Allied Health & Imaging · Independent exam preparation
Certified Clinical Data Manager (CCDM®) Examination Study Guide
Practise defensible clinical data decisions: design clear forms, reconcile transfers, test study systems, write SQL checks, and verify lock and reporting readiness.
By CertMedbry · Updated 31 October 2026
Learn to distinguish missing data from inapplicable data, identify unsafe record matches, evaluate query evidence, test study workflows, and verify that reported results reflect controlled data handling. These 60 selected concepts connect study design with processing, quality control, database lock, and reporting. Use SCDM’s complete official task outline to plan full preparation; this guide develops selected decisions across all seven tested domains.
Try 20 free practice questionsWhat is the exam format?
Check the question types, timing and outline before planning your revision.
Which certification and pathway does this guide cover?
It covers the Society for Clinical Data Management’s Certified Clinical Data Manager (CCDM®) examination. CCDA and CCDS are separately named SCDM credentials; their pathways are not combined here. CCDM addresses professional clinical data management rather than a clinical practice license, nursing examination, or imaging modality. The supplied handbook describes the CCDM examination as originally released in September 2005, with subsequent revisions in 2008, 2010, and 2019; it does not identify a historical replacement examination name.
How many questions are there, and how long is the examination?
The July 2026 handbook specifies 150 multiple-choice questions, 3.5 hours, and English administration. Unscheduled breaks are permitted, but the examination clock continues. The supplied official evidence does not establish a scored-versus-pretest breakdown, the number of answer options on every examination item, or a numerical passing score.
How is the examination delivered, and is it adaptive?
The handbook describes live remote proctoring through Prolydian in partnership with Proctor360, including identity and room checks. It requires a compatible desktop or laptop, webcam, microphone, and a computer-readiness check. The supplied evidence does not establish adaptive delivery. The practice questions here are independent four-option, single-best-answer drills and do not reproduce the examination interface or predict a passing result.
Which outline and dates apply to this guide?
The scope follows SCDM’s currently linked CCDM Exam Study Guide V2, dated 22 July 2025, alongside the Certification Handbook updated 16 July 2026, as verified on 31 October 2026. The study guide supplies domains, tasks, exclusions, and preparation references. It does not state a separate blueprint effective date. The handbook revision date should not be treated as proof that every examination task changed on that date.
CCDM Certification · CCDM Exam Study Guide · CCDM® Certification Handbook
What is the full tested scope, and what is excluded?
The full scope is the official task list across Design Tasks, Training Tasks, Data Processing Tasks, Programming Tasks, Testing Tasks, Coordination and Project Management Tasks, and Review Tasks. Only Programming task P.5, ad hoc SQL queries, is tested within Programming. Personnel Management is excluded, together with D.14, D.18, DP.6, C.7, C.9, C.10, C.14, C.15, and R.3. Exclusions are task-specific: for example, DP.15 still includes curating and archiving or sharing data, although coordinating archival or sharing under C.9 is excluded.
Are official domain percentages available?
Neither supplied official document states domain percentage weights. This selective guide therefore uses an editorial allocation, not an estimated official blueprint: 16 Design, 5 Training, 16 Data Processing, 3 Programming, 6 Testing, 9 Coordination and Project Management, and 5 Review concepts. The questions and flashcards each allocate 5, 2, 5, 1, 2, 3, and 2 items to those domains respectively. The complete official outline remains the authority for full preparation.
Who is eligible, and do requirements vary by jurisdiction?
The handbook lists these eligibility routes: a bachelor’s degree or higher plus at least two years of full-time CDM experience; an associate’s degree plus at least three years; or at least four years of full-time CDM experience. Equivalent part-time experience may qualify. These routes lead to the same CCDM credential. The supplied evidence does not establish separate jurisdiction-specific CCDM examinations. Applicable study regulations and regional fees should be checked separately through SCDM.
How should the supporting references be used?
SCDM identifies GCDMP and JSCDM as foundational preparation resources, and its study guide is supplemental rather than a replacement for GCDMP review. Practice references support the data management principles taught here. Current ICH, FDA, CDISC, MedDRA, and SQL references clarify specific mechanisms where cited; their inclusion does not establish that SCDM formally revised its examination to every newer publication. FDA guidance addresses its regulatory context and should not be generalized into a universal legal requirement.
This guide addresses SCDM’s Certified Clinical Data Manager (CCDM®), a professional clinical data management credential. The workbook label “Clinical Research Data Manager” maps to that credential. These independently authored learning materials are not endorsed by SCDM.
Outline used: Verified on 2026-10-13 against SCDM's currently linked CCDM® Certification Handbook, updated 16 July 2026, and CCDM Exam Study Guide V2, dated 22 July 2025. The study guide supplies exam domains, tasks and exclusions but no percentage weights or separate blueprint effective date. Workbook label 'Clinical Research Data Manager (CCDM)' maps to SCDM's official Certified Clinical Data Manager examination; the assigned slug is preserved.
60 key concepts: understand, apply and avoid the traps
Work through the questions by exam domain. Explain the answer in your own words, try the example, then check the common pitfall.
Design Tasksnot stated
1.Which data belong on a study collection form?
Start with protocol objectives, safety requirements, eligibility criteria, and the information needed to interpret results. Each proposed field should have a defined purpose and an identified collection source. Map endpoints to their supporting measurements so that essential timing, units, and context are captured. A field that seems interesting but has no planned use adds burden and opportunities for error. Conversely, collecting an endpoint label without its necessary components can make the endpoint unusable.
Apply it
A team proposes collecting every historical diagnosis, although the protocol requires only specified conditions affecting eligibility and safety interpretation. Map each proposed field to a requirement, retain necessary conditions, and remove unsupported collection.
Common pitfall: Do not equate a larger database with a more useful study. Extra fields can increase missingness while diverting attention from essential endpoint and safety information.
2.What makes a study data element unambiguous?
A data element needs a clear meaning, data type, permissible values, precision, units where relevant, and rules for missing or inapplicable responses. Define the observation’s context, including whose information is recorded and when it applies. Distinguish collection conventions from later derivations. For example, an assessment date and a visit label describe different properties. Explicit definitions allow sites, database designers, reviewers, and analysts to interpret the same value consistently across systems.
Apply it
Two sites interpret “assessment time” as either appointment time or specimen collection time. Revise the definition to identify the required event, specify the time convention, and align the form and completion instructions.
Common pitfall: A familiar field label is not a complete definition. Unstated units, time conventions, or permissible values can produce apparently compatible records with different meanings.
3.How should form layout follow study procedures?
Organize forms around the workflow of the person recording data and the sequence of protocol procedures. Clear grouping, concise prompts, and appropriate conditional sections reduce navigation errors. Avoid collecting the same fact repeatedly when a controlled relationship can supply it. If repeated measurements are intentionally used to assess validity, they must be obtained independently. Reusing standard forms is useful only after confirming that their content and sequence fit the current study.
Apply it
A visit form asks staff to enter a specimen identifier in three sections. Replace unnecessary repeated entry with a controlled link, while keeping independently collected measurements separate when the protocol requires them for comparison.
Common pitfall: Copying a previous study’s form can preserve irrelevant fields and unsuitable visit logic. Standardization still requires a documented check against the current protocol.
4.How should completion guidance distinguish missing responses?
Completion guidance should explain both general conventions and field-specific decisions from the user’s perspective. Define distinctions such as unknown, not done, and not applicable, using only values the form supports. Explain when each applies and how partial information is recorded. These states represent different circumstances and should remain distinguishable. Guidance should clarify recording behavior without suggesting a clinical answer, and changes affecting completion require controlled updates and distribution to relevant users.
