FDA COA Evidence Dossier in 2026: A Planning Crosswalk and Draft-Status Review

TL;DR

FDA's September 2025 COA evidence dossier document remains a discussion draft in the sources reviewed here, prominently marked “DRAFT- Not for Implementation. For Discussion Only.” Its six-section outline can inform an internal evidence inventory and questions for an agency meeting. It should not be presented as an effective submission requirement or used to override a program's agreed evidence strategy. [1]

FDA's 30 September 2026 priorities paper says it intends to finalize the template, with a recommended format and flexibility for sponsors. Replacement of the 2009 PRO guidance appendix is a future step contingent on finalization. The paper publishes neither a revised outline nor a disposition of individual docket requests. [2]

The substantive finding is about evidence location. Content validity is present in section IV, despite the absence of that familiar heading. Appendix E houses psychometric plans and reports; Component G in the body addresses measurement error, while Appendix G addresses usability. Comments about restoring headings concern findability and organization, not proof that the evidence disappeared. [1]

The useful preparation is a traceable crosswalk: evidence item, proposed draft location, existing document, responsible team, version and an unresolved question. Preserve qualitative reports, linguistic-validation records and digital-technology evidence even when the printed appendix list gives them no dedicated title. The draft expressly allows additional material needed to support the evidence. [1]

Document status comes before dossier design

The September 2025 file is titled Public Meeting Discussion Documents: Draft COA Evidence Outline / Draft COA Evidence Dossier Template. FDA shared it to discuss how to organize evidence supporting clinical outcome assessments. Its status banner governs how the wording should be read: references below to what the draft “asks” describe proposals inside a discussion document. They are neither new legal requirements nor a recommendation to implement the draft as binding policy. [1]

In its 30 September 2026 priorities paper, FDA says it will work toward a final template using workshop-docket comments. FDA describes the intended result as a recommended organizational format that supports fit-for-purpose review while preserving sponsor flexibility. It says the finalized template will replace the appendix titled Information on a PRO Reviewed by the FDA in the 2009 PRO guidance. The conditional sequence matters: a work plan announces intended work; the final document would establish the eventual text. [2]

We checked the live workshop page and PFDD series page on 4 October 2026, together with the two PDFs. The workshop page still links the discussion document and describes a draft. The series page labels Guidance 3 final and Guidance 4 draft. No revised dossier was located in those checked materials. This is a bounded document check, not a guarantee that every FDA page has been searched. Before a submission, recheck the current documents and confirm program-specific expectations with the review division. [3] [4]

The priorities paper sets out future work rather than individual comment dispositions. Acceptance or rejection of a proposed heading would require an affirmative agency statement. The draft also expressly permits relevant evidence beyond its named sections. These distinctions keep the crosswalk flexible and faithful to the documents. [1] [2]

PFDD Guidance 3 addresses selection, development and modification of fit-for-purpose measures. The priorities paper separately discusses finalizing Guidance 4 on COA endpoints and ICH E22 on patient preference studies. The dossier template is another work item. Instrument fitness, endpoint interpretation and presentation of evidence are connected decisions with different source documents. Our fit-for-purpose COA overview addresses selection; this report examines evidence organization. [2] [3]

The workshop took place on 18–19 September 2025. Its Federal Register notice set an 18 November 2025 comment deadline, at 11:59 p.m. Eastern Time, under docket FDA-2025-N-2368. The retrieved comment index contained 24 entries. The comparison below uses the 17 substantive request letters within that captured set; a worked example and six other or partially retrieved entries are excluded. Counts describe this reviewed collection, not industry prevalence, agreement or a complete measure of every theme in the docket. [5] [6]

Where did the September 2025 file put the 2009 appendix?

FDA's mapping table places 2009 content validity in section IV and other measurement properties in sections IV and V and Appendices E, F, G, and I. The six headings provide a useful reference for a provisional internal crosswalk.

