considering a pilot. A research group may need to decide which population, comparator, and outcomes belong in a trial.

That decision defines the scope. It also prevents the analysis from becoming a sales document. An assessment commissioned to choose the next study should not be presented as proof of future adoption.

The broader guide to health technology assessment explains how later comparative, economic, organizational, ethical, and social questions connect to policy. Early HTA uses many of the same disciplines for a different immediate purpose.

Define the problem before estimating value#

An innovation has no value in the abstract. Value depends on the people affected, the current pathway, realistic alternatives, outcomes that matter, resources used, and the perspective of the decision maker.

A useful early assessment describes the unmet need without assuming the proposed technology is the answer. How large is the problem? Where does the present pathway fail? Who bears the burden? Are the relevant delays, harms, costs, or access barriers measured, or merely asserted?

The “proof-of-problem” approach is helpful here. Before refining a solution, the team tests whether clinicians, patients, and system leaders recognize the same problem and whether the available room for improvement is meaningful. An elegant device aimed at a minor inconvenience may offer less value than a simple process change.

The comparator must also be plausible. Comparing a new test with “no testing” is misleading when the real alternative is a well-established pathway. Comparing a digital service with an idealized clinic hides the question of how care is actually delivered.

Potential value is a conditional claim#

Early evidence is sparse by definition. A model may rely on small studies, analogous products, expert elicitation, early technical performance, or assumptions about price and uptake. Those inputs can support a decision if their limitations remain visible.

The most informative result is often conditional: the technology could be valuable if sensitivity exceeds a threshold, if it prevents a specified number of admissions, if its unit cost stays below a ceiling, or if enough eligible people can use it safely.

Threshold analysis reverses the usual question. Instead of claiming a likely answer from uncertain inputs, it asks what performance would be required to change the decision. Scenario analysis shows what happens under coherent optimistic, central, and pessimistic cases. Probabilistic analysis can display how uncertainty across inputs affects the result, provided the assumed distributions are defensible.

False precision is a warning sign. A model built on speculative inputs does not become certain because it produces three decimal places. Report the assumptions that drive the conclusion and the range in which the decision changes.

Evidence planning is a central product#

Early HTA can connect a development program with the evidence later decision makers will need. It can clarify the target population, place in the pathway, comparator, patient-relevant outcomes, follow-up period, resource measures, and subgroups.

This is not the same as designing a trial solely to satisfy a future economic model. Clinical validity, safety, ethics, regulatory requirements, feasibility, and scientific importance remain separate responsibilities. The benefit is alignment: a study can answer its scientific question while avoiding obvious gaps in comparative or implementation evidence.

Value-of-information methods ask whether reducing a particular uncertainty could improve decisions enough to justify more research. The method can help choose between a larger trial, a targeted usability study, a registry, a smaller technical study, or no further work. Its result depends on the model and decision context, so it should guide judgment rather than replace it.

The guide to first-in-human dose selection illustrates why technical, ethical, and clinical evidence questions cannot be collapsed into one economic estimate.

Users and systems belong in the assessment#

Economic modeling receives much attention, but early HTA is multidisciplinary. Interviews, observation, process mapping, discrete-choice work, usability testing, and stakeholder workshops can reveal whether the proposed pathway is understandable and feasible.

A home diagnostic may reduce travel while shifting work to patients and caregivers. A decision-support tool may improve one metric but create alert burden. A device may perform well in a laboratory and fail when cleaning, calibration, storage, connectivity, or training is considered. A service may help digitally confident users while excluding people who need language support or accessible formats.

These findings can change the design before expensive testing begins. They can also show that the technology should not proceed. Stopping an unpromising pathway is a legitimate result, not a failed assessment.

Participation needs a stated role. Asking patients or clinicians for comments after the product requirements are fixed is not co-design. The assessment should record who contributed, which decision their input affected, and what remained outside their authority.

Keep early HTA separate from neighboring activities#

Early HTA is related to horizon scanning, early scientific advice, and market-access planning, but the terms are not interchangeable.

Horizon scanning identifies technologies that may affect a system and supports preparation. Early dialogue or scientific advice usually involves structured discussion about an evidence plan with regulators, HTA bodies, or both. Market-access planning may focus on the route to price, coverage, procurement, and adoption. Early HTA is the assessment itself, conducted to inform a defined development, research, or investment decision.

The same project may use all four activities. Clear labels matter because each has different participants, authority, confidentiality, and outputs. Advice from one body does not bind another, and an early assessment does not prejudge a later formal decision.

Reassessment is part of the design#

An early model becomes stale quickly. Prototype performance changes. A trial reports. A competitor enters. The intended population narrows. Manufacturing cost moves. A clinical guideline changes the comparator.

A sound report names triggers for reassessment and preserves enough detail to update the analysis. It distinguishes evidence from assumptions and records the date of each input. It also states whether findings are transferable to another setting, because prices, staff capacity, pathways, and social values vary.

The assessment should become more demanding as the technology matures. Early scenarios may be broad. Later work should replace assumptions with measured data, validate models, examine distributional effects, and compare the technology with current care.

A practical early HTA report#

Read an early HTA by tracing five elements:

1. the decision and the person or body making it;
2. the health problem, population, pathway, and comparator;
3. the evidence and assumptions used for each important input;
4. the uncertainty, thresholds, and implementation constraints; and
5. the action the analysis supports now.

The strongest report makes it easy to disagree. It shows which assumption drives the result and what new information would change the recommendation. The useful endpoint is not a favorable number. It is a better next decision.

References#

1. HTAi Early HTA Terminology Working Group
2. Consensus definition of early health technology assessment
3. Supporting translation with early health technology assessment
4. WHO overview of health technology assessment
5. WHO Health Technology Assessment of Medical Devices, second edition
6. Using HTA to set priorities, inform target product profiles, and design studies

This article is educational, not regulatory, reimbursement, procurement, investment, or product-development advice. Requirements and decision authority vary by setting.