Apply it
An assessment was required but not performed; another assessment was outside the participant’s applicable schedule. Provide separate approved recording instructions for these situations, rather than directing staff to use the same missing-value notation.
Common pitfall: A blank field does not reliably distinguish an omitted entry from an unavailable result. Do not let sites invent local missing-value codes or interpretations.
5.Why annotate collection forms?
Annotations connect collected questions and responses with their intended data representation. They make mappings visible for database specification, standards implementation, review, and downstream traceability. An annotation should identify the relevant variable or mapping without obscuring the form’s collection meaning. Collection standards and tabulation standards serve related but different purposes: CDASH supports standardized collection and traceability toward SDTM. Mapping still requires attention to context, controlled terminology, and information needed for later representation.
Apply it
A standard adverse-event form contains onset and resolution dates. Annotate each item to its intended data representation, then check that partial-date conventions and ongoing-event status remain interpretable throughout the planned mapping and review process.
Common pitfall: An annotation does not prove that every downstream variable is collected directly. Some variables are derived or supplied through controlled mappings and metadata.
6.What should a study data-flow diagram reveal?
A useful data-flow diagram identifies data origins, receiving systems, transfers, transformations, reconciliation points, and responsible parties. It should reveal where records can be lost, duplicated, altered, or exposed to unauthorized users. Include externally collected data rather than treating EDC as the whole study database. Distinguish a transfer from an authoritative correction process. The diagram becomes actionable when each interface has defined ownership, acceptance controls, timing, and a route for resolving failures.
Apply it
Imaging results reach the sponsor through a central reviewer, while visit dates come from EDC. Draw both paths, identify the shared matching fields, and assign responsibility for unmatched records and corrected reviewer files.
Common pitfall: A diagram showing only system names misses the decisions between them. Unowned reconciliation and correction steps can remain hidden despite apparently complete technical connections.
7.When should a data management plan be established?
The DMP prospectively defines study-specific data handling from initiation through database closeout. Establish a draft before the first participant is enrolled, and obtain appropriate approval before undertaking the work it describes. Define responsibilities, collection and processing methods, quality controls, reconciliation, coding, and lock arrangements. Maintain the plan as processes change, with version history and stakeholder awareness. Its value is operational: another qualified person should understand how the study’s data were managed.
Apply it
A vendor begins sending files before reconciliation responsibilities are agreed. Resolve the process and ownership prospectively, approve the applicable DMP content, and align the transfer specification before routine processing creates inconsistent local practices.
Common pitfall: Writing the DMP only at study end converts a prospective control into a retrospective narrative. A generic template also cannot replace study-specific decisions.
8.How do SOPs and study procedures support reconstruction?
An organizational SOP defines a controlled process that applies across relevant work, while study procedures explain how that process is implemented for a particular study. Together they should identify roles, inputs, decisions, outputs, and retained evidence at sufficient detail for traceability. Reference controlled documents rather than creating conflicting instructions. If a study needs a departure from an SOP, document the rationale and obtain the required review through the organization’s quality process.
Apply it
A study uses an unusual correction-file workflow not described by the standard transfer procedure. Document the study-specific steps, responsibilities, and evidence, and route the planned procedural departure through the established quality review process.
Common pitfall: An email agreement does not automatically replace a controlled procedure. Undocumented local workarounds make the actual process difficult to reconstruct and evaluate consistently.
9.Which keys distinguish repeated study records?
Specify the level represented by each database record before choosing its key. A participant identifier may identify a person but cannot uniquely identify multiple visits, specimens, or repeated assessments. Composite keys should include the fields needed to distinguish legitimate repetitions, with explicit rules for missing components and duplicates. Define relationships between tables and the expected cardinality. These decisions prevent accidental overwriting and support reliable integration without requiring candidates to program database tables.
Apply it
A participant has two samples at the same visit. A table keyed only by participant and visit would collide. Add the approved specimen or collection identifier and document how corrected records retain their identity.
Common pitfall: Adding a row number can make rows technically unique without establishing meaningful record identity. A key must support matching, corrections, and interpretation across transfers.
10.How can entry screens prevent errors without biasing responses?
Entry screens should express approved collection requirements through suitable field types, permissible values, clear units, and tested conditional logic. They should guide users while allowing legitimate protocol-relevant observations to be recorded. A validation warning can invite confirmation when a value is unusual; an unjustified restriction can suppress truthful data. Distinguish impossible formats from clinically unexpected values. Screen specifications should also explain what happens to existing answers when a controlling response changes.
Apply it
A screen rejects every measurement outside a typical range, although the protocol permits abnormal values. Specify an appropriate confirmation or query workflow and retain a way to record valid unusual observations with necessary context.
Common pitfall: An attractive screen can still distort data if its response choices are incomplete. Do not force a plausible answer merely to satisfy a restrictive interface.
Electronic Data Capture—Study Implementation and Start-up · CCDM Exam Study Guide
11.Which edit checks deserve greatest design attention?
Prioritize checks protecting primary endpoints, important safety information, and data supporting their interpretation. Link checks to protocol requirements and credible error mechanisms rather than testing every conceivable relationship. Automated checks are useful for defined, reproducible conditions; clinical interpretation may require human review. Consider the burden of false signals and overlapping outputs. A check is valuable when its findings can lead to a meaningful clarification or correction, and its performance should be evaluated during use.
Apply it
A team can implement either a cross-source endpoint consistency check or several cosmetic formatting checks. Assess the effect of each error on study interpretation and prioritize the endpoint check while retaining proportionate formatting controls.
Common pitfall: The number of programmed checks is not a quality measure. Numerous low-value or duplicate queries can delay review of consequential safety and efficacy discrepancies.
12.What makes an edit-check specification testable?
A testable specification states the relevant fields, triggering condition, handling of missing or partial values, applicable records, and expected output. Define comparison boundaries and any timing or status dependencies explicitly. “Flag abnormal dates” is insufficient because different implementers can reasonably interpret it differently. Link the rule to its purpose and give representative examples. Controlled review and versioning allow testers to compare observed behavior with an approved expectation rather than an informal understanding.
Apply it
A rule should flag assessments before consent, but partial dates are possible. Specify which complete-date comparisons are valid and how indeterminate comparisons are handled, then provide triggering, nontriggering, and missing-information test examples.
Common pitfall: A query message alone is not a rule specification. Without precise conditions, a correctly worded message may be generated for the wrong records.
Edit Check Design Principles · Electronic Data Capture—Study Implementation and Start-up
13.How should a data-review report be specified?
Define the report’s intended decision, audience, population, record level, selection rules, displayed fields, and refresh timing. Include identifiers and context sufficient for a reviewer to investigate a finding without guessing how rows were selected. Specify sorting, status definitions, and whether results represent participants, visits, records, or queries. A report used for operational decisions also needs a clear data cut. The specification should make both the output and its interpretation reproducible.
Apply it
A missing-assessment report includes participants whose visits are not yet due. Define due-date and visit-status criteria, retain the relevant scheduling context, and state the reporting cut so reviewers can distinguish overdue data from future work.
Common pitfall: A report title such as “missing data” does not define its denominator or eligibility rules. Misclassified future assessments can create unnecessary site follow-up.
14.What belongs in an external data-transfer specification?
Agree on file structure, field definitions, formats, mandatory values, record keys, transfer timing, and acceptance checks before routine delivery. Specify whether each file is cumulative or incremental and how additions, corrections, deletions, and retransmissions are represented. Identify responsible contacts and the discrepancy process. Include protections for blinded information where relevant. A complete specification controls record meaning and lifecycle, allowing the receiver to distinguish a legitimate update from a duplicate or unexplained disappearance.
