22

For my master's thesis, I completed a systematic literature review and supplemented it with surveys and patient interviews. The topics required no medical information of patients.

My supervisor was aware of both the surveys and interviews and ethics board was never brought up.

All participants completed a consent form, but I understand now an ethics review may be needed. My submission is soon ... is this an auto fail?

Alex T
  • 221
  • 2
  • 4
  • 4
    Please don’t write answers in comments. It bypasses our quality measures by not having voting (both up and down) available on comments, as well as having other problems detailed on meta. Comments are for clarifying and improving the question; please don’t use them for other purposes. Existing answers in comments and other discussion has been moved to chat. – cag51 Jun 30 '21 at 05:13
  • 3
    Two important clarifying questions for @AlexT: (1) How did you come to understand a review may be needed? Did your supervisor tell you or somebody else? (2) If you are comfortable sharing, what country are you in? – jakebeal Jun 30 '21 at 10:28

5 Answers5

32

I need to disagree with the hard-line answers saying that you are just plain out of luck. An ethics approval is not always required with human subjects research.

In the United States, DHHS maintains a set of flow charts for decision-making about human subjects research. Most relevant to this question, there is an exemption that specifically addresses "survey procedures, interview procedures", as long as the subjects are not identifiable or there is no reasonable risk associated with disclosure of identity. Similar exemptions may apply in other countries as well.

From what you have written, it sounds like your supervisor believed that such an exemption applied to the work that you are doing. Otherwise, from sheer self-protective reflex they would almost certainly have brought up institutional review processes.

I would suggest that you bring this up with your supervisor from this perspective.

  • If you can document that your work is already covered under an appropriate exemption and complies with your institution's processes, then your worries may be over and you can simply go on using your nice, consented data set. From what you have written, this sounds possible, though it is impossible to know without knowing the full details of your circumstances and your institution's regulations. Your advisor might even have gone through an institutional process without telling you, if it was just a routine exemption.
  • If the exemption does not apply, but there was a reasonable case to believe that it did when you began, then it may still be possible to bring it to the institutional review board as an emergent situation, again depending on the particular circumstances.
  • If both of these fail, you may indeed need to discard your data and begin anew, but your advisor should work with you to see if the situation can be salvaged first.
jakebeal
  • 187,714
  • 41
  • 655
  • 920
  • for sure ----- we might note some universities and publishers do still want you to file with the review board and get their official determination of exemption – Mike M Jun 29 '21 at 12:28
  • @MikeM Certainly: and we cannot know the details without knowing the institution. – jakebeal Jun 29 '21 at 12:58
  • +1 - it is important to properly implement ethical review, but this seems like a very good candidate for exemption. Collecting survey data that could have no foreseeable negative impact on the respondent even if linked back to their identity is often not a problem. It's easy (but incorrect) to skip the formal review when dealing with something that intuitively seems quite far from the realm of "unethical", but formally determining exemption in a retrospective manner can confirm the initial impressions on the ethics. – Nuclear Hoagie Jun 29 '21 at 14:36
  • 12
    Research that is "exempt" from these rules still needs to be submitted to an institutions IRB. The IRB determines that something is "exempt", not the PI. Now, it is entirely possible that the IRB will determine retroactively that the research done here is exempt and provide retroactive approval, but this is not guaranteed and, most importantly, it is not up to the PI to determine that whatever was done is "exempt". – Wolfgang Bangerth Jun 29 '21 at 15:42
  • 6
    @WolfgangBangerth That depends on the regulations of the institution. I have had an experience, for example, with an institution that needed to use contract IRBs rather than hosting its own "free" IRB, in which the policy was to determine if the exemption was obvious before deciding whether it was necessary to pay a contract IRB for assessment. Some things might also be assessed at the project scale rather than at the scale of individual studies. All of this boils down to the same base: it depends on specific regulations, and the professor needs to navigate those regulations. – jakebeal Jun 29 '21 at 16:09
  • 2
    OP suggests that a review was needed in their case. I suppose it's possible they are mistaken, but it seems more likely to me that even if their research is "exempt" it does fall on some sort of review process to determine that, whether that's a full IRB process or something truncated. It sounds like they're dealing with patients as well as providers; the former is more likely to require scrutiny. OP says there's no medical information, but also suggested in a comment that their research is related to a "treatment intervention" which suggests the survey might have been targeted to people... – Bryan Krause Jun 29 '21 at 18:31
  • 2
    ...based on medical information. I think "If you can document that your work is under an appropriate exemption, then your worries are over" is not necessarily true, because if the institution has a process for determining exemption that was not followed, then although it's not as bad as if it were not exempt, it's not simply "everything okay" either. – Bryan Krause Jun 29 '21 at 18:32
  • 2
    Not only is this wrong as @WolfgangBangerth has explained, we have a whole question about how it's wrong. https://academia.stackexchange.com/questions/143701/my-human-subjects-research-is-exempt-under-us-regulations-does-that-mean-i-ca – Anonymous Physicist Jun 30 '21 at 05:20
  • Also, I have never heard of an institution that had students which used outside IRB routinely. – Anonymous Physicist Jun 30 '21 at 05:23
  • I have added clarifications to the answer and made shifted the fail case from implicit to explicit. – jakebeal Jun 30 '21 at 14:04
  • We don't talk to patients unless we have privacy training. In fact, whoever identified the patients as eligible may have violated privacy policy – Scott Seidman Jun 30 '21 at 17:08
  • The situation strikes me as that of a mentor not used to doing human subjects research who just didn't know policy, more than an investigator who thought the work was exempt. It happens. – Scott Seidman Jun 30 '21 at 17:10
  • Jake, it seems like you're in engineering. Have you ever worked on humans subjects research? Or passed an IRB? I ask because your answer contradicts mine and Brian's, and I know we both do work in human subjects. – Azor Ahai -him- Jun 30 '21 at 17:27
  • @AzorAhai-him- Yes, I have dealt with this before. I actually think that your answer and mine are compatible --- the main difference is that I spell out how to try to salvage things if they are salvageable. – jakebeal Jun 30 '21 at 17:44
  • I am very unhappy with the edited answer. The question is clear that the asker should have applied for an exemption but did not. The first paragraph is highly misleading. This answer is much clearer: https://academia.stackexchange.com/a/170499/13240 – Anonymous Physicist Jun 30 '21 at 17:45
24