The outline printed before the template instructions is:

• I. Executive Summary
• II. Context of Use Information
• III. Meaningful Aspect(s) of Health (MAH)
• IV. Description and Justification for COA
• V. Description and Interpretation of COA-Based Endpoint(s)
• VI. Appendices A through J [1]

The proposed template would support interpretation of COA scores and corresponding endpoints. It allows one dossier covering relevant COAs or a separate dossier for each, with the extent of evidence varying by measure and development stage. It proposes retaining all main sections and explaining inapplicability with an NA entry, while allowing unused method-specific subsections to be marked N/A. The introduction also invites relevant material for novel approaches or other evidence gaps. These are the discussion draft's proposed organizational instructions. [1]

Read those instructions with the document's status and its express allowance for novel or otherwise missing evidence. A provisional inventory can preserve the six headings and mark inapplicable topics with reasons, while adding clear labels for relevant sponsor material. The draft's hypothetical rationale table also allows additional components. There is no basis here to prohibit a useful glossary or an additional supporting report simply because it lacks a printed appendix letter. [1]

For a CDER or CBER marketing application, the discussion draft proposes that "the evidence for the COA measure should be submitted in Module 5.3.5.3" of the eCTD, with a reviewer's guide that explains the module and links to protocols, study reports, and data folders with the define.xml files. Full articles go in Module 5.4. A footnote says device submissions should follow least-burdensome principles. Hyperlinks should be active, and the PDF should be searchable, with a table of contents and lists of tables and figures. [1]

Sections I and II: the trial and the claim

Section I.A, "Summary of COA-Based Endpoint Approach," covers the context of use, the meaningful aspect of health, and the concept of interest when it differs from the health aspect. Table 1, "Summary of COAs by Context of Use," has columns for the meaningful aspect of health, the concept of interest, COA type, name, score, COA-based endpoint, and endpoint positioning. One table per context of use. The table "should include all relevant COAs in the program." [1]

Section I.B asks for a rationale that "explains how and why the COA is expected or intended to work," then a table of components, a high-level summary of the evidence, statistical results where they apply, and a link to section IV. The printed Table 2 is "a hypothetical example" for illustration. Sponsors "should include components that are relevant to their development program using language that is clearest for their situation." The example rows run from A through H, plus "Other components needed to justify interpreting scores as measures of the concept of interest." [1] Rows E through H organize different parts of the measurement argument. Their topics and Appendix E's supporting documents need to be read together.

Section I.C asks the executive summary for "a summary of results of efficacy analyses of COA-based endpoint(s) and link to clinical study report(s) (CSR)," and for a conclusion on clinical meaningfulness, including the meaningful-score-difference or meaningful-score-region range and a qualitative summary. [1] The CSR link and the efficacy summary are the same numbered item. Letters below ask FDA to keep the link and drop the summary. That is an edit to I.C. The template already names the CSR.

Section II is "specific to the registrational and/or pivotal trial(s) intended for regulatory decision-making": population, with a protocol link; design, including the COA schedule and links to the schema and the full schedule of assessments; the COA objective or hypothesis; endpoint definition and positioning; and the desired claim, with a link to proposed United States prescribing information when one applies. Sponsors should "explicitly note whether multiple endpoints are being used to support inference about a single meaningful aspect of health and specify whether each endpoint is controlled for multiplicity." [1] The 2009 rows for targeted claims and the endpoint model both map to section II. The 2009 row for trial and analysis plans maps to sections II and V and to "links and directions to specific portions of submission(s)." [1]

Sections III and IV: the health aspect, then the measure

The 2009 conceptual framework maps to section III, and also to section IV "if concept of interest (COI) is different from MAH." [1] Section IV.A asks for the relationship between the two. When they differ, the draft asks for a diagram "showing relationships among patient experiences, items and tasks, domains, and total score," aligned to the protocol endpoints and the proposed claims, plus evidence from patients, caregivers, clinicians, or others. [1]

