Generally speaking, peer reviewed journals will not publish research on human subjects that did not receive IRB approval.

If a study was done without the proper IRB approval, but the study itself was ethical and clearly would have passed IRB approval if it had been properly submitted, is it generally possible to get a post-facto IRB approval that is sufficient to publish in most journals, or is the work forever "tainted"? Is the fact that post-facto IRB approval is not often considered an option because it literally isn't allowed, or because research conducted without IRB approval almost always contains ethical infirmities that would have prevented IRB approval from the start? In other words, is the requirement to obtain IRB approval truly a "You absolutely must have prior IRB approval", or is it really more like "You absolutely must not conduct research unethically, and obtaining prior IRB approval is part of Best Practices in accomplishing this, because nearly all proposals that you or I might come up with are actually unethical and we need the wisdom of the IRB to show us the light and put us on the right path. If you forgo prior approval, you are wading in crocodile-infested waters, and you will have no one to blame but yourself, not the IRB, not your mom, not your advisor, but only you yourself, if you get your leg bitten off"?

To be clear, I'm not asking whether it is possible to "whitewash" unethical research to make it look like no one was harmed when in fact people were harmed, but about institutional procedural violations in otherwise ethical research. I am well aware that one of the primary reasons for seeking prior IRB approval (other than because journals expect it) is to avoid committing unethical acts, so for the purpose of this question, assume that the researchers either just get lucky or are just so competent in ethical practice that their actual research demonstrates sublime virtue (just without IRB approval).

Clarification: Some people have seen fit to warn me that I am in danger of losing my job. I am not currently facing this issue. I'm simply curious as to what would happen or what typically happens in cases like these - that is, whether there are Best Practices on how to mitigate the fallout or even how serious it truly is - whether this is more of a "Calm down, it happens all the time, all you have to do is write an apology letter, sit through a five hour 'naughty, naughty bad boy, sit in the corner' lecture, be ritually spanked on the rear by your department head or the Dean, accept a 5% penalty pay cut for six months, and go through a retrospective IRB review." or a "Your career is doomed, don't dream of ever teaching or doing research ever again, you will be lucky if you ever find work cleaning floors."

  • 16
    It might be relevant why IRB approval wasn't obtained. Did the researchers believe, reasonably and in good faith, that it wasn't needed? Did they try to get it, but unforeseen circumstances prevented the IRB from considering the case or giving approval? Or is it the case that they knew, or should have known, that it was required, but intentionally went ahead without it? – Nate Eldredge Oct 17 at 2:20
  • 15
    Yes it is, but that is not your issue. At this point you should not be worried about publishing the work, you should be worried about losing your job and/or PI privileges. – StrongBad Oct 17 at 2:45
  • 2
    Right- the common rule doesn't allow an IRB to give approval after the fact. – Brian Borchers Oct 17 at 2:58
  • 7
    Can we please drop the "you're going to lose your job" talk and stop speculating on what the OP's circumstances are? I received similar comments once, and it's upsetting and feels somewhat threatening. – Rebecca J. Stones Oct 17 at 4:35
  • 14
    @RebeccaJ.Stones your question was about wanting to make sure you did not break the rules, so suggesting you could get fired if you break the rules, seems unfair. Here, the rules have been broken and the OP mentions one relatively minor penalty. I thought it was important to point out there is a much larger potential penalty for breaking the rules. – StrongBad Oct 17 at 13:35

The human research ethics committee is in place to ensure that human research is undertaken in a way that safeguards the dignity, rights, health, safety and privacy of those involved. This assessment is difficult to make by the investigator because he or she has a vested interest in proceeding. Thus, an independent body is required.

In the institutions in which I've been involved, conducting human research without ethical review is, quite simply, research misconduct and subject to investigations and disciplinary actions. A priori ethical review is required. Regulations are in place that specifically disallow approval of research that has already been conducted.

I've been a member of institutional review boards and have chaired three -- one based in a hospital and two in research-intensive universities. I've been involved in these investigations in the past. As the original poster implies, most are found at the point in which the journal requires evidence from the IRB and the author quickly shoot an email to the office requesting one. In these investigations, it is often the case that there is a lack of training in the expectations and regulations of human research and offenders usually fall in three categories:

  • inexperienced researchers, including students (often poorly supervised) or practitioners with no research training
  • researchers trained "elsewhere", including those from countries with little to no requirements for ethical review
  • researchers from fields that don't usually involve human subjects

My worst case was a person from an electrical engineering background who collected data on body image for an app he was developing. During these sessions, the person gave advice that drove one participant to self harm. During the investigation, the offender did not think that ethical review was necessary because the questions were "innocuous (his term)".

