The Template, Not the Exception: Ojemda and the Self‐Confirming Logic of the Joint Clinical Assessment

1. Introduction

The Ojemda report is not a surprise [2]. It delivered what we had anticipated across earlier publications: it implemented the methodological guidance and guidelines with great rigour [3,4,5,6,7,8], it ran into the precise tensions we had foreseen [9,10], and, exactly as predicted, it failed to deliver a degree of certainty of relative effectiveness, confirming a lack of being fit for purpose [10]. The interest of the case therefore lies not in any unexpected finding, but in the fact that the outcome was foreseeable from the framework’s design. This paper sets out why each result was predictable, and what that predictability implies. Our central contention is that the Joint Clinical Assessment (JCA) has become a process-focused instrument for procedurally informing pricing decisions, rather than a result-focused instrument for genuinely informing public-health decisions — and that Ojemda is simply the first clear demonstration of it.

2. The JCA as an Audit, and the Problem of the Missing Gold Standard

At its core, the JCA is an audit: it confronts the research delivered by the health technology developer (HTD), whose objective is to estimate the degree of certainty of relative effect, against a gold-standard set of methodological referentials. An audit of this kind presupposes that the gold standard exists and is implementable. This creates the following three difficulties.

First, for the JCA the referential is a corpus of methodological documents, guidance and guidelines, developed by EUnetHTA and endorsed by the HTA Coordination Group [3,4,5,6,7,8]. That corpus has been heavily criticised by parts of the scientific community, on both its scientific validity and its development process, to the point of being characterised as “eminence-based” rather than evidence-based [11,12,13,14,15,16,17,18,19,20]. A contested corpus cannot serve as a gold standard.

Second, the documents embed an epistemological commitment that is never made explicit and that, in our reading, reflects a single institutional culture advanced through soft power and disproportionate investment, which challenges their legitimacy as a neutral benchmark [21,22].

Third, not all audits are alike. Accounting or IT program audits may demand full alignment with a fixed standard and the closure of every gap; security audits for offices or public events must instead balance maximum protection against practical feasibility. The only way to eliminate hacking risk entirely is to keep a computer off any network, rarely workable in practice. An empirical clinical study is the second kind of case: it must trade ideal experimental conditions against feasibility while preserving the key outcomes so they can be inferred with substantial confidence [23,24,25]. Auditing clinical evidence as though it were an accounting ledger guarantees that gaps will be found [11].

3. The Interpretive Gap: An Audit That Omits Meaning

Even granting a broad and objective audit, the JCA misses the essential task: interpreting what the outcomes actually imply. That interpretation cannot be reached through purely factual, context-free analysis; it requires contextualisation and informed judgment [26,27,28,29]. In the manner of IQWiG, the JCA nevertheless treats factual, judgment- and context-free assessment as feasible, and assumes that appraisal can be hived off to an independent body downstream [21,30]. This contradicts the epistemological foundations of the exercise. Separating assessment from appraisal is, in itself, not unusual; what is singular is the belief that assessment can be conducted free of judgment and context. That model is specific to IQWiG, it matches a particular national culture, anthropology and philosophy of evaluation, and has not been reproduced elsewhere. Because any comparison between an ideal standard and real-world conditions will inevitably reveal gaps, the consequence is structural: no product will reach the ideal of a very high degree of certainty of relative effect, and the vast majority will end at a very low degree. We have published on this prediction before; Ojemda confirms it [9,10].

4. The Single-Arm Predicament

We anticipated specifically that the JCA would not be fit for purpose for products supported by single-arm trials, and that such products would be assessed very negatively whenever the audit framework was applied [9]. Ojemda bears this out [2,31,32,33]. From it follows a proposal we have already advanced: products supported only by single-arm evidence should arguably not be eligible for the JCA at all [9,15,16]. Subjecting them to it is a large, inefficient and ultimately irrelevant effort, because the conclusion is effectively known before the process begins. When the result of an assessment is foreseeable at the outset, the assessment is no longer measuring the product; it is reproducing its own design. At the same time, we do not advocate single-arm designs, while acknowledging that they are sometimes inevitable.

5. The Retreat from Evidence-Based Medicine

The JCA claims to adhere strictly to evidence-based medicine (EBM) [34,35,36]; this first report shows how dramatically it departs from it — again, unsurprisingly. The aims diverge at the root: the JCA performs a single audit to inform payers, whereas EBM seeks to identify the best management practice for the patient. EBM rests on a tripod of evidence, patients and clinicians, weighted as equally important [34]; the JCA leans disproportionately on evidence and leaves little or no room for the contributions of clinicians and patients [37,38,39,40,41,42]. Even within the evidence leg, the JCA is highly selective where EBM is inclusive. EBM, in principle, considers all available evidence irrespective of its quality and then grades the strength of the resulting recommendation accordingly, from weak to strong [20]; the JCA, by contrast, filters the evidence base selectively at the outset. The non-inclusiveness we have described at length is confirmed here in two concrete ways. First, the restriction of the report-review stage to a factual check forecloses any challenge to the assessment team’s reasoning, in effect, that reasoning cannot be found wrong. Second, only factual input from carers and clinicians is taken up, not their experience [37]. The JCA is, characteristically, restrictive to a single perspective.

6. Negative Rhetoric and the Certainty That Never Comes

As anticipated, the JCA generates its own difficulty by multiplying PICOs, subgroups and outcomes [5,43,44], which leaves the assessment largely fed by post-hoc analyses, analyses the report then criticises systematically as post hoc, subject to multiplicity, and reliant on nominal p-values [5,44]. In effect, the JCA creates these limitations by design and then criticises them as poor evidence generation, an untenable contradiction. Each criticism is technically true. But the report displays a consistent disposition to minimise positive information and to amplify every limitation; the asymmetry in tone is difficult to miss. Finally, and most consequentially, the inability to conclude on the degree of certainty of relative effectiveness was entirely foreseeable [10]. With no framework to move from documented uncertainty to a stated certainty, no clear, transparent and reproducible certainty conclusion is possible [10,20,28,29,45]. The report therefore fails at its highest-priority objective and cannot, on this analysis, support Member State HTA appraisal as intended.

7. Conclusion

Nothing in the Ojemda report should surprise anyone who has followed how the framework was built. It conducted its audit rigorously against a contested referential [11], omitted the interpretive step that gives an assessment meaning [26], leaned on evidence to the near-exclusion of patients and clinicians [34,37], and, predictably, could not state a degree of certainty [10,45]. The conformity to expectation is itself the finding. When the outcome of a process can be known before it begins, the process is not assessing the product so much as reflecting its own design assumptions back at it. For single-arm orphan evidence in particular, the JCA is, on this showing, an elaborate confirmation of a foregone conclusion. Two implications follow with some force: that products resting on single-arm evidence should be considered outside the JCA’s remit [9], and that the framework cannot claim fitness for purpose until it builds the missing, transparent bridge from uncertainty to certainty [10]. Until then, Ojemda will not be an exception. It will be the template.