Section IV.B is the instrument. It asks for the final version and the type, "e.g., PRO, ObsRO." The exact copy "as it was and/or will be administered in the trial (e.g., screenshots of eCOA, photocopy of paper)" goes in Appendix A and is linked here. Copies "should be of high resolution and legible." The same subsection asks for additional administration instructions, the recall period, response options and any numeric values, "Mode(s) of administration," and a brief scoring algorithm, including missing-data rules and the score range. [1]

Section IV.C is the rationale. Evidence "should be specific to the planned clinical trial population and indication (i.e., the COU)." A related context of use is a discussion with the review division before submission. A result that serves more than one component "may be included once and referenced." [1]

Component A is the COA type, filled in as PRO, ObsRO, ClinRO, or PerfO. Expansion "may not always be needed but would be expected when there are multiple COA types being used to assess a given MAH." The DHT sentence is the next sentence in that component. [1] Component B is the content-validity material. FDA asks for evidence that the COA "captures all of the most clinically important concepts and items, and that items are complete and relevant," from literature, expert input, and qualitative protocols, guides, analysis plans, summaries, and transcripts, plus item origin in the Appendix C matrix and "Qualitative study summaries that support validity of item content." Transcripts go in Appendix D. [1] Component C is administration as the developer intended: manuals in Appendix B, standardized training, in-trial observation, and rater certification when raters are used. Component D is whether respondents understand the items and can select a matching answer. [1]

Components E, F, and G are rationale rows. Appendix E is the psychometric plan and report. Appendix G is "Results of Usability Testing."

Component E is scoring: the measurement model, with classical test theory, item response theory, and Rasch as the printed examples; the missing-data rule; and E4. The E4 heading says scores "are sufficiently sensitive to reflect changes." The evidence printed under it is "Results of quantitative evaluation of floor and ceiling effects at the COA score level." [1] Component F is other influences. F1 is linguistic and demographic interpretation, including translation, patient testing, adapted copies, and the equivalence rationale. F3 is fatigue or burden, and the trial-battery steps "should be included in trial protocol(s)." F4 is implementation and, for electronic capture, the device, windows, and "interaction between the eCOA and the user." [1] Component G is measurement error: test-retest reliability in clinically stable patients, and inter-rater and intra-rater reliability, with a standalone protocol in Appendix F. Component H is prespecified convergent, divergent, and known-groups evidence. [1]

The mapping table sends content validity to section IV only, and other measurement properties to sections IV and V and Appendices E, F, G, and I. Modifications are "Not a standalone section. Evidence should be provided in IV.C., as appropriate/as needed," with prior versions in Appendix A. [1] This exposes a practical navigation risk: body Component G and Appendix G have different subjects. A reliability report should be linked through the measurement-error rationale and the psychometric plan/report, while usability evidence has its own appendix.

Section V, including the arm-level display

Section V.A is computation of the endpoint. V.B is why that endpoint was selected, including clinical benefit, why the COA is fit-for-purpose in the planned trial, and a pointer that PFDD Guidance 3 "discusses developing an evidence-based rationale for interpreting the COA score as a measure of the concept of interest" and that this rationale "is part of the overall rationale to support the use of the endpoint." V.B also asks about precision given the design, sample size, and analysis plan. "Well-established relevant outcomes such as organ failure and death" do not need additional importance evidence. [1]

V.C evaluates meaningfulness. It separates derivation of a meaningful score difference or region from application of that range. Derivation may be anchor-based, qualitative, bookmarking, literature, or another approach. One timing example for interviews is "after completing the main portion of the study before unblinding." [1] Item 3 is the trial application:

Details and discussion of how derived final range of score interpretation metrics (i.e., MSDs, MSRs) were applied to evaluate the meaningfulness of the COA endpoint scores in clinical trial(s). [1]

The displays include "eCDF curves of COA endpoint scores by treatment arm annotated to aid interpretation (e.g., sample size, median, reference line(s) for proposed range of thresholds)." For an MSD and the horizontal gap, the draft also asks for "a figure illustrating the observed treatment effect with corresponding 95% confidence interval and MSD values." The vertical gap asks for the percentage of patients in each treatment arm who met the threshold. Item 4 is an "overall conclusion regarding meaningfulness of COA-based endpoint scores in the context of the trial(s)." [1] The 2009 interpretation-of-scores row maps to section V. [1] Deriving a threshold and applying it to treatment-arm curves are adjacent items in V.C. The letters below question their placement in the same dossier.

