CCDM Certified Clinical Data Manager Questions and Answers

Thebest methodto identify sites withhigh subject attritionis to calculate theproportion of patients for which two visit periods have passed without data, by site.

According to theGCDMP (Chapter: Data Quality Assurance and Control), subject attrition is an important performance indicator for data completeness and site compliance. Evaluating missing or delayed data acrossmultiple consecutive visit periodsallows for early detection of potential dropouts or site-level operational issues.

By assessing this proportion at thesite level, the Data Manager can distinguish between random missing data and systematic site underperformance. Counting or proportioning late visits (options B and C) identifies scheduling delays, not attrition. Looking at missing data without site context (option D) fails to identify site-specific patterns, limiting corrective action.

This metric aligns withrisk-based monitoring (RBM)practices recommended byICH E6 (R2)andFDA RBM Guidance, which promote proactive identification of sites at risk of data loss.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Data Quality Assurance and Control, Section 5.4 – Site Performance Metrics

ICH E6 (R2) Good Clinical Practice, Section 5.18 – Monitoring and Site Performance Evaluation

FDA Guidance for Industry: Oversight of Clinical Investigations – Risk-Based Monitoring, Section 6 – Site Performance Metrics

According to theGood Clinical Data Management Practices (GCDMP), thedatabase lock (DBL)process signifies the formal closure of the clinical trial database, ensuring that no further changes can be made to the data before statistical analysis. This process must be documented, controlled, and approved by key study stakeholders to ensuredata accuracy, completeness, and readiness for analysis.

The GCDMP specifies that database lock should occur only after all data cleaning, discrepancy resolution, and reconciliation activities are complete. The lock authorization typically requires the approval of theClinical/Scientific Representative (to confirm clinical completeness), the Data Manager (to confirm data integrity and query closure), and the Biostatistician (to confirm readiness for statistical analysis).

This tri-party approval ensures that the database reflects final, verified data consistent with the clinical protocol, and that the statistical analysis dataset derived from the database is accurate and auditable. The approval process is documented via aDatabase Lock Authorization Form or Sign-off Log, which becomes part of the permanent trial master file (TMF).

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Database Lock and Archiving, Section 7.1 – Lock Procedures and Approvals

ICH E6 (R2) GCP, Section 5.5.3 – Data Handling and Record Keeping

FDA Guidance for Industry: Computerized Systems Used in Clinical Investigations – Section on Database Closure

To monitorsite variability in eligibility screening, you must analyzethe number of subjects screenedversusthe number of subjects enrolledat each site. This allows identification of sites that are over- or under-screening relative to their enrollment yield.

TheGCDMP (Chapter: Data Quality Assurance and Metrics)emphasizes thatscreening-to-enrollment ratiosare critical indicators of protocol compliance and data quality. Sites with unusually low conversion rates may have unclear understanding of inclusion/exclusion criteria, requiring targeted training or monitoring.

Other options (A, C, D) provide enrollment metrics but do not revealscreening efficiency or variability, which depend on bothscreening and enrollment data.

Thus,option Bcorrectly identifies the data necessary for monitoring eligibility screening performance across sites.

Reference (CCDM-Verified Sources):

SCDM GCDMP, Chapter: Data Quality Assurance and Metrics, Section 5.4 – Site Performance Metrics

ICH E6(R2) GCP, Section 5.18 – Monitoring and Site Oversight Requirements

InElectronic Data Capture (EDC)system validation,User Acceptance Testing (UAT)is a mandatory phase that must becompleted before data collection begins in the production environment.

According to theGCDMP (Chapter: Database Design, Validation, and Testing)andFDA 21 CFR Part 11, UAT ensures that the EDC system meets allprotocol-specific, functional, and regulatory requirementsbefore it is deployed for live use. The goal is to verify that the system performs exactly as intended by simulating real-world user interactions withtest datain avalidated test environment.

Data collection prior to UAT completion would violate validation requirements and risk noncompliance withICH E6 (R2) GCP Section 5.5.3, which mandates that all computerized systems be validated and tested before use.

While options A and C describe correct components of testing strategy,the key regulatory requirementis thatUAT must be completed and approved before live data entry begins. Option D is incorrect — risk-based UAT is an accepted modern validation approach under bothFDA and GAMP5principles.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Database Design and Validation, Section 5.3 – User Acceptance Testing

FDA 21 CFR Part 11 – Validation of Electronic Systems (Section 11.10(a))

ICH E6 (R2) GCP, Section 5.5.3 – Validation Before Use in Production Environment

Database auditsare conducted to ensureongoing data accuracy, completeness, and compliancethroughout the lifecycle of a clinical trial. According to theGood Clinical Data Management Practices (GCDMP, Chapter: Data Quality Assurance and Control), quality audits are most effective when performedperiodically during study conduct, rather than waiting until study completion.

Performing audits periodically allows early detection of data entry errors, protocol deviations, and system inconsistencies, thereby reducing the risk of large-scale data issues before database lock. This proactive approach aligns withrisk-based quality management principlesoutlined inICH E6(R2)and ensures corrective actions are implemented in real time.

Options A and B represent reactive quality control, which occurs too late to prevent data issues. Option C (after first few cases) provides initial validation but does not ensure continuous oversight.

Therefore,option D — “Periodically throughout the study”— represents the optimal and compliant timing for quality audits of the database.

Reference (CCDM-Verified Sources):

SCDM GCDMP, Chapter: Data Quality Assurance and Control, Section 5.3 – Ongoing Quality Control and Auditing

ICH E6(R2) GCP, Section 5.1.1 – Quality Management System and Risk-Based Monitoring

FDA Guidance for Industry: Computerized Systems Used in Clinical Investigations, Section 6.5 – Data Review and Auditing Practices

The correct action is toconsult the CRO’s Lead Data Manager and CRO’s Project Leader(Option C) to ensure the issue is addressed through the appropriate oversight and escalation process.

According to theGCDMP (Chapter: Project Management and Communication), when a sponsor Data Manager identifies significant data management issues under CRO oversight — such as aging queries or site performance disparities — communication must followthe established governance and escalation pathwaydefined in the Scope of Work (SOW) and Data Management Plan (DMP).