It is impossible to say because we don't know the rules of your institution. But the answer is it probably should be. You need to immediately tell your advisor and be prepared for bad news. I'm sorry your program failed you on this.

Azor Ahai -him-
  • 30,111
  • 9
  • 88
  • 115
  • 20
    I'm not inclined to agree with this. It's often possible to retrospectively get ethics board approval. Obviously, it's not ideal but as long as the end result is an approval, then no harm no foul. – ScottishTapWater Jun 29 '21 at 10:03
  • 7
    @persistence I disagree that it is often possible – Azor Ahai -him- Jun 29 '21 at 12:17
  • 3
    @persistence Can you link me to your institution's (or any's) process for retroactive approval? – Azor Ahai -him- Jun 29 '21 at 22:16
  • I was told retrospective approval is forbidden in the USA. I have never heard of a successful application for retrospective approval. – Anonymous Physicist Jun 30 '21 at 05:18
  • @AnonymousPhysicist I suppose it's not quite the same thing, but the in the hypocrite commits controversy, the researches apparently obtained an exemption from their IRB after the research was conducted (and the paper submitted to IEEE). – muru Jun 30 '21 at 07:49
  • @muru I do not find your source to be convincing support for Persistence's claim. But it's an interesting case. – Anonymous Physicist Jun 30 '21 at 07:57
  • 1
    @AnonymousPhysicist indeed, this case is very much not a case of "no harm, no foul" - I think the exemption made things worse for the researchers and the university too. – muru Jun 30 '21 at 08:13
8

If you were in a field that tends to do research with humans you probably would have taken some sort of course that explains to you research ethics procedures and such, and so you would have known this was a needed step. I had dozens of versions of this training as a student even though I was only doing animal research at that time. It seems like you have at least some familiarity with research ethics if you've gathered informed consent and such, but as you are now finding that is not enough.