Section VI: the ten appendices

The printed appendix list is A, exact copies, including anchors, and prior versions; B, user manual, scoring manual, patient training, investigator training, and other training; C, item tracking matrix; D, transcripts; E, "Psychometric Analysis Plans (PAPs) and Psychometric Analysis Report"; F, standalone or observational protocols; G, usability results; H, regulatory history; I, additional analyses; J, literature copies. [1] The 2009 user manual, item-tracking matrix, and transcripts map to B, C, and D, and the 2009 references map to J, with the articles themselves in Module 5.4 for a CDER or CBER application. [1]

Qualitative study reports, translation certificates, a glossary and a DHT validation report lack dedicated titles in that appendix list. Component B nevertheless requests qualitative protocols and summaries, with the item matrix in C and transcripts in D. Relevant additional material can be included and clearly cross-referenced. [1]

What the docket letters actually dispute

The comparison below codes three narrow themes across 17 substantive letters. Each listed author counts once per row; a letter can appear in more than one row. The themes overlap, precluding an additive total. The mix of consultants, sponsors, trade associations and service providers is a self-selected group.

The request is the unit being coded, rather than the presence of a keyword. For example, EMD Serono sought clearer content-validity wording within the existing body section; it is outside the first row's narrower organization criterion. OPEN Health questioned concept-of-interest versus meaningful-aspect-of-health wording around an existing DHT sentence; it is outside the second row. PhRMA discussed some digital-measure questions as future workshop topics. A different, broader coding rule could produce different totals, so the contributors are shown alongside each count. [7] [8] [9]

Two further letters in the collection are Otsuka's request for examples and an analysis-planning template, and ICON's questions about abbreviations, embedding documents and copyrighted literature. Their requests sit outside the three narrowly defined themes above. The Clinical Outcomes Solutions submission is a worked example, not a request letter; its table of contents demonstrates an alternative organization, without establishing an FDA-endorsed format. Its full body was not evaluated for this comparison. [10] [11] [12]

Familiar titles versus evidence coverage

IQVIA asked for dedicated content-validity and psychometric sections, or explicit cross-references and examples. That request addresses an organizational problem. The draft's mapping table already sends content validity to section IV and other measurement properties to sections IV and V and several appendices. The two observations are compatible: the evidence is present, but the familiar signposts are less visible. [13] [1]

Gwaltney argued that the rationale-based organization makes established categories harder to find and asked for flexibility to retain a legacy narrative. Slagle proposed a hybrid: a rationale table directing reviewers to qualitative evidence, quantitative psychometric evidence and meaningful-change work. AbbVie similarly proposed connecting the rationale table to established evidence categories. Ultragenyx favored using the 2009 structure as a foundation; Johnson & Johnson wanted qualitative and psychometric summaries retained in the body. These are six related proposals, not one unanimously agreed replacement outline. [14] [15] [16] [17] [18]

PPD Evidera wanted the relocation table to identify new requests and their implications for supporting work. That is a useful limit on this article's own crosswalk. Matching an old topic to a new location is a navigation exercise; evidence burden requires a separate assessment of the detailed content. A program must inspect the detailed requests and the relevance of each to its context of use. [19]

For an internal inventory, maintain both labels: the scientific evidence type and the draft destination. A row can say “content validity — qualitative interview report — section IV.C, Component B.” The evidence remains recognizable to a measurement specialist while the proposed destination remains visible to a document author. This is our editorial document-control recommendation.

Psychometric results need their own interpretation

The draft's E4 heading refers to sensitivity to change, but the listed evidence there is score-level floor and ceiling effects. IQVIA asked FDA to clarify whether responsiveness belongs in the dossier or whether E4 should be retitled to match the listed analysis. These two concepts should remain distinct in a sponsor's evidence map: longitudinal responsiveness requires evidence beyond limited floor or ceiling effects. [1] [13]

Where a program has justified responsiveness analyses, preserve them in its measurement-evidence plan and report and explain their relationship to the rationale. Do not drop them because a draft bullet is narrower, or relabel a floor/ceiling analysis as the full responsiveness result. Appendix E offers a proposed destination for plans and reports, and the draft expressly permits needed material beyond the printed outline. [1]

The same discipline applies to burden. IQVIA argued for assessing patient burden at the trial level. The draft itself distinguishes protocol measures addressing the overall assessment battery from cognitive-interview or usability evidence about respondents' experience with the instrument. A study team can map both without treating successful usability testing as proof that a repeated, multi-instrument trial schedule is unburdensome. The exact evidentiary approach should be justified for that study. [13] [1]

Digital technology: one rationale sentence, several different requests

Component A contains a conditional request for a rationale when digital health technology collects data used to assess the meaningful aspect of health. It gives a location for that rationale, rather than a complete digital-measure validation checklist. Other implementation details also appear in IV.B and F4, so counting only occurrences of the abbreviation DHT would understate the draft's operational content. [1]

IQVIA requested more detail on placement, device characteristics, technical validation, data capture and interpretation. Boehringer Ingelheim asked how evidence expectations for DHTs relate to those for COAs. BIO sought clarification for passive and sensor-derived measures and proposed an additional COA category. Critical Path Institute welcomed the DHT reference while questioning how exact copies and administration modes work for telephone and sensor technologies. AbbVie asked whether digitally derived COAs would require an evidence dossier. These are the five letters in the table; their specific requests differ. [13] [20] [21] [22] [16]

The priorities paper lists digital health technologies and digitally derived endpoints among cross-FDA collaborations. That establishes continued attention, not a finalized new dossier section. For internal preparation, gather relevant technical and measurement evidence and cross-reference it to the rationale and implementation descriptions. Add clearly labeled supporting material where useful, while checking the applicable DHT guidance and program agreement separately. Fitness for purpose depends on the supporting evidence for that specific sensor-derived endpoint. [2]

Deriving an interpretation range and applying it to treatment arms

Three draft passages create the unblinded-data question: the introduction mentions blinded and unblinded analyses once trial data exist; I.C requests an efficacy summary plus a CSR link; and V.C.3 requests application of meaningful-score metrics to trial results, including treatment-arm displays. The proposed content is broader than a report establishing measurement properties alone. [1]

IQVIA and OPEN Health favored cross-references to efficacy analyses instead of duplication. Vector Psychometric Group distinguished derivation of score-interpretation thresholds from their application to a particular trial, stressing the independence of psychometric work from treatment assignment. Slagle expressed a related concern about choosing meaningful-change values after seeing treatment results. Gwaltney, Ultragenyx, Johnson & Johnson and AbbVie also favored separating efficacy findings or retaining them in the CSR and related submission documents. Critical Path Institute's position was narrower: allow unblinded material as an option for a labeling-focused NDA submission. [13] [8] [23] [15] [14] [17] [18] [16] [22]

Our operational recommendation is to keep the analysis stages and access rights explicit. A team preparing a provisional dossier inventory should identify who defines score-interpretation metrics, which data those analysts can access, when the plan is finalized and who later applies the metrics to treatment-arm results. The trial’s approved blinding controls remain in force throughout that work.