Directly contacting the site (Option B) bypasses the CRO’s chain of command and violates communication protocols. Notifying only the Clinical Leader (Option A) is insufficient, and ignoring the issue (Option D) jeopardizes theData Safety Monitoring Board (DSMB)review timeline.

Therefore,Option Censures a documented, collaborative approach to problem resolution within the contractual oversight structure.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Project Management and Communication, Section 7.1 – Oversight of CRO Data Management Activities

ICH E6 (R2) GCP, Section 5.2 – Contract Research Organization Responsibilities

FDA Guidance for Industry: Oversight of Clinical Investigations – Sponsor and CRO Roles and Communication Pathways

TheInformed Consent Form (ICF)is the document that explicitly describes what study subjects can expect regardingdata disclosure, privacy, and confidentialityduring and after participation in a clinical trial. According toICH E6 (R2) Good Clinical PracticeandFDA Human Subject Protection Regulations (21 CFR Parts 50 and 56), participants must be fully informed about how their personal and clinical data will be collected, used, stored, and shared — both during the study and in any subsequent data-sharing or publication activities.

TheGCDMPreiterates that clinical data managers must ensure that all data handling practices align with the privacy commitments made in the ICF. This includes compliance withdata protection regulationssuch as HIPAA (in the U.S.) and GDPR (in the EU). The ICF defines the permissible scope of data use, ensuring ethical management and subject protection.

Documents like theprotocolordata sharing planmay outline procedures and responsibilities but do not directly inform participants of their rights and data use expectations. Only theICFis designed for that ethical communication purpose.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Ethics, Privacy, and Data Security

ICH E6 (R2) Good Clinical Practice, Sections 4.8.10 & 4.8.12

FDA 21 CFR Part 50 – Protection of Human Subjects, Informed Consent Requirements

InElectronic Data Capture (EDC)systems, free text fields should be usedonly when a predefined list of acceptable responses cannot accommodate the full variability of input data— most notably forAdverse Event (AE) verbatim terms.

According to theGood Clinical Data Management Practices (GCDMP, Chapter: CRF Design and Data Collection), AE verbatim terms are initially entered asfree textby site staff to accurately capture the investigator’s exact medical description of the event. These verbatim terms are latercodedusing standardized dictionaries such asMedDRAduring medical coding, ensuring both flexibility and standardization in reporting.

Conversely, fields such asurine sedimentation rate (A),date of birth (C), andBody Mass Index (D)requirestructured numeric or date formatsto enable validation, range checks, and consistency across datasets. Free text would compromise data integrity, accuracy, and validation efficiency for these structured data elements.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: CRF Design and Data Collection, Section 4.3 – Use of Free Text and Coded Fields

ICH E6 (R2) Good Clinical Practice, Section 5.5.3 – Data Structure and Validation

MedDRA Introductory Guide, Section 2.3 – Verbatim Entry and Coding Requirements

In studies utilizingElectronic Data Capture (EDC)systems, many traditional paper-based processes such astracking and retrieving CRFsare eliminated or automated. However,query management and resolutionremain essential because discrepancies, missing data, and protocol deviations still require clarification and correction, regardless of the data collection medium.

According to theGCDMP (Chapter: Data Validation and Cleaning),data queriesare generated automatically or manually when inconsistencies are detected by edit checks. Sites must still respond to these queries electronically to ensure the integrity and completeness of data.

AandDare obsolete with EDC (no physical CRFs).

Brefers to manual data entry updates, which are replaced by direct EDC entry.

C (Resolving queries)continues as a key part of the data management workflow, even in fully electronic environments.

Thus,option Cis correct.

Reference (CCDM-Verified Sources):

SCDM GCDMP, Chapter: Data Validation and Cleaning, Section 5.4 – Query Generation and Resolution in EDC Systems

ICH E6(R2) GCP, Section 5.5.3 – Data Review and Query Resolution Requirements

FDA 21 CFR Part 11 – Electronic Records: Audit Trails and Query Documentation

C

When designing or configuringdata entry screenswithin an Electronic Data Capture (EDC) system, three critical components are required for each field:

Data Type– Defines the nature of the data (e.g., text, numeric, date).

Prompt– The label or question displayed to the user.

Response Format– Specifies how the user enters or selects data (e.g., free text, drop-down, checkbox).

According to theGCDMP (Chapter: EDC Systems and Database Design), these three attributes form thelogical data structurerequired to build and validate data entry interfaces. They ensure consistency in how information is captured, displayed, and validated during data entry.

Whileuser roles (A)andhelp text (D)are system-level configurations, not field-level specifications,page numbers (C)relate to printed CRFs rather than digital data screens.

Therefore,option B (Data type, prompt, and response format)correctly identifies the essential information needed to define data entry screens.

Reference (CCDM-Verified Sources):

SCDM GCDMP, Chapter: EDC Systems and Database Design, Section 4.3 – Screen Design Specifications

CDISC CDASH Implementation Guide, Section 3.2 – Data Field Attributes

ICH E6(R2) GCP, Section 5.5.3 – Data Capture and Input Standards

Inpaper-based clinical studies,front-end data checks(those performed during data entry) are intentionally kept minimal to ensure thatdata are entered exactly as recorded on the paper CRF. This principle ensuresdata integrityby maintaining fidelity between source and electronic records before any cleaning or edit validation occurs.

TheGCDMP (Chapter: Data Validation and Cleaning)explains that data entry operators should input values as written, even if they appear incorrect or inconsistent, because the purpose of front-end checks is not to interpret but to capture data faithfully. Theback-end edit checks—performed later by data managers—are designed to identify inconsistencies, out-of-range values, or logical errors that require clarification through queries.

This approach separatesdata capturefromdata cleaning, minimizing bias and preserving original investigator input. Hence,option Aaccurately states the rationale for keeping front-end checks minimal in paper-based studies.

