DSCAdapt

IDAlert Project · Horizon Europe
Grantham Research Institute · LSE

How does climate adaptation affect infectious disease risk?

Describe an adaptation measure, a location, and a sector. DSCAdapt analyses the co-benefits and trade-offs using IDAlert's curated evidence base.

Policymaker Researcher
Evidence base
Response length: standard (200–400 words)

Or start with a common scenario

Enter to send · Shift + Enter for a new line

Currently in the evidence base

Water managementFlood protectionGreen & blue spacesBuilding standardsRetrofittingCoastal & marineSustainable farmingFood safetyTourismTransportBarcelona case study
Review for country:
— or paste the text —

Pathway explorer

Trace how climate adaptation measures connect to infectious disease risk. Start with a policy sector — each selection reveals the next column in the chain.

Download full diagram (PDF)
§ 01 · Overview

Pathways, not predictions.

DSCAdapt is a retrieval-augmented language tool for European policymakers working on climate adaptation. It reasons in terms of pathways — the ecological and infrastructural mechanisms that link a given measure to disease transmission, exposure, or vulnerability — rather than generating forecasts or recommendations.

Every claim is tied back to the IDAlert evidence base or current web sources through explicit references. The tool uses conditional language because these are plausible scenarios shaped by local context, not deterministic outcomes. Where evidence is thin or conflicting, the tool says so.

Who it is for

Municipal and national climate adaptation officers, public health policy leads, and researchers across the IDAlert consortium.

What it isn't

Not medical advice. Not epidemiological surveillance. Not a substitute for local expertise. An evidence-synthesis tool meant to sit alongside those sources.

§ 02 · Architecture

How a question becomes an answer.

Five stages: a routing model classifies the question, retrieval pulls relevant evidence, an optional web-search agent gathers current context, a generation model synthesises the response, and the output is rendered or exported.

01 · Route

Classify

Rewrites follow-ups, enforces scope, picks retrieval paths. Out-of-scope queries redirected before generation.

ministral-8b
02 · Retrieve

Pull evidence

Hybrid cosine (85%) + IDF keyword (15%), deduplicates per section, returns top-k chunks with references.

mistral-embed
03 · Context

Current conditions

Web-search agent pulls supplementary current legislation, policy, or geographic context. Kept separate from curated evidence.

mistral-medium
04 · Generate

Synthesise

Generation model produces the response against a system prompt enforcing scope, citations, and the co-benefit/trade-off split.

mistral-small
05 · Deliver

Read or export

Responses in chat or exported as formatted PDF briefings with references preserved.

PDF
§ 03 · Limitations

What it will and won't do.

Will

Surface co-benefits and trade-offs

Reasons through both positive and negative implications for infectious disease risk, with references.

Won't

Predict outbreaks

A retrieval-augmented language system, not a forecasting model.

Will

Work in multiple EU languages

Responses are returned in the language of the query.

Won't

Give medical advice

Operates at the population and policy level, not the clinical level.

Will

Acknowledge its own gaps

Where the evidence base doesn't cover a topic, the tool says so explicitly.

Won't

Make recommendations

Lays out evidence and trade-offs. The judgement belongs to policymakers.

Will

Cite its sources

Every response includes a reference list tracing claims back to pathway documents or web sources.

Won't

Replace local expertise

Epidemiological judgement and ground knowledge from public health authorities remain essential.

§ 05 · Provenance

Built at LSE under IDAlert.

DSCAdapt is developed at the Grantham Research Institute on Climate Change and the Environment, London School of Economics, as part of the IDAlert project (Infectious Disease decision-support tools and Alert systems), funded by the European Union's Horizon Europe research and innovation programme under grant agreement No. 101057554.

The tool is part of IDAlert's Work Package 2, which focuses on indicators for monitoring climate-induced infectious disease risk and emergence. The wider consortium includes 18 partner institutions across seven EU member states, plus the United Kingdom and Bangladesh.

For access, collaboration enquiries, or bug reports: s.aggarwal28@lse.ac.uk · The tool is in research preview and access is currently limited to IDAlert consortium members and invited European policymakers.