VIEW THIS AS

Auto mode follows the Route Engine until you choose a viewpoint.

YOU ARE HERE

ROUTE CHECK

CONNECTED TO

WHAT NEXT

Use the canonical route for this room, or HELP if you are unsure.

Why English? | Writing a Clinical Trial Protocol Deviation Report

Three learners review open books together at a classroom table, with stacks of textbooks, stationery and a whiteboard in the bright room.

WHY ENGLISH?

Protect the participant, then make the departure learnable

Use the routes to identify the approved requirement, reconstruct the event, assess participant and data impact, complete required notifications, and connect cause, corrective action and prevention without hiding uncertainty.

Open the full contents · See the How English Works hub

Full contents

Anchor protocol
  1. Identify the trial
  2. Identify the site
  3. Identify protocol version
  4. Identify related documents
  5. Identify the requirement
  6. Protect participant identity
Rebuild event
  1. Identify the event date
  2. Identify discovery date
  3. Identify reporter and roles
  4. Prioritise immediate safety
  5. State what should have happened
  6. State what actually happened
Assess impact
  1. Build the chronology
  2. Identify source evidence
  3. Separate fact from inference
  4. Describe participant context
  5. Describe protocol window
  6. Describe eligibility departures
Notify correctly
  1. Describe consent departures
  2. Describe treatment departures
  3. Describe investigational-product departures
  4. Describe assessment departures
  5. Describe specimen departures
  6. Describe randomisation departures
Prevent repeat
  1. Describe unblinding events
  2. Classify without minimising
  3. Avoid classification inflation
  4. Assess participant rights
  5. Assess participant safety
  6. Assess participant well-being
Practice and next steps
  1. Assess data reliability
  2. Assess primary-endpoint impact
  3. Assess blinding impact
  4. Assess analysis-set impact
  5. Assess recurrence
  6. Assess scope
Practice and next steps
  1. Take correction
  2. Protect source records
  3. Update trial systems
  4. Notify the principal investigator
  5. Notify the sponsor
  6. Notify the IRB where required
Practice and next steps
  1. Notify HSA where required
  2. Distinguish safety reporting
  3. Distinguish amendment
  4. Handle urgent hazards
  5. Use neutral causal language
  6. Perform root-cause analysis
Practice and next steps
  1. Differentiate contributing factors
  2. Design corrective action
  3. Design preventive action
  4. Assign CAPA owners
  5. Set CAPA deadlines
  6. Define verification
Practice and next steps
  1. Define effectiveness
  2. Document training
  3. Trend deviations
  4. Escalate systemic noncompliance
  5. Preserve essential records
  6. Control report versions
Practice and next steps
  1. Close only with evidence
  2. Write a bounded conclusion
  3. Learn the transferable skill
  4. A worked example
  5. A practical checklist
  6. Advice for students, parents and young adults
Practice and next steps
  1. Frequently asked questions
  2. The deeper English lesson
  3. Useful next reading

A clinical trial protocol deviation report documents a departure from the approved protocol, applicable requirements or controlled trial procedures. English matters because the record must tell the truth quickly enough to protect a participant, precisely enough to support oversight, and neutrally enough to distinguish what happened from why it happened and what should change.

People searching for protocol deviation report, clinical trial deviation log, major deviation, GCP noncompliance, CAPA or IRB deviation reporting often want a universal template. The classification and notification route depend on the protocol, sponsor procedures, ethics requirements, applicable law, participant impact and jurisdiction. A clear report supports—not replaces—that decision process.

Singapore's Health Sciences Authority states that the Principles and Annex 1 of ICH E6(R3) Good Clinical Practice took effect in Singapore on 1 January 2026, while Annex 2 is scheduled for 1 January 2027. HSA's clinical-trial template forms include a deviation and noncompliance tracking log, and its essential records guidance explains that trial records collectively support evaluation of methods, factors, actions, reliability and GCP compliance. Current trial-specific procedures and authority requirements control real reporting.

Use the deviation map trial, protocol, amendment, approval, site, participant code, visit, procedure, eligibility, consent, randomisation, treatment, investigational product, specimen, assessment, window, departure, discovery, source record, chronology, immediate action, medical assessment, rights, safety, well-being, data reliability, classification, recurrence, notification, sponsor, investigator, IRB, authority, privacy, root cause, correction, corrective action, preventive action, owner, deadline, verification, effectiveness, closure, trend and essential record.