Reference (CCDM-Verified Sources):

SCDM GCDMP, Chapter: Data Validation and Cleaning, Section 4.2 – Data Entry, Edit Checks, and Query Process

ICH E6(R2) GCP, Section 5.5.3 – Data Handling and System Controls

FDA Guidance for Industry: Computerized Systems Used in Clinical Investigations, Section 6.1 – Data Entry and Verification Processes

In clinical data management, aderived fieldrefers to any variable that is not directly collected from the Case Report Form (CRF) but is instead calculated or inferred from one or more collected variables (for example, calculating an average blood pressure from multiple readings). Proper documentation of derived fields is essential for ensuringdata traceability, transparency, and compliancewith both FDA and SCDM guidelines.

According to theGood Clinical Data Management Practices (GCDMP, May 2007), all derivations and transformations applied to clinical data must beclearly defined and documentedin metadata such as thedataset definition file (also referred to as data specifications, variable definition tables, or Define.xml files). The derivation algorithm should be explicitly stated in this documentation to allow independent verification, regulatory review, and reproducibility of results.

TheFDA Guidance for Industry (April 2006)on electronic submissions further emphasizes thatderived fields must be supported by comprehensive metadatathat defines the computational method used. This documentation enables the FDA or any regulatory body to audit andreproduce analytical results without ambiguity. Annotating or describing derivations directly on the CRF (as in options A, B, or D) isnot sufficient, as CRFs represent data collection instruments—not analytical documentation.

Therefore, the correct and regulatory-compliant practice isto provide the derivation algorithm for a calculated field within the dataset definition file, aligning with bothFDAandGCDMPexpectations for data integrity and auditability.

Reference (CCDM-Verified Sources):

Society for Clinical Data Management (SCDM), Good Clinical Data Management Practices (GCDMP), Chapter: Data Handling and Processing – Derived and Calculated Data Fields, Section 5.3.3

FDA Guidance for Industry: Providing Regulatory Submissions in Electronic Format, April 2006, Section 3.2 on Dataset Documentation Requirements

CDISC Define.xml Implementation Guide – Metadata and Algorithm Documentation for Derived Variables

In a discrepancy management workflow,“Received from site and not yet reviewed”isnot a resolution type— it represents astatus, not a final resolution outcome.

According to theGCDMP (Chapter: Data Validation and Cleaning), resolution types describe how a data discrepancy was finalized or addressed, such as:

Resolved with data correction,

Confirmed as correct (no change required),

Self-evident correction applied by data management, or

Unresolvable discrepancies documented.

In contrast,statusesdescribe the stage of the query (e.g., open, sent, answered, pending review, closed). “Received from site and not yet reviewed” indicates anintermediate workflow statewhere the response awaits validation by data management.

Proper classification of resolution types is essential for performance reporting, audit readiness, and ensuring the traceability of query management actions underICH E6 (R2)andFDA 21 CFR Part 11.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Data Validation and Cleaning, Section 5.3 – Discrepancy Resolution Lifecycle

ICH E6 (R2) Good Clinical Practice, Section 5.5.3 – Data Handling and Record Management

FDA 21 CFR Part 11 – Electronic Records; Audit Trails and Discrepancy Tracking Requirements

AData Transfer Agreement (DTA)defines the operational and technical details for transferring data between a sponsor and an external vendor (e.g., central lab, ECG vendor). It is a formalized, controlled document specifyingwhat data will be sent, when transfers will occur, the transfer method, file structure, encryption or security protocols, and the recipients of the data.

The DTA is developed jointly by the sponsor and vendor before production data transfers begin. According to theGCDMP, Chapter on External Data Transfers, this agreement ensures both parties share a clear understanding of timing, responsibility, and data content to minimize errors and ensure regulatory compliance.

TheData Management Plan (DMP)outlines general data handling processes but does not capture the technical specifics of vendor data transfer logistics. TheProject Plan (A)andCommunication Plan (B)are broader operational tools and not specific to data transfer protocols.

Hence,option C (Data Transfer Agreement)is the correct answer, as it precisely governs the procedural and technical framework of vendor data exchange.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: External Data Transfers, Section 4.1 – Data Transfer Agreements and Specifications

ICH E6(R2) Good Clinical Practice, Section 5.5 – Trial Management, Data Handling, and Record Keeping

TheScope of Work (SOW)is a contractual document that outlines thespecific deliverables, responsibilities, timelines, and budgetary detailsfor a given project between the sponsor and the contract research organization (CRO).

According to theGood Clinical Data Management Practices (GCDMP, Chapter: Project Management and Communication), the SOW defineswhat work will be performed,how many resources are allocated, andthe expected deliverables. This includes detailed information such as:

The number of database builds or migrations,

Timelines for deliverables (e.g., database lock),

Responsibility distribution between sponsor and CRO, and

Budget parameters for defined activities.

Therefore, if a Data Manager needs to determinehow many database migrations are budgeted for a project, theSOWis the correct document to reference.

Information such as PK sample scheduling (option A), site monitoring dates (option B), or staff contact details (option D) would be found in operational plans or contact lists, not in the SOW.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Project Management and Communication, Section 6.2 – Scope of Work Definition and Deliverables

ICH E6 (R2) GCP, Section 5.5.3 – Documentation and Responsibilities for Data Management Tasks

FDA Guidance for Industry: Oversight of Clinical Investigations – Sponsor and CRO Agreements

In the context of aninvestigator-initiated trial (IIT)whereStandard Operating Procedures (SOPs)are not available, the most appropriate and compliant approach is todevelop a Data Management Plan (DMP) templateand then create astudy-specific DMPbased on that template (Option C).

According to theGood Clinical Data Management Practices (GCDMP, Chapter on Data Management Planning and Study Start-up), theDMPis the central document that defines all processes, responsibilities, systems, and quality controls related to data collection, processing, validation, and database management throughout the clinical study. The DMP serves as a formal framework for ensuringdata integrity,traceability, andregulatory compliance, especially in the absence of established institutional SOPs.

While SOPs provide organizational-level standards, the DMP providesstudy-specific operational detail. In an investigator-initiated setting, researchers often lack institutional data management infrastructure, so the DMP must substitute for SOP guidance by detailing:

Data entry and validation procedures

Query management and resolution processes

CRF design and data flow specifications

Database design, backup, and security

Responsibilities of study personnel (investigator, data manager, statistician)

Quality control and audit trail practices

OptionA(“Data handling should be documented in a DMP”) is correct in principle but incomplete—without aDMP template, there is no standardized format or consistency across studies.

