Living HTA

Living HTA in Reimbursement and Reassessment

A living HTA keeps the whole assessment of a technology current, from the evidence review through the economic model to the recommendation, instead of treating the first appraisal as final. Thokala and colleagues, who proposed the working definition in 2023, describe a full assessment planned from the outset for updates at regular intervals or at specific trigger points. The idea matters to health economists because cost-effectiveness depends on inputs that change after launch, such as longer trial follow-up, new comparators, real-world data and prices. It costs more than a one-off appraisal, so the open questions are when it is worth doing and how updates feed into reimbursement.

Why a one-off appraisal goes out of date

A conventional health technology assessment rests on a snapshot of the evidence, usually fixed at the date of the last literature search. Thokala and colleagues note that agencies which do update assessments usually do so a few years apart. Shields and colleagues group the reasons an economic evaluation loses relevance into changes to the decision problem, such as new comparators, changes to parameters such as costs, utilities and resource use, and methodological updates.

The evidence synthesis ages when new trials or longer follow-up appear. The health economic model ages when a new technology becomes standard of care, when survival data extend beyond the trial period, or when methods change, for example the threshold, discount rate or perspective. The report's conclusions age whenever either input does.

From living reviews to living guidelines and living HTA

The living approach began with evidence synthesis. Elliott and colleagues proposed living systematic reviews in 2014 as high quality, up-to-date online summaries updated as new research becomes available. Elliott and colleagues also warned that repeated naive testing of an updated meta-analysis inflates false-positive findings, the type I error problem.

Living guidelines followed. Akl and colleagues define them as an optimisation of guideline development that allows individual recommendations to be updated as soon as new relevant evidence becomes available, so the unit of update is the recommendation. During COVID-19 a living network meta-analysis of drug treatments in The BMJ informed the World Health Organization's living guideline, and Elvidge and Dawoud report that NICE and WHO adopted a living approach for their COVID-19 treatment guidelines.

Living HTA extends the same logic to the whole assessment. Thokala and colleagues stress that a living review or living model alone does not cover the full process from searching to reporting.

How a living HTA is planned and run

Thokala and colleagues describe two ways to run one. In a manual living HTA, researchers update and combine each part by hand, helped by standard templates for passing data between steps. In a semi-automated version, scripts in R or Python link searching, screening, meta-analysis, modelling and reporting, with analysts checking each stage. At the time of writing, HTA agencies did not accept automated processes in submissions, and the authors knew of no published living HTA.

Several design choices are fixed at the start.

1. When to update. Updates may run at set intervals, for example every 12 months, or at triggers such as a new treatment entering the pathway, major new evidence, a population expansion or a company leaving the market. Updates that are too frequent burden stakeholders and may confuse users if a treatment moves in and out of favour.
2. What to update. If the model structure still holds, new parameter values are entered. A changed care pathway may need a new structure or comparators, which is much harder to automate than a new data cut.
3. How long to stay living. A living HTA need not run indefinitely and can close once the evidence base stabilises.

Smith, Schneider and Mohammed built an R pipeline in which an automated workflow sends the model to an interface hosted by the data owner, so sensitive data stay with the owner, and the analysis and report can run monthly or when the underlying data change.

Prices, confidentiality and the clean analysis

Confidential prices complicate every update. Thokala and colleagues note that frequent changes in commercial discounts for a technology and its comparators can make each reassessment less transparent than the last. They suggest a consistent clean analysis at each update, using all technologies' public list prices as at the first assessment. It cannot support the decision, since list prices are not the true cost, but it shows how much of any change comes from the evidence. Cheyne and colleagues add that reimbursement and procurement systems may not be designed for frequent price changes, and that renegotiation needs to allow price increases as well as decreases or disinvestment.

Worked example separating new evidence from a new price

All numbers in this example are illustrative. A drug is appraised at an illustrative cost-effectiveness threshold of 30,000 GBP per QALY, and a living HTA updates it two years later after a longer data cut and a new confidential discount. The comparison uses net monetary benefit per patient.

$$\mathrm{NMB} = \lambda \, \Delta Q - \Delta C$$

where $\lambda$ is the threshold in GBP per QALY, $\Delta Q$ is the incremental QALYs per patient and $\Delta C$ is the incremental cost per patient in GBP.

