Discussion: View Thread

Hello. I'm just a year into my budding solo consulting business (finally retiring from day job in 2 weeks!) and already I face an ethical dilemma.

A fellow at another medical university asked me to help with a project. Fellows in this program are required to do some sort of scholarly project. Fellow's deadline (end of fellowship) is December. The attending physician is the PI. Their goals are (1) submission to a conference by mid-October, and (2) hopefully publication thereafter.  By the fellow's telling, the in-house statistical consultants did not have the bandwith to help.

Briefly, this would be a historical cohort study based on retrospective review of electronic medical records.

Last we'd left it, Fellow was awaiting IRB approval. I said great, get back in touch when that's obtained, and we can get started.

Some months went by. Fellow and Attending just got back in touch with me. Turns out they had assembled a team and had already abstracted records and stored data in a very messy Excel spreadsheet (of course!). That was done as IRB process was underway, but before IRB approval. (I suspect the IRB submission was rather delayed.) Institutional policy is as customary: "all research must be sumbitted . . . before the study is conducted, including recruitment and data collection."  The gist of the IRB determination was "We don't offer retrospective approvals.  If we'd had the opportunity to review it prospectively, we believe we would have found it Exempt. . . . Any publication must include a statement that the study did not receive prospective IRB review."

There has also been brief reference to grant funding, so I wonder what the grantor's IRB requirements are.

I'm skeptical about getting an abstract accepted for a conference, or a manuscript accepted for a journal, given the IRB determination above.

I have many reservations about participating in this.  I don't want to get involved with a project that did not adhere to ethical principles and regulations, both internal and external. It also just seems a bit "loose" in execution. On the other hand, I want to support the Fellow. I don't know what sort of guidance and training Fellow received in how to execute this project. Fellow should not suffer if that was inadequate.

Appreciate any thoughts/perspectives.

Thanks.

The best way to support the fellow is to stand your ground and refuse to participate on the grounds of ethical issues. How else is the fellow supposed to learn that the IRB actually matters and that ethical review by an external body for human subjects research is a necessary step?

Just my two cents - hurts in the short term, but you can look in the mirror at night.

Best,

Jana

They could still get IRB approval and claim they had approval even if they started without it - so they could still theoretically end up published!

Dave--

The institution's IRB does in fact review these types of studies---when the proposal is submitted before the research starts. If it's a study based on retrospective review of existing medical records, they consider the start of records review and data abstraction to be the start of the research. I think that's appropriate. By that definition, these researchers started their research before IRB approval. 

It probably would have been more clear if the IRB had written "Stop, you cannot proceed with this, since you did not follow the rules." Instead they said, basically, "Well, we don't vouch for this. If you do it, you need this disclaimer." An awkward bit of fence-straddling.

The trouble with IRBs, and they are legion, is that there is no real standards about how they operate and they make it up as they go along.  Here the study obviously does not rise to any serious issues, since it is using that already has been collected.  The only issue is that they seem unhappy that you did not contact them ahead of time, even though they think it should be exempt from review.  Sounds to me like they are simply exerting their power, and are stopping what very likely is reasonable research.  Is there any "higher authority" to which you can go.  Run away IRBs exist all over the US.  Here it seem to me that your only actual potential harm is breaching of confidentiality of medical records.  

At CUNY this would go up a level to the administrator who overseas the IRBs.

Andy

Andy Beveridge, Professor of Sociology Emeritus, Queens College and Grad Center CUNY.

Thanks Shah and Jana.  Those are my instincts as well, but I will continue to appreciate all perspectives from this group.

Could get ugly, but a clean professional conscience is important.

I hope to explore with the attending physician (PI) what guidance was provided the Fellow. If a scholarly activity is required, and if this one isn't going to happen, perhaps some learning about research ethics and execution, and a presentation to other residents/fellows, would fulfill this Fellow's requirement. Make it a learning experience. And maybe head off future problems. 

I have been a consultant in a solo practice since 1977.

Short answer: DON'T DO IT!

  I have a slightly different take than the rest of the group but I largely share and completely understand their concerns.