Identify the trial

Anchor protocol title number phase and sponsor to the approved protocol and controlled trial context.

The departure cannot be assessed when the project nickname substitutes for controlled identifiers.

Lock identifiers and use the trial's formal identity.

Protocol example. Protocol ABC-201 phase II is stated in full.

Did You Know? HSA implemented the ICH E6(R3) GCP Principles and Annex 1 in Singapore on 1 January 2026, with Annex 2 scheduled for 1 January 2027, so a report must date the framework it applies. This is why protocol title number phase and sponsor must remain attached to purpose, evidence and scope.


Identify the site

Reconstruct institution and site number where the event occurred from source evidence and time.

Fact and explanation merge when multi-site context is hidden.

Protect identity and name the responsible site.

Event example. Site 004 documents the deviation.

Expected action, actual action, detection and immediate response belong in separate traceable statements.


Identify protocol version

Assess the approved version in effect at the event across rights, safety, well-being and data reliability.

Impact is minimised when the latest amendment is applied retrospectively.

Use qualified judgement and record version date and approval status.

Impact example. Version 5 dated March governed the visit.

No observed injury does not automatically mean no participant or trial impact.


Route manuals plans standard procedures and ethics conditions relevant to the event through the required oversight channels.

Compliance fragments when only the main protocol is checked.

Check sponsor, investigator, IRB and list each controlling source.

Notification example. The pharmacy manual governs temperature handling.

Deviation, safety, amendment and regulatory processes may run in parallel.


Identify the requirement

Turn the exact clause visit window criterion or procedure expected into verified system learning.

Closure is premature when the report says protocol not followed without showing what was required.

Connect cause, action and cite the bounded rule.

CAPA example. Section 7.3 requires sampling before dosing.

Corrective action is credible when implementation and effectiveness are both visible.


Protect participant identity

Anchor coded identifiers and minimum necessary personal information to the approved protocol and controlled trial context.

The departure cannot be assessed when names or identifying combinations enter the report.

Lock identifiers and use the trial code and privacy rules.

Protocol example. Participant 004-017 is used instead of a name.

The applicable version and requirement define the deviation before any classification begins.


Identify the event date

Reconstruct when the departure occurred from source evidence and time.

Fact and explanation merge when discovery date replaces occurrence date.

Protect identity and record both dates.

Event example. The missed assessment occurred Monday and was found Wednesday.

Did You Know? HSA implemented the ICH E6(R3) GCP Principles and Annex 1 in Singapore on 1 January 2026, with Annex 2 scheduled for 1 January 2027, so a report must date the framework it applies. This is why when the departure occurred must remain attached to purpose, evidence and scope.


Identify discovery date

Assess when and how the deviation became known across rights, safety, well-being and data reliability.

Impact is minimised when late discovery remains unexplained.

Use qualified judgement and state the detection route.

Impact example. A monitor found the issue during source review.

No observed injury does not automatically mean no participant or trial impact.


Identify reporter and roles

Route who observed documented assessed and approved the record through the required oversight channels.

Compliance fragments when anonymous passive voice hides responsibility.

Check sponsor, investigator, IRB and use roles and signatures.

Notification example. The study coordinator documented and the PI assessed.

Deviation, safety, amendment and regulatory processes may run in parallel.


Prioritise immediate safety

Turn medical or operational action needed before paperwork into verified system learning.

Closure is premature when form completion delays participant care.

Connect cause, action and state care first.

CAPA example. The investigator contacts the participant for urgent assessment.

Corrective action is credible when implementation and effectiveness are both visible.


State what should have happened

Anchor the planned protocol-compliant sequence to the approved protocol and controlled trial context.

The departure cannot be assessed when the ideal pathway is implied but not written.

Lock identifiers and describe the requirement plainly.

Protocol example. Consent confirmation should precede the procedure.

The applicable version and requirement define the deviation before any classification begins.


State what actually happened

Reconstruct observable departure in neutral language from source evidence and time.

Fact and explanation merge when blame or intention replaces facts.

Protect identity and use actor action time and source.

Event example. The sample was collected forty minutes outside the window.

Expected action, actual action, detection and immediate response belong in separate traceable statements.