Retain a trace of the proposed I.C and V.C.3 content in the crosswalk, but label its disposition as a program-specific discussion point. If trial results are not yet available, record that status honestly. Use controlled links to the planned CSR location rather than creating placeholder conclusions. Once results exist, follow the agreed analysis, blinding and submission plans. The dispute over duplicated summaries and displays remains open in the 2026 priorities paper. [1] [2]

Scope, flexibility and proprietary evidence

Several comments address scope or proportionality, but their requests are too different to collapse into a single precise “mandatory dossier” vote. BioMarin and BIO asked FDA to confirm that a centralized collated dossier is unnecessary. Boehringer Ingelheim raised agency alignment and Compendium-listed measures. PhRMA proposed a streamlined approach for established, qualified or guidance-recommended measures. EMD Serono distinguished new from qualified measures and asked about exploratory endpoints. [24] [21] [20] [9] [7]

Slagle favored a recommended template hosted on FDA's website; Critical Path Institute urged flexibility; PPD Evidera asked for clearer distinction between encouragement and requirement. AbbVie focused on the time and organizational burden involved in restructuring existing dossiers. Those positions illuminate implementation concerns, but one contributor's interpretation of the current position remains that contributor's position. [15] [22] [19] [16]

For a sponsor, the consequence is proportional planning. An exploratory instrument, a qualified measure and a novel primary endpoint can have different evidence needs. Discuss the intended use, existing evidence and remaining gaps with the appropriate review team. The discussion draft alone establishes neither a blanket exemption nor a new universal obligation to prepare a full dossier for every COA. Its own text varies the extent of evidence by measure and development stage. [1]

Glossary and transition requests are also separate questions. EMD Serono, Boehringer Ingelheim, Johnson & Johnson, PhRMA and Ultragenyx proposed additional definitions in or around the dossier materials. ICON asked about abbreviations. Timing proposals included a transition period and earlier alignment during development. These requests can inform a sponsor's meeting questions without being represented as published deadlines. [7] [20] [18] [9] [17] [11]

The priorities paper commits to updating the PFDD Glossary with terms from the planned E22 and COA-endpoint guidances. That is a different statement from a glossary inside a finalized dossier template. A sponsor can still maintain a clearly labeled working glossary to support consistent internal use. The absence of a named template glossary is no reason to leave a study team uncertain about its terminology. [2]

For proprietary measures, FDA identifies Drug Master Files as a possible mechanism for confidential instrument-level supporting information and says the subject remains under discussion. Its feasibility, reviewer access arrangements and instrument-reproduction rights still need confirmation. Confirm the available access route and rights with the developer and agency. Keep the evidence locator, owner and any access conditions visible in the preparation record. [2]

A practical crosswalk for internal preparation

The table is an editorial planning aid based on the discussion draft. “Proposed location” means a location in that draft, not a binding filing instruction. Add relevant material and adapt the organization where scientifically useful, while preserving clear navigation and the program's agreed submission approach. [1]

Worked preparation example: a fixed-form electronic measure

Consider a hypothetical multilingual registrational study using a fixed-form patient-reported instrument. The example illustrates document control only. The first inventory entry names the licensed version, the intended population and the endpoint. If those three items disagree across the protocol, license and build specification, resolve the mismatch before formatting the dossier.

Next, connect the version to the collection system. A screenshot identifies the implemented wording and response options, while configuration records explain permitted completion windows, reminders and scoring. Keep the instrument's scoring rule separate from the study endpoint's computation. A rule for handling a missing item within a score and a rule for handling a missing visit in an endpoint analysis answer different questions.

For each language, connect the approved text to its adaptation evidence and the release actually deployed. The inventory should make a late wording change visible to both the linguistic-validation owner and the study-build owner. A translation certificate by itself leaves open whether the corresponding version was used in the trial. Conversely, a screenshot by itself leaves open whether the wording was linguistically validated.

