From Protocol to CSR: Mapping the Full Clinical Document Lifecycle

Introduction

The first draft of a clinical trial protocol is rarely the last. By the time a study reaches database lock, the original document will have been amended, referenced in an investigator's brochure, reflected in a consent form, operationalized through a statistical analysis plan, reviewed in annual safety reports, and ultimately reconstructed in a clinical study report running to hundreds of pages. Each document is distinct. Each carries its own regulatory standard. Yet they all share one governing constraint: the CSR must reconcile the protocol, all amendments, the SAP, the TLFs, the safety records, and the conduct documentation into a single coherent account of what was planned and what actually happened.

That chain of accountability is not accidental. It is the design. Regulatory agencies across ICH member regions accept clinical trial data on the premise that the documents governing a study form a coherent, internally consistent record of how the trial was planned, conducted, and reported. When that consistency breaks down, when the protocol says one thing and the CSR describes another, the consequences range from time-consuming regulatory queries to submission timelines that slip by months.

This article maps the full document lifecycle: what each document is, when it is produced, which regulatory standards govern it, and how each feeds into the next. Understanding this sequence matters for sponsors, CROs, medical writers, and regulatory teams who want to move from first-in-human studies to marketing authorization without losing months to documentation gaps.

Why Document Architecture Matters More Than Any Single Document

There is a tendency in clinical development to treat regulatory documents as milestones to be completed and filed rather than as an integrated system. A protocol team finalizes the study design. A separate medical writing group drafts the informed consent. A statistics department writes the SAP. A safety team manages the DSUR. By the time the CSR is commissioned, the people writing it may never have seen the original protocol.

The result is a predictable category of errors: CSRs that describe endpoint definitions inconsistent with the SAP, consent forms that list risks not captured in the IB, DSUR narratives that use different adverse event terminology than the protocol's safety monitoring section. These are not trivial. ICH E6(R3), finalized January 6, 2025 and adopted by the FDA in September 2025, makes explicit that sponsors bear responsibility for the integrity of clinical trial documentation across the entire trial lifecycle ^[1].

The argument for treating clinical documents as a system rather than a checklist is also economic. A 2024 Tufts Center for the Study of Drug Development benchmark study (analyzing data on 950 protocols and 2,188 amendments from 16 pharmaceutical companies and CROs) found that the prevalence of protocols with at least one amendment rose from 57% in 2015 to 76% across Phase I through IV trials, with the mean number of amendments per protocol increasing 60% to 3.3 ^[2]. The same study found that the time from identifying the need to amend a protocol to receiving last oversight approval now averages 260 days, and that investigative sites operate with different protocol versions for a mean of 215 days during that period ^[2].

Document 1: The Clinical Trial Protocol

Every subsequent document in the lifecycle is downstream of the protocol. The protocol defines the study's scientific rationale, objectives, design, methodology, statistical approach, safety monitoring plan, and the specific procedures required of investigative sites and participants. ICH E6(R3) specifies that the protocol must "describe the objectives, design, methodology, statistical considerations, and organization of a trial," and that it must address factors critical to participant safety and the reliability of trial results ^[1].

In practice, the protocol serves three simultaneous audiences: the regulatory agency reviewing the IND or CTA submission, the ethics committee evaluating participant safety and consent requirements, and the investigative sites determining whether and how they can execute the study. A protocol that satisfies one audience but not another creates downstream problems. Overly restrictive eligibility criteria may satisfy a regulatory reviewer's desire for a clean study population while making enrollment impossible at sites. That pressure drives amendments.

ICH E8(R1), adopted October 6, 2021, introduced a structured approach to protocol quality with direct implications for document design ^[4]. The guidance frames protocol development around "critical to quality factors," meaning attributes whose compromise would undermine participant safety, data integrity, or the reliability of a study's conclusions. Teams are expected to identify these factors during protocol design and build monitoring and documentation requirements around them, rather than applying the same oversight intensity to every study element regardless of risk ^[4].

The protocol must be finalized before the IND or CTA is submitted. Under 21 CFR 312.23, the protocol submitted as part of an IND application must include the study's objectives, population, design, dosing schedule, procedures to minimize bias, and clinical procedures to monitor safety ^[5].

