Tests of model-dictated eligibility lawfulness (QCovid score example)

These compliance documents are live pre-released drafts that are specifically not structured for scientific audiences. Specific scientific referencing can be provided in consulting contexts, not limited to practice management, claims management, underwriting or litigation support for professionals or organisations.

Copyright and redistribution rights are governed by our terms of site usage. Communication of factual or typographical errata or other suggestions is welcome. This content was originally contained here (external link), has been split into separate compliance sections on our website and is in need of significant ongoing update.

A journalistic summary of these documents for lay audiences will soon be published.

Identically, models cannot excuse discrimination that is substantiated in law. The simple clinical judgement lawfulness tests above apply to the lawfulness of model driven decision-making. Replacing the word “clinician” with “model” in the tests arrives at identical conclusions: a priori assumption of ineligibility/insufficient eligibility of ME/CFS patient cohorts or individuals is unlawful.

Identically, models cannot excuse discrimination that is substantiated in law. Model-driven scoring replicates, scales and accelerates decision-making. It does not replace the clinician, particularly where reasonable clinical limitations are highlighted

The purpose of model driven scoring in any medical context is to replicate, scale and accelerate decision-making. It does not replace the clinician – it is applied by medical practice.

Where a model is known not to account for either medical, policy or requirements, its application must be selective to be lawful. Where model limitations have been communicated, the tenability of delaying or obfuscating terminal obligation/liability is negated further.

Example: Q Model assessment tool and Group 6 vaccination scheme eligibility

As demonstrated regarding tests of clinical judgement unlawfulness, medical requirements dictated by Joint Committee on Vaccination and Immunisation (JCVI) policy cannot be applied in a way that substantiates incidents of discrimination in law or breaches duties of care.

As a specifically relevant example, the Q Model tool made available to UK specifically predicts probability of death and hospitalisation. It does not account for exacerbation of existing illness and thus does not account for:

  • the Green Book and JCVI requirement of clinical judgement
  • the consensus ME/CFS literature specificities outlined above regarding the impact of viral infection, the unusual severity of ME/CFS or its prevalence
  • the requirement that clinical judgement avoids discrimination in order to be lawful
  • the requirement that clinical judgement accounts for duties of care in order to be lawful

These tests are related to the tests of clinical judgement lawfulness: