What, If Anything, It Changed

In June 2026, I noted that the entire patient voice in Europe's first Joint Clinical Assessment was carried by one carer submitting written comments, and I promised to come back later in the summer with reflections on this.

‍

Patient experience data can fail to count because it was never collected, because it never entered the record, or because it was never weighed. Years before a dossier is submitted, someone decides whether the way patients live with a disease is worth measuring, how to measure that experience, and how to include it in a protocol. Or no one does. That decision is the first weighing, and who is in the room for it matters a great deal.

‍

Stakeholders have different obligations. The developer's is to design patient experience research so that it can be weighed: for a regulator, evidence bearing on benefit and risk; for an assessor, the same evidence made comparative, against the right comparator, in the right population, robustly; for a negotiation, evidence bearing on unmet need and therapeutic alternatives. The system's obligation is to say what weighing the evidence had.

‍

‍

Where patient evidence stops

A 2025 comparative analysis scored 56 health technology assessment systems worldwide and found that almost 93% have some form of patient participation, however, the average HTA system scores 3.61 out of 10 on the depth of it, and participation is weakest at scoping, the stage that decides what question gets asked.

‍

The scope for that first Joint Clinical Assessment asked for patient-relevant evidence: health-related quality of life measured by PedsQL and EORTC instruments, symptoms across six domains including neurological function, vision, balance and hearing, and fatigue, all as relative effect estimates. The pivotal trial did collect quality of life, as an exploratory objective, in a single-arm study, and what reached the assessors was data the report sets aside in a sentence: these data are not discussed, because they do not provide information on the relative effects against the comparators defined in the scope. In the summary report the rows for quality of life, symptoms and fatigue read "relative effectiveness results not available."

‍

Then there is the consultation, where the number that gets quoted for the first Joint Clinical Assessment is the one carer. In a pediatric brain tumor indication, a parent report is not a shortfall; it is often the only accurate way to reach the experience. However, across the first three published Joint Clinical Assessments four external contributors have represented the patient voice so far, the reports describe all four as carers, and two of the three assessments are in adult small cell lung cancer.

‍

When nine patient advocates from EURORDIS member organizations scored that first report they graded it a C on involvement of external experts, and their criticism concentrated on integration and transparency: the carer contribution is documented, but its effect on the scope and conclusions is not clear, and the report never explains how the external experts were identified and selected.

‍

Years before a dossier is submitted, someone decides whether the way patients live with a disease is worth measuring. Or no one does.

‍

The Joint Clinical Assessment operates under a statutory constraint, since Article 9 confines the joint report to relative effects and the certainty around them and bars any conclusion on overall added value. Appraisal is a national accountability. Germany shows what can happen nationally, and it is a useful example as the largest market in the Union, often among the first to rule, and one of the more developed systems in its treatment of patient-relevant endpoints. Patient representatives have sat in the Federal Joint Committee since 2004 with the right to take part in deliberations, the right to be present when the vote is taken, and no vote. In the same system, quality of life is named in the benefit assessment ordinance as a statutory category of patient-relevant benefit, and when it arrives as a measured endpoint it enters the assessment itself.

‍

Then it is graded, and the developer's obligation returns. German assessments routinely discount quality-of-life data for high risk of bias, for declining questionnaire return rates and incomplete capture, and for response criteria that were not validated or prespecified. IQWiG named the underlying problem in March 2026: patient-reported endpoints, repeatedly, do not hold the same standing in study planning and conduct as clinical parameters.

‍

‍

From being heard to being weighed

The right of the patient voice to be heard has been built almost everywhere, with some exceptions like the Medicare negotiation program, where patient input is not among the factors Congress instructed the Secretary to consider at all.

‍

The duty to acknowledge the patient voice in the record also exists. In the United States, the Food and Drug Administration has published a patient experience data table in 252 of 277 drug approvals since 2019, and in 85 of those 277, 30%, the data reached the benefit-risk assessment. The authors of that analysis, all from industry, concluded that more transparency is needed for anyone outside the agency to understand how patient experience data is used.

‍

What remains underdeveloped is attribution: a statement of what weight the input carried and what, if anything, it changed. This is not an exotic demand. European institutions operate under a general duty to state reasons, and United States agencies are expected to respond to significant comments and to show they considered the important aspects of a problem. It is a normal public-law expectation this domain has not yet absorbed.

‍

‍

What the next protocol needs to know

In the case of ritlecitinib in severe alopecia areata the sponsor performed a discrete choice experiment in adults and adolescents across the United States and Europe, trading scalp, eyebrow and eyelash regrowth against three-year risks of serious infection, cancer and blood clots. Scalp regrowth dominated, carrying 42.1% of the weight adults placed on any attribute, and those preferences were then combined with the trial results in a quantitative benefit-risk analysis.

‍

The European Medicines Agency's assessment report found the studies well performed, the patients representative, and noted that the preference analysis showed no patient preference for a lower dose, in the paragraph immediately before concluding that a maintenance dose of 50 mg is considered appropriate. The preference evidence visibly shaped the dose, though not the indication, which the committee credits to the efficacy data.

‍

In FDA's review of the same molecule, the patient experience data table carries a line for patient preference studies and the box is not ticked, and the meeting record describes the study as informative, not determinative for regulatory decisions. Same evidence, two regulators, different weights. All of this is signal. A developer weighing whether to fund the next patient preference study is investing in evidence of uncertain impact in decision-making, and silence is informative in its own way, since it tells the next protocol nothing at all.

‍

Which is why I keep pulling this conversation upstream. I will not pretend that is simple: bringing a patient-relevant endpoint into the hierarchy means alpha you allocate, powering you buy, and the better part of two years of qualitative work before an instrument is fit for purpose, and a measure that is fielded and misses is worse than none, because it sits in the study report and the assessor will read it. When I work with teams on this, that is the conversation, and it belongs at protocol design.

‍

And patient written accounts and testimony serve important functions. In ultra-rare disease, patient and carer submissions are frequently the only source of natural history, treatment burden, and caregiver impact in the entire dossier. Some of my thinking here returns to the work I did co-chairing the Caregiver Pathways Task Force, a National Alliance for Caregiving and LEAD Coalition initiative, where Mousumi Bose, a scientist and the mother of a son with a rare disease, described becoming an expert on the disease and, more important, the expert on her son. No instrument captures that, and where a determination has no route to weigh it, the shortfall is in the determination.

‍

Which brings me to August 14, 2026. FDA published its PDUFA VIII reauthorization materials that morning, and among the proposed commitments, under the heading of transparency and consistency for patient experience data, is an undertaking to publish case studies demonstrating how patient experience data was considered in specific regulatory decisions. That is attribution, in draft, and it is the same thing that industry analysis asked for. The public meeting is September 16, 2026, and comments on this proposal close October 16, 2026.

‍

So, as summer begins to yield to fall, the question I would put to any team building toward an assessment, a national decision, or a negotiated price is not whether patients were consulted. It is which patient-relevant endpoint was cut at protocol lock, who decided, and whether anyone in the building intends to file a comment by October 16.

‍