Economic evaluation of using polygenic risk score to guide risk screening and interventions for the prevention of type 2 diabetes in individuals with high overall baseline risk

1️⃣ ANALYTICAL SUMMARY

Context and issues — a question of resource allocation, not field clinic. Type 2 diabetes (T2D) is progressing globally; the study cites the IDF (2021) which estimates the burden at around several hundred million people by 2030 (data cited, not produced by the study — see §preliminary note) (p. 2). In the face of constrained resources, the authors raise the issue of targeting: how to identify, at controlled cost, the individuals who would benefit most from intensified prevention? The modeled population: Finnish adults aged 30–79 years, without T2D, already classified as “high risk” (10–20% at 10 years) by the FINDRISC questionnaire — representing 10.5% of the adult population, ~313,000 people, average age 61.5 years, 63.9% men (p. 3).

Operational contributions — an economic demonstration, to be handled with caution. The study constructs an individual state transition model (healthy / T2D / T2D with complications / death) over a lifetime horizon, and compares two strategies: usual identification alone, or the addition of a polygenic score in a second step to identify individuals at very high risk (> 20%) eligible for a lifestyle intervention complemented by medical treatment (p. 2–4). Study-specific result: the strategy with the score is “dominant” (less costly and more effective), with an average saving of €253 per person and +0.022 QALY (p. 7). The contribution for the field actor is not a tool, but a framework for reasoning on the cost/benefit/uncertainty ratio of targeted screening.

2️⃣ KEY POINTS OF THE DOCUMENT

1️⃣ A two-step strategy, not a widespread genetic test. The polygenic score is only applied after a pre-screening by FINDRISC, on the subgroup already at high risk. The authors emphasise that genetic screening without this pre-filter would be less cost-effective and more burdensome budget-wise (p. 3 ; p. 8).

2️⃣ Measured effect: a reclassification of 12.4 percentage points. The score moves 12.4% of individuals into the “very high risk” category who would have remained in “high risk” with usual screening — approximately 38,000 people out of the 313,000 targeted (p. 7 ; p. 8).

3️⃣ Economic outcome: dominant strategy in the base case. Costs: €13,373 vs €13,619 (usual); QALY: 12.15 vs 12.13; thus −€253 and +0.022 QALY per person, lifetime horizon, discounting 3% (p. 7, Table 4).

4️⃣ The result relies on a low test price — and a precise threshold. The test is assumed to be €50; the “cost neutrality point” (where the economic advantage cancels out) is reached at a test price of €303 in the base case, but drops to €19 over a 10-year horizon (p. 4 ; p. 7, Table 5). The time horizon thus weighs heavily.

5️⃣ An explicitly assumed uncertainty. At a willingness to pay of 0 €/QALY, the probability that the strategy is dominant is only 63 %, and the expected value of perfect information (EVPI) is 243 € per person (p. 7). In other words: the meaning of the result is robust, but its magnitude remains uncertain.

3️⃣ ACTION TRACKS FOR LOCAL STAKEHOLDERS

Reserve (R5): this document does not contain any field tool, checklist or mobilisable protocol. The tracks below are therefore argumentative and reflective, not procedures. Unmet needs are indicated in point 6.

1️⃣ Defend the logic of tiered detection. Use the demonstration (p. 3, 8) to argue, in a local project, in favour of a simple pre-detection before any costly testing — a principle transferable beyond genetics (do not test everyone, target first).

2️⃣ Objectify a debate on genomic screening. Mobilise the uncertainty figures (63 %, EVPI 243 €; p. 7) to temper promotional discourse on genetic tests, showing that 'economically favourable' does not mean 'certain'.

3️⃣ Question the transferability in your territory. Rely on the explicit limit of the study (score not transferable between ethnic groups; p. 8) to question any genomic device in diverse populations — central equity issue in the French context.

4️⃣ Reposition the intervention on lifestyle. Remind that, in the model, the effect is entirely through a modest weight loss (2.5–4.9 % over one year) resulting from diet and physical activity (p. 4): the lever remains behavioural, the score only guides the targeting.

5️⃣ Alert on the actual availability of interventions. The authors note that lifestyle support is not systematically available in the Finnish system (p. 8): to transpose locally, the question "do we have the offer downstream of the identification?" must precede any discussion on the identification tool.

6️⃣ Cross-reference with the French framework. Do not import Finnish thresholds: confront with French recommendations for screening/management of T2D (see §4) before any operational translation.

4️⃣ ADDITIONAL REFERENCES

1. HAS — Therapeutic strategy for patients living with type 2 diabetes (May 2024). French reference framework for management, prioritising non-pharmacological therapies. — https://www.has-sante.fr/jcms/p_3191108/fr/strategie-therapeutique-du-patient-vivant-avec-un-diabete-de-type-2
2. HAS — Care pathway for adult patients living with type 2 diabetes (June 2025). Includes tool sheets (initial assessment checklist, follow-up checklist, dietary advice, risk foot prevention) — it is the most "field-based" resource of the lot. — https://www.has-sante.fr/jcms/p_3634754/fr/parcours-de-soins-du-patient-adulte-vivant-avec-un-diabete-de-type-2
3. Nature Communications (2025) — Polygenic risk score for type 2 diabetes shows context-dependent effects across populations. Directly addresses the central limit of the analysed article (transferability of the score between populations of diverse ancestries). In English. — https://www.nature.com/articles/s41467-025-63546-4

To be monitored (status to be confirmed): the HAS has initiated a update of the entry criteria for type 2 diabetes screening (page https://www.has-sante.fr/jcms/p_3826415/ ) — check if the report is published before citing it as a stabilised reference.