OptionB(developing full SOPs) is not practical for a single IIT; SOPs are organizational-level documents requiring longer development and approval cycles.

OptionD(briefly describing data management in the protocol) is insufficient, as the protocol should reference data management activities but not serve as the operational manual for them.

Therefore,Option Cprovides the most comprehensive, regulatory-compliant, and practical solution—ensuring structured documentation of all data management activities while maintaining flexibility for investigator-led research.

Reference (CCDM-Verified Sources):

Society for Clinical Data Management (SCDM), Good Clinical Data Management Practices (GCDMP), Chapter: Data Management Planning and Study Start-up, Section 5.2 – Data Management Plan (DMP) Development and Maintenance

ICH E6 (R2) Good Clinical Practice, Section 5.1 – Quality Management and Documentation Requirements

FDA Guidance for Industry: Computerized Systems Used in Clinical Investigations, Section 4 – Data Management and Documentation Practices

SCDM GCDMP, Chapter: Project Management in Data Management – Study-Specific Documentation and Planning in Investigator-Initiated Trials

When an investigator retires mid-study and the replacement refuses to use theElectronic Data Capture (EDC)system, thedata managermust not take unilateral action but rathercollaborate with the study teamto explore acceptable solutions.

Per theGCDMP (Chapter: Project Management in Data Management), any deviation from the established data capture method — particularly a change that affects regulatory compliance, data consistency, or site operations — requires a cross-functional assessment. The study team, which includes clinical operations, project management, regulatory affairs, and data management, should evaluate feasible alternatives such as:

Allowing paper CRF entry followed by centralized data transcription,

Retraining site staff on EDC use, or

Temporarily suspending data entry until compliance can be restored.

Immediate site closure (option A) or unilateral decisions by data management (options C and D) violate escalation and communication protocols. Collaborative decision-making ensures continuity, compliance, and data integrity, in line withICH E6 (R2) GCPandFDA 21 CFR Part 11.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Project Management and Communication, Section 5.2 – Handling Site and Investigator Changes

ICH E6 (R2) Good Clinical Practice, Section 4.1 – Investigator Responsibilities

FDA Guidance for Industry: Computerized Systems Used in Clinical Investigations – Section on EDC Operations and Site Management

TheIndividual Case Safety Report (ICSR)is the standard format used globally for the exchange ofSerious Adverse Event (SAE)data between clinical data management systems (EDC) and safety management systems.

According toICH E2B(R3)andGood Clinical Data Management Practices (GCDMP, Chapter: Safety Data Management and SAE Reconciliation), the ICSR provides thedata structure and content standardsfor electronic transmission of safety data, including patient demographics, event details, outcomes, and product information. It ensures interoperability between systems by defining standardized message elements and controlled terminologies.

Other options are not applicable:

A. Medical Document for Regulatory Activities (MDRA)is not a recognized standard.

B. Biomedical Research Domain Model (BRIDG)provides conceptual modeling but not data exchange guidance.

D. SDTMis used for regulatory submission datasets, not real-time SAE exchange.

Thus,option C (Individual Case Safety Report)is correct, as it defines the internationally accepted electronic format for SAE data exchange between safety and clinical databases.

Reference (CCDM-Verified Sources):

SCDM GCDMP, Chapter: Safety Data Management and SAE Reconciliation, Section 4.3 – SAE Data Exchange and Standards

ICH E2B(R3): Electronic Transmission of Individual Case Safety Reports

FDA Guidance for Industry: Providing Regulatory Submissions in Electronic Format — Postmarketing ICSRs (2014)

Inpaper-based trials, site staff (e.g., study coordinators) record data manually on paper Case Report Forms (CRFs), which are later transcribed by data entry personnel into an electronic database.

However, inEDC-based studies, thesite coordinatoris directly responsible forentering data into the EDC system. This eliminates the need for centralized double data entry and shortens data cleaning timelines.

TheGCDMP (Chapter: Electronic Data Capture Systems)states that EDC systems shift certain tasks, includingdata entry, initial query response, and source verification preparation, to the site level. Yet,data entryremains the most significant additional responsibility compared to paper-based studies.

Option A (Query resolution)is performed in both EDC and paper-based systems.

Option C (Data curation)is typically a Data Management function.

Option D (Medical record abstraction)is part of source documentation, not specific to EDC.

Thus,option B (Data entry)is correct — it is the additional site coordinator duty unique to EDC environments.

Reference (CCDM-Verified Sources):

SCDM GCDMP, Chapter: Electronic Data Capture (EDC) Systems, Section 5.3 – Site Responsibilities and Workflow Changes

ICH E6(R2) GCP, Section 5.5.3 – Data Entry and Role Delegation in Computerized Systems

FDA Guidance for Industry: Computerized Systems Used in Clinical Investigations, Section 6.2 – Site-Level Data Entry Controls

When theprimary efficacy endpointin a clinical trial is3-month survival, the key data element required is thedeath date. This is because the survival endpoint is determined by calculating whether the subject lived or died within a defined time frame from study enrollment or randomization.

According to theGCDMP (Chapter: Data Management Planning and Study Start-up), the Clinical Data Manager (CDM) must identify and ensure the capture of allcritical data elementsnecessary to evaluate the study endpoints. For time-to-event analyses (e.g., survival studies), accurateevent dates (death date)are essential for endpoint derivation and statistical analysis.

Other data elements such as cause of death or date of autopsy (options B and C) may support secondary analyses or safety reviews but are not necessary to determine the survival endpoint itself. Similarly,birth date(option D) contributes to demographic data but is unrelated to the primary efficacy outcome.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Data Management Planning and Study Start-up, Section 4.4 – Critical Data Identification for Endpoints

ICH E9 – Statistical Principles for Clinical Trials, Section 2.2.3 – Time-to-Event Data Considerations

FDA Guidance for Industry: Clinical Trial Endpoints for Drug Development

In international or multicenter clinical studies,laboratory dataoften originate from different laboratories that use varying measurement units (e.g., mg/dL vs. mmol/L). TheGood Clinical Data Management Practices (GCDMP, Chapter on CRF Design and Data Collection)provides clear guidance on managing this variability to ensuredata consistency,traceability, andminimized transcription errors.