The draft's evidence locations help distribute those records across an internal preparation map, while the measure's own documentation and the study's analysis plans determine the substantive details. Appendix A can identify the administered copy; F1 can link adaptation evidence; F4 can describe electronic implementation; Appendix E can point to the psychometric plan and report. Each destination still requires substantive assessment of the evidence it holds. [1]

Finally, give each unresolved item an owner and a decision point. An unavailable trial-arm display is “pending trial results,” not a favorable result and not necessarily an inapplicable analysis. An unavailable proprietary algorithm creates an access issue requiring resolution; identify the documentation owner and the permitted review route. This distinction makes the inventory useful before all documents exist.

What the eCOA handoff can supply

The instrument, study system and dossier have related but different responsibilities. EClinCloud describes eCOA as the capture layer, scale management as a separate service covering due diligence, licensing, translation and linguistic validation, and eLearning as a training workflow. These descriptions identify possible records and services; instrument fitness and submission acceptability remain separate scientific and regulatory determinations. [25]

A project can ask the capture team for the administered configuration, relevant screenshots and user-interaction description; the scale-management team for the approved language/version and rights record; and the training team for the applicable completion evidence. The sponsor's measurement and regulatory teams still need to connect those records to the rationale for the specific context of use. EClinCloud's eCOA offering is the relevant starting point for that implementation discussion.

Fixed-form screenshots are only one example. A computerized adaptive assessment presents different items to different respondents; a sensor may have no respondent-facing form at all. Vector Psychometric Group and Critical Path Institute raised these limitations in their letters. For such cases, describe the actual administration and data-collection process and agree how reviewers will inspect it. Treating one screenshot as the complete instrument representation would conceal important behavior. [23] [22]

Related EClinCloud reports address registry outcome descriptions, instrument names and CDISC QRS, and multilingual eCOA implementation. This crosswalk adds the evidence-location layer. A registry name, terminology mapping, license, completed build and well-organized dossier each answer a different question.

Limits and the next decision

This review compares the September 2025 discussion document, the 30 September 2026 priorities paper and a captured set of public comments. It evaluates document wording and evidence placement. Acceptance rates, future comment dispositions and validation of particular COAs fall outside its scope. Theme counts reflect the stated coding rule and the material read; they should not be generalized to stakeholders who did not submit a letter.

The publication decision is therefore modest and actionable: preserve the substantive evidence, make it easy to locate, and keep draft status visible. Use the crosswalk to prepare questions about scope, missing evidence, proprietary access and analysis separation. Recheck the eventual final template and agency agreement before treating any proposed organization as the program's submission plan.

Sources

1. U.S. Food and Drug Administration, Public Meeting Discussion Documents: Draft COA Evidence Outline / Draft COA Evidence Dossier Template, September 2025. Accessed 4 October 2026. Banner; introduction and "What has changed?"; table mapping the 2009 appendix to the draft; outline sections I–VI; dossier instructions, including Module 5.3.5.3; Tables 1 and 2; section IV.C Components A–H; section V.C; Appendices A–J.

2. U.S. Food and Drug Administration, Methodologic and Other Challenges Related to Patient Experience Data: Priorities for Future Work, cover date 30 September 2026. Accessed 4 October 2026. Evidence Dossier Template section; Finalizing Guidance Documents section; proprietary COA and Drug Master File paragraph; DHT bullet under Continuing Collaboration Across FDA.

3. U.S. Food and Drug Administration, FDA Patient-Focused Drug Development Guidance Series for Enhancing the Incorporation of the Patient's Voice in Medical Product Development and Regulatory Decision Making. Accessed 4 October 2026. Series heading for Guidance 3.

4. U.S. Food and Drug Administration, Patient-Focused Drug Development: Workshop #2 to Discuss Methodologic and Other Challenges Related to Patient Experience Data. Accessed 4 October 2026. The page retrieved that day still described a draft evidence dossier template.

5. Food and Drug Administration, Patient-Focused Drug Development: Workshop #2 To Discuss Methodologic and Other Challenges Related to Patient Experience Data; Public Workshop; Request for Comments, 90 FR 42017, 28 August 2025, docket FDA-2025-N-2368. Comment deadline 18 November 2025, 11:59 p.m. Eastern Time.

