Methodological Constraints on the Evidence
1 The Qualitative Participation Filter
The evidence reviewed in this chapter — qualitative studies, surveys, and co-production research — systematically underrepresents the most severely affected patients. Severely and very severely ill patients (housebound, bedbound, unable to tolerate sensory stimulation) are the patient subgroup with the most extreme healthcare-system needs and the least capacity to participate in research. Their exclusion means the strongest claims about healthcare-system failure (diagnostic delay, invalidation, access barriers) are documented primarily from the experiences of mild-to-moderate patients — the patients most able to navigate the system despite its failures. The gap between the reported experience and the full-severity experience is unmeasured and systematically unmeasurable by standard research methods.
Consequence: The paper’s claim that healthcare systems fail ME/CFS patients is strengthened, not weakened, by the participation filter: the patients who can tell us about the system’s failures are the ones the system has not yet completely failed. The most severe patients, who have been fully failed, cannot participate in the research that documents the failure. (Origin: brainstorm)
2 The Guideline Evidence Paradox
The Smith 2014 AHRQ review found zero adequately validated diagnostic methods and low-to-moderate treatment evidence — but that review used studies conducted under pre-IOM, pre-NICE clinical frameworks. The NICE 2021 guideline is a response to that evidence landscape, and the guideline itself changes the standard of care against which future evidence will be judged. This creates a paradox: if NICE 2021 is correct that GET harms ME/CFS patients and PEM must be the central clinical feature, then all pre-2021 treatment evidence (conducted under different diagnostic criteria and with GET as comparator) is methodologically compromised — not because the studies were poorly executed, but because they studied the wrong outcomes in potentially the wrong patient populations. This is not a call to disregard historical evidence. It is an acknowledgment that a paradigm shift in clinical guidance invalidates some of the evidence base that informed the previous paradigm.
Consequence: Future systematic reviews of ME/CFS healthcare must use diagnostic criteria and outcome measures consistent with the post-NICE-2021 clinical framework. The Smith 2014 AHRQ finding of “zero adequately validated diagnostic methods” remains a valid null result for the pre-2015 era; the open question is whether a 2026 AHRQ update would produce the same finding. (Origin: brainstorm)
3 The Null Hypotheses: What If Reform Doesn’t Change Outcomes?
If all identified barriers — medical education deficits, diagnostic delay, guideline inconsistencies, specialist access gaps, disability determination mismatches, and healthcare invalidation — were removed, would ME/CFS patient outcomes improve? The universal access null hypothesis says: reform removes barriers but doesn’t change the disease trajectory, because the disease trajectory is driven by pathophysiology, not healthcare access.
This null is partly refuted by available evidence: reducing diagnostic delay almost certainly improves outcomes by preventing inappropriate exercise recommendations that worsen PEM, and guideline-consistent care (avoiding GET) almost certainly prevents iatrogenic harm. But the null is not fully refuted: the specialist-clinic evidence gap means we do not know whether any specific care model changes long-term functional outcomes; the medical education gap means we do not know whether better-educated physicians produce better patient outcomes. The null hypothesis distinguishes between removing harm (preventing iatrogenic injury) and providing benefit (improving function beyond pre-reform baseline). The former is almost certainly achievable; the latter is untested.
Consequence: Reform advocacy should distinguish between “stop the harm” (remove GET recommendations, end healthcare invalidation, fix disability assessments) — which has strong evidence — and “improve outcomes” (specialist care, co-produced services, multidisciplinary teams) — which has no evidence. The moral case for stopping harm does not require evidence of benefit; it requires evidence of harm, which this chapter provides. (Origin: brainstorm)
4 Cross-Sectional Design Constraint
The healthcare-systems evidence base is overwhelmingly cross-sectional. Qualitative studies capture patient experiences at one point in time; surveys capture a single snapshot of diagnostic rates or provider attitudes; guideline comparisons are synchronic (published in year X, compared to guideline published in year Y). But the healthcare-system response to ME/CFS is evolving in real time — NICE 2021 was published only four years ago; the post-COVID diagnostic surge is still unfolding; and the slow generational turnover of medical school curricula means the Muirhead 2021 finding (41% of UK schools don’t teach ME/CFS) is still current. The cross-sectional evidence base is a series of still frames of a moving target. What looks like institutional stasis may be institutional latency — guidelines take years to disseminate, curricula take decades to update, and physician behavior takes generations to change.
Consequence: The evidence documents failure convincingly. It does not distinguish between failure-by-design (the system is designed to exclude ME/CFS) and failure-by-latency (the system is changing too slowly to be measured as change by cross-sectional methods). Both are failure. The reform prescription may differ: design failure requires structural change; latency failure may be accelerated — but not replaced — by structural change. (Origin: brainstorm)