Apply it
A laboratory plans to resend corrected results in weekly files. Require stable record identifiers and explicit replacement rules, establish how missing records are interpreted, and test a correction followed by retransmission before approving routine transfers.
Common pitfall: Agreeing only on column names leaves critical lifecycle behavior undefined. A repeated file can duplicate data when the receiving process cannot distinguish retransmission from new records.
15.How should standards be selected and governed?
Select standards according to their purpose, applicable regulatory expectations, study needs, and implementation guidance. Record the version and controlled terminology used, and apply them consistently through forms, metadata, and mappings. CDASH supports standardized collection and traceability to SDTM; it does not eliminate study-specific interpretation. Manage proposed departures through controlled governance rather than silently altering a published standard. Assess the downstream effect before adopting a newer version during an active study.
Apply it
A standard terminology update changes a permitted response during recruitment. Assess compatibility with collected data, mappings, and documentation, then approve a controlled implementation decision rather than changing the live form solely because a newer release exists.
Common pitfall: “Use the latest version” is incomplete advice. Suitability, applicable requirements, interoperability, and controlled transition matter when a study already depends on a defined standard.
16.How do preventive quality assurance and in-process QC differ?
Quality assurance establishes processes intended to prevent errors, while quality control evaluates whether work and data meet defined expectations during execution. A useful QC plan identifies what is checked, when, by whom, how findings are classified, and what action follows. Focus controls on consequential errors and apply them at relevant processing stages. Document findings and responses so recurring defects can improve the underlying process rather than merely producing repeated corrections to individual records.
Apply it
Weekly transfer review repeatedly finds the same unit-label defect. Correct affected records through the approved process, investigate the upstream specification or training failure, and modify preventive controls while checking whether the corrective action reduces recurrence.
Common pitfall: A final database inspection cannot substitute for controls throughout processing. Repeatedly correcting outputs without addressing the process allows the same error mechanism to persist.
Training Tasksnot stated
17.How should training differ by user role?
Base training on the tasks, decisions, and system privileges each role actually requires. Site coordinators may need detailed entry and query workflows, while a read-only reviewer needs interpretation and navigation skills. Identify prior knowledge and study-specific gaps before selecting content. Cross-functional education should explain how one team’s actions affect downstream data quality. Role-specific training is effective when learners can perform their assigned work safely and consistently, rather than merely recognize system terminology.
Apply it
Safety reviewers and site coordinators join the same study. Provide shared instruction on data flow, then separate practice for review escalation versus entry and query response, matching each exercise to the user’s actual responsibilities.
Common pitfall: Giving everyone the same exhaustive presentation can hide critical role-specific decisions. Access level alone also does not establish that a user understands the relevant workflow.
18.What makes a learning objective measurable?
A measurable objective describes observable performance under relevant conditions. Replace “understand query management” with a task such as identifying a discrepancy, drafting a neutral query, and documenting its resolution in a practice case. Align content, exercises, and evaluation with that objective. This connection makes the training’s success assessable and exposes missing instructional steps. Objectives should reflect current procedures and realistic work, including decisions learners must make when information is incomplete.
Apply it
A course aims to improve laboratory reconciliation. Set an objective requiring learners to classify matched, missing, duplicate, and conflicting records in a sample transfer, then assess their classifications and proposed follow-up against predefined criteria.
Common pitfall: Course satisfaction and attendance do not demonstrate task mastery. An objective using only “know” or “understand” leaves acceptable performance and evaluation criteria unclear.
19.Why use a separate training environment?
A separate training environment lets learners practise representative tasks without altering production study data. It should simulate relevant system behavior, permissions, forms, and workflows closely enough for skills to transfer. Use study-relevant examples and include unusual but credible situations, not only ideal records. Select instructor-led, online, or combined delivery according to the task and audience, and provide support when learners encounter difficulty. Training cases also reveal whether instructions are usable.
Apply it
New coordinators practise correcting an entered value and responding to a query in a training study. The instructor observes their actions, checks that they preserve the intended correction history, and addresses confusion before production access.
Common pitfall: Practising in production can introduce fictitious records or unintended changes. A training system with unrelated forms or permissions may also teach behavior that fails in actual work.
20.What evidence should training records preserve?
Training records should establish who received which instruction, when it occurred, its objectives, and who delivered it. Identify the relevant course or controlled material so the record can be associated with the procedure or system version in use. Retain evidence according to the training process, including applicable external training. Completion records support authorization decisions, but competence may require separate assessment. Updated responsibilities or workflows can create new training needs despite an earlier completion record.
Apply it
A user completed initial EDC instruction six months ago, but a new reconciliation workflow now applies. Record targeted update training and its applicable materials, rather than assuming the original attendance entry covers the changed responsibility.
Common pitfall: A name on an attendance sheet may be insufficient to establish what was taught. Do not treat a historical course entry as permanent authorization for changed tasks.
Training · Electronic Data Capture—Study Conduct, Maintenance and Closeout
21.When should query trends trigger retraining?
Use recurring errors, help-desk issues, monitoring findings, and workflow changes to identify training needs. First determine whether the cause is a knowledge gap, unclear instructions, poor form design, or a system defect. Match the intervention to the cause, then evaluate whether performance improves. For training-related problems, targeted practice and feedback are more informative than simply repeating a broad presentation. Track comparable post-training outcomes so reduced error counts are not confused with reduced activity.
Apply it
One site repeatedly confuses unknown with not applicable. Confirm that the form and guidance distinguish them clearly, deliver focused examples, and compare the proportion of misclassified responses before and after training using similar record populations.
Common pitfall: Retraining is not the automatic solution to every discrepancy. Repeated instruction will not correct a defective rule, an impossible screen restriction, or contradictory completion guidance.
Data Processing Tasksnot stated
22.How should received and expected data be tracked?
Track receipt against what should exist, using participant, visit, form, specimen, or file identifiers appropriate to the source. Record receipt status and processing progress without confusing them with completeness or acceptance. A delivered file can still be rejected, and an entered form can still lack required information. Expected-data tracking identifies missing records that ordinary field checks cannot detect because no row exists. Reconcile expectations with legitimate schedule changes and documented exceptions.
Apply it
A vendor delivers a file successfully, but ten expected specimens have no records. Mark the transfer as received while separately investigating missing specimen results; do not classify the data stream as complete from delivery status alone.
Common pitfall: Counting files received measures transport activity, not participant-level completeness. A clean import log cannot reveal all expected records unless it is compared with an independent expectation.
23.When should an incoming transfer be accepted?
Apply agreed acceptance checks before treating incoming data as ready for routine use. Check file identity, expected structure, required fields, record counts, key validity, duplicate behavior, and relevant content constraints. Retain evidence of receipt, acceptance or rejection, and follow-up. A technical load is only one stage: successful parsing does not demonstrate correct participant linkage or meaningful values. Handle failed checks through the specified discrepancy process rather than making undocumented receiver-side repairs.
Apply it
A transfer loads without errors, but every record has an unexpected study identifier. Hold the affected delivery for investigation, retain its receipt evidence, and obtain a controlled correction or documented resolution before downstream integration.
Common pitfall: A loader’s success message confirms only the conditions that loader checked. Do not assume it establishes semantic validity, completeness, or compliance with the transfer agreement.
24.How can records from different sources be linked safely?
