01 · The Question
What if the Study You Are Conducting Is No Longer Quite the Study That Was Approved?
No one deliberately redesigned the study. Recruitment was simply difficult, so the team added another channel. Interviews regularly ran longer than expected, so everyone stopped worrying about the approved duration. A few questions proved unhelpful, so interviewers began skipping them. Other questions were added informally because they produced interesting answers.
Each adjustment may have seemed reasonable on its own. Months later, however, the actual study can look noticeably different from the approved protocol.
This kind of gradual divergence is especially easy to miss because there may be no single dramatic event. The problem emerges cumulatively: informal workarounds become routine, exceptions become standard practice, and the protocol gradually stops describing the research participants are actually experiencing.
03 · What You Need to Know
Protocol Drift Usually Develops Through Accumulated Small Departures
Protocol Drift Is a Practical Description, Not a Universal Regulatory Category
“Protocol drift” is a useful way to describe a study that has gradually moved away from its approved procedures, but researchers should not assume that every ethics committee or regulator formally uses that term.
The individual events may instead be classified as protocol deviations, noncompliance, unauthorized modifications, or other categories under the framework governing the study. What matters is the underlying condition: approved research and actual research are no longer aligned.
For clinical trials, ICH E6(R3) states that investigators should comply with the protocol, Good Clinical Practice, and applicable regulatory requirements, document protocol deviations, review deviations relevant to their conduct of the trial, and take appropriate measures to prevent recurrence of important deviations where applicable.
Drift Often Begins With Reasonable-Sounding Workarounds
Most drift does not begin with a research team announcing that it intends to ignore the approved protocol. More commonly, people adapt to practical difficulties.
An interview takes longer than planned. A recruitment method produces too few participants. A survey question confuses respondents. One study site develops a slightly different workflow. Staff begin accepting documents a few days outside the required window because doing so seems harmless.
A single departure may be manageable as an isolated deviation. The warning sign appears when the workaround becomes the normal way the study operates.
Repeated Deviations Can Reveal That the Protocol and Reality No Longer Fit
Suppose the protocol requires interviews to last 30 to 45 minutes, but almost every interview lasts 75 minutes. Recording 25 separate “long interview” deviations without asking why they keep happening misses the larger problem.
Perhaps the interview guide is too long. Perhaps the approved duration was unrealistic. Perhaps interviewers are systematically adding questions. Each explanation points to a different corrective response.
ICH E6(R3) takes a similarly systemic approach to significant noncompliance in clinical trials. It calls for root-cause analysis, appropriate corrective and preventive actions, and confirmation that those actions are adequate when noncompliance significantly affects or could significantly affect participant rights, safety, well-being, or reliability of results.
Drift Can Occur in Recruitment
A study approved to recruit through institutional email may gradually begin using social-media posts, personal referrals, direct messages, or recruitment through supervisors because the original strategy is slow.
The target population may remain unchanged, but privacy, voluntariness, recruitment materials, and methods of first contact may no longer match what was approved.
If this has occurred, reconstruct which methods were actually used and when. Then determine whether the team has accumulated recruitment deviations and whether a prospective change to the recruitment strategy is appropriate.
Drift Can Occur in Data Collection
Researchers may begin dropping questions that “never produce anything useful,” adding informal probes, changing response options, collecting additional variables, or administering an instrument differently across participants.
These changes can alter participant burden, data comparability, sensitivity of collected information, and the study's scientific design. If an interview guide or survey has evolved substantially in practice, compare what was actually administered with the approved research instrument.
Drift Can Occur Across Research Sites
Multisite studies are particularly vulnerable because local adaptations can accumulate independently. One site may use a different recruitment script. Another may interpret eligibility criteria differently. A third may store or transfer data using a locally convenient workflow.
The protocol may still appear uniform on paper while participants receive materially different procedures across sites.
This is one reason monitoring should look beyond whether required documents exist. It should ask whether actual research conduct corresponds to those documents.
Drift Can Affect Eligibility
Perhaps one eligibility criterion proves difficult to apply, so staff begin interpreting it more loosely. Over time, the operational definition used by the team may become different from the approved criterion.