The approach that results infewer transcription errorsis toallow sites to enter lab values exactly as recorded in the source document (original lab report)and to requireexplicit selection of the corresponding unitfrom a predefined list on the data collection form or within the electronic data capture (EDC) system. This method (Option B) preserves the original source data integrity while enabling centralized or automated unit conversion later during data cleaning or statistical processing.

Option B also supports compliance withICH E6 (R2) Good Clinical Practice (GCP), which mandates that transcribed data must remain consistent with the source documents. Attempting to derive units automatically (Option A) can lead to logical errors, while forcing sites to manually convert units (Option D) introduces unnecessary complexity and increases the risk of miscalculation or inconsistent conversions. Printing only standard units on the CRF (Option C) ignores local lab practices and can lead to discrepancies between CRF entries and source records, triggering numerous data queries.

TheGCDMPemphasizes that CRF design must account for local variations in measurement systems and ensure thatunit selection is structured (dropdowns, controlled lists)rather than free-text to prevent typographical errors and facilitate standardization during data transformation.

Therefore, OptionB—“Allow values to be entered as they are in the source and the selection of units on the data collection form”—is the most compliant, accurate, and efficient strategy for minimizing transcription errors in international lab data collection.

Reference (CCDM-Verified Sources):

Society for Clinical Data Management (SCDM), Good Clinical Data Management Practices (GCDMP), Chapter: CRF Design and Data Collection, Section 5.4 – Laboratory Data Management and Unit Handling

ICH E6 (R2) Good Clinical Practice, Section 5.18 – Data Handling and Record Retention

CDISC SDTM Implementation Guide, Section 6.3 – Handling of Laboratory Data and Standardized Units

FDA Guidance for Industry: Computerized Systems Used in Clinical Investigations, Section 6 – Source Data and Accuracy of Data Entry

When integratingexternal environmental datasuch asair quality readingswith clinical study data, it is essential to uselocation and time identifiersto properly align the environmental data with subject-level data.

According to theGood Clinical Data Management Practices (GCDMP, Chapter: Data Management Planning and Study Start-up), external data sources (like national weather or pollution databases) must be merged usingcommon linkage variablesthat allow synchronization without breaching subject confidentiality. In this case:

Location identifiers(e.g., city, postal code, or region) align the subject’s study site or residential area with the environmental dataset.

Time identifiers(e.g., date and time of data collection) ensure that the environmental readings correspond to the same period as the subject’s clinical observations.

Including subject identifiers (option C or D) is unnecessary and would poseprivacy and data protection risks. Instead, linkage is typically done at theaggregate (site or regional) level, maintaining compliance withHIPAAandGDPR.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Data Integration and External Data Handling, Section 4.3 – Linking External Data Sources

ICH E6 (R2) GCP, Section 5.5.3 – Data Traceability and External Data Management

FDA Guidance for Industry: Use of Electronic Health Record Data in Clinical Investigations, Section 5.2 – Linking and Integration Principles

When a system validation test fails duringElectronic Patient-Reported Outcome (ePRO)system testing, thevalidation documentationmust record theexpected results(what should have occurred) and theactual results(what occurred).

According to theGCDMP (Chapter: Database Validation and Testing), proper system validation documentation ensurestraceability, reproducibility, and compliancewithFDA 21 CFR Part 11andICH E6 (R2). Each test case must include:

Test objective,

Preconditions,

Test steps,

Expected results,

Actual results, and

Pass/fail status.

If a test fails, this documentation provides the objective evidence necessary for deviation handling, issue resolution, and re-testing. While a separateroot cause analysismay be performed later (option D), the validation record itself must focus on verifying outcomes against predefined expectations.

Therefore, the correct answer isB – Expected and actual results.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Database Validation and Testing, Section 4.4 – Documentation of Test Results

FDA 21 CFR Part 11 – Validation Requirements (Section 11.10(a))

ICH E6 (R2) GCP, Section 5.5.3 – Computer System Validation and Documentation

Before implementing anEDC system upgrade, thefirst taskof the Data Manager is toassess the impact on the data.

According to theGCDMP (Chapter: Electronic Data Capture Systems)andFDA 21 CFR Part 11, any system upgrade must undergoimpact assessmentto determine how the change might affectdata integrity, functionality, validation, and ongoing study operations. This assessment ensures that no data are lost, corrupted, or rendered inconsistent during or after the upgrade.

The Data Manager should evaluate:

Potential effects on existing data, edit checks, and reports,

System functionality impacting current workflows, and

Any revalidation requirements.

Only after the impact is understood should the Data Manager proceed to communicate with sites (option A), update documentation (option B), or modify CRFs if required (option D).

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Electronic Data Capture Systems, Section 7.3 – System Upgrades and Change Control

FDA 21 CFR Part 11 – Change Control and Validation Requirements

ICH E6 (R2) Good Clinical Practice, Section 5.5.3 – Change Impact on Data Integrity and System Validation

Enrolling subjects before theElectronic Data Capture (EDC) systemis ready poses majordata integrity and compliance risks. The primary issue is thatdata cannot be accurately captured, validated, or monitoredwithout the system in place.

Per theGCDMP (Chapter: Data Management Planning and Study Start-up), data collection systems must befully validated, tested, and releasedbefore enrollment begins to ensure:

Real-time data entry and quality control

Proper tracking of adverse events (AEs/SAEs)

Audit trails and traceability for regulatory compliance

Option A highlights the most critical consequence — without an operational EDC, data collection and verification processes cannot occur, compromising data quality and study oversight.

While options B, C, and D may be partially true, they aresecondary effects. The fundamental consideration is data capture capability and monitoring control, makingoption Acorrect.