Protocol Amendments: The Document Lifecycle Within a Document

Not every protocol change triggers an IND amendment submission. Under 21 CFR 312.30, a sponsor must submit a protocol amendment to the FDA for any change in a Phase I protocol that significantly affects the safety of subjects, or for any Phase II or III change that significantly affects the safety of subjects, the scope of the investigation, or the scientific quality of the study ^[5]. Examples the regulation identifies include significant dose increases, addition or removal of a control group, and changes to safety monitoring procedures ^[5]. Administrative or editorial changes (correcting a typographical error, clarifying a procedure description without altering its substance) do not necessarily meet this threshold, though sponsors must document the basis for that determination.

When an amendment is warranted, the documentation cascade that follows is substantial. The amendment must be version-controlled and submitted to the relevant regulatory authority. It must also be reviewed by the IRB or IEC before implementation at sites in most circumstances. Any modifications to protocol-specified procedures made before formal amendment approval should be documented and assessed as protocol deviations under FDA draft guidance ^[6]. And critically, the amendment triggers reviews of related documents: a change to an eligibility criterion typically requires an update to the informed consent form, re-consent of enrolled participants where applicable, and potentially a revision to the IB if the amendment was driven by a new safety signal.

A 2016 Tufts CSDD study of 836 Phase I through IV protocols, the original source for amendment cost benchmarks, found the median direct cost to implement a substantial amendment (defined as one requiring internal sponsor approval followed by regulatory or ethics board review on a global level) was $141,000 for Phase II protocols and $535,000 for Phase III protocols ^[3]. Protocols that underwent at least one substantial amendment also took an average of three months longer to complete their study conduct period than protocols that did not ^[3]. These figures represent direct costs only and do not capture indirect costs from site disruptions, staff time, or delayed enrollment.

Document 2: The Investigator's Brochure

The investigator's brochure (IB) connects the sponsor's knowledge about an investigational product to the investigators using it. It compiles clinical and nonclinical data on the product: pharmacological properties, mode of action, prior human experience, and known or potential risks. ICH E6(R3) Appendix A specifies that the IB must be reviewed at least annually and revised whenever clinically significant new information becomes available, and that more frequent revision may be appropriate depending on the stage of development ^[1].

During an active development program, the IB may be revised multiple times within a single year if safety signals emerge, new nonclinical data are generated, or results from other ongoing studies alter the product's known risk-benefit profile. Under FDA regulations at 21 CFR 312.23(a)(5), sponsors must include current IB content in their IND submission and update it when material changes occur ^[5].

The IB's significance extends beyond site staff orientation. It serves as the primary reference document for determining whether an adverse event is expected or unexpected, a determination that governs SUSAR (Suspected Unexpected Serious Adverse Reaction) reporting timelines. ICH E2F specifies that the IB in effect at the start of a reporting period serves as the reference safety information (RSI) for DSUR purposes; if an event appears in the RSI, it is expected, and if not, expedited reporting requirements apply ^[7]. An IB that does not reflect the current state of safety knowledge, or that describes the product's risk profile less precisely than the accumulated data warrant, distorts the SUSAR determination process at every site using it.

Document 3: The Informed Consent Form

The informed consent form (ICF) is the participant-facing document and is often treated as primarily an ethical rather than a regulatory document. This framing undersells its regulatory significance. The ICF must accurately reflect the current protocol version, including its risks, procedures, and participant burden. When a protocol amendment changes any element that affects participant safety or the nature of participation, the ICF will typically need to be updated and re-approved by the IRB or IEC. Whether existing participants must be re-consented depends on the materiality of the change and the requirements of the applicable IRB, IEC, or local regulatory authority.

ICH E6(R3) specifies that the consent process must be conducted in a way that ensures participants understand the trial's objectives, procedures, foreseeable risks, and their right to withdraw at any time without penalty ^[1]. The FDA's September 2025 adoption of E6(R3) clarified that remote and electronic consent modalities are acceptable within this framework, provided the process ensures genuine comprehension and voluntariness ^[8].

The downstream implication for document consistency is direct. An ICF written against protocol version 1.0 but never updated to reflect version 3.0 (perhaps because the amendment team considered the change minor) creates a consent record that does not accurately represent what participants experienced. Regulatory inspections routinely identify this discrepancy. The severity of the consequence depends on whether the un-updated element was material to participant safety or to data integrity.