Step 1 (first appraisal). Dividing incremental cost by incremental QALYs gives the ICER, $\mathrm{ICER} = 24{,}000 / 0.90 = 26{,}667$ GBP per QALY, and $\mathrm{NMB} = 0.90 \times 30{,}000 - 24{,}000 = 3{,}000$ GBP.

Step 2 (evidence effect). The longer follow-up cuts the QALY gain to 0.60. At the original list price $\mathrm{ICER} = 24{,}000 / 0.60 = 40{,}000$ GBP per QALY and $\mathrm{NMB} = 0.60 \times 30{,}000 - 24{,}000 = -6{,}000$ GBP, a fall of 9,000 GBP per patient.

Step 3 (price effect). The discount lowers the incremental cost by 7,200 GBP to 16,800 GBP, giving $\mathrm{ICER} = 16{,}800 / 0.60 = 28{,}000$ GBP per QALY and $\mathrm{NMB} = 0.60 \times 30{,}000 - 16{,}800 = 1{,}200$ GBP.

Step 4 (interpretation). The net change $\Delta \mathrm{NMB} = 1{,}200 - 3{,}000 = -1{,}800$ GBP equals the evidence effect of minus 9,000 GBP plus the price effect of 7,200 GBP. The split is exact here only because the new data changed QALYs and not costs. Without the discount, 2,000 patients a year would lose $\text{net QALYs} = 2{,}000 \times 6{,}000 / 30{,}000 = 400$ a year at the threshold. The final ICER alone, 28,000 against 26,667 GBP per QALY, would hide that the clinical case weakened considerably and that the price now carries the recommendation.

Decision use, value of information and managed access

Claxton argued that adoption should depend only on expected net benefit, while the distribution of net benefit matters for whether more information is needed. Value of information analysis puts a value on that second question, and so tests whether a scheduled update or further data collection is worth its cost. Eckermann and Willan showed that costs of reversing an adoption decision reduce the expected value of information, and Walker and colleagues weigh the benefit of coverage against the chance of reversal and the risk that coverage hinders evidence generation. Sarri and colleagues describe a Swedish Parkinson's disease treatment withdrawn after unconvincing evidence at reassessment and reimbursed again after follow-up data were re-evaluated.

For life-cycle HTA, Kirwin and colleagues propose risk-based pricing and research-oriented managed access, with value of information methods to inform those agreements and contractual decision rules that simplify reassessment. The international definition of HTA agreed in 2020 states that HTA can apply at different points in a technology's lifecycle, through to disinvestment.

NICE's technology appraisal manual (PMG36) contains partial elements of this. A committee can recommend a medicine with managed access when it has the plausible potential to be cost effective at the currently agreed price but the evidence is too uncertain, and the new data could feasibly be collected within up to 5 years without undue burden. NICE updates the guidance at the end of the data collection period, may consider an earlier update if the joint NICE and NHS England and NHS Improvement managed access governance group agrees, and cannot recommend the medicine with managed access again at that update. Unless the data collection arrangement says otherwise, the update uses the process and methods in place when NICE invites participation, so a methods change can move the result as well as new data. Outside managed access, PMG36 updates guidance when changes to the evidence base, clinical pathway or economic case are likely to have a material effect on the recommendations. That is trigger-based surveillance rather than an assessment planned from the outset to be living, and the funding route is described under managed access agreement.

Boundaries with nearby ideas

A living review is the evidence-synthesis component, and a living guideline updates clinical recommendations, which Cheyne and colleagues judge easier to change than an HTA decision because HTA also involves unpublished data, confidential prices and reimbursement rules. Life-cycle HTA is the policy framework of assessment from development to disinvestment, and a living HTA is one way to deliver it. Early health technology assessment happens before launch, while a living HTA usually starts around the first reimbursement decision. Reassessment and managed access updates are single revisits rather than a maintained assessment.

Limitations and common misreadings

The evidence base is mostly conceptual. Cheyne and colleagues and Sarri and colleagues both call for pilots, and the confidentiality of prices and of some clinical inputs limits how far updates can be shared or automated.

