This blog is closed to new posts due to inactivity. The post remains here as part of the network’s archive of useful research information. We hope you'll join the conversation by posting to an open topic or starting a new one.
I have a case like this: The protocol requests patients to come back for F/U visits on D30, D45 and D60. One patient missed the visit on D30 but came back on D45 and D60. It seems in many of our researches, it could be considered as a protocol deviation or violation depending on its consequences. However, some says this event occurs but not due to a fault of study staff so does not meet the definition (A protocol violation occurs whenever a study staff person performs any action that does not adhere to the research protocol.) so cannot be considered as protocol deviation/violation. It seems the definition of protocol violation/deviation is not widely understood and accepted. Please share your advices! Thanks a lot, colleagues!
Well, I agree that a lot is said on violation / deviation (sometimes preceded with minor or major)... so here's my understanding.
First, I'd say that deviation = violation, to start with. I prefer the term deviation as it is less aggressive and scary. And also it reflects that there's a deviation from the study protocol.
However, and this is where minor and major may step in: it is important to know (or guestimate) how the deviation will plausibly affect the measurement of what is being studied.
Ideally this should be mentioned in the study protocol along with the implication during the statistical analysis.
Example: (from malaria efficacy trials) an anaemis slighly below the inclusion criteria can be considered a minor deviation as it should not affect the drug efficacy ==> the patient can be analysed as a normal participant.
Other example: a patient takes a powerful antimalarial drug during follow-up is a major deviation as we can't judge the efficacy of the investigational product anymore. This patient should not be analysed like other participants (major deviation from study protocol).
So... is a D30 missed visit a minor or major deviation (your question)? It's up to the Protocol Development Team to answer to this question in relation to the effect it can have on the patient's efficacy outcome.
In summary, the effect of a deviation on the efficacy outcome defines if it is minor or major (and hence how it should handled during analysis of the primary objective).
Just an opnion to feed the debate.
All the best!
Thank you for sharing this piece and indeed it is one that can be confusing at times. Allow me to share my perspectives to this discussion point.
Firstly, there is a very fine line between protocol violation and deviation. It often depends on specific protocols which should define what constitutes violation or deviation. I do not think in general that the definitions have to do with whether it is the investigator or participant who has to have done something out side protocol. But as you rightly say, it is the implications of the event that matter. In general violations tend to be ones whose consequence is removal of the subject or particular variables from protocol analysis. Deviations would then be "lesser crimes" if you like.
For example, if your protocol has its primary outcome specifically derived from day Day 30; then missing that visit may constitute a violation of the protocol schedule and therefore the subject may not be evaluable. But it may well be a routine visit say for safety follow up and missing the day may only constitute "missing variables" in data and therefore could be deemed a deviation.
So in summary, to the best of my knowledge and experience, there is NO fixed definition for deviation or violations; they are protocol specific issues.
In the UK we have some guidance on this issue from our regulatory authority (the MHRA). They accept that minor deviations occur frequently and should be documented as they may affect analysis (e.g. a participant comes to clinic on day 9 rather than day 7) but do not provide any cause for concern.
The MHRA are much more concerned with serious breaches, which may or may not be deliberate, but: "... is likely to effect to a significant degree:
(a) the safety or physical or mental integrity of the subjects of the trial; or
(b) the scientific value of the trial."
See http://www.mhra.gov.uk/home/groups/is-insp/documents/websiteresources/con060111.pdf for more detail and examples.
While this is UK-only it does provide a sensible recognition that there is a big difference between a single missed blood sample and deliberate fraud or lack of informed consent.
Thank you all so much! That's really helpful!
Thanks for sharing experiences, It is the scenario which I also had to discuss several time with the clinical staff during my monitoring visits, and I adhere completely to the explanations of the collegues above.
Good Day Tran,
Thank you for sharing.
I think sometimes as researchers we see protocol deviation as a certificate of guilt or blame. In reality it is simply a method to document the reason/s that protocol did not hold for a certain case.
Of course this does not imply that anything goes just so long as the paper trail is complete however, I would add, when in doubt document and submit. Rarely, if ever, have I been in a situation where I was told not to document or report an incident.
I know this advice does not hold perfectly as you asked specifically if the event should be documented as a "protocol deviation" (not to say that it would not be documented in another manner). It is a case where the protocol did not hold and should be noted though corrective action on your part would be difficult, if it is even possible.
The powers that write the protocol should know to write in procedures in the event that D30 is missed in the future.
My humble opinion
At your service,
Hi Tran, thanks for initiating an interesting discussion. I also take the view that protocol deviation=protocol violation though the latter has a more serious connotation thus you find this term used less often. PDs/PVs can result from commission e.g. taking more blood than required by protocol and stated in consent form, or omission e.g. a procedure/test not done.
The sponsor usually provides guidance on how these events should be captured and reported.
Well done and thank you everyone for contributing to this. This is exactly the type of exchange that Global Health Trials was set up to encourage. Did you know that we have over 40,000 using this now? Most people read and don't post (we hope this sort of thing encourages more people to get involved!). So you should know that 1000's of trialists will read this exchange and learn from it. So we really appreciate you taking the time and sharing your experience. We need more of these real examples so please keep them coming! As this is such a good exchange we would like to turn this into an article.So we will be writing to you all and requesting you as authors for this and turning this into a formal summary of the problem and your collective opinions on the solution.