As I said, in most cases the consequences are far less severe than this. For example, a common scenario is a young doctor who collects information from patient files, performs a few statistical tests and write up the results of this audit and sends the publication to a journal that then requests evidence of IRB approval. In cases like this, IRB approval is not provided and the doctor is placed under supervision.

  • 3
    This is a great answer, but it leaves out non-research human usage which can lead to data that are used in retrospective studies, which is like getting approval for research that has already been conducted. – StrongBad Oct 17 at 13:08
  • 2
    I'm struggling to see how this answers the question in the title. – thosphor Oct 17 at 15:48
  • 10
    @thosphor First, the question in the title is just intended as a summary. We should be answering the question, not the title. Second, the second paragraph is quite clear that IRB approval cannot be granted after the research has been done, according to the regulations of the multiple institutions that St. Inkbug has worked on IRBs for. How is that not an answer to the title question? – David Richerby Oct 17 at 16:21
  • Your example about the young doctor is actually a borderline case, at least in the US. If the existing information collected is public or not personally identifiable, i.e. the records were anonymized, then no IRB approval is necessary. (45 CFR 46.101(b)(4)) – user71659 Oct 17 at 17:49
  • 8
    @user71659 I would note at many institutions (including every one I have ever been at), making that determination is the IRB's prerogative, not the researcher's. – Fomite Oct 17 at 22:57

As StrongBad has noted, publication is the least of one's worries in that particular context - you're talking about the kind of thing that people get fired for.

There's really no mechanism for "Oh, we totally would have approved this" post-hoc IRB approval. Indeed, there's an argument to be made that conducting the research without an IRB approval is inherently ethically suspect - in the same way that driving without a seatbelt on isn't only dangerous if you get in a car wreck. You are skipping the step meant to protect study subjects. That alone is a risk.

At this point in your scenario, the only real option is to seek the guidance of the IRB, and see what paths are available. This is absolutely one of those "This gets worse the longer it goes" situations.

  • 4
    "Indeed, there's an argument to be made that conducting the research without an IRB approval is inherently ethically suspect" That has been bugging me after reading the question. You can claim a study is ethical all you want, but if you're trying to get IRB approval after the fact, regardless of whether you would've got it had you done things in proper order, the study is suspicious already. It's done. – Mast Oct 18 at 7:05

The main purpose of an IRB is to protect human research subjects. If you are doing research, even a seemingly harmless questionnaire needs approval. If you are not doing research, you can give humans that same questionnaire (and a lot worse). IRBs do not have a role in the use of humans in improving "process".

If you want to make your students happy, a department can give a questionnaire, swap out the water cooler with coffee and sugar and heavy metal music, repeat the questionnaire, then swap out with decaf coffee, artificial sweeteners, and classical music, and repeat the questionnaire again. This is a lot like a research project with key differences often being the lack of a randomized control group.

Under the guise of improved process, you can get away with a lot of things as long as you do not violate any laws or OSHA regulations. Companies do in house marketing studies all the time (do people like red or green or sweet or savory food, is this device too loud/quiet). As it is not research, it does not fall under the remit of an IRB.

Sometimes the data collected during this process evaluation look publishable. You might be able to publish them straight up, sometimes a journal will ask for IRB approval. In this case an IRB might grant you approval for a retrospective study. If the data however are deidentified, do not include any PHI/PII or biological, it might not be considered human use, and therefore the IRB would not have a role to play. The IRB might also decide that your process evaluation was in fact research. At that point they may decide you engaged in research misconduct. If that is the case, you could be in for a world of hurt.

  • 2
    "This is a lot like a research project with key differences often being the lack of a randomized control group." Are you suggesting a lack of a control group indicates IRB approval is not required? A bit confused by that whole paragraph actually. – Azor Ahai Oct 17 at 17:05
  • The common rule includes an exception for research based on educational records that have already been collected for educational purposes (stuff like grades, course evaluation surveys, etc. could be used in this way.) At the point that you decide you want to use the data for research rather than program assessment/improvement, you can file an IRB request asking for permission to use this already collected data in research. The reasoning behind this is that the data would have been collected anyway for the educational purpose and that collection would have been exempt from the common rule. – Brian Borchers Oct 17 at 17:12
  • 6
    @AzorAhai it is confusing because the definition of what is research is confusing. Two people can do the exact same thing to a set of individuals and in one case it is research and in the other it is not. A teacher giving an algebra test and recording the scores in a spread sheet and calculating the mean and standard deviation is not doing research. A developmental psychologist doing this would quite possibly be classed as research. – StrongBad Oct 17 at 18:00
  • 1
    Some of this depends on who is conducting the (non-)research. Reality television gets away with all kinds of things that any halfway-competent IRB would shut down immediately...which recently allowed some actual researchers to use data obtained through methods they never could have gotten approved up-front. – 1006a Oct 18 at 1:00
  • @StrongBad I'm afraid I don't understand your point any more – Azor Ahai Oct 18 at 15:49

This could possibly happen when data collected for internal use is repurposed as part of a research study.