Build the chronology

Assess ordered actions communications and decisions across rights, safety, well-being and data reliability.

Impact is minimised when events are grouped by department rather than time.

Use qualified judgement and use timestamps from source records.

Impact example. Dose time precedes the laboratory result review.

Did You Know? HSA implemented the ICH E6(R3) GCP Principles and Annex 1 in Singapore on 1 January 2026, with Annex 2 scheduled for 1 January 2027, so a report must date the framework it applies. This is why ordered actions communications and decisions must remain attached to purpose, evidence and scope.


Identify source evidence

Route medical records logs systems emails and device data supporting the account through the required oversight channels.

Compliance fragments when memory becomes the only record.

Check sponsor, investigator, IRB and cite traceable records.

Notification example. The electronic case report form and clinic log confirm timing.

Deviation, safety, amendment and regulatory processes may run in parallel.


Separate fact from inference

Turn observed information versus interpretation into verified system learning.

Closure is premature when the assumed cause enters the event description.

Connect cause, action and label analysis separately.

CAPA example. The timestamp is fact while workload contribution is analysed later.

Corrective action is credible when implementation and effectiveness are both visible.


Describe participant context

Anchor relevant visit treatment and clinical circumstances without excess identifiers to the approved protocol and controlled trial context.

The departure cannot be assessed when context is either missing or over-disclosed.

Lock identifiers and include only decision-relevant facts.

Protocol example. The participant was asymptomatic at the scheduled visit.

The applicable version and requirement define the deviation before any classification begins.


Describe protocol window

Reconstruct allowed timing and actual timing from source evidence and time.

Fact and explanation merge when outside window is stated without magnitude.

Protect identity and show boundary and variance.

Event example. The visit occurred two days beyond the plus-or-minus three-day window.

Expected action, actual action, detection and immediate response belong in separate traceable statements.


Describe eligibility departures

Assess inclusion or exclusion criteria not met or confirmed across rights, safety, well-being and data reliability.

Impact is minimised when enrolment is called erroneous without naming the criterion.

Use qualified judgement and state the criterion and evidence.

Impact example. A required laboratory threshold was unavailable at randomisation.

No observed injury does not automatically mean no participant or trial impact.


Route timing version process signature or reconsent issue through the required oversight channels.

Compliance fragments when consent paperwork is treated as purely administrative.

Check sponsor, investigator, IRB and assess rights and understanding.

Notification example. The outdated consent version omitted a newly approved risk.

Did You Know? HSA implemented the ICH E6(R3) GCP Principles and Annex 1 in Singapore on 1 January 2026, with Annex 2 scheduled for 1 January 2027, so a report must date the framework it applies. This is why timing version process signature or reconsent issue must remain attached to purpose, evidence and scope.


Describe treatment departures

Turn dose route schedule interruption or prohibited therapy into verified system learning.

Closure is premature when all dosing differences are grouped together.

Connect cause, action and state expected and actual exposure.

CAPA example. One dose was delayed six hours.

Corrective action is credible when implementation and effectiveness are both visible.


Describe investigational-product departures

Anchor storage dispensing accountability labelling or return issue to the approved protocol and controlled trial context.

The departure cannot be assessed when product control and participant exposure are merged.

Lock identifiers and trace batch condition and use.

Protocol example. A temperature excursion affected quarantined stock not dispensed units.

The applicable version and requirement define the deviation before any classification begins.


Describe assessment departures

Reconstruct missed late early incomplete or incorrect study procedure from source evidence and time.

Fact and explanation merge when the word missed hides what data are absent.

Protect identity and name assessment and consequence.

Event example. The week-eight ECG was not performed.

Expected action, actual action, detection and immediate response belong in separate traceable statements.


Describe specimen departures

Assess collection processing labelling storage shipment or analysis issue across rights, safety, well-being and data reliability.

Impact is minimised when sample problems are blamed on the laboratory generically.

Use qualified judgement and trace the specimen chain.

Impact example. The tube was centrifuged after the protocol limit.

No observed injury does not automatically mean no participant or trial impact.


Describe randomisation departures

Route allocation process timing or system use outside protocol through the required oversight channels.

Compliance fragments when treatment assignment is discussed without blinding impact.

Check sponsor, investigator, IRB and record system and access.

Notification example. The participant was randomised before final eligibility confirmation.