Document 4: The Statistical Analysis Plan

The statistical analysis plan (SAP) specifies, before database lock, how the trial's data will be analyzed. It operationalizes the protocol's primary and secondary endpoints into concrete statistical procedures, defines analysis populations (ITT, per-protocol, safety), pre-specifies the handling of missing data and protocol deviations, and includes shells for all tables, listings, and figures (TLFs) that will appear in the CSR.

The SAP is governed by ICH E9, the statistical principles guidance for clinical trials, and its 2019 addendum, ICH E9(R1), which introduced the estimands framework ^[9]. The estimands framework addresses a longstanding gap: the disconnect between what a trial's primary objective specifies and what the primary statistical analysis actually estimates, particularly when intercurrent events occur: treatment discontinuation, rescue medication use, or death before the primary endpoint assessment. ICH E9(R1) requires that the estimand be defined precisely: specifying the population, endpoint, handling of intercurrent events, and the summary measure, then linking that directly to the analysis method that estimates it ^[9].

The SAP must be finalized before database lock. A SAP finalized after the team has seen unblinded data, or modified after preliminary results were reviewed, introduces potential outcome reporting bias that undermines the study's inferential validity. Regulatory reviewers examine version histories and timestamps when assessing whether pre-specified analyses were genuinely pre-specified.

The SAP's relationship to the protocol is one of specificity without contradiction. The SAP expands on the protocol's statistical section with detail that a protocol cannot always include, but it cannot introduce endpoints, analysis populations, or decision rules that the protocol does not support. When the two documents diverge, the protocol governs. Unresolved discrepancies become the subject of regulatory queries during CSR review.

Document 5: The Development Safety Update Report

The development safety update report (DSUR) is the annual safety review document submitted to regulatory agencies during an active clinical program. It consolidates safety information from ongoing and completed clinical trials of an investigational product during the reporting period and evaluates whether the safety profile has changed materially from what was known at the start of that period.

ICH E2F specifies that the DSUR must include a cumulative summary of relevant safety data, an analysis of serious adverse events across the development program, and an evaluation of whether the data warrant changes to the IB or any ongoing clinical procedures ^[7]. The DSUR's reference date is typically the development international birth date (DIBD), which is the date of first authorization to conduct clinical trials globally, creating a uniform annual reporting cycle.

Reporting obligations under ICH E2F apply to investigational drugs with at least one ongoing clinical trial or that are under clinical development in any ICH region. The specific submission requirements, including which regulatory authorities must receive the DSUR and in what form, depend on national or regional regulatory law. Sponsors conducting multi-country trials typically align on a single global DSUR and supplement it with country-specific formats where required.

The FDA has proposed replacing its current IND annual report requirement with an FDA DSUR aligned to the ICH E2F format. A December 2022 proposed rule cited the DSUR's additional safety evaluation requirements as an improvement over the existing IND annual report structure, particularly its systematic evaluation of whether IB changes are warranted ^[10]. That proposed rule had not been finalized as of mid-2026.

The DSUR's connection to other lifecycle documents is systematic. The IB version in effect at the start of the reporting period serves as the RSI for expectedness determinations during that period ^[7]. This means a DSUR cannot be written accurately without access to the correct version-controlled IB. Changes documented in the DSUR (new signals, revised expectedness determinations, protocol modifications driven by safety data) may trigger cascading updates to the IB, protocol, and consent forms before the next reporting cycle.

Document 6: The Clinical Study Report

The clinical study report (CSR) is the final document in the trial lifecycle and, in terms of regulatory weight, the most consequential. It describes the complete methods and results of the study in sufficient detail for a regulatory reviewer to reconstruct the analysis, assess the data independently, and evaluate the study's contribution to the benefit-risk profile of the investigational product.

The CSR is written according to ICH E3, the guideline for the structure and content of clinical study reports, adopted November 30, 1995 and accepted by regulatory authorities across ICH member regions ^[11]. ICH E3 specifies a standard structure: title page, synopsis, introduction, study objectives, investigational plan, study patients, efficacy evaluation, safety evaluation, discussion, references, and appendices. This structure is flexible, explicitly allowing flexibility in presentation where scientific clarity benefits from a different arrangement ^[11]. A CSR conforming to ICH E3 can be submitted to the FDA, EMA, PMDA, and other ICH-member regulators without material reformatting.