Ethics reviews for research aren't an easy process, and you definitely need some mentorship your first time through, so it's a failure by your advisor to not shepherd you through things and at a minimum to let you know you need this training.

It's likely your work can't be published if it wasn't conducted under proper ethical procedures, but that doesn't necessarily impact your graduation. Whether you "pass" or "fail" is completely up to some academic unit - your advisor/committee/department. We won't be able to tell you what their decisions will be, you'll have to work with them. Start working with them as soon as possible to find out what your next steps are.

Good luck!

Bryan Krause
  • 114,149
  • 27
  • 331
  • 420
  • 4
    A lot of questions like this we get seem to often be people in non-(human|animal) departments who decided to collect a survey for their modeling or whatever. Perhaps that's what happened here. – Azor Ahai -him- Jun 28 '21 at 21:06
  • @AzorAhai-him- Yeah, could be. Likely that will influence the committee in deciding what's next for OP and could help them. I am kind of perplexed that those departments don't also push this sort of training, though; I'll try to find what the norms are at my institution for, say, CS students. – Bryan Krause Jun 28 '21 at 21:07
  • Thanks for the input! My program is a bit complex (joint degree between 4 universities). Unfortunately, there was no formal class on research ethics. My project is on the economic modelling of a treatment intervention, completed a SLR but the committee said it wasn't sufficient so then the qualitative component was added. – Alex T Jun 28 '21 at 22:49
  • 2
    @AlexT Not sure what SLR is, not an acronym I've heard before. Yes, that sounds like a circumstance (joint between 4 universities?!) where some things can be missed, but I'm very surprised you got anywhere near patients without a formal review. I think if they decide it wasn't your fault you're in better shape, but be upfront and don't try to hide anything. Talk to your advisor(s) sooner rather than later. Don't delay, and don't do anything to make results public. – Bryan Krause Jun 28 '21 at 22:55
  • 6
    downvoted for victim blaming in the first para. As OP has since said in the comments, your presumption of their getting teaching on this was incorrect, and even had they taken such courses these things end up being a question of culture and precedent in reality. Your second para is therefore the correct one, OP's supervisor failed to do their job. (If leaving ethics to guided precedent seems fraught I totally agree, but I guess it's not the supervisors' degrees on the line is it?) – benxyzzy Jun 29 '21 at 05:59
  • 1
    I am from computer science, and I can confirm that ERB is not part of our usual training, since most computer science graduates never do experiments with human subjects. I myself only learned about ERB after I won a grant that required ERB approval. Faculty members who do not win such grants might be totally unaware of what ERB is. – Erel Segal-Halevi Jun 29 '21 at 15:39
  • 4
    @benxyzzy I softened it a bit. You're right, this training may simply not be standard for the OP, I just find it really hard to believe someone could be in a field where they'd possibly be interviewing patients and not have university-required human subjects training. But the fault is possibly more with the universities and advisor. – Bryan Krause Jun 29 '21 at 16:05
  • @BryanKrause Agreed something like what AzorAhai-him- said must have happened. falling between the gaps. They hammer it home in the human-focussed fields I've experienced. When I've seen a similar failure to gain approval occur, they were coming into the lab from Computer Science – benxyzzy Jun 29 '21 at 16:30
  • 1
    At least for CS, one of the reasons departments don't push for this sort of training, is they really don't have room on the courses. There are already so many things in security, operations, networking, systems design, logic, set and group theory, hardware etc that departments struggle to teach in 4 years, that research ethics, something which is relevant to a tiny proportion of the students, really is down on the list of things that are missed. – user1937198 Jun 29 '21 at 22:00
  • 1
    @user1937198 I'm not necessarily thinking in terms of full-blown courses for credit, but rather training courses that take an hour or two at most that ensure people at least know the basics. They aren't nearly enough to get someone through the whole IRB process, but it's enough to get them to know that it exists and that they have to go there if they plan to do research with humans. – Bryan Krause Jun 29 '21 at 22:04
  • @BryanKrause Does your university have a system where students would take such courses if their advisors weren't aware of them and encourage them to take it outside the normal courses? Because at least under my experiance, they would just be ignored, unless the advisor was already aware, meaning no real improvement over the current situation. Unless its something required as part of a course, I doubt any students would take them. And then its competing with the hundreds of other things, like threat modelling, which would also like said hour. – user1937198 Jun 29 '21 at 22:10
  • @user1937198 It's required for everyone in certain departments/schools, and anyone receiving funds from NIH or NSF regardless of department. Required as in you get emails that start polite and then eventually get scary if you don't do it. As others have pointed out, students in areas that don't typically interact with people as research subjects might not get prompted to take them. – Bryan Krause Jun 29 '21 at 22:15
  • 3
    @BryanKrause 'I'm very surprised you got anywhere near patients without a formal review' This is actually quite important: there are understandable reasons why OP and OP's supervisor might have been unaware of the need for ethical review, but whoever was in possession of patients' contact details and handed them to OP really should have known, and should have demanded documentation showing the ethical review had been completed. – Daniel Hatton Jun 30 '21 at 11:18
  • 1
    @ErelSegal-Halevi it probably varies a lot per institution, and countries. I was in computer science and electrical engineering in Australia. I was given no training on Ethics Approvals, but all PhD students in the university were told about needing them in orientation, and require to submit write-ups on if they would need them (and details if so) as part of their PhD proposals. (and even as an undergrad in those coursesI was made aware of them as one of our professors had to get them before the class could have option to interview people at companies during projects on real world practices) – Frames Catherine White Jun 30 '21 at 20:07
1