Use agreed identifiers and contextual keys that identify the same real-world observation across sources. Check identifier formats, leading zeros, visit conventions, specimen identity, and expected relationship cardinality before combining records. Investigate unmatched and multiply matched records explicitly. Preserve source provenance so an integrated value can be traced back to its origin. A seemingly plausible match is insufficient when repeated visits or specimens could associate a result with the wrong observation.
Apply it
Two laboratory records share a participant and visit, but correspond to different collection times. Use the agreed specimen key to link them to EDC; route unresolved matches for clarification rather than selecting the first available row.
Common pitfall: Matching on participant alone can attach valid data to the wrong visit or specimen. Removing duplicates afterward does not repair an incorrect relationship between source observations.
25.What does reconciliation between safety and study data establish?
Reconciliation compares corresponding information across independently managed sources to identify missing records and meaningful disagreements. For safety and study databases, agree which events and fields should correspond, the matching approach, ownership, timing, and discrepancy resolution process. Differences can reflect reporting stages or database purpose rather than an error, so evaluate context. Resolve discrepancies through authorized source processes and retain evidence. Reconciliation supports consistency; it does not replace either system’s required reporting or review activities.
Apply it
A serious event appears in the safety system but has no corresponding study record. Confirm participant and event identity, investigate the expected collection requirement, and route the discrepancy to the responsible parties without simply copying values across systems.
Common pitfall: Reconciliation does not require every field in two databases to be identical. Different purposes and update timing require agreed comparison rules and documented explanations for legitimate differences.
26.How should a data transformation preserve meaning?
A transformation applies a defined rule to create a different representation of existing information, such as standardized units or a controlled date format. Retain original values and sufficient metadata to reproduce the result. Specify eligible inputs, conversion logic, precision, missing-value behavior, and quality checks. Distinguish transformation from inventing unavailable information. Unexpected input should trigger controlled handling rather than an unapproved assumption, especially when a conversion could affect endpoint or safety interpretation.
Apply it
A study’s approved rule converts a measurement by multiplying by 0.5. For an original value of 18, the standardized value is 9; retain the original value, original unit, factor, and output unit alongside the result.
Common pitfall: Changing a unit label without changing the numeric value is not a valid conversion. Filling a missing observation with an estimate is also outside simple transformation.
27.Which laboratory reference range applies to a result?
Link a result to the reference range applicable to its laboratory, test, method where relevant, units, participant characteristics, and effective period. Maintain range versions and implement updates in a controlled manner. A current range is not automatically the correct range for an older specimen. When standardizing units, keep the result and comparison limits compatible. Missing or substitute reference information must be identified and documented so reviewers understand the limitations of any abnormality assessment.
Apply it
A laboratory changes its range on 1 June. A May specimen arrives in a July transfer. Match it using the applicable collection-period rules and laboratory metadata, rather than applying the newest range simply because it was received most recently.
Common pitfall: Comparing a converted result with unconverted limits can create a false abnormal flag. Reference-range selection must preserve both temporal applicability and unit compatibility.
28.How should reported medical information be coded?
Code the reported information faithfully using the selected dictionary and applicable term-selection conventions. For MedDRA, choose a current lowest level term that best represents the reported concept without adding an inferred diagnosis. Preserve the verbatim information and seek clarification when ambiguity materially affects coding. A recognizable pattern of symptoms does not authorize a coder to supply a diagnosis absent from the report. Consistent coding supports analysis while preserving the distinction between reported facts and interpretation.
Apply it
A report states “itchy rash” without a diagnosis. Select terminology representing the reported information under the coding convention, or clarify an important ambiguity; do not replace the description with a specific allergic diagnosis based on personal interpretation.
Common pitfall: Medical plausibility does not justify coding information that was never reported. Choosing a more specific diagnosis can misrepresent the source and distort grouped safety summaries.
29.Why control coding dictionary versions?
Dictionary versions can change available terms and their relationships, affecting coding and grouped analyses even when verbatim reports stay unchanged. Record the dictionary name and version, retain used versions, and control associated synonym or support-table changes. Before upgrading, assess effects on previously coded records and integrated reporting. Use trained, appropriately authorized personnel and maintain relevant licensing. A version transition needs a documented decision about recoding and consistency rather than a silent software update.
Apply it
Two studies are being integrated but were coded with different dictionary versions. Assess term and hierarchy differences, agree a controlled harmonization approach, and retain traceability between original coding, revised coding, and the dictionary versions used.
Common pitfall: Identical displayed terms do not guarantee identical hierarchy or grouping across versions. Silently updating a dictionary can change summaries without any change to the reported events.
CCDM Exam Study Guide · MedDRA Term Selection: Points to Consider, Release 4.26
30.Which discrepancies require contextual review?
A discrepancy is a potential inconsistency, omission, or implausible value requiring evaluation against the protocol, collection guidance, and related information. An automated flag identifies a condition, not a proven error. Review whether a legitimate explanation already exists and whether the issue matters to data use. Some discrepancies need site clarification; others require vendor correction, specification review, or documented acceptance. Apply consistent rules so unusual but valid observations are not forced into expected patterns.
Apply it
An assessment date falls outside its planned visit window, but the record documents an unscheduled assessment. Check the protocol and visit conventions before querying; the date may be valid while its visit classification needs review.
Common pitfall: A value outside an expected range is not automatically wrong. Repeatedly querying confirmed valid observations can pressure users to change accurate data and waste review capacity.
31.How should a neutral data query be written?
State the observed discrepancy, identify the relevant field or records, and request verification or clarification without supplying an unsupported answer. Include enough context for the responsible source to investigate. A query should permit confirmation of a valid value as well as an authorized correction. Keep the request focused and actionable, and use the established communication route. Neutral wording protects source accuracy and reduces the chance that a reviewer’s expectation becomes substituted for reported information.
Apply it
A recorded end date precedes the start date. Ask the site to verify both dates against its records and correct or explain as appropriate, rather than stating which date should be changed or proposing a replacement.
Common pitfall: “Please change the end date to Friday” assumes facts the reviewer may not possess. A leading query can manufacture agreement instead of resolving the underlying discrepancy.
32.When is an answered query ready to close?
An answered query is ready for closure when its response and any authorized data change resolve the stated issue or provide an acceptable documented explanation. Review the resulting data, not merely the presence of a reply. A response can introduce another inconsistency or leave the original question unanswered. Apply the defined closure responsibilities and retain the communication history. Where other parties close queries, data management should have the review oversight established by the study process.
Apply it
A site replies “corrected,” but the date remains unchanged and the inconsistency persists. Review the record and response, then seek focused clarification through the query workflow rather than closing the issue because a reply exists.
Common pitfall: Answered and resolved are different statuses. Closing every replied query can hide unresolved discrepancies and make operational reports appear healthier than the underlying data justify.
Electronic Data Capture—Study Conduct, Maintenance and Closeout · CCDM Exam Study Guide
33.What should an electronic data correction preserve?
A controlled correction should preserve the original entry and a traceable history of the change, including who made it and when, with the reason where applicable. Authorized users should act through the defined correction process rather than overwrite records outside system controls. Audit trails support reconstruction but do not themselves establish that a change was justified. Review the source basis and relevant workflow, including whether changed data require renewed investigator review or signature.
Apply it
A laboratory resubmits a corrected result after identifying a transcription error. Process it through the approved update mechanism, retain the prior value and correction evidence, and assess any affected review or signature status.
Common pitfall: A spreadsheet of final values cannot substitute for the history of changes. A technically logged update also remains questionable if the user lacked authority or source support.