If people have already been enrolled under the informal interpretation, determine whether participants who did not meet the approved eligibility criteria have entered the study. Do not simply amend the wording retrospectively and treat the earlier enrollments as though they had always complied.
Drift Can Occur Through Document-Version Problems
A protocol amendment may be approved centrally while one site continues using an old consent form or questionnaire. Researchers may believe the study has changed correctly because approval was obtained, yet actual implementation remains inconsistent.
Approval of a revised document is therefore only part of change control. The research team must also ensure that superseded versions leave active use. If they do not, the problem becomes one of using an incorrect approved document version.
Protocol Drift Can Be Methodological as Well as Ethical
Repeated informal adaptations can make the final study scientifically difficult to interpret. If participants receive different instruments, recruitment procedures, interventions, follow-up schedules, or eligibility decisions, the study may no longer represent a consistent implementation of its intended design.
The concern is therefore not merely administrative compliance. Scientific validity can be affected when the study described in the methods section is not the study that produced the data.
For clinical trials, ICH E6(R3) explicitly connects serious noncompliance not only with participant rights, safety, and well-being but also with the reliability of trial results.
The First Response Is to Map the Drift
When cumulative divergence is suspected, avoid correcting only the most recently discovered event. Reconstruct what has actually happened.
Compare the approved protocol with current practice across recruitment, consent, eligibility, interventions or procedures, instruments, visit timing, data collection, data handling, study sites, and analysis where relevant. Identify when each divergence began and which participants or records were affected.
This gap analysis turns a vague concern that “we have not been following the protocol exactly” into an assessable set of facts.
Past Deviations and Future Amendments Must Be Kept Separate
Once the team recognizes that current practice may actually be preferable, it can propose a permanent change. But the sequence matters.
Conduct that has already departed from the approved protocol should be documented and handled according to applicable deviation or noncompliance procedures. The desired future procedure can then be submitted through the appropriate amendment process.
The distinction between a protocol deviation and a protocol amendment is especially important here. Approval of the revised procedure does not retrospectively transform earlier unauthorized conduct into approved conduct.
Corrective Action Should Address the Cause, Not Just the Symptoms
If researchers keep deviating because the protocol is impractical, retraining everyone to follow an impractical procedure may produce another stack of deviations next month. If the protocol is workable but staff do not understand it, an amendment may be unnecessary while training and supervision need improvement.
Root-cause analysis asks why the divergence developed. Possible causes include ambiguous procedures, inadequate training, poor document control, unrealistic timelines, inconsistent supervision, local site practices, software limitations, or a protocol that genuinely needs revision.
The corrective response should match the cause.
Watch Out
Do not “clean up” protocol drift by editing study records so that actual conduct appears to match the approved protocol. Preserve the history accurately. Correct the study going forward, report past departures where required, and amend the protocol prospectively when a permanent change is justified.
04 · A Practical Example
When Small Interview Adjustments Accumulate Into a Different Study
Hypothetical Example
An approved interview study gradually changes over six months
A qualitative study is approved to conduct 45-minute interviews with university faculty using a 12-question interview guide. Over time, interviewers routinely omit three questions they consider repetitive, add several questions about workplace conflict, allow interviews to run for 90 minutes, and begin recruiting through participant referrals even though referral recruitment was not part of the approved plan.
Recognize the pattern
The issue is no longer one unusually long interview or one spontaneous follow-up question. Several aspects of routine study conduct differ from the approved research.
Map the differences
The team compares actual recruitment, interview content, duration, and participant experience with the approved protocol and documents.
Identify affected participants
Researchers establish when each practice began and which interviews or participants were affected.
Assess consequences
Participant burden, consent information, confidentiality, recruitment privacy, data comparability, and scientific implications are evaluated.
Address past conduct
The team follows applicable deviation and reporting procedures for what has already occurred.
Decide future conduct
If the revised approach is scientifically and ethically justified, the team submits the necessary prospective amendments before making it the authorized study procedure.
The useful question is no longer “Was that one extra probe allowed?” It is whether the cumulative adaptations have produced a materially different research experience from the one the ethics committee reviewed.