Deviation, safety, amendment and regulatory processes may run in parallel.


Describe unblinding events

Turn who learned treatment assignment when and why into verified system learning.

Closure is premature when all knowledge is called a protocol deviation automatically.

Connect cause, action and check authorised emergency procedures.

CAPA example. The investigator used the emergency code for clinical management.

Did You Know? HSA implemented the ICH E6(R3) GCP Principles and Annex 1 in Singapore on 1 January 2026, with Annex 2 scheduled for 1 January 2027, so a report must date the framework it applies. This is why who learned treatment assignment when and why must remain attached to purpose, evidence and scope.


Classify without minimising

Anchor the sponsor or institution's defined category based on impact and requirements to the approved protocol and controlled trial context.

The departure cannot be assessed when minor major significant and serious are used as ordinary adjectives.

Lock identifiers and apply controlled definitions.

Protocol example. The event meets the sponsor's important deviation criteria.

The applicable version and requirement define the deviation before any classification begins.


Avoid classification inflation

Reconstruct language that implies a category before evidence is assessed from source evidence and time.

Fact and explanation merge when worst-case labels replace evaluation.

Protect identity and state provisional status if needed.

Event example. Classification remains pending sponsor review.

Expected action, actual action, detection and immediate response belong in separate traceable statements.


Assess participant rights

Assess informed choice privacy and fair treatment across rights, safety, well-being and data reliability.

Impact is minimised when no physical injury is equated with no impact.

Use qualified judgement and consider consent and confidentiality.

Impact example. The outdated consent may affect informed understanding.

No observed injury does not automatically mean no participant or trial impact.


Assess participant safety

Route actual and potential clinical consequence through the required oversight channels.

Compliance fragments when absence of symptoms proves zero risk.

Check sponsor, investigator, IRB and document medical judgement and follow-up.

Notification example. The PI finds no immediate harm but schedules monitoring.

Deviation, safety, amendment and regulatory processes may run in parallel.


Assess participant well-being

Turn burden distress inconvenience and continuing support into verified system learning.

Closure is premature when only laboratory values count.

Connect cause, action and include relevant experience.

CAPA example. An unscheduled return visit creates additional burden.

Corrective action is credible when implementation and effectiveness are both visible.


Assess data reliability

Anchor whether endpoints eligibility treatment or analysis are affected to the approved protocol and controlled trial context.

The departure cannot be assessed when all deviations are said to invalidate the trial.

Lock identifiers and specify variables and records.

Protocol example. The late scan affects one secondary time-point.

Did You Know? HSA implemented the ICH E6(R3) GCP Principles and Annex 1 in Singapore on 1 January 2026, with Annex 2 scheduled for 1 January 2027, so a report must date the framework it applies. This is why whether endpoints eligibility treatment or analysis are affected must remain attached to purpose, evidence and scope.


Assess primary-endpoint impact

Reconstruct connection to the main trial question from source evidence and time.

Fact and explanation merge when any missed procedure is assumed central.

Protect identity and map the endpoint definition.

Event example. The missing assessment is not part of the primary endpoint.

Expected action, actual action, detection and immediate response belong in separate traceable statements.


Assess blinding impact

Assess whether allocation knowledge could bias conduct or assessment across rights, safety, well-being and data reliability.

Impact is minimised when unblinding is described without who knew.

Use qualified judgement and name person timing and affected decisions.

Impact example. Only the treating physician learned assignment after the visit.

No observed injury does not automatically mean no participant or trial impact.


Assess analysis-set impact

Route how the event may be handled in planned analysis through the required oversight channels.

Compliance fragments when the site decides exclusion independently.

Check sponsor, investigator, IRB and defer to statistical and sponsor procedures.

Notification example. The participant remains enrolled while analysis impact is reviewed.

Deviation, safety, amendment and regulatory processes may run in parallel.


Assess recurrence

Turn isolated event pattern or systemic condition into verified system learning.

Closure is premature when first detected means first occurred.

Connect cause, action and search comparable records.

CAPA example. Three late visits share the same scheduling process.

Corrective action is credible when implementation and effectiveness are both visible.


Assess scope

Anchor participants visits sites systems and periods potentially affected to the approved protocol and controlled trial context.

The departure cannot be assessed when the report stays limited to the discovered record.

Lock identifiers and perform a bounded retrospective check.

Protocol example. All visits scheduled by the same workflow are reviewed.

The applicable version and requirement define the deviation before any classification begins.


Take correction

Reconstruct action addressing the immediate event or record from source evidence and time.

Fact and explanation merge when correction is called prevention.

Protect identity and describe the direct fix.

Event example. The overdue assessment is completed when clinically appropriate.

Did You Know? HSA implemented the ICH E6(R3) GCP Principles and Annex 1 in Singapore on 1 January 2026, with Annex 2 scheduled for 1 January 2027, so a report must date the framework it applies. This is why action addressing the immediate event or record must remain attached to purpose, evidence and scope.


Protect source records

Assess contemporaneous documentation and audit trail across rights, safety, well-being and data reliability.

Impact is minimised when records are rewritten to look compliant.

Use qualified judgement and correct transparently under data-integrity rules.

Impact example. A dated clarification explains the original entry.

No observed injury does not automatically mean no participant or trial impact.


Update trial systems

Route case report form queries logs and tracking records through the required oversight channels.

Compliance fragments when the narrative and database disagree.

Check sponsor, investigator, IRB and reconcile controlled records.

Notification example. The deviation log links to the relevant data query.

Deviation, safety, amendment and regulatory processes may run in parallel.


Notify the principal investigator

Turn clinical and site oversight by the PI into verified system learning.

Closure is premature when coordinators close events without PI assessment.

Connect cause, action and obtain timely documented review.

CAPA example. The PI signs the safety and impact assessment.

Corrective action is credible when implementation and effectiveness are both visible.


Notify the sponsor

Anchor communication under protocol contract and sponsor procedures to the approved protocol and controlled trial context.

The departure cannot be assessed when site staff wait for routine monitoring.

Lock identifiers and follow specified urgency and channel.

Protocol example. The sponsor receives the initial report within one business day.

The applicable version and requirement define the deviation before any classification begins.


Notify the IRB where required

Reconstruct ethics oversight under local policy and event classification from source evidence and time.

Fact and explanation merge when all deviations are either never reported or reported identically.

Protect identity and follow current IRB rules.

Event example. A rights-affecting consent deviation is submitted promptly.

Expected action, actual action, detection and immediate response belong in separate traceable statements.


Notify HSA where required

Assess regulatory reporting according to applicable Singapore requirements across rights, safety, well-being and data reliability.

Impact is minimised when a site template is mistaken for the statutory rule.

Use qualified judgement and check the current authority route.

Impact example. The sponsor determines whether the event triggers a regulatory submission.

Did You Know? HSA implemented the ICH E6(R3) GCP Principles and Annex 1 in Singapore on 1 January 2026, with Annex 2 scheduled for 1 January 2027, so a report must date the framework it applies. This is why regulatory reporting according to applicable Singapore requirements must remain attached to purpose, evidence and scope.


Distinguish safety reporting

Route adverse event and serious adverse event duties from protocol-deviation documentation through the required oversight channels.

Compliance fragments when one report is assumed to satisfy every obligation.

Check sponsor, investigator, IRB and run each required process.

Notification example. An SAE is reported separately while its associated deviation is tracked.

Deviation, safety, amendment and regulatory processes may run in parallel.


Distinguish amendment

Turn planned prospective protocol change versus unplanned departure into verified system learning.

Closure is premature when a recurring deviation is normalised without amendment review.

Connect cause, action and use formal change control.

CAPA example. A visit window change requires approved amendment before routine use.

Corrective action is credible when implementation and effectiveness are both visible.


Handle urgent hazards

Anchor departure needed to protect a participant from immediate risk to the approved protocol and controlled trial context.

The departure cannot be assessed when staff wait for prior approval during an emergency.

Lock identifiers and follow applicable urgent-safety provisions and notify afterward.

Protocol example. The investigator acts immediately to prevent harm.

The applicable version and requirement define the deviation before any classification begins.


Use neutral causal language

Reconstruct systems conditions decisions and evidence rather than personal blame from source evidence and time.

Fact and explanation merge when careless staff becomes the root cause.

Protect identity and describe mechanism.

Event example. The scheduling alert failed after a workflow migration.

Expected action, actual action, detection and immediate response belong in separate traceable statements.


Perform root-cause analysis

Assess why controls did not prevent or detect the event across rights, safety, well-being and data reliability.