Reference (CCDM-Verified Sources):

SCDM GCDMP, Chapter: Data Management Planning and Study Start-up, Section 4.2 – EDC Readiness and System Validation

ICH E6(R2) GCP, Section 5.5.3 – Computerized Systems Validation Before Use

FDA Guidance for Industry: Computerized Systems Used in Clinical Investigations, Section 6.1 – System Qualification Prior to Data Entry

The most critical consideration is thatdata cleaning is an iterative process, and completing all necessary steps — includingquery generation, site resolution, and second-pass validation— cannot realistically be accomplished within two weeks after study close.

According to theGood Clinical Data Management Practices (GCDMP, Chapter: Data Validation and Cleaning), data cleaning must occurcontinuously throughout the study, not only at the end. Post-database lock activities typically include running final validation checks, resolving outstanding queries, performing reconciliation (e.g., SAEs, labs, coding), and conducting final quality review.

Even in small studies,query turnaround and response cyclesfrom sites take time — typically2–4 weeks per iteration— making a two-week total cleaning period unrealistic.

Therefore,Option Dis correct: it would take more than two weeks to handle second-round (follow-up) queries and confirm final resolutions prior to database lock.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Data Validation and Cleaning, Section 5.4 – Ongoing vs. End-of-Study Data Cleaning

ICH E6 (R2) Good Clinical Practice, Section 5.5.3 – Data Quality and Timeliness

FDA Guidance for Industry: Computerized Systems Used in Clinical Investigations – Data Management and Cleaning

Themost common mode of data entryinElectronic Data Capture (EDC)systems issingle data entry.

According to theGCDMP (Chapter: Electronic Data Capture Systems), EDC systems have built-inedit checks, validation rules, and audit trailsthat ensure data accuracy and integrity at the point of entry. These real-time validation capabilities makedouble data entry(a legacy practice from paper studies) unnecessary.

EDC systems automatically verify data as they are entered by site staff, generating queries for inconsistencies or out-of-range values immediately.Blind verification (option B)andthird-party comparisons (option D)are not standard data entry modes but may be used for specialized reconciliation or external data imports.

Thus,single data entry (Option C)is the industry standard approach, ensuring both efficiency and compliance withFDA 21 CFR Part 11andICH E6 (R2)data integrity requirements.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Electronic Data Capture (EDC) Systems, Section 5.4 – Data Entry and Verification Processes

ICH E6 (R2) Good Clinical Practice, Section 5.5.3 – Computerized Systems and Data Validation

FDA 21 CFR Part 11 – Electronic Records and Electronic Signatures: Validation and Data Entry Requirements

When using acopyrighted or validated rating scale(e.g., Hamilton Depression Scale, Visual Analog Pain Scale), anymodification to the original instrument, including preprinting database codes on the CRF, must beapproved by the instrument’s owner or licensing authorityto ensure thevalidity and reliabilityof the instrument are not compromised.

According to theGCDMP (Chapter: CRF Design and Data Collection), validated rating scales are psychometrically tested tools. Any visual or structural modification (such as adding codes, changing layout, or rewording questions) can invalidate prior validation results. Therefore, the CRF designer mustconsult the independent source (copyright holder)for approval anddocument that the validity of the tool remains intact.

Merely consulting statisticians (option B) or verifying database alignment (option D) does not ensure compliance. Thus,Option Censures scientific and regulatory integrity.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: CRF Design and Data Collection, Section 6.1 – Use of Validated Instruments and Rating Scales

ICH E6 (R2) GCP, Section 5.5.3 – Validation of Instruments and Data Capture Tools

FDA Guidance for Industry: Patient-Reported Outcome Measures – Use in Medical Product Development to Support Labeling Claims, Section 4 – Instrument Modification and Validation

In initial vendor discussions forexternal data integration(e.g., central lab, ECG, imaging vendors), themost critical and foundational topicis defining theacceptable record, field, and file formats.

According to theGCDMP (Chapter: External Data Transfers and Integration), establishing theData Transfer Specifications (DTS)early in the process ensures consistent structure, proper mapping, and compatibility between the vendor’s system and the sponsor’s database. These specifications define:

Data structure (variable names, formats, delimiters)

File naming conventions

Frequency of transfers

Methods of secure data transmission

Discussing formats first allows later alignment on data validation, quality metrics, and dictionary standards (which occur in subsequent stages). Without format agreement, all downstream processes risk misalignment, resulting in data incompatibility and rework.

Thus,option C (Acceptable record, field, and file formats)correctly represents the foundational focus of initial vendor discussions for ensuring data quality and integration efficiency.

Reference (CCDM-Verified Sources):

SCDM GCDMP, Chapter: External Data Transfers and Integration, Section 4.1 – Data Transfer Planning and Specification Development

ICH E6(R2) GCP, Section 5.5.3 – Data Handling and System Validation

FDA Guidance: Computerized Systems Used in Clinical Investigations, Section 6.3 – Data Import and Format Control

TheInvestigatorholds theprimary responsibilityfor ensuring the accuracy, completeness, and timeliness of Case Report Form (CRF) entries. This responsibility is mandated by regulatory requirements underICH E6(R2) Good Clinical Practice (GCP).

The investigator may delegate CRF completion to aqualified designee (e.g., site coordinator), but the ultimate accountability remains with the investigator. The investigator’s signature (electronic or manual) on the CRF serves as certification that the data accurately reflect the source documents and the patient’s participation.

TheGCDMP (Chapter: CRF Design and Data Collection)reinforces this by stating that while data managers ensure design quality and CRAs verify consistency with source data,the investigator is legally responsible for CRF accuracy.

Thus,option D (Investigator)is correct, as it aligns with both GCP and CCDM standards.

Reference (CCDM-Verified Sources):

ICH E6(R2) GCP, Section 4.9 – Records and Reports (Investigator Responsibilities)

SCDM GCDMP, Chapter: CRF Design and Data Collection, Section 5.1 – Investigator’s Role in Data Accuracy

FDA 21 CFR Part 312.62 – Investigator Recordkeeping and Record Retention

Astandardized data collection elementhas well-defined metadata, consistent naming conventions, and controlled terminology to ensure uniform data collection and interoperability across studies.

