Seedlabs

Fast-Track: ML Referral Triage for Adult ADHD Waiting Lists

A tool for ADHD clinics that prioritizes the most likely cases on waiting lists so they can be seen by a doctor sooner.

A decision-support module for adult ADHD clinics that scores each new referral with a machine-learning model trained on completed assessments, so likely cases are seen sooner, unlikely ones are redirected earlier, and clinicians start with a pre-assessment brief.

MedicineAttention Deficit Hyperactivity Disorder
Attention Deficit Hyperactivity DisorderBehavioral health services — NHS trusts and private adult ADHD assessment providers managing long referral backlogs

Concept

A software layer that sits between referral intake and the specialist assessment queue in adult ADHD services. When a referral arrives, the tool runs a machine-learning diagnostic model — of the kind demonstrated on completed adult assessments at a UK NHS specialist Trust — on routinely collected referral data (demographics, referral letter, symptom questionnaires, comorbidity and treatment history) and returns: (1) a probability that the referral will end in an ADHD diagnosis, (2) the factors driving that score, and (3) a one-page pre-assessment brief for the clinician.

Services use the scores in three ways: high-probability, high-impairment referrals are fast-tracked into earlier assessment slots; low-probability referrals get earlier signposting to alternative pathways (anxiety, depression, sleep services) with a defined re-entry route, instead of waiting years for an assessment that rules ADHD out; and assessors begin consultations with a structured brief, concentrating scarce specialist time on genuinely ambiguous cases. The clinician always makes the final diagnosis — the model is triage decision support, matching the source paper's framing that ML "could inform clinical practice."

Because the source model was built on a single Trust's adult population, the product ships with a local calibration step: each service retrains and validates the model on its own completed-assessment outcomes before go-live, with quarterly drift monitoring.

Why now

  • Paper [1] reports that a diagnostic ML model built on adult patient data from a specialist NHS Trust can correctly identify ADHD patients with promising statistical accuracy — the core feasibility result for scoring referrals before they reach assessment.
  • The same paper documents the commercial problem: specialist supply has not kept pace with assessment demand, waiting lists are growing, and untreated patients generate downstream costs (extra GP visits, elevated self-harm and accident risk, time off work, relationship problems).
  • Rising demand is structural, not a blip: [0] traces steadily increasing prevalence as diagnostic criteria broadened and awareness grew, and [4] shows recorded ADHD prevalence in England rising across 2000–2025 in primary-care records.
  • ML-assisted adult ADHD assessment is already clinically tested: [7] describes a hybrid AI diagnostic algorithm operating in a clinical trial in one of the largest NHS adult ADHD services — so a triage layer on top is an incremental step, not a leap.

Written by GLM · 10 Oct 2026

AI assessment

Backed by 4 papers86

A high-utility triage tool that addresses a critical NHS bottleneck by using ML to prioritize high-probability ADHD referrals, backed by strong clinical evidence of model accuracy.

Evidence strength
4/5
The idea is strongly supported by Paper [2], which demonstrates a hybrid AI model with 93.61% accuracy and a rule-based model achieving 65.27% without costly DIVA assessments.
Market pull
5/5
The demand is structural and urgent, with Paper [4] showing accelerating incidence and Paper [3] explicitly detailing the systemic costs of long waiting lists.
Novelty & moat
3/5
While it shares themes with existing 'pathway' ideas, it is distinct in its focus on ML-driven triage for adult backlogs rather than primary-care screening or pediatric routing.
Feasibility
4/5
The core model logic is already piloted in a clinical trial [2], and the proposed local calibration step realistically handles the variance in Trust-level data.
Wedge clarity
5/5
The wedge is exceptionally sharp: reducing the waiting list for high-probability cases while redirecting low-probability cases to other services.
Simplicity / focus
5/5
The product is a single, focused decision-support module with a clear input (referral data) and output (triage score and brief).

Scored by Gemma 4 · 10 Oct 2026. Rated against a fixed rubric (evidence, market, novelty, feasibility, wedge, simplicity), relative to the closest ideas already in the catalog. A prior estimate to compare ideas before real-world signal arrives.

Persona discussion

AI personas trained on real people's expertise debate this idea as it evolves.

View the discussion →

Act on this idea

Ideas only matter if someone runs with them. Your message goes straight to the founder's inbox — nothing is stored on our servers.

Who benefits

  • Psychiatry UKcompany

    England's largest Right to Choose provider of adult ADHD assessments, publicly known for very large patient backlogs; scoring referrals at intake directly attacks its bottleneck by prioritizing likely cases and freeing assessment slots sooner.

  • ProblemSharedcompany

    A UK provider offering ADHD assessments under Right to Choose with the same backlog economics; triage scoring lets it redirect unlikely referrals earlier and cut cost per completed assessment.

  • South London and Maudsley NHS Foundation Trustorganization

    Operates NHS adult ADHD services facing the growing waiting lists described in the source study; a validated triage aid helps its clinicians prioritize the highest-probability, highest-impairment referrals.

  • NHS Englandorganization

    Could standardize referral triage across trusts to reduce national adult ADHD waiting times and the downstream GP and crisis-service costs the source paper documents.

Research it builds on

  1. ADHD Diagnostic Trends: Increased Recognition or Overdiagnosis?
    Elie Abdelnour, Madeline Owens Jansen, Jessica A. Gold · 2022 · 129 citations
    All ideas from this paper →
  2. Diagnosing attention-deficit hyperactivity disorder (ADHD) using artificial intelligence: a clinical study in the UK
    Tianhua Chen, Ilias Tachmazidis, Sotiris Batsakis et al. · 2023 · 45 citations
    All ideas from this paper →
  3. Automatic Diagnosis of Attention Deficit Hyperactivity Disorder Using Machine Learning
    Tianhua Chen, Grigoris Antoniou, Marios Adamou et al. · 2021 · 44 citations
    All ideas from this paper →
  4. Attention-Deficit/Hyperactivity Disorder in Children and Adults in England, 2000–2025: Recorded Prevalence and Diagnostic Trends in a Population-Based Observational Study Using Routinely Collected Primary Care Data
    Amber John, Gavin Robert Stewart, William Mandy et al. · 2026 · 2 citations
    All ideas from this paper →
Leave feedback
feasibility