• If you say no to them, they will probably just email some other consultants or maybe rope in a grad student unpaid, and never mention the IRB thing to them. You have a chance to intervene and stop the problem from spreading.
• So I would probably still help them, but at an arms length - especially if it's an easy problem, ie clean the data and do a chi-square test, I might show them how to do that ON THEIR COMPUTER. Or maybe work it out with simulated data. I would /not/ run it my self on the actual data.
• I would /not/ let them add me as a coauthor or otherwise lend them my credibility.
• The publication component ought to be out-of-scope in this case - they could have IRB'd earlier, they can figure it out what that means for them. Maybe they go find a third-tier journal that doesn't care, I don't know. Not my problem.
• I would /absolutely not/ let them share (ie email) the dataset with me - if they can't follow IRB, then HIPPA and the other hospital security etc is probably not being followed either, you don't want the liability in the worst case scenario. 
• I /would/ up-charge them for the hassle, however much time you spent on figuring out the IRB mess, at least an extra hour or two. Email time is still time worked.

Just my two cents. You don't have to think about it in terms of all-or-nothing. As a private consultant, it's your job to help the client, not to be the IRB police, and if you do decide to tell them no (and to be clear, I think that would be an entirely reasonable thing to do), it will go down a lot smoother if you can offer them some other options or advice. Or at least, that's how I approach the business.

First of all, congratulations on the upcoming retirement :-)

My rule of thumb around ethics has always been: if it does not feel right, it probably is not (especially if the feel is of nagging nature). This has not failed me. So, I would not proceed as is if you're not comfortable.

That said, there is also an opportunity here as a few people have already pointed out, be it an opportunity to educate, course-correct, come up with alternative solutions, etc. In one of the section's Pathways to Promotions webinars last year, I talked about presenting options rather than answering with a no as a singular response. What are the things you are willing to do? Then I would think along the lines of "I will not do this, but here are some things we can do (and still preserve my integrity)." This can make the working relationship a little easier by showing that you are willing to work with them and that you've given that a careful thought. If they reject that, they were not likely to change their mind anyway; it's a different problem, and you move on. I have walked away from a number of engagements this way, and I am at peace having protected my own integrity.

FWIW: As a Certified Management Consultant, I abide by and am evaluated on ethics by the standards by the Institute of Management Consultants - its Code of Ethics is a useful guide for matters that are not strictly statistical in nature, so I post it here: https://imcusa.org/about/ethics/code-of-ethics/

I am re-visiting the ASA statement on ethical statistical practice and finding much helpful guidance in it.

As an academic family physician for my first 30 working years,

1) I always try to make encounters educational,

2) I try to look out for the interests of the "lowest" folks on the med ed hiearchy, which in this case is the Fellow. Imbalances of power are real in med ed.

3) and I always come back to a touchstone of clinical practice: "all care is negotiated."  I think this fits well with the theme expressed by several discussants, about figuring out mutually-agreeable solutions and finding a way forward, while preserving limits.

In the interim, I had an interesting conversation with the IRB official. Seems this is not this particular PI's first go-round with these issues.. They have been impervious to previous educational interventions by the research admin folks.

Interesting -- 

I think the regulatory guidances here are pretty black and white. 'No research procedures prior to IRB approval of research procedures' seems crystal clear. 

In addition, since the research involves review of personally identifiable medical records, it seems very unlikely that this 'retrospective chart review' would qualify for 'exemption' (there are a very limited number of 'exemption' types and very specific rules), or be classified as 'not involving human subjects'. The latter situation can occur when a third party ('honest broker') retrieves and completely de-identifies data or biospecimens. Based on your description, it seems more likely that the research would qualify for a 'waiver of the requirement for informed consent'.

I work at an institution that requires that I also conform to these rules, and I would not participate in the research. I am surprised, based on your description and the possible continuing nature of the problem, that research compliance is not involved. 

I agree with the advice , paraphrasing "if it looks like a duck , walks like a duck, quacks like a duck..."  There is no undoing ethical taint. however, re-re-reading your note, the IRB was moving slowly. Probably because they're busy.  IN the pharma world, FDA has a guidance as to who and how an IRB selects members.