Electronic Systems, Electronic Records, and Electronic Signatures in Clinical Investigations: Questions and Answers · ICH E6(R3) Guideline for Good Clinical Practice
34.How should a data-quality error rate be interpreted?
Define the unit of inspection and denominator before calculating an error rate. Errors per inspected field, erroneous records per reviewed record, and participants with discrepancies measure different things. Document selection methods, critical-variable treatment, assessment timing, and corrective action. Generalizing sample results requires appropriate sampling and statistical methods. Compare rates only when definitions and populations are sufficiently comparable, and distinguish confirmed errors from queries that merely requested clarification of valid information.
Apply it
A review finds 12 confirmed field errors among 600 inspected fields, giving 2%. Another team reports 12 affected records among 200 records, giving 6%; explain that these rates use different units and cannot be compared directly.
Common pitfall: A query count is not an error rate without a meaningful denominator and classification. High-query sites may have more data, more scrutiny, or more complex participants.
35.How can data patterns reveal operational problems?
Analyze patterns by source, site, variable, time, and process stage to distinguish isolated errors from systematic problems. Useful signals include recurring missingness, unusually long delays, unexpected distributions, and repeated corrections after transfers. Verify the underlying data and consider exposure or workload differences before interpreting a signal. Analytics identifies where investigation is warranted; it does not independently establish misconduct or clinical error. Link findings to a responsible owner, investigation, and documented follow-up.
Apply it
One site has many more open queries than others. Compare queries per comparable volume, age distribution, and recent enrollment before deciding whether the cause is workload, delayed responses, unclear instructions, or a site-specific process failure.
Common pitfall: An outlier is a signal for investigation, not a conclusion. Raw comparisons can unfairly penalize a busy site or overlook a smaller site with a high error proportion.
CCDM Exam Study Guide · Project Management for Clinical Data Managers
36.When should system privileges change?
Grant privileges according to authorized responsibilities and required training, and maintain identifiable users with documented access decisions. Review access when roles change, users leave, study status changes, or privileges are no longer required. Separate permissions for viewing, entering, correcting, and administering data as appropriate. Prompt revocation and controlled updates reduce unauthorized activity while preserving historical accountability. Access records should show the rights granted and relevant authorization, modification, and removal dates.
Apply it
A vendor reviewer leaves the project but remains employed by the vendor. Remove the person’s study access through the approved process, document the change, and authorize any replacement individually after confirming the required training and role.
Common pitfall: Vendor employment does not automatically justify continued study access. Shared accounts also obscure responsibility and weaken the usefulness of audit trails and access reviews.
Electronic Data Capture—Study Conduct, Maintenance and Closeout · Electronic Systems, Electronic Records, and Electronic Signatures in Clinical Investigations: Questions and Answers
37.What makes an archived study reconstructable?
An archive needs more than final data values. Preserve relevant metadata, controlled specifications, audit trails, dictionary versions, processing documentation, validation evidence, and other records needed to understand data origin and handling. Use indexed, accessible formats and identify the hardware or software dependencies where relevant. Verify that records remain readable after major system changes. Clearly assigned custody and access controls support later review while protecting confidentiality and the integrity of retained study information.
Apply it
A retired EDC system is being replaced. Before losing access, verify that exported data, metadata, change history, forms, and essential documentation can be located and interpreted independently, then test readability using the planned archive access method.
Common pitfall: A successful export is not proof of a usable archive. Files may omit relationships, signatures, metadata, or software context needed to interpret the final dataset.
CCDM Exam Study Guide · Electronic Data Capture—Study Conduct, Maintenance and Closeout
Programming Tasksnot stated
38.How do you construct SQL queries for unmatched records, grouped counts, and aggregate filters?
Choose joins and filters from the review question. An INNER JOIN returns matched combinations; a LEFT JOIN also preserves unmatched left-table rows, supplying NULL for right-table columns. Join on identifiers that represent the intended relationship and check its cardinality. In the example below, participant_id uniquely identifies a participant within one study, while several laboratory records may belong to that participant. The aliases p and l identify columns from participants and labs. COUNT(l.lab_id) counts matched laboratory records because lab_id cannot be NULL in the input. WHERE filters rows before grouping; HAVING filters completed groups. These queries count laboratory records, not distinct participants or completed visits.
Apply it
Use these hypothetical input tables for all three complete queries. All displayed columns are nonnullable. participants has primary key participant_id; labs has primary key lab_id and references participants through participant_id. L01 and L02 represent different tests at the same visit. participants(participant_id, site_id, status): ('P01', 'A', 'ENROLLED') ('P02', 'A', 'ENROLLED') ('P03', 'B', 'ENROLLED') ('P04', 'B', 'SCREENED') labs(lab_id, participant_id, visit): ('L01', 'P01', 'W4') ('L02', 'P01', 'W4') ('L03', 'P02', 'BASELINE') ('L04', 'P04', 'W4') 1. Find enrolled participants without any W4 laboratory record: SELECT p.participant_id FROM participants AS p LEFT JOIN labs AS l ON p.participant_id = l.participant_id AND l.visit = 'W4' WHERE p.status = 'ENROLLED' AND l.lab_id IS NULL ORDER BY p.participant_id; Expected output, participant_id: P02; P03. P02's baseline record does not satisfy the W4 join condition. P03 has no laboratory records. P01 has W4 matches, and P04 is excluded by enrollment status. Testing the nonnullable lab_id reliably identifies the rows supplied by the outer join. This identifies absence of a W4 record; deciding whether W4 is overdue would additionally require visit expectations and timing. 2. Count W4 laboratory records for enrolled participants by site, retaining sites with zero matches: SELECT p.site_id, COUNT(l.lab_id) AS matched_lab_count FROM participants AS p LEFT JOIN labs AS l ON p.participant_id = l.participant_id AND l.visit = 'W4' WHERE p.status = 'ENROLLED' GROUP BY p.site_id ORDER BY p.site_id; Expected output, (site_id, matched_lab_count): ('A', 2); ('B', 0). Site A has two matched records for P01 and a NULL-extended row for P02. Site B has a NULL-extended row for P03. COUNT(l.lab_id) ignores those NULL values. COUNT(*) would instead return 3 for site A and 1 for site B because it counts every joined row. 3. Return only sites with at least two matched W4 laboratory records among enrolled participants: SELECT p.site_id, COUNT(l.lab_id) AS matched_lab_count FROM participants AS p LEFT JOIN labs AS l ON p.participant_id = l.participant_id AND l.visit = 'W4' WHERE p.status = 'ENROLLED' GROUP BY p.site_id HAVING COUNT(l.lab_id) >= 2 ORDER BY p.site_id; Expected output, (site_id, matched_lab_count): ('A', 2). WHERE excludes screened participants before counts are calculated. HAVING then excludes site B because its count is zero. An aggregate condition such as COUNT(l.lab_id) >= 2 belongs in HAVING rather than WHERE.
Common pitfall: Debugging exercise: in the first query, move l.visit = 'W4' from ON into WHERE while retaining l.lab_id IS NULL. Predict the result and repair the query. Answer: no rows qualify. Matched W4 records have nonnull identifiers; unmatched rows have NULL visit values, so the comparison with 'W4' is unknown and fails the WHERE condition. Restore the visit condition to ON, as shown, to preserve participants without W4 matches. Use IS NULL rather than = NULL. Joining only on participant_id would also conceal P02's missing W4 record because its baseline record would match. DISTINCT cannot repair that incorrect visit scope.
PostgreSQL Documentation: Table Expressions · PostgreSQL Documentation: Aggregate Functions · PostgreSQL: Documentation: 18: 9.2. Comparison Functions and Operators
39.Why do COUNT expressions return different totals?
