Why the Distinction Matters
An adaptive trial permits specified changes as the trial proceeds, but “adaptive” does not mean that trial conduct can be changed informally whenever circumstances suggest doing so. The distinction between a planned adaptation and a protocol deviation depends on what was specified and authorized in advance, what action occurred, and which approved provisions applied at that time. The same action—for example, pausing enrollment—could be a planned adaptation in one trial and a deviation in another.
The practical question is not simply whether a change was sensible. It is whether the action was allowed by the applicable trial documents and followed the stated decision and authorization process. A departure can be inadvertent and still be a deviation; conversely, a properly executed adaptation is not a deviation merely because it changes the trial’s conduct.
The governing details may be set out in the protocol itself or in a controlled document that the protocol properly incorporates or references. The exact document structure and reporting terminology vary among institutions, sponsors, and regulators. For a trial action that changes participant-facing procedures or other protocol requirements, a separate analysis document alone should not be assumed to authorize the change. The applicable governance process determines what documentation and approvals are needed.
Classify the Action Against the Rules in Force
A sound classification separates three questions. First, was the possibility of this change specified before the decision was made? Second, did the situation meet the stated conditions and was the action within the permitted options? Third, did the person or body with the required authority make and implement the decision under the applicable process? If any required element is missing, the action may not qualify as execution of the planned adaptation.
For a particular action, identify when it actually occurred and the protocol version then in force. Apply any explicit transition provision in an amendment, such as a rule about assessments performed after an effective date or visits scheduled before it. The scheduled date, the decision date, and the date the action occurred can differ; they should not be treated as interchangeable. If the transition rule or relevant dates are not provided, the classification may be undetermined rather than safely inferred.
State what was done, to whom or what it applied, and when it occurred. Avoid classifying a general idea such as “the trial changed.”
Identify the approved protocol version, any incorporated controlled documents, their effective dates, and any transition rules relevant to this action.
Determine whether the action was prospectively permitted, whether its conditions were met, and whether the required decision-maker and process were used.
A compliant action under a planned rule is an implemented adaptation. A departure from the applicable requirements is a deviation. If key timing, documentation, or authorization facts are missing, record that the classification is unresolved.
Worked Scenarios
Worked Example: A Specified Pause Is Triggered
Hypothetical trial. The approved protocol states that if a predefined safety review criterion is met, enrollment will pause while the designated independent committee reviews the information. It specifies who confirms that the criterion has been met and who is authorized to implement the pause. The criterion is met, confirmation and authorization are documented, and enrollment is paused as specified.
Classification. This is implementation of a planned adaptation, not a protocol deviation. The protocol prospectively described the condition, the permitted action, and the required process. The stated condition occurred and the authorized process was followed.
Why the details matter. If the pause had been initiated by an unauthorized person, before the specified condition was met, or for a different reason not covered by the protocol, the facts would no longer establish compliant implementation of this planned rule. The action might then constitute a deviation or require another classification under the trial’s governance procedures. Calling any pause “safety-related” would not, by itself, make it an authorized planned adaptation.
Worked Example: An Assessment Outside the Current Window
Hypothetical trial. The protocol in force requires an assessment within 14 days after a scheduled visit. The scheduled visit is 20 August, and the assessment is performed 18 days later, on 7 September. No exception or amendment is stated.
Calculation of the assessment date. Eighteen days after 20 August is 7 September. The assessment is therefore 18 days after the scheduled visit, which is 4 days beyond the 14-day limit: \(18-14=4\).
Classification on the stated facts. If the 14-day requirement remains applicable to this assessment and no other provision authorizes it, the assessment is outside the required window and is a protocol deviation. This conclusion follows from the date the assessment occurred relative to the stated limit, not just from the scheduled-visit date.
Important qualification. If an amendment took effect on 1 September and explicitly extended the window to 21 days for assessments performed on or after that date, the 7 September assessment would fall within that amended window: \(18\leq21\). If instead the amendment said that visits scheduled before 1 September remained subject to the 14-day limit, the assessment would still be outside the applicable window. Without an applicable transition rule, or other facts establishing which provision governs, a confident classification is not justified.
Worked Example: A Helpful Change Made Without a Planned Rule
Hypothetical trial. A site coordinator believes that participants would find an extra check-in helpful and begins offering it to some participants. The current approved protocol does not allow this check-in, no prospective rule describes it as an adaptation, and no required amendment or authorization has been completed.
Classification. The extra check-in is not a planned adaptation on these facts: it was not prospectively specified and authorized through the applicable process. If offering it departs from the protocol requirements in force, it is a protocol deviation. Whether it must be reported under a particular category, and what corrective action is required, depend on the trial’s procedures and applicable oversight requirements.