1. Living HTA means updating constantly. The plan sets intervals or triggers, because frequent changes carry costs.
2. A living model is a living HTA. A model updated alone does not refresh the evidence review, the comparators or the recommendation.
3. Updates only ever widen access. A living approach implies a commitment to disinvest, or to restrict use, when a technology stops being cost effective.

Sources

• Thokala P, Srivastava T, Smith R, Ren S, Whittington MD, Elvidge J, Wong R, Uttley L. Living health technology assessment: issues, challenges and opportunities. PharmacoEconomics. 2023;41(3):227-237. https://doi.org/10.1007/s40273-022-01229-4
• Cheyne S, Chakraborty S, Lewis S, Campbell S, Turner T, Norris S. What could health technology assessment learn from living clinical practice guidelines? Frontiers in Pharmacology. 2023;14:1234414. https://doi.org/10.3389/fphar.2023.1234414
• Sarri G, Forsythe A, Elvidge J, Dawoud D. Living health technology assessments: how close to living reality? BMJ Evidence-Based Medicine. 2023;28(6):369-371. https://doi.org/10.1136/bmjebm-2022-112152
• Shields GE, Pennington B, Bullement A, Wright S, Elvidge J. Out of date or best before? A commentary on the relevance of economic evaluations over time. PharmacoEconomics. 2022;40(3):249-256. https://doi.org/10.1007/s40273-021-01116-4
• Elliott JH, Turner T, Clavisi O, Thomas J, Higgins JPT, Mavergames C, Gruen RL. Living systematic reviews: an emerging opportunity to narrow the evidence-practice gap. PLoS Medicine. 2014;11(2):e1001603. https://doi.org/10.1371/journal.pmed.1001603
• Akl EA, Meerpohl JJ, Elliott J, Kahale LA, Schünemann HJ; Living Systematic Review Network. Living systematic reviews: 4. Living guideline recommendations. Journal of Clinical Epidemiology. 2017;91:47-53. https://doi.org/10.1016/j.jclinepi.2017.08.009
• Siemieniuk RA, Bartoszko JJ, Zeraatkar D, Kum E, Qasim A, Díaz Martinez JP, et al. Drug treatments for covid-19: living systematic review and network meta-analysis. BMJ. 2020;370:m2980. https://doi.org/10.1136/bmj.m2980
• Elvidge J, Dawoud D. Assessing technologies for COVID-19: what are the challenges for health technology assessment agencies? Findings from a survey and roundtable workshop. PharmacoEconomics. 2021;39(12):1455-1463. https://doi.org/10.1007/s40273-021-01097-4
• Smith RA, Schneider PP, Mohammed W. Living HTA: automating health economic evaluation with R. Wellcome Open Research. 2022;7:194. https://doi.org/10.12688/wellcomeopenres.17933.2
• Claxton K. The irrelevance of inference: a decision-making approach to the stochastic evaluation of health care technologies. Journal of Health Economics. 1999;18(3):341-364. https://doi.org/10.1016/S0167-6296(98)00039-3
• Eckermann S, Willan AR. Expected value of information and decision making in HTA. Health Economics. 2007;16(2):195-209. https://doi.org/10.1002/hec.1161
• Walker S, Sculpher M, Claxton K, Palmer S. Coverage with evidence development, only in research, risk sharing, or patient access scheme? A framework for coverage decisions. Value in Health. 2012;15(3):570-579. https://doi.org/10.1016/j.jval.2011.12.013
• Kirwin E, Round J, Bond K, McCabe C. A conceptual framework for life-cycle health technology assessment. Value in Health. 2022;25(7):1116-1123. https://doi.org/10.1016/j.jval.2021.11.1373
• O'Rourke B, Oortwijn W, Schuller T; International Joint Task Group. The new definition of health technology assessment: a milestone in international collaboration. International Journal of Technology Assessment in Health Care. 2020;36(3):187-190. https://doi.org/10.1017/S0266462320000215
• National Institute for Health and Care Excellence. NICE technology appraisal and highly specialised technologies guidance: the manual (PMG36). Published 31 January 2022; last updated 13 October 2026. https://www.nice.org.uk/process/pmg36