Talk to your committee and your advisor. Submit to IRB now.

It sounds like you were genuinely unaware that your University's IRB board had to sign off on the experiment. Talk with your advisor and figure out the best way to submit to IRB after the fact.

You may be able to get after-the-fact approval. This is a situation where your advisor shoulders a good portion of the blame. It's their job to shepard you through the process, of which IRB is a part of.

The worst case is you'll have to leave the survey info out of any published work. If your field doesn't regularly involve humans as subjects in tests, it's possible your advisor genuinely didn't know about IRB approval. IRB boards are made up of individuals - they may be convinced this was an honest mistake.

Do not publish on this work without IRB approval

sevensevens
  • 7,461
  • 20
  • 39
-2

As stated, conducting research on humans without prior ethical approval is a grave transgression. The work becomes unpublishable because no journal would touch it (as a matter of universal policy) and you are open to formal censure.

However, this depends on nuances of your country, institute/department and possible categorisation of your work. Not all human research is ethically equal, and this is why a supervisor's guidance is crucial. Their academic well-being should be on the line over the activities of their mentees, and yours failed to do their job, but that's a rant for another time.

What follows does ultimately hinge on supportive engagement of your supervisor, so it is doubly dismaying to hear what sounds like distancing noises from them. If they are jettisoning this very professionally hazardous mess on to you then I am sorry, and I wish I could say it was rare.

In UK Psychology departments provision is often made for "generalised" ethical approval based off the Principal Investigator (your supervisor), or prior approved research from which the present does not materially deviate in ethical terms. There might well be rules which say no, or only cursory further approval is required for a study which essentially extends a previously approved study, i.e. by deploying the same methods to study the same topic/cohort such that no new ethical considerations are raised. If your supervisor is lucky (these days), they might have wangled blanket ethical approval for a whole class of prospective experiments on this rough basis: same kinds of methods on the same wholly mundane participants for roughly the same ends. Some kind of special remit is also typically extended to final year undergraduate project students, given there are so many and their projects are often trivial recapitulations of wholly harmless designs reused yearly by the supervisor in this way. (Note that those are usually very strongly stipulated conditions of being able to go this quicker route, which still entails prior ethical approval just through a faster cut-down process, and a signed statement of taking responsibility from said supervisor).