COUNT(*) counts input rows, whereas COUNT(expression) counts rows where that expression is not null. COUNT(DISTINCT expression) counts distinct nonnull values. After a left join, an unmatched left record still contributes an output row, so COUNT(*) can overstate the number of matching right-side records. Count a suitable nonnull right-side identifier to count matches. Always identify whether the requested measure concerns rows, events, or participants before choosing the aggregate expression.
Apply it
A participant has no matching laboratory record after a left join. In that participant’s group, COUNT(*) returns one, but COUNT(lab_id) returns zero when lab_id is nonnull for real laboratory records. Use the latter for matched-record counts.
Common pitfall: COUNT(DISTINCT participant_id) measures participants, not events. It cannot replace a record count when repeated events are legitimate and the operational question concerns their total.
PostgreSQL Documentation: Aggregate Functions · PostgreSQL Documentation: Table Expressions
40.When should SQL use WHERE, HAVING, and IS NULL?
WHERE filters individual input rows before grouping; HAVING filters groups after aggregation. To identify participants with more than one relevant event, first restrict the event population, group by participant, and apply a count condition in HAVING. Null testing requires IS NULL or IS NOT NULL rather than equality to NULL. Ordinary comparisons involving null do not evaluate as true, so they can silently exclude missing values from a review unless missingness is addressed explicitly.
Apply it
You need participants with at least two unresolved queries. Use WHERE to select unresolved query rows, GROUP BY participant, and HAVING COUNT(*) >= 2. Review missing status values separately rather than assuming they satisfy an ordinary status comparison.
Common pitfall: Filtering an aggregate in WHERE confuses row selection with group selection. Likewise, status = NULL does not identify missing status and can produce an empty review set.
PostgreSQL Documentation: Table Expressions · PostgreSQL Documentation: Aggregate Functions
Testing Tasksnot stated
41.How should testing trace back to requirements?
A test plan defines scope, methods, test data, expected results, acceptance criteria, responsibilities, and defect handling. Link test cases to approved requirements so reviewers can see what has been demonstrated and where gaps remain. Scale validation effort to intended use and risks to participant protection and data reliability. Supplier evidence can inform the approach, but study-specific configuration and workflows still need appropriate verification. Testing should demonstrate intended behavior rather than merely successful system operation.
Apply it
An EDC vendor provides platform validation documentation. Map the study’s critical forms, permissions, edit checks, and transfer workflows to appropriate study-specific tests, using supplier evidence where justified while addressing configuration-specific risks before release.
Common pitfall: A vendor certificate does not establish that a particular study build meets its requirements. Conversely, repeating every supplier test may add effort without addressing study-specific risks.
Electronic Data Capture—Study Implementation and Start-up · ICH E6(R3) Guideline for Good Clinical Practice
42.Which values expose boundary errors?
Boundary testing checks values immediately below, exactly at, and immediately above a specified threshold, along with relevant missing or invalid inputs. It distinguishes inclusive from exclusive comparisons and exposes off-by-one errors. Expected outcomes must come from the approved requirement, not the current implementation. Consider precision and data type when choosing neighboring values. A test that only uses a clearly ordinary value can pass while the rule behaves incorrectly at its decisive boundary.
Apply it
A study rule flags a measurement strictly greater than 10. Test 9.9, 10, and 10.1 when that precision is permitted. The expected outcomes are no flag, no flag, and flag, with missing-input behavior tested separately.
Common pitfall: Testing 5 and 20 alone cannot distinguish “greater than 10” from “greater than or equal to 10.” Include the exact threshold and justified neighboring values.
Electronic Data Capture—Study Implementation and Start-up · Edit Check Design Principles
43.Why test invalid inputs and changing workflow states?
Positive tests show that allowed actions succeed; negative tests show that disallowed inputs or actions are handled as specified. Include missing fields, invalid formats, unauthorized roles, and conflicting states. Workflow testing should also change earlier answers and verify the effect on dependent sections, queries, review status, and signatures where relevant. A form can work on first entry yet fail after correction. Expected handling must preserve valid information and prevent unauthorized or misleading state transitions.
Apply it
A user changes “assessment performed” from yes to no after entering results. Test the approved behavior for existing results, visibility, queries, and review status, rather than checking only whether the dependent section disappears from view.
Common pitfall: A hidden field may still contain data. Visual disappearance alone does not prove that the system handles dependent records or downstream extracts according to the specification.
Electronic Data Capture—Study Implementation and Start-up · CCDM Exam Study Guide
44.How should an interface be tested end to end?
End-to-end interface testing follows representative records from source delivery through receipt, loading, matching, transformation, reconciliation, and usable output. Include new records, corrections, duplicates, rejected files, and retransmissions according to the transfer specification. Check record identity and provenance as well as counts. A structurally valid file may still produce incorrect units or participant associations. Test blind-protection controls where relevant, and confirm that failures generate actionable evidence and reach the responsible party.
Apply it
A test file contains one new result, one corrected result, and a repeated delivery. Verify the intended final record state, retained history, reconciliation output, and error notifications; do not approve the interface solely because the row count is plausible.
Common pitfall: Separate successful tests of sender export and receiver import do not prove the complete interface works. Matching and update behavior can fail between otherwise functioning components.
External Data Transfers · Electronic Data Capture—Study Implementation and Start-up
45.What evidence makes test execution reviewable?
Record the tested requirement, test case, input conditions, system or build version, expected result, actual result, executor, and execution date. Preserve appropriate objective evidence and classify departures through defect handling. A failed test needs a documented resolution and relevant retest rather than an unexplained status change. Summarize results against acceptance criteria and obtain required approval before release. Evidence should allow another reviewer to understand what happened without relying on the tester’s memory.
Apply it
A tester marks a rule as passed but retains no input values or observed output. Re-execute with controlled test data and document the comparison to expected behavior so the result supports an independently reviewable acceptance decision.
Common pitfall: A screenshot without context may not identify the tested build, conditions, or expectation. A pass label is a conclusion, not sufficient evidence of the underlying test.
Electronic Data Capture—Study Implementation and Start-up · Electronic Systems, Electronic Records, and Electronic Signatures in Clinical Investigations: Questions and Answers
46.When does a system change require regression testing?
Assess each change for its effect on intended use, critical data, interfaces, permissions, and dependent functions. Test the changed behavior and select regression tests for functions that could be affected, documenting the rationale and results. Risk-based change control avoids both automatic full retesting and unsupported assumptions that a small change is harmless. Preserve version and approval records, and update relevant specifications and training. Release should depend on defined acceptance criteria and resolved or justified findings.
Apply it
A visit-date change affects window calculations and external matching. Test the revised field behavior, relevant edit checks, matching logic, and extracts identified by impact assessment, then document acceptance before applying the change to the live study.
Common pitfall: The number of changed lines or fields does not measure impact. A small shared-definition change can affect many reports, checks, and interfaces through dependent logic.
ICH E6(R3) Guideline for Good Clinical Practice · Electronic Data Capture—Study Implementation and Start-up
Coordination and Project Management Tasksnot stated
47.How should a change in data management scope be controlled?
Define deliverables, responsibilities, assumptions, exclusions, and acceptance expectations before estimating work. When requirements change, assess effects on effort, cost, timing, quality controls, and dependent activities. Obtain the required agreement and update controlled plans rather than absorbing work invisibly. Scope control supports realistic commitments and prevents a new deliverable from bypassing specification, testing, or review. It should preserve a clear account of what changed, why it changed, and who accepted the consequences.