Key attributes, as perGCDMPandCDISC standards, include:

Aclear definitionof meaning (A)

Acontrolled set of response values(C)

Metadata specificationslike variable names, formats, and data types (D)

However, thephysical positioningof a data element on a case report form (B) is a matter ofform layout design, not a characteristic of data standardization. While consistent form structure aids usability, it is not part of data standardization or metadata management principles.

Hence,option Bis correct —form positioning is not a standardized data element attribute.

Reference (CCDM-Verified Sources):

SCDM GCDMP, Chapter: Standards and Data Integration, Section 4.1 – Data Element Standardization

CDISC CDASH Implementation Guide, Section 3.2 – Standardized Data Collection Elements and Metadata

ICH E6(R2) GCP, Section 5.5.3 – Data Handling and Standardization

According toICH E6(R2) Good Clinical Practice (GCP), Section 5.18.1, theSponsoris ultimatelyresponsible for developing and implementing the Monitoring Plan.

The Monitoring Plan defines:

Theextent and nature of monitoring(e.g., on-site, remote, risk-based).

Theresponsibilities of monitors.

Thecommunication and escalation proceduresfor data quality and protocol compliance.

While theCRO (B)orMonitor (D)may perform monitoring activities under delegation, theSponsorretains legal accountability for ensuring a compliant and effective plan is developed and maintained. TheData Manager (C)may contribute by outlining data review workflows, but is not responsible for authoring or owning the plan.

Therefore,option A (Sponsor)is the correct answer.

Reference (CCDM-Verified Sources):

ICH E6(R2) GCP, Section 5.18.1 – Purpose and Responsibilities for Monitoring

SCDM GCDMP, Chapter: Regulatory Compliance and Oversight, Section 5.3 – Sponsor Responsibilities in Monitoring and Quality Assurance

FDA Guidance for Industry: Oversight of Clinical Investigations – Sponsor Responsibilities (2013)

Adding multiple new sites late in the enrollment period creates aconcentrated influx of new datanear the end of the study. These sites typically start enrolling patients later, resulting in a“bolus” of Case Report Forms (CRFs)that must be entered, validated, and cleaned within a shorter timeframe to meet database lock deadlines.

According to theGood Clinical Data Management Practices (GCDMP, Chapter: Project Management and Data Flow), late site activation compresses the timeline for data management tasks, necessitating increased resources fordata entry, query management, and cleaning. Data management teams must anticipate this surge and plan accordingly—either by increasing staffing or revising timelines to prevent bottlenecks and maintain quality.

Whileoption D(increased query rates) can occur, it is a secondary effect. Themost direct and consistent impactis the surge in data volume requiring expedited processing near study end.

Reference (CCDM-Verified Sources):

SCDM GCDMP, Chapter: Project Management, Section 5.3 – Managing Changes in Site Activation and Data Flow

ICH E6(R2) GCP, Section 5.1 – Quality Management and Oversight

If a clinical site continues to provideinconsistent or illogical dataafter multiple queries, the correct course of action is toescalate the issue to the appropriate site contact personnel, typically theClinical Research Associate (CRA)orSite Monitor.

According to theGood Clinical Data Management Practices (GCDMP), persistent data discrepancies often indicate a misunderstanding of the protocol, CRF instructions, or data entry procedures at the site level. Repeatedly re-querying the same data without escalation wastes time and risks introducing bias or error. By escalating through formal communication channels, the issue can be clarified through re-training, documentation review, or site monitoring visits.

The GCDMP emphasizes that escalation ensuresdata accuracy, site accountability, and protocol adherence, maintaining both data quality and regulatory compliance. Data managers must document the escalation process in the Data Management Plan (DMP) and ensure proper follow-up resolution is achieved.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Communication and Issue Escalation, Section 4.2 – Handling Persistent Data Discrepancies

ICH E6 (R2) Good Clinical Practice, Section 5.18 – Monitoring and Site Communication

FDA Guidance for Industry: Oversight of Clinical Investigations – Risk-Based Monitoring, Section on Issue Escalation

When a clinical study usesblood pressure (BP)as anefficacy endpoint, the most reliable and standardized method of data collection is throughstudy-provisioned equipment.

According to theGCDMP (Chapter: CRF Design and Data Collection), data collected for primary efficacy endpoints must beconsistent, accurate, and standardized across all investigative sites. Usingstudy-provided calibrated equipmentensures that measurements are taken under uniform conditions, eliminating inter-site variability due to differences in devices, calibration, or measurement methods.

Collecting BP data from medical records (option A) risks inconsistent timing and techniques. Using each site’s own equipment (option B) introduces variability, while patient self-reports (option D) lack reliability and objectivity.

Thus, thebest practiceis to provision and standardize all equipment used to collect endpoint-related physiological data, ensuring regulatory-quality results suitable for analysis.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: CRF Design and Data Collection, Section 5.1 – Standardization of Clinical Measurements

ICH E6 (R2) GCP, Section 5.5.3 – Data Accuracy and Equipment Standardization

FDA Guidance for Industry: Electronic Source Data in Clinical Investigations, Section 4.3 – Data Capture and Standardization Requirements

Once a clinical database islocked, it becomesread-only— no further data modifications can be made by any users, including site personnel. This ensures that the data arefinalized, consistent, and auditablefor statistical analysis and regulatory submission.

According to theGCDMP (Chapter: Database Lock and Archiving), the lock process involves freezing the database to prevent accidental or unauthorized changes. After lock, access permissions are restricted, and all edit and update functions are disabled. If any corrections are required post-lock, the database must beunlocked under controlled procedures(with full audit trail documentation).

Thus,option C—The site study coordinator is not able to make the change— correctly reflects standard EDC functionality and regulatory compliance.

Reference (CCDM-Verified Sources):

SCDM GCDMP, Chapter: Database Lock and Archiving, Section 5.2 – Database Lock Procedures and Controls

ICH E6(R2) GCP, Section 5.5.3 – Data Integrity and Audit Trail Requirements

FDA 21 CFR Part 11 – Controls for Electronic Records and System Lock Functions

TheScope of Work (SOW)is a project management document that defineswhat services are includedin the work agreement between the sponsor and the CRO or vendor. It outlines deliverables, responsibilities, timelines, and budget allocations.