Impact is minimised when the last human action ends analysis.

Use qualified judgement and examine process technology training workload and oversight.

Impact example. The calendar template excluded amended visit windows.

No observed injury does not automatically mean no participant or trial impact.


Differentiate contributing factors

Route conditions increasing likelihood without being the sole cause through the required oversight channels.

Compliance fragments when every observation is labelled root cause.

Check sponsor, investigator, IRB and rank evidence and connection.

Notification example. High workload contributed but did not create the incorrect template.

Did You Know? HSA implemented the ICH E6(R3) GCP Principles and Annex 1 in Singapore on 1 January 2026, with Annex 2 scheduled for 1 January 2027, so a report must date the framework it applies. This is why conditions increasing likelihood without being the sole cause must remain attached to purpose, evidence and scope.


Design corrective action

Turn change removing the identified cause or restoring control into verified system learning.

Closure is premature when retraining is selected automatically.

Connect cause, action and match action to mechanism.

CAPA example. The scheduling template is corrected and independently verified.

Corrective action is credible when implementation and effectiveness are both visible.


Design preventive action

Anchor system change reducing recurrence beyond the single record to the approved protocol and controlled trial context.

The departure cannot be assessed when the participant-specific correction is called CAPA.

Lock identifiers and address broader scope.

Protocol example. A version-controlled visit calculator is deployed across sites.

The applicable version and requirement define the deviation before any classification begins.


Assign CAPA owners

Reconstruct named roles responsible for each action from source evidence and time.

Fact and explanation merge when the team owns it means nobody owns it.

Protect identity and record accountable person.

Event example. Clinical operations owns the template change.

Expected action, actual action, detection and immediate response belong in separate traceable statements.


Set CAPA deadlines

Assess realistic completion dates based on risk across rights, safety, well-being and data reliability.

Impact is minimised when open actions have no time control.

Use qualified judgement and calendar and escalate overdue work.

Impact example. High-risk consent actions close within the specified period.

No observed injury does not automatically mean no participant or trial impact.


Define verification

Route evidence that the action was implemented as intended through the required oversight channels.

Compliance fragments when uploading a training slide proves completion.

Check sponsor, investigator, IRB and check the changed control.

Notification example. System audit confirms the new visit logic is active.

Deviation, safety, amendment and regulatory processes may run in parallel.


Define effectiveness

Turn evidence that the action prevents or detects recurrence into verified system learning.

Closure is premature when implementation and effectiveness are merged.

Connect cause, action and select follow-up metric and period.

CAPA example. No window errors occur across the next twenty eligible visits.

Did You Know? HSA implemented the ICH E6(R3) GCP Principles and Annex 1 in Singapore on 1 January 2026, with Annex 2 scheduled for 1 January 2027, so a report must date the framework it applies. This is why evidence that the action prevents or detects recurrence must remain attached to purpose, evidence and scope.


Document training

Anchor who needed what role-specific learning and when to the approved protocol and controlled trial context.

The departure cannot be assessed when everyone receives generic retraining.

Lock identifiers and target the changed task.

Protocol example. Schedulers practise the amended-window workflow.

The applicable version and requirement define the deviation before any classification begins.


Trend deviations

Reconstruct categories sites processes and recurrence over time from source evidence and time.

Fact and explanation merge when each event is closed in isolation.

Protect identity and use meaningful coded analysis.

Event example. Consent-version deviations rise after document updates.

Expected action, actual action, detection and immediate response belong in separate traceable statements.


Escalate systemic noncompliance

Assess patterns or serious failures requiring wider action across rights, safety, well-being and data reliability.

Impact is minimised when repeat events are relabelled as isolated.

Use qualified judgement and use sponsor quality and governance routes.

Impact example. Multiple sites show the same randomisation-control failure.

No observed injury does not automatically mean no participant or trial impact.


Preserve essential records

Route complete records and metadata allowing trial reconstruction through the required oversight channels.

Compliance fragments when the deviation form is stored without linked evidence.

Check sponsor, investigator, IRB and index related records.

Notification example. The report links protocol clause source note notifications and CAPA.

Deviation, safety, amendment and regulatory processes may run in parallel.


Control report versions

Turn draft review signed final and amended records into verified system learning.

Closure is premature when later edits erase the original assessment.