Writing a CSR for a moderately complex Phase III study is operationally demanding. A survey of medical writers conducted over five years found a mean of 16.9 days from receipt of final TLFs to delivery of the first draft, and 25.7 days from first draft to the final draft routed for review ^[12]. From database lock to CSR completion, the mean was 83 days across the survey respondents ^[12]. Industry estimates for manual Phase III CSR preparation typically range from three to six months depending on study complexity, figures cited by commercial vendors as a basis for describing automation time savings, and therefore classified here as vendor-stated rather than independently validated ^[13].

The CSR's dependency on every prior document in the lifecycle is absolute. The synopsis must match the protocol's stated objectives. The efficacy section must apply exactly the analysis methods and population definitions from the SAP. The safety narrative must categorize adverse events using the same terminology and expectedness determinations that governed safety reporting during the trial. Every protocol amendment must be described with its rationale in the study conduct section. Any deviation from the SAP must be explained and disclosed. The CSR does not simply report results. It accounts for the entire study as it was actually conducted, against the plan as it was pre-specified.

The Cross-Document Consistency Problem

The document lifecycle described above creates a specific risk distinct from any individual document error: cross-document inconsistency. The risk accumulates across a study's duration as documents are produced by different teams, at different times, and often without a shared repository that enforces version alignment.

Common examples include eligibility criteria defined in the protocol but described differently in the ICF; DSUR safety narratives using MedDRA coding conventions applied inconsistently in the trial database; SAP analysis populations that do not match the protocol's definitions precisely enough to be unambiguous; and CSR sections that reference the wrong protocol version or omit description of an amendment that affected a material proportion of enrolled participants.

None of these is necessarily fatal in isolation. Each can generate a regulatory query or request for clarification. When queries cluster across multiple sections of a pivotal NDA or BLA submission, they extend the review clock substantially. FDA's refuse-to-file (RTF) criteria, set out in the agency's Good Review Practice guidance, cover applications that are materially incomplete, inadequately organized, or not reviewable in a timely and complete way ^[15]. Ordinary cross-document inconsistencies do not automatically trigger an RTF, but when they are pervasive enough to prevent substantive review (for instance, because the CSR cannot be reconciled with the SAP or the protocol), they can contribute to filing concerns. The upstream remedy is governance, not remediation.

The upstream solution is document planning: mapping dependencies at the protocol stage and establishing governance processes that require cross-review whenever any document is amended. The downstream solution, when inconsistencies are found late in development, is a systematic reconciliation process before CSR finalization: comparing each CSR section against the protocol, all amendments, the SAP, and the TLFs it references.

Document Dependency Map

The table below summarizes the primary cross-document dependencies, the consistency risk each creates, and the typical trigger for a downstream update.

Document Governance Checklist

The dependency map above becomes operational through a governance process. The following checklist covers the minimum steps teams should complete at each major lifecycle transition.

A Concrete Inconsistency Example

Consider a Phase III trial where the protocol defines the primary endpoint as "mean change from baseline in HbA1c at Week 24, analyzed using a mixed model for repeated measures (MMRM) in the intent-to-treat (ITT) population." The SAP, written several months after the protocol, defines the ITT population as "all randomized participants who received at least one dose of study drug and had at least one post-baseline efficacy assessment." The protocol did not include the "at least one post-baseline efficacy assessment" qualifier; it defined ITT simply as all randomized participants who received at least one dose.

During CSR drafting, the medical writer follows the SAP definition. The efficacy section therefore excludes five participants who were randomized, dosed, but never attended a post-baseline visit. In the protocol-defined ITT, those five participants would be included with their baseline data carried forward under the MMRM missing data assumptions.

The discrepancy is small in absolute terms. But a regulatory reviewer comparing the protocol's ITT definition to the SAP and then to the CSR's analysis dataset will flag it. The sponsor must then provide a post-hoc explanation, an additional sensitivity analysis using the protocol-defined population, and a written justification for why the SAP deviation did not materially affect the primary conclusion. All of that delays the review clock.

The fix is not complicated. An explicit footnote in the SAP acknowledging the refinement of the ITT definition relative to the protocol, with the sponsor's documented rationale. Written prospectively, this is a minor administrative step. Identified during CSR review or regulatory inspection, it is weeks of remediation.

Regulatory and Documentation Considerations

ICH E6(R3) and the Trial Master File

ICH E6(R3) reorganized essential document requirements, specifically the Trial Master File (TMF), to reflect the reality of electronic systems and decentralized operations ^[1]. The essential documents listed in E6(R3) span the entire trial lifecycle and include, at different stages, the protocol and all amendments, the IB and all revisions, signed consent forms, the SAP, safety reports, and the final CSR. The TMF must be current, organized, and retrievable for inspection at any point during or after the trial ^[1].

E6(R3) also introduced explicit requirements around data governance that affect every document in the lifecycle. Electronic systems used in documentation must be validated, access-controlled, and audit-trailed. Changes to documents stored in electronic systems must be attributable, dated, and non-destructive of prior versions ^[1].

21 CFR Part 11 and Electronic Records

For US-regulated studies, documents maintained in electronic form are subject to 21 CFR Part 11 ^[14]. Specifically, 21 CFR 11.10 requires that electronic record systems be validated to ensure accuracy, reliability, and consistent performance; include audit trails that record when entries are created, modified, or deleted; limit system access to authorized individuals; and ensure that the authorship of each record is attributable to a specific individual ^[14]. Electronic signatures are governed under 21 CFR 11.50, which requires that each electronic signature be linked to its respective record and include the signer's name, the date and time of signing, and the meaning of the signature ^[14]. These requirements apply to protocols, amendments, consent forms maintained electronically, and the systems used to prepare and archive the CSR.

AI and Automation Across the Document Lifecycle

Generative AI has entered clinical document production at several points in the lifecycle, with varying levels of maturity and appropriate caution. The clearest current application is in structured document drafting: systems that use protocol text, prior TLF outputs, and reference documents to generate draft CSR sections, DSUR narratives, and IB revisions. These tools do not eliminate the medical writing function; they change what medical writers spend time doing, shifting effort from initial drafting toward review, scientific judgment, and cross-document consistency verification.

The cross-document consistency problem described earlier is one where AI-assisted tools may add meaningful value. Systems designed to flag when a CSR section contradicts the protocol or SAP, or when an IB revision introduces a term not yet reflected in the consent form, address a problem that is difficult to solve through human review alone when documents run to hundreds of pages across multiple versions. The prerequisite is structured, version-controlled document inputs, and that returns the conversation to document governance practices.

AI performance claims in this space require careful evaluation. Estimates that tools can generate "up to 70% of CSR content automatically" reflect vendor-stated design intent under specific conditions, not independently validated performance across diverse study types ^[13]. The 3-to-6-month manual baseline those claims are measured against also comes from vendor sources rather than controlled studies. Human medical writer review remains essential for scientific accuracy, regulatory compliance, and narrative coherence, and ICH E6(R3)'s emphasis on sponsor accountability for trial documentation makes that oversight a regulatory expectation, not merely a best practice.

How Kitsa Fits Into This Problem

Kitsa describes KScribe as an AI-native document generation platform designed to support the production of clinical regulatory documents including protocols, IBs, ICFs, DSURs, SAPs, and CSRs, with cross-document consistency as an explicit design goal. The platform is built to draw from a shared, study-specific knowledge base so that definitions, terminology, and cross-references that appear in one document carry through to others, rather than being re-entered independently by separate teams. For sponsors and CROs managing multiple simultaneous studies, this architecture addresses a real operational risk in the document lifecycle described throughout this article.

If cross-document consistency is a pressure point in your regulatory submissions, the KScribe product page covers how Kitsa approaches document generation across the full lifecycle.

Key Takeaways

• The six core clinical documents (protocol, IB, ICF, SAP, DSUR, and CSR) are not independent deliverables. Each is governed by an ICH guideline and depends structurally on the documents that precede it.
• Protocol amendment prevalence has risen from 57% to 76% of Phase I-IV trials since 2015, with the average number of amendments per protocol now at 3.3. The time from identifying the need to amend to receiving last oversight approval now averages 260 days, and sites operate with different protocol versions for an average of 215 days during that period (Tufts CSDD, 2024).
• Under 21 CFR 312.30, protocol amendments to the IND are required for changes that significantly affect participant safety, the scope of the investigation, or the scientific quality of the study. Not all protocol changes meet this threshold.
• The SAP must be finalized before database lock. Any deviation from its pre-specified methods must be disclosed and justified in the CSR.
• The IB version in effect at the start of a DSUR reporting period governs expectedness determinations for that period, meaning IB version control directly affects SUSAR reporting accuracy across the program.
• Cross-document inconsistency, rather than errors within a single document, is a practically significant source of regulatory queries during CSR and submission review, and one that governance planning at the protocol stage is well positioned to prevent.
• A moderately complex Phase III CSR averages 83 days from database lock to completion under traditional medical writing workflows; planning document governance from the protocol stage reduces the reconciliation burden that accumulates at that final stage.

Frequently Asked Questions

References

1. [1]International Council for Harmonisation. "ICH E6(R3) Good Clinical Practice Guideline." Step 4, January 6, 2025. https://database.ich.org/sites/default/files/ICH_E6%28R3%29_Step4_FinalGuideline_2025_0106.pdf
2. [2]Getz K, Smith Z, Botto E, Murphy E, Dauchy A. "New Benchmarks on Protocol Amendment Practices, Trends and Their Impact on Clinical Trial Performance." Therapeutic Innovation & Regulatory Science 58(3):539-548, 2024. DOI: 10.1007/s43441-024-00622-9. PMID: 38438658. https://pubmed.ncbi.nlm.nih.gov/38438658/
3. [3]Getz K, Dunn D, et al. "The Impact of Protocol Amendments on Clinical Trial Performance and Cost." Therapeutic Innovation & Regulatory Science 50(4):436-441, 2016. DOI: 10.1177/2168479016632271.
4. [4]International Council for Harmonisation. "ICH E8(R1): General Considerations for Clinical Studies." Step 4, October 6, 2021. https://database.ich.org/sites/default/files/E8-R1_Guideline_Step4_2021_1006.pdf
5. [5]U.S. Food and Drug Administration. "21 CFR Part 312: Investigational New Drug Application." Electronic Code of Federal Regulations. https://www.ecfr.gov/current/title-21/chapter-I/subchapter-D/part-312
6. [6]U.S. Food and Drug Administration. "Protocol Deviations for Clinical Investigations: Guidance for Industry." Draft Guidance, 2024. https://www.fda.gov/media/184745/download
7. [7]International Council for Harmonisation. "ICH E2F: Development Safety Update Report." Step 5, September 2011. https://www.ema.europa.eu/en/documents/scientific-guideline/ich-guideline-e2f-development-safety-update-report-step-5_en.pdf
8. [8]U.S. Food and Drug Administration. "E6(R3) Good Clinical Practice: Guidance for Industry." September 2025. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/e6r3-good-clinical-practice-gcp
9. [9]International Council for Harmonisation. "ICH E9(R1): Addendum on Estimands and Sensitivity Analysis in Clinical Trials." Step 5, November 2019. https://database.ich.org/sites/default/files/E9-R1_Step4_Guideline_2019_1203.pdf
10. [10]U.S. Food and Drug Administration / Federal Register. "Investigational New Drug Application Annual Reporting: Proposed Rule." Federal Register Vol. 87, No. 235, December 9, 2022. https://www.federalregister.gov/documents/2022/12/09/2022-26731/investigational-new-drug-application-annual-reporting
11. [11]International Council for Harmonisation. "ICH E3: Structure and Content of Clinical Study Reports." Step 4, November 30, 1995. https://database.ich.org/sites/default/files/E3_Guideline.pdf
12. [12]Comis K. "Clinical Study Reports 101: Tips and Tricks for the Novice." Association of Clinical Research Professionals (ACRP), September 15, 2020. https://acrpnet.org/2020/09/15/clinical-study-reports-101-tips-and-tricks-for-the-novice
13. [13]Clinion Editorial Team. "Clinical Study Report (CSR): Structure, ICH E3 Format and Submission." Clinion, December 2025. [Vendor-stated figures; cited for illustrative purposes only and not independently validated.] https://www.clinion.com/insight/clinical-study-reports-csr-complete-guide/
14. [14]U.S. Food and Drug Administration. "21 CFR Part 11: Electronic Records; Electronic Signatures." Electronic Code of Federal Regulations. https://www.ecfr.gov/current/title-21/chapter-I/subchapter-A/part-11
15. [15]U.S. Food and Drug Administration. "Good Review Practice: Refuse to File." Manual of Policies and Procedures (MAPP), October 2013. https://www.fda.gov/files/about%20fda/published/Good-Review-Practice---Refuse-to-File.pdf

Suggested internal links: kitsa.ai/regulatory-document-generation | ICH E6(R3) Implications for Sponsors | Protocol Amendments: Causes, Costs, and Prevention | What Is a DSUR? (Glossary)