According to theGCDMP (Chapter: Project Management in Data Management), the SOW includes specifications such as:

The number and frequency ofdata transfers,

Database build and lockmilestones,

Quality control deliverables, and

Resource allocationfor data management tasks.

The SOW does not cover operational site-level details such as monitoring schedules (B), protocol sampling details (C), or personnel contact lists (D).

Therefore,option A (number of data transfers budgeted for a project)correctly identifies a use case directly addressed in the SOW.

Reference (CCDM-Verified Sources):

SCDM GCDMP, Chapter: Project Management, Section 4.1 – Scope of Work and Resource Planning

ICH E6(R2) GCP, Section 5.5 – Sponsor Oversight and Data Management Responsibilities

PMI Project Management Framework – Scope Definition and Deliverable Specifications

Although clerical staff can technically perform data entry,data entry in clinical research requires study-specific training, particularly in the use of theElectronic Data Capture (EDC) systemand understandingdata discrepancy resolutionprocedures.

According to theGood Clinical Data Management Practices (GCDMP, Chapter: CRF Design and Data Collection)andICH E6 (R2), individuals responsible for data entry at clinical sites must bequalified by education, training, and experience. This includes understanding how to navigate the EDC system, enter data according to CRF Completion Guidelines, and appropriately respond to queries or system-generated edit checks.

Untrained clerical personnel may inadvertently introduce errors, violate Good Clinical Practice (GCP) standards, or fail to recognize protocol-relevant data. Therefore, theData Managermust ensure that site users receivestudy-specific and system trainingbefore gaining access to the EDC environment.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: CRF Design and Data Collection, Section 5.2 – Investigator Site Training and Data Entry Requirements

ICH E6 (R2) Good Clinical Practice, Section 4.1.5 – Qualified Personnel and Training Requirements

FDA 21 CFR Part 11 – User Access and Training Provisions for Electronic Records

RACIis aproject management and governance frameworkused to defineroles and responsibilitieswithin a project. Each letter represents a distinct role type:

Responsible (R):The person(s) who perform the work or execute the task.

Accountable (A):The individual ultimately answerable for the task’s completion and success (only one per activity).

Consulted (C):Subject matter experts who provide input or guidance before decisions are made.

Informed (I):Individuals kept up to date on progress or outcomes but not directly involved in execution.

The RACI model ensuresclarity in ownership and accountability, preventing duplication of effort or responsibility confusion. It is a key component of theGCDMP (Chapter: Project Management in Data Management)for ensuring clear delegation and communication within clinical data management teams.

Hence,option Dis correct.

Reference (CCDM-Verified Sources):

SCDM GCDMP, Chapter: Project Management in Data Management, Section 5.1 – Roles, Responsibilities, and RACI Matrices

Project Management Institute (PMI) Framework – Responsibility Assignment Matrices (RACI)

ICH E6(R2) GCP, Section 5.1.1 – Defined Roles and Quality Oversight Responsibilities

UnderFDA 21 CFR Part 11, organizations using electronic systems must ensure thatall system users are trainedto perform their assigned functions before gaining access to the system. The regulation requires documented evidence of training but does not specify how it should be conducted (e.g., exam-based, in person, or language-specific).

TheGCDMP (Chapter: Computerized Systems and Compliance)further clarifies that personnel training should include instruction onsystem functionality, audit trails, data entry procedures, and electronic signaturesto maintain compliance and data integrity. Training must beperformed and documentedbut does not require a specific format or delivery method.

Therefore,option A—Training must be performed—is correct, as it reflects theminimum regulatory expectationper FDA and SCDM standards.

Reference (CCDM-Verified Sources):

FDA 21 CFR Part 11, Section 11.10(i) – Personnel Training Requirements

SCDM GCDMP, Chapter: Computerized Systems and Compliance, Section 5.4 – System Training and Documentation

ICH E6(R2) GCP, Section 2.8 – Qualified Personnel and Training Requirements

TheStatisticianis responsible for assessing theoverall impact of data errors on the analysis and study results.

According to theGood Clinical Data Management Practices (GCDMP, Chapter: Data Quality Assurance and Control)andICH E9 (Statistical Principles for Clinical Trials), while theData Managerensures data accuracy and completeness through cleaning and validation, theStatisticiandetermines whether the observed data discrepancies are statistically significant or if they may affect thevalidity, power, or interpretabilityof the study’s outcomes.

TheQuality Auditor (C)identifies and reports issues but does not quantify analytical impact. TheInvestigator (D)is responsible for clinical oversight, not statistical assessment. Thus, after a database audit, theStatistician (B)performs a formal evaluation to determine whether the magnitude and nature of the errors could bias results or require reanalysis.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Data Quality Assurance and Control, Section 7.3 – Data Audit and Impact Assessment

ICH E9 – Statistical Principles for Clinical Trials, Section 3.2 – Data Quality and Analysis Impact Assessment

FDA Guidance for Industry: Computerized Systems Used in Clinical Investigations – Data Validation and Analysis Review

The best metric to identifymalfunctioning or ineffective edit checksis thecount by edit check of the number of times the check fired. This allows data managers to assess whether specific edit checks are performing as intended.

According to theGCDMP, Chapter: Data Validation and Cleaning, edit checks are programmed logic conditions that identify data inconsistencies or potential errors during data entry. A properly functioning edit check should trigger only when data falls outside acceptable or logical limits. If an edit check fires too frequently or not at all, it may indicate alogic errorin the check’s programming or configuration.

By analyzing counts by individual edit checks, data managers can:

Identify checks that never trigger (potentially inactive or incorrectly written),

Detect overactive checks (poorly designed parameters causing excessive false positives), and

Optimize system performance and review efficiency.

This metric supports continuous improvement in data validation logic and contributes to cleaner, higher-quality clinical databases.

Reference (CCDM-Verified Sources):

SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Data Validation and Cleaning, Section 6.2 – Edit Check Design and Performance Metrics

FDA Guidance: Computerized Systems Used in Clinical Investigations – Section on Validation of Electronic Data Systems