Apply it
A study adds a new external endpoint source after startup. Assess transfer specification, integration, reconciliation, testing, training, and review effort, then revise the agreed scope and timeline rather than treating receipt of another file as the entire change.
Common pitfall: A technically simple request can create substantial operational work. Accepting it informally may conceal dependencies and leave necessary testing or reconciliation without assigned time or ownership.
Project Management for Clinical Data Managers · CCDM Exam Study Guide
48.How can processing workload be estimated?
Estimate workload from expected volumes, task complexity, observed processing rates, review cycles, and available capacity. Separate productive processing time from other commitments and allow for variability, rework, and dependencies. Update estimates as enrollment and actual throughput become clearer. Arithmetic provides a baseline, not a guarantee: averages can conceal concentrated peaks or specialist bottlenecks. Use workload estimates to adjust resources and sequencing while protecting required quality controls and acceptance criteria.
Apply it
A team must review 480 records at an observed rate of 12 records per productive hour. The baseline is 40 productive hours; then account for available reviewer time, expected rework, and whether all records are equally complex.
Common pitfall: Forty productive hours are not automatically one calendar week. Availability, parallel work, specialist review, and arrival patterns determine when the workload can actually be completed.
Project Management for Clinical Data Managers · CCDM Exam Study Guide
49.Why do startup timelines need explicit dependencies?
A realistic timeline connects tasks to the prerequisites that enable them. Protocol and form decisions affect database specifications; specifications support testing; accepted systems and trained users support deployment. Some development can overlap, but unresolved inputs create rework and release risk. Define owners, required dates, acceptance points, and dependencies across sponsor, site, and vendor activities. Monitor actual progress and revise forecasts transparently when a prerequisite moves rather than preserving an unsupported target date.
Apply it
The first participant visit is approaching while a critical transfer specification remains unsigned. Identify which testing and integration activities depend on it, resolve the decision promptly, and reassess readiness instead of counting incomplete parallel tasks as finished startup work.
Common pitfall: Starting every task early does not remove dependencies. Overlapping work based on unstable requirements can increase rework and leave too little time for meaningful acceptance testing.
Project Management for Clinical Data Managers · CCDM Exam Study Guide
50.Which evidence supports vendor qualification?
Evaluate a prospective vendor’s relevant capabilities, capacity, qualifications, experience, quality processes, regulatory suitability, and organizational stability before contracting. Match the depth of evaluation to the service’s importance and risks. Agree confidentiality protections before exchanging proprietary information. Qualification should examine whether the vendor can deliver this study’s requirements, including interfaces and documentation, rather than relying only on reputation. Define expectations, responsibilities, applicable procedures, and deliverables clearly when establishing the relationship.
Apply it
A vendor offers an attractive price for endpoint processing but has limited experience with the required transfer format. Review capability and capacity evidence, investigate the interface risk, and require clear deliverables and acceptance criteria before selecting the service.
Common pitfall: Low cost and a polished demonstration do not establish qualification. Past success on another service also may not demonstrate capacity or competence for the current study’s critical work.
51.How should sponsor oversight continue after outsourcing?
Outsourcing requires continuing oversight of agreed responsibilities, quality, timeliness, risks, and issue resolution. Establish meaningful performance measures, review deliverable evidence, and follow significant problems through corrective action. Specify which sponsor or vendor procedures govern each activity and maintain accessible contacts and escalation routes. Oversight should reflect service risk and actual performance. A contract establishes expectations, but it does not demonstrate that processing remains controlled or that emerging problems have been addressed.
Apply it
A qualified vendor begins missing transfer dates and sending recurring duplicate records. Review the trend with the vendor, assess downstream risk, agree corrective action and ownership, and verify improvement instead of relying on the original qualification assessment.
Common pitfall: Vendor qualification is not permanent evidence of satisfactory performance. Delegating an activity does not remove the need for sponsor oversight of its quality and study impact.
Vendor Selection & Management · ICH E6(R3) Guideline for Good Clinical Practice
52.When is a database ready for final lock?
Readiness depends on the study’s predefined lock criteria and documented completion or approved disposition of required activities. Review collection completeness, discrepancies, coding, reconciliation, quality checks, investigator review, and required approvals as applicable. Notify relevant parties and remove privileges permitting data changes according to the lock procedure. Define interim, soft, and final lock terms locally. If a later change is necessary, use controlled authorization, impact assessment, unlocking, correction, and relocking rather than informal editing.
Apply it
The target lock date arrives with an unresolved endpoint discrepancy. Assess it against approved criteria and obtain the required documented resolution or disposition; do not declare readiness solely because the milestone date has been reached.
Common pitfall: A database status flag does not prove lock readiness. Likewise, “zero open queries” can conceal unresolved missing records, incomplete reconciliation, or unreviewed answered queries.
Electronic Data Capture—Study Conduct, Maintenance and Closeout · CCDM Exam Study Guide
53.How should metrics support data-risk management?
Identify data-related risks, their potential consequences, owners, warning signals, and response strategies. Select metrics that reveal whether controls are working, with documented definitions, denominators, data cuts, thresholds, and actions. Use both progress measures and signals of emerging problems. Review the underlying metric data for accuracy and timeliness, and update responses as the study evolves. A dashboard is useful when findings lead to investigation and decisions, rather than simply displaying favorable totals.
Apply it
Transfer delays threaten endpoint review before lock. Track delay relative to the agreed schedule, identify affected participants and milestones, assign an escalation owner, and activate a contingency when the defined threshold is crossed.
Common pitfall: A threshold without an owner or response is only a display convention. Improving a headline count can also conceal worsening age or criticality of unresolved work.
Project Management for Clinical Data Managers · CCDM Exam Study Guide
54.What makes a project communication plan actionable?
Specify who needs information, what they need, the communication method and frequency, and how decisions and records are retained. Include issue ownership, severity-based escalation, response expectations, and alternate contacts. Match the plan to users across functions and regions rather than broadcasting every detail to everyone. Routine updates support coordination, while urgent issues need a defined route outside the routine cycle. Maintain the plan when team membership, vendors, or project risks change.
Apply it
A critical transfer failure occurs when the primary vendor contact is unavailable. Use the documented alternate and escalation route, communicate the affected deliverables to responsible stakeholders, and retain the decision and follow-up record.
Common pitfall: A contact list alone does not specify when or how to escalate. Waiting for a routine update can allow a consequential data issue to outlast its response window.
Project Management for Clinical Data Managers · CCDM Exam Study Guide
55.How should audit evidence connect procedures with practice?
Prepare evidence that links approved procedures and study decisions with what was actually done. Relevant records may include DMP versions, specifications, training, access approvals, validation results, transfers, discrepancy histories, and corrective actions. Index records so the study process can be reconstructed efficiently. For findings, distinguish immediate correction from root-cause investigation and prevention, assign ownership, and verify effectiveness. Audit preparation should expose documentation gaps early rather than create unsupported retrospective accounts of past activities.
Apply it
An auditor asks how corrected vendor records entered the final database. Trace the approved transfer rule to receipt evidence, discrepancy resolution, controlled updates, and verification results, identifying any gap instead of offering an unsupported verbal assurance.
Common pitfall: Producing a procedure does not prove it was followed. A corrective-action promise also remains incomplete until implementation and effectiveness are supported by appropriate evidence.
CCDM Exam Study Guide · Project Management for Clinical Data Managers
Review Tasksnot stated
56.How should a protocol amendment be reviewed for data impact?