Connect cause, action and use audit trails and reasons.

CAPA example. A corrected report retains prior version and approval history.

Corrective action is credible when implementation and effectiveness are both visible.


Close only with evidence

Anchor required notifications actions records and reviews complete to the approved protocol and controlled trial context.

The departure cannot be assessed when administrative closure hides open risk.

Lock identifiers and use a closure checklist.

Protocol example. The event closes after CAPA effectiveness is confirmed.

Did You Know? HSA implemented the ICH E6(R3) GCP Principles and Annex 1 in Singapore on 1 January 2026, with Annex 2 scheduled for 1 January 2027, so a report must date the framework it applies. This is why required notifications actions records and reviews complete must remain attached to purpose, evidence and scope.


Write a bounded conclusion

Reconstruct what occurred impact classification actions and remaining uncertainty from source evidence and time.

Fact and explanation merge when the report says no impact without reasoning.

Protect identity and state the supported scope.

Event example. The late secondary assessment caused no identified safety harm but reduces completeness at one time-point.

Expected action, actual action, detection and immediate response belong in separate traceable statements.


Learn the transferable skill

Assess honest reconstruction that protects people and evidence across rights, safety, well-being and data reliability.

Impact is minimised when deviation reporting is treated as confession or blame.

Use qualified judgement and practise fact impact action and learning.

Impact example. Students can analyse a missed laboratory step using the same layers.

No observed injury does not automatically mean no participant or trial impact.


A worked example

A site writes that a coordinator accidentally missed a visit and that there was no impact. The record omits the protocol window, actual date, participant context, source evidence, investigator assessment, missing endpoint data, notification decisions and whether similar visits were affected.

A defensible report locks the protocol version, reconstructs expected and actual events, protects the participant's identity, records immediate care, assesses rights, safety and data reliability, applies controlled classification, completes required notifications, investigates system causes and verifies CAPA effectiveness.

English keeps protection and learning in the same record: this approved requirement applied, this departure occurred, this participant and evidence were affected in these bounded ways, these actions followed, and these controls will be tested before closure.


A practical checklist

  1. Trial site protocol version and clause exact
  2. Participant identity protected
  3. Occurrence and discovery dates separated
  4. Expected and actual sequence reconstructed
  5. Source evidence linked
  6. Immediate safety action recorded
  7. Rights safety well-being and data impact assessed
  8. Classification uses controlled definitions
  9. PI sponsor IRB HSA and safety routes checked
  10. Cause and contributing factors evidence-based
  11. CAPA owners deadlines verification and effectiveness defined
  12. Essential records and audit trail complete

Advice for students, parents and young adults

Students can practise with a fictional study visit: write what the plan required, what occurred, how the participant and data might be affected, what action is immediate and what system change prevents recurrence.

Parents can use the topic to discuss why good research records protect people: clear documentation is not about blaming a person; it makes oversight, correction and trust possible.

Real clinical-trial reporting is regulated and participant-centred. Use the approved protocol, sponsor procedures, IRB requirements, current HSA rules, qualified investigators and clinical judgement.


Frequently asked questions

What is a protocol deviation?

A departure from the approved protocol or applicable controlled trial requirements. Use the definitions adopted by the sponsor, institution and authorities.

Is every deviation serious?

No. Impact and classification vary, but every event should be assessed under the applicable process rather than minimised informally.

Does no injury mean no impact?

No. Rights, informed consent, privacy, well-being, endpoint data and trial reliability may be affected even without physical harm.

Who must be notified?

That depends on the event and requirements. Check the principal investigator, sponsor, IRB, HSA and separate safety-reporting routes.

Can a source record be changed to look compliant?

No. Corrections must preserve the original record and audit trail according to data-integrity requirements.

When can the deviation close?

After required assessment, notifications, records and actions are complete, with CAPA effectiveness demonstrated where applicable.


The deeper English lesson

Protocol-deviation English is participant-to-learning language. It joins the approved requirement, actual event, source evidence, rights and safety, data reliability, oversight duties and tested preventive action without converting a quality record into either blame or reassurance.


Useful next reading

Continue with writing a research ethics application, writing a laboratory method validation report, writing a laboratory notebook entry, explaining scientific evidence, and the How English Works.

Discover more from eduKate Singapore

Subscribe now to keep reading and get access to the full archive.

Continue reading