Institutional Review Boards are supposed to oversee research involving human subjects. Per 45 CFR §46.102, the relevant regulation in the US, defines research as:

(l) Research means a systematic investigation, including research development, testing, and evaluation, designed to develop or contribute to generalizable knowledge

The "generalizable" part means that organizations can collect data that might otherwise be covered if it's part of their regular operations. For example, a hospital might track post-operative infections by operating theatre for quality control. They could use this data to determine if staff need to be retrained, or if a specific operating room might be colonized with some bacteria. The critical point is that this information is specific to a particular place, surgical team, etc.

If you had a lot of this data, you might try to make more general statements. For example, you might want to build a regression model that predicts surgical outcomes so that you could argue that this type of surgical team, or that cleaning procedure should be used more wildly. This crosses the line into a research activity and needs IRB approval. Likewise, if the data isn't already being collected, you'll need IRB approval to start collecting it.

Nevertheless, you can request IRB approval to examine any existing records. This can be done after the records have already been collected for their non-research purpose, but should be done before starting the research part of the project.

This is the only example I can think of where an IRB retroactively "blesses" the use of previously collected. There are also some categories of "Exempt Research" per 45 CFR §46.104. However, our IRB--and I think this is pretty standard--claims that the IRB itself needs to determine that a proposed project matches one of those categories; you can't just opt yourself out.

  • 1
    The IRB needs to approve everything is a stupid model. If you want to do a computer simulation of where to store surgical supplies to minimize walking, my institution supposedly requires IRB approval since it is research. They then classify it exempt since it is not human subject research. Imagine what would happen if every math research project required an IRB determination. – StrongBad Oct 18 at 1:03
  • When I wrote "The IRB oversees research", I meant to contrast it to QA (etc) activities. The IRB doesn't oversee all research, only research that involves human subjects; your optimization project seems like it shouldn't ever cross their desk. – Matt Oct 18 at 1:29

Many US journals demand a statement to the effect of "This research was conducted under the ethical principles of the Belmont Report", and European journals would ask for adherence to the Declaration of Helsinki. At the very least, Journals often ask for the name of the IRB, and IRB's hold to Belmont or Helsinki standards for medical research.

My understanding is that both guidelines require outside review before you "do stuff" to human subjects, and that if the outside review was not conducted, you don't meet the guidelines. You can't just go to the IRB for a do-over -- the violation has already happened.

I recommend a conversation with your IRB. There will be grey areas. I know one facility that required IRB review for beta testing software (!!!!!) but most IRBs would scoff at the idea.

  • 2
    I've never seen anything about the Belmont Report before. It's usually "This research was approved by the Institutional Review Board at the University of X" – Azor Ahai Oct 17 at 18:02
  • I think this is a massive over statement. Not all human subject research is medical so the DoH does not apply. The DoH 2013 requires preregistration, which most fields are not equipped to deal with. Even in medical fields it is not uncommon to see research complying with DoH 2013 except for pre registration ... or new research only complying with DoH 2008. – StrongBad Oct 17 at 19:20
  • @AzorAhai First Journal I tried: Brain Research "Policy and ethics The work described in your article must have been carried out in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans" – Scott Seidman Oct 17 at 20:03
  • Second Journal, J Neurophys: The research described in papers submitted to any of the APS publications that involve the use of human beings, including healthy volunteers, should be conducted according to high standards of safety and ethics. We expect that all studies adhere to the requirements of the U.S. Federal Policy for the Protection of Human Subjects (45 CFR, Part 46), and support the general ethical principles of the Declaration of Helsinki. – Scott Seidman Oct 17 at 20:03
  • @ScottSeidman I'm not saying you're wrong. I'm just saying statements I see in publications more often are "This research was approved by the Institutional Review Board at the University of X" rather than anything about the Belmont Report or Declaration of Helsinki. – Azor Ahai Oct 17 at 20:05

Your Answer


By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.