Compare changed protocol requirements with the current collection and handling design. Assess forms, visit schedules, definitions, checks, transfers, reconciliation, derived outputs, training, and documentation as relevant. Determine when changes apply and how records collected under different versions remain interpretable. Coordinate the impact assessment with appropriate functions before implementation. A wording change can alter endpoint meaning or eligibility logic, so review substance and dependencies rather than judging impact by the amendment’s page count.
Apply it
An amendment changes an endpoint assessment window. Review scheduled visits, date checks, missing-data reports, transfer expectations, and affected reporting definitions, and document how participants under earlier requirements will be distinguished from those under the new version.
Common pitfall: Updating only the visible form can leave incompatible checks and reports in use. An unchanged field label does not establish that its underlying definition stayed the same.
57.Why review consent documents for data-management implications?
Consent documents can affect what information may be collected, accessed, retained, transferred, or used, alongside the protocol and applicable requirements. Review the implications for data flows and operational controls, involving appropriate ethics, legal, privacy, and study personnel where interpretation is required. Distinguish stopping future participation from the handling of information already collected; these are not automatically identical decisions. Document the applicable study-specific approach rather than assuming a universal consequence for every withdrawal.
Apply it
A revised consent introduces an optional future-use component. Review how consent status will be captured and communicated to affected data users, and ensure the handling process distinguishes core study data from the optional use.
Common pitfall: Do not assume that withdrawal always requires deletion or that all previously collected data may always be reused. The decision depends on consent and applicable requirements.
ICH E6(R3) Guideline for Good Clinical Practice · CCDM Exam Study Guide
58.How should safety-table numerators and denominators be checked?
First identify the defined analysis population and whether the display counts participants, events, or another unit. Participant incidence usually counts each affected participant once within the specified category, even when that participant has repeated events. Event counts answer a different question. Verify treatment attribution, data cut, grouping, and percentage denominator against approved definitions. Reconcile tables with supporting listings while recognizing that a participant can contribute to multiple categories without implying an error.
Apply it
Among 50 participants in a defined safety group, eight participants experience 13 events in one category. A participant-incidence display should show eight participants and 16%, while a separate event-count display may correctly report 13 events.
Common pitfall: Dividing event count by participant count does not produce participant incidence. Category totals may also exceed the overall affected-participant count because participants can appear in several categories.
59.How can a reported result be traced to controlled data handling?
Trace a result through its analysis population, source records, transformations, coding, grouping, and data cut using approved definitions and metadata. Check that displays use the intended units, timing conventions, dictionary versions, and treatment assignments. A numerical match alone can be misleading if two errors offset each other. Review representative records and meaningful edge cases to understand the mechanism producing the output, and route discrepancies to the responsible function for controlled resolution.
Apply it
A laboratory summary differs from its listing. Trace the population, applicable ranges, unit conversion, and data cut for selected records before concluding that the difference is rounding or changing the displayed total to match the listing.
Common pitfall: Forcing two outputs to agree can conceal incompatible definitions. Agreement is persuasive only when both outputs use the intended data, populations, units, and handling rules.
ICH E3: Structure and Content of Clinical Study Reports · Laboratory Data Handling
60.What should be verified in a clinical study report’s data-handling description?
Check that descriptions of collection, coding, data handling, study changes, and analysis-related conventions accurately reflect approved processes and their actual implementation. Compare the report with controlled study documentation and relevant outputs. Distinguish planned methods from changes introduced during conduct or analysis, and ensure important changes are described appropriately. Review definitions as carefully as totals: a correct number can still mislead when the report describes the wrong population, unit, timing rule, or coding convention.
Apply it
A draft report says all studies used one dictionary version, but the coding records show a documented version transition. Confirm the actual handling and harmonization approach, then correct the narrative and affected descriptions through the report review process.
Common pitfall: Copying a standard methods paragraph can misstate actual practice. Do not approve familiar wording without checking study-specific decisions, amendments, and the documentation supporting the final outputs.
ICH E3: Structure and Content of Clinical Study Reports · CCDM Exam Study Guide
What to focus on during revision
The following are CertMedbry study suggestions. Use the official outline to check the full scope of your examination.
- Map protocol endpoints and safety requirements to collection fields, definitions, checks, and review outputs.
- Keep source meaning intact through transfers, record matching, transformations, laboratory ranges, and medical coding.
- Distinguish a discrepancy signal from a confirmed error, an answered query from a resolved query, and receipt from acceptance.
- Practise P.5 SQL at the level of joins, cardinality, null handling, filters, grouping, and counts.
- Connect requirements to test cases, boundary values, objective evidence, defect resolution, and change-impact testing.
- Use workload, dependencies, vendor oversight, actionable metrics, and documented criteria to judge lock readiness.
- Verify reporting populations, counting units, definitions, and the correspondence between described and actual data handling.
A practical preparation plan
1. Build a task-based study map
Read the complete official study-guide task list and mark tested tasks and explicit exclusions. Map your experience and knowledge gaps to all seven tested domains; use these 60 concepts as selected explanations rather than a replacement for the full outline.
2. Design one coherent study workflow
Choose a simple hypothetical study. Map an endpoint from protocol requirement to field definition, form, completion guidance, annotation, edit check, transfer specification, and review report. Explain why each component is necessary.
3. Work through controlled processing decisions
Practise a missing transfer record, a repeated specimen, a corrected laboratory result, an ambiguous coding term, and a neutral query. For each, identify source authority, record identity, retained evidence, and the criterion for resolution.
4. Practise SQL and testing together
Predict join outputs and COUNT results using small tables that include unmatched records and repeated observations. Then design tests for boundaries, missing inputs, correction files, permissions, and dependent workflow states.
5. Rehearse readiness and reporting reviews
Use a hypothetical lock checklist to evaluate incomplete reconciliation, answered queries, signatures, access changes, and approvals. Check a participant-incidence table against event listings and verify that a report’s methods match documented handling.
6. Turn practice errors into targeted review
Complete the 20 drills, explain why each alternative is less appropriate, and classify mistakes by decision mechanism. Use the flashcards for spaced recall, then revisit the cited practice reference and the corresponding official task.
Use your sample score as a starting point
The free set contains 20 original practice questions. It is a small learning sample, not an official examination or a validated pass prediction. Review each explanation and its source before moving to another topic.
Start the free sampleOfficial sources and updates
- CCDM Certification
- CCDM® Certification Handbook
- CCDM Exam Study Guide
- CDASH
- CRF Completion Guidelines
- Data Management Plan
- External Data Transfers
- Edit Check Design Principles
- Training
- MedDRA Term Selection: Points to Consider, Release 4.26
- Laboratory Data Handling
- Electronic Data Capture—Study Conduct, Maintenance and Closeout
- PostgreSQL Documentation: Table Expressions
- PostgreSQL Documentation: Aggregate Functions
- Electronic Data Capture—Study Implementation and Start-up
- ICH E6(R3) Guideline for Good Clinical Practice
- Vendor Selection & Management
- Project Management for Clinical Data Managers
- ICH E3: Structure and Content of Clinical Study Reports
- Electronic Systems, Electronic Records, and Electronic Signatures in Clinical Investigations: Questions and Answers
- PostgreSQL: Documentation: 18: 9.2. Comparison Functions and Operators
- Foundational | CDISC
- | Electronic Data Capture-Study Implementation and Start-up |Journal of the Society for Clinical Data Management
- Edgerton | Project Management for Clinical Data Managers |Journal of the Society for Clinical Data Management
Confirm current eligibility, fees, scheduling and policies directly with the examination provider. CertMedbry is independent and is not endorsed by these organisations. How we handle content and corrections.