5️⃣ FREQUENTLY ASKED QUESTIONS (FAQ)

1️⃣ What is a polygenic risk score (PRS) and what does it measure? It is an indicator that summarises hundreds of thousands of genetic variants — increasing or decreasing risk — into a single measure of susceptibility to a disease. A high score indicates a genetic risk above the population average (p. 2).

2️⃣ How does the PRS relate to the FINDRISC questionnaire in this study? In two stages: the FINDRISC (age, BMI, physical activity, family history…) first identifies high-risk individuals; the PRS is only applied afterwards, to this subgroup, to distinguish those at very high risk (p. 2–3).

3️⃣ What is the main economic outcome? Over a lifetime horizon, the strategy with PRS is ‘dominant’: less costly (−£253 per person) and more effective (+0.022 QALY) than usual screening (p. 7, Table 4).

4️⃣ How many people would be affected by the reclassification? The PRS reclassifies 12.4 percentage points of people to very high risk, which is about 38,000 people out of the 313,000 targeted in Finland (p. 7–8).

5️⃣ At what price does the test stop being advantageous? At the "cost neutrality point": €303 per test in the base case. But over a 10-year horizon, this threshold drops to €19 — the chosen horizon changes everything (p. 7, Table 5).

6️⃣ Can the result be considered certain? No. The probability that the strategy is actually dominant is 63% (at a willingness to pay of €0/QALY); the residual uncertainty is estimated at €243 per person (EVPI) (p. 7).

7️⃣ Are these results transferable outside of Finland? Strong caution. The authors note that the score is not transferable between ethnic groups (results possibly limited to the Finnish population), that they assumed full adherence to the interventions, and that the provision of support is not guaranteed (p. 8).

6️⃣ REWRITING IN PLAIN LANGUAGE

What the study says

• The study comes from Finland. It was published in 2022.
• Type 2 diabetes is a common disease. It affects more and more people.
• This disease can be avoided. To do this, it is necessary to identify at-risk individuals early.
• There is a genetic test. It is called "polygenic risk score". It looks at a person's genes. It indicates whether their risk is high.

How it works

• First, a simple questionnaire is used. It asks questions about age, weight, and physical activity.
• This questionnaire identifies at-risk individuals.
• Then, only for these individuals, we conduct the genetic test.
• The test identifies those who are at the highest risk.
• These individuals receive support. They change their diet. They move more. Sometimes, they take medication.

What the study found

• The genetic test helps to better identify individuals at very high risk.
• Approximately 38,000 more individuals would be identified in Finland.
• This approach costs a little less. And it protects health a little better.
• But be careful: this result is not 100% certain. There is a degree of uncertainty.
• The cost of the test matters a lot. If the test is expensive, the advantage disappears.

What to take away with caution

• This study refers to Finland. The results may not apply elsewhere.
• The genetic test does not work the same for all population groups.
• The real lever remains simple: eat better and move more.

7️⃣ CROSS-SECTIONAL ANALYSIS — VALUES OF HEALTH PRACTICES

• Literacy: low. The document is technical and aimed at decision-makers; it does not provide any tools suitable for varying levels of understanding (no accessible version, no user support).
• Empowerment: absent. The individuals concerned are subjects of modelling, not stakeholders; no involvement in the design or evaluation is described.
• Participation: no co-construction mechanism is described.
• Community health: collective dimension presented only in the population/economic sense (resource allocation at the national level), not in the sense of community health on the ground.
• Ethics : partially addressed. The authors explicitly mention the ethical and psychosocial issues of communicating a genetic risk (p. 8), without elaborating.
• Human rights / equity : the issue of equity surfaces through the non-transferability limit of the score between ethnic groups (p. 8), but is not treated as a guiding principle.
• Intersectorality : not addressed.
• Partnership : no formalised collaboration model (the discussion is methodological).
• Combating discrimination : not addressed as such; the risk of genetic discrimination is not discussed, even though it is prominent for this type of device (limitation of the document).

8️⃣ EVALUATION OF THE RELIABILITY OF THE RESOURCE

Scientific relevance — solid, with current reservations. Robust methodology and compliant with HTA standards: individual state transition model, probabilistic sensitivity analysis, cost-effectiveness acceptability curve, EVPI, discounting and horizon explained, representative Finnish data (FINRISK, THL registers) (p. 2–7). Abundant and traceable sources. Three reservations : (1) declared conflict of interest — the lead author (JM) is a founding associate of ESiOR Oy and a director of Siltana Oy (p. 8); to be mentioned for a cost-effectiveness study. (2) Current — data 2017/prices 2021, publication 2022; the economic result depends on the price of the genetic test, in continuous decline. The authors themselves anticipate that the “dominance” would improve with the decrease in prices (p. 8): the meaning of the conclusion is therefore rather reinforced by time, but the precise figures (50 €, 303 €, 253 €) are dated. (3) Strong assumptions — full adherence assumed, intervention effects assumed to be independent of genetic risk, adverse treatment effects not modelled (p. 8).

Operational relevance — low in direct, real in indirect. No direct usability on the ground: neither tool, nor protocol, nor user support. Its value for a prevention actor is reflective and argumentative (staged targeting logic, critical reading of genomic screening, issue of equity/transferrability). For real operational use, refer to the French framework (§4).

9️⃣ STRATEGIC HASHTAGS

#HealthPractices #Prevention #Type2Diabetes #PolygenicRiskScore #TargetedScreening #MedicalEconomicEvaluation #GenomicHealth #HealthEquity