Ultimately then, the slow and onerous ethics committee approval process will probably include allowances to make specified cases of especially low-risk research more expedient, and although the bounds of these are very tightly specified because of the gravity of ethical risk, many PIs are wont to exploit the letter of the law (academics are professional arguers after all). As I've said they typically do so for two reasons: to circumvent the sheer onerousness and slowness of the full approval process, or to retrospectively render disorganised and messily planned research ethically approved. I have seen these rules creatively interpreted in the service of either goal far more often than those writing such rules would like.

Your case might fall into the second category, and it should be clear here why the benevolence and alliance of your supervisor is key. I have seen supervisors protect their juniors (and their findings) from the mess that you are in, following similar disorganisation and poor mentorship, by finding a way to frame the research conducted as coming under the remit of previously granted ethical approval. Perhaps this rogue study can in fact be argued as a mere extension of a previously approved study. Perhaps this study's ethical ramifications are sufficiently trivial that it could fall under some aforementioned triviality type allowance. Significantly, all of this is meant to be obviated by the iron imperative to resolve any such approval ambiguity before conducting the research - but perhaps there are only one or two, non- or back-dated bits of paperwork that need be collected to bring the study back into defensible territory.

None of this is at all on the table unless the supervisor wants it to be however, and if they don't, the mere suggestion of this to them will be responded to as a grave professional hazard. It sounds unfortunately like your supervisor specifically might prefer what is now smeared on you, to remain smeared on you alone.

I saw this kind of thing happen as the not-unusual solution to a full approval process being too long for certain deadlines, valuable juniors (or their findings) being jeopardised by the kind of situation you've been allowed to get into, other forms of craven programme disorganisation, or combinations thereof. Is this whole thing very uncomfortable if not inappropriate? I invite you to read some other responses on this site regarding how much inappropriateness is forgiven, when it is not a mere junior in the firing line.

benxyzzy
  • 743
  • 4
  • 11
  • Where are you seeing an indication of "distancing" or "jettisoning" by OP's supervisor? – nanoman Jun 29 '21 at 10:22
  • @nanoman you're right, I have read into that perhaps too far. I am referring to OP's reply to CaptainEmacs 's question on the original question. It sounds like the supervisor left a big gap in their otherwise clear instructions to OP, and I read (perhaps too far) into that a subsequent reluctance of supervisor to fix their mess. It sounds like a stark if not sustained mismatch between OP's and supervisor's expectations here, as if some washing of hands might be occurring. Mainly I was trying to emphasise the risk of approaching the supervisor with request to bend rules, but I could reword – benxyzzy Jun 29 '21 at 10:36
  • Sorry, I had to DV, because the blanket statement in the first sentence is just not generally true. – henning Jun 29 '21 at 11:17
  • @henning I am sure I will get downvoted for various actual reasons, but please do give an example of human research without ethical approval being non-transgressive – benxyzzy Jun 29 '21 at 12:57
  • 2
    @benxyzzy surveys, expert interviews, partcipant observation, if no sensitive personal data is collected and the data is pseudonymized/anonymized. In my country, ethics board approval is not required in these cases. – henning Jun 29 '21 at 13:02
  • 2
    @henning Then I guess our countries take very different approaches to ethics, because all of those things would count here ("participant observation"??) Juniors - masters, PhDs, undergraduate final year project students - have the importance of ethical approval and the liability of breaching it heavily impressed upon them, in formal mandatory workshops from members of the dept approval board. This is probably why OP is so concerned, and why several other answers confirm the potential seriousness. My answer describes how it can actually play out with the benefit of a PI's power and latitude. – benxyzzy Jun 29 '21 at 13:20
  • 1
    @benxyzzy I suppose this is another instance of https://academia.meta.stackexchange.com/questions/4471/academia-varies-more-than-you-think-it-does-the-movie And with that caveat, i found your answer quite useful. – henning Jun 29 '21 at 13:24
  • 5
    If OP's supervisor trained in an environment like the one @henning is used to, and is now working in an environment like the one benxyzzy is used to, that could be the origin of the problem. – Daniel Hatton Jun 29 '21 at 16:57