This is one of several guidances FDA has for IRB. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/institutional-review-boards-frequently-asked-questions

google search for the other IRB guidances from FDA. And  in pharma, there are many times when the clinical people make arrangements for an expedited review of a research protocol by an IRB. IRB's usually have a process for expediting a review.  IRB's operate to protect innocent humans from unnecessary or possibly harmful experiments. on some occasions the clinical people will arrange for a "central IRB" rather than at each institution. There is a fee ($$$) for the central IRB services.

You do have an option (you might not prefer to use it) for a "speak softly and carry a big stick" . There should be a chief compliance officer. and there may be a way to submit a report anonymously. Better for your colleagues to get them to follow the procedures and contact the IRB themselves. They are medical fellows, so you could contact the chief of medicine or similar. Last but not least, They will probably need to destroy the records they collected so far, and collect them again. 

your best negotiation strategy may be to sit down with then and layout the options - and have them contact the IRB.

The more I learned, the more concerning the situation became. I have bowed out. Thank you everyone for the thoughtful and educational discussion.

Incidentally, speaking of the FDA, I see that last month they issued a warning letter to MIT's IRB.    They have 15 days to respond or risk "regulatory action without further notice to you".  (The issues at stake were not statistical.)  Perhaps it's just a reminder that working with an IRB that doesn't have its act together can be a problem too.

https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/warning-letters/massachusetts-institute-technology-mit-663218-06212024

In the spirit of being helpful, hopefully maintaining a positive relationship with a potential customer, and contributing to the education of the Fellows, I did offer the following:

a couple talks for the Fellows, covering the following areas:

- how to work with a statistical consultant
- the sequence of steps involved in executing a project properly
- safe and efficient data collection, storage, and management (most importantly, use your university's REDCap platform; never use an Excel spreadsheet.)

Followed by, if desired, consultation on individual projects, as long as I could be involved from their very beginnings. This could include building REDCap databases to support chart abstraction, for retrospective studies.

We'll see what happens.

Very Interesting.

As the IRB member/Chair, also according to my consulting experience, for my points of view, could you give the Fellow suggestions as follows?

·          The IRB usually does not accept an already formal study on or close to complete, but, if the formal study carried out prior to the pilot study, for example, at the moment, the fellow just did a pilot study which was for the feasibility assessment (including data detection), the fellow can apply for a formal study of the IRB review. It is important to note that "We don't offer retrospective approvals." does not mean "We don't accept retrospective studies."

·          the IRB review focuses on the protection of patients' rights and interests, and the rationality and feasibility of the study protocol. For retrospective studies, special attention will be paid to data security, that is, personal privacy protection

·          A formal study protocol should be submitted to the IRB for ethical review. Scheme in addition to the general study content, should be focused on the data source and data curation plan. The data curation includes, but is not limited to, the following aspects: data extraction (including multiple data sources), data security processing (de-identification or anonymization,  and protection from data corruption, leaking, theft, tampering, or unauthorized access),  data cleaning (edit check and outliers processing, data completeness processing), data conversion (common data models,  normalization, natural language processing, medical coding, derived variable calculation), data quality control,  data transmission and storage.

·          For a retrospective study, the person can apply for exemption which means the exemption with the consent of patients.

·          In addition, as far as I am concerned, IRB review in journals is usually seen as "... was approved by the IRB of … Hospital ", but rarely seen. a statement that the study did not receive prospective IRB review.

Best,

Pingyan Chen

Professor

Department of Biostatistics，Southern Medical University

Gername to the excellent discussion here, and by  a coincidence, an example of a major concern about (lack of ) IRB involvement and Informed consent. An article appeared (today) on my reading list about informed consent for the "Human Genome Project"   

Honor your reservations about not participating. Secondly, consider explaining to the Fellow that per your last conversation, you asked that the Fellow get back to you "after IRB approval was obtained". Had you been kept in the loop that the Fellow and "team" were considering to collect data prior to IRB approval, you would have explained the consequences of collecting data prior to IRB approval and advised against such action. For ethical reasons, you're unable to participate.

Furthermore, the Fellow is already demonstrating that you are not part of the team; otherwise, the Fellow would have checked in with you prior to abstracting data. So, not only is this an ethical issue but a possible indicator of other working relationship incidents to come.