6. Regulations.gov, comments on FDA-2025-N-2368-0001. Accessed 4 October 2026. Page status: closed; comment period ended 18 November 2025 at 11:59 p.m. EST; Document Comments 24; displaying 1–24 of 24.

7. EMD Serono, comment FDA-2025-N-2368-0023. Accessed 4 October 2026. Glossary, qualified and exploratory scope, transition period, and the line-230 content-validity wording.

8. OPEN Health, comment FDA-2025-N-2368-0024. Accessed 4 October 2026. Comments on discussion-document lines 43–44, 208–212, and 529–547.

9. Pharmaceutical Research and Manufacturers of America, comment FDA-2025-N-2368-0016. Accessed 4 October 2026. Applicability and glossary comments; DHT remarks read as workshop topics.

10. Otsuka. Comment FDA-2025-N-2368-0010. Requests for completed examples and analysis-planning material. Accessed 4 October 2026.

11. ICON. Comment FDA-2025-N-2368-0025. Abbreviation, document-embedding and literature-access questions. Accessed 4 October 2026.

12. Clinical Outcomes Solutions, worked example posted as FDA-2025-N-2368-0022. Accessed 4 October 2026. Cover and table of contents only.

13. IQVIA, comment FDA-2025-N-2368-0020, 18 November 2025. Accessed 4 October 2026. General comments on dedicated sections and DHT direction; specific comments on unblinded analyses and patient burden.

14. Chad Gwaltney, Gwaltney Consulting, comment FDA-2025-N-2368-0021, 18 November 2025. Accessed 4 October 2026.

15. Ashley F. Slagle, Aspen Consulting, comment FDA-2025-N-2368-0018. Accessed 4 October 2026. Proposed hybrid outline; unblinded-data comment; website-hosting comment.

16. AbbVie, comment FDA-2025-N-2368-0026. Accessed 4 October 2026. Bridge onto the 2009 organization; digitally derived COAs; CSR comment. The "Proposed Guidance Document" header is AbbVie's table label.

17. Ultragenyx, comment FDA-2025-N-2368-0012. Accessed 4 October 2026. 2009-foundation comment; meaningful-aspect-of-health glossary comment; unblinded-data comment.

18. Johnson & Johnson, comment FDA-2025-N-2368-0013. Accessed 4 October 2026. Qualitative and psychometric summaries; CSR link; glossary; qualitative-report location; implementation timeline.

19. PPD Evidera / Thermo Fisher Scientific, comment FDA-2025-N-2368-0019. Accessed 4 October 2026. Mapping-table comment; required-versus-encouraged comment.

20. Boehringer Ingelheim, comment FDA-2025-N-2368-0011, 17 November 2025. Accessed 4 October 2026. The letter header prints Docket No. FDA-2025-D-2368. The posting ID is FDA-2025-N-2368-0011.

21. Biotechnology Innovation Organization, comment FDA-2025-N-2368-0014. Accessed 4 October 2026. Essential-versus-optional elements; DHT and passive monitoring; collated-dossier sentence; transition plan.

22. Critical Path Institute, comment FDA-2025-N-2368-0004. Accessed 4 October 2026. Nonbinding-guidance comment; unblinded analyses as a labeling option; screenshots, mode, and the DHT sentence.

23. Vector Psychometric Group, comment FDA-2025-N-2368-0017. Accessed 4 October 2026. Section V.C.3 comment; exact-copy comment on computerized adaptive tests.

24. BioMarin, comment FDA-2025-N-2368-0015. Accessed 4 October 2026. Collated-dossier comment.

25. EClinCloud, Scale Management, and the eCOA and eLearning product descriptions on eclincloud.com. Scale management accessed as a distinct service on 6 September 2026. Capture, licensing, translation, and training workflows are separate from a determination that a COA is fit-for-purpose.