What would change the classification? If the check-in were first assessed through the required governance process, prospectively documented in an approved amendment or other properly controlled document, and implemented only once its terms became effective, actions within those terms could be compliant conduct under the amended requirements. A later amendment does not automatically establish that an earlier action was authorized; its effective date and transition provisions must be checked.
A Brief Justification for the Decision Process
The classification follows from comparing an observed action with the requirements that governed it. Let \(a\) denote the action, \(t\) its occurrence time, and \(P(t)\) the approved protocol provisions applicable at that time, including any valid transition rules. If \(a\) satisfies the applicable requirements, it is not a deviation on that basis. If it fails a requirement that applied at \(t\), it is a departure and should be assessed as a possible deviation. Whether that action is a planned adaptation requires a further check: the relevant change must have been specified prospectively and carried out under the required conditions and authority.
This is a classification framework, not a claim that every operational detail has one universal regulatory label. Trial procedures define how deviations are documented and assessed. Nor does the framework imply that a departure is automatically serious: significance and reporting obligations require a separate assessment under the applicable procedures. The central justification is narrower: calling an action an adaptation cannot replace checking whether it was allowed and properly implemented.
Documentation, Diagnostics, and Common Pitfalls
Good records make the classification auditable. They should allow a reviewer to identify the action and date, the relevant protocol version, any applicable amendment and transition language, the prespecified condition and permitted action, and the decision and authorization record. If the action departed from requirements, the record should also follow the trial’s process for documenting and assessing deviations. Missing documentation can make it impossible to establish whether a planned rule was followed; absence of evidence is not proof that an action was compliant.
- Calling an improvised change an adaptation. Good intentions or a reasonable clinical judgment do not establish prospective specification. Check the controlled documents and authorization process.
- Classifying by the wrong date. A scheduled visit date may not determine which rule governs an assessment performed later. Apply the amendment’s transition terms to the date and type of action they address.
- Assuming every amendment applies to all pending events. An amendment’s effective date and transition language govern its reach. If those terms are absent or ambiguous, do not invent a rule.
- Treating every deviation as misconduct. A deviation is a classification of conduct against requirements, not a conclusion about intent. It may be inadvertent; its seriousness and implications must be assessed separately.
- Assuming a deviation invalidates the whole trial. A departure requires appropriate documentation and assessment, but its effect depends on the facts and applicable standards. It does not by itself establish that the trial’s results are invalid.
An urgent action taken to address an immediate safety concern may be governed by special procedures. Those procedures, including any required notification, documentation, and subsequent review, should be checked rather than assumed. Urgency does not justify relabeling an improvised action as a prespecified adaptation.
How This Classification Differs from Other Approaches
This approach is different from judging a change only by its scientific rationale or apparent benefit. A rationale can explain why a change was considered, but it does not show that the change was permitted or authorized. It also differs from classifying solely by intention: an action meant to help may still depart from the protocol, while a change in conduct may be fully compliant when it follows a specified rule.
The main practical alternative is a case-by-case judgment made after the action, without a documented rule. Such judgment may be necessary to assess an unexpected event, but it cannot create prospective planning retroactively. A planned adaptation offers a traceable basis for action; an unplanned change must instead be handled through the applicable deviation, amendment, safety, or oversight processes. The exact route depends on what changed and the trial’s governing procedures.
Check Your Understanding
For each situation, identify what additional fact, if any, is needed and give the best-supported classification.
- A protocol prospectively permits a specified change after a defined condition is confirmed by an authorized reviewer. The condition is confirmed and the authorized reviewer directs the change as documented. Is this a planned adaptation or a deviation?
- An assessment is required within 10 days of a scheduled visit. It occurs 12 days after that visit. No exception or amendment is given. What is the classification, and by how many days was the limit exceeded?
- An amendment takes effect on 1 October and states that a revised procedure applies to actions performed on or after that date. A participant’s action occurs on 3 October, although the related visit was scheduled for 28 September. What governs the classification?
- A site begins a new procedure on 30 September. An amendment allowing it takes effect on 1 October, with no stated provision authorizing earlier use. Does the later amendment by itself establish that the 30 September action was a planned adaptation?
- A coordinator inadvertently misses a protocol requirement. Does the word “deviation” alone establish that the coordinator acted improperly or that the trial results are invalid?
Solutions. (1) It is implementation of a planned adaptation: the condition, authorization, and specified action are all present. (2) It is outside the stated window and is a potential protocol deviation; \(12-10=2\), so it is 2 days late. (3) The stated transition rule applies to actions performed on or after 1 October. Since the action occurred on 3 October, assess it under the amended procedure; the earlier scheduled date does not override that explicit rule. (4) No. The action occurred before the amendment’s effective date, and the stated facts do not authorize earlier use. Assess it against the provisions in force on 30 September; if it departed from them, it may be a deviation. (5) No. A deviation describes a departure from requirements, not intent or seriousness. The trial’s procedures must guide assessment of the event and its implications.