IRB Opinion on Randomise Me Trial for Epic Delirium Order Set

I thought it might be fun to let you guys know how my Institutional Review Board (IRB) evaluation went for my proposed randomized trial regarding our Epic Delirium Order Set quality improvement study, which I’ve posted about previously. A couple of my colleagues thought the safest course would be to make sure this project wouldn’t qualify as human research.

HawkIRBSo, for the first time, I completed a HawkIRB online form, which starts with the title of our project.

Epic Delirium Order Set Usage: Does It Reduce the Number of Psychiatry Consultation Requests?

“My colleagues’ opinions differed about whether I needed to contact the IRB about this. Although this might not clearly be a research project and I got the idea from a website, and wrote a blog about it,

The idea really is to promote the use of the Epic Delirium Order Set by non-psychiatric clinicians so as to encourage the development of greater comfort with assessing and managing delirium in general medical units. I would like to track how often the order set is used by internal medicine and see whether that influences the number of psychiatric consultation requests. Currently, many patients are misdiagnosed with a primary mental disorder. Psychiatry is consulted often when the most I can do is to identify what the patient doesn’t have, i.e., a primary mental disorder. Since delirium is, by one definition, a medical emergency, scope of practice issues prevent me from tracking down the medical etiologies since internal medicine is not my specialty.

The outline would be:

Create a Trial Question: Does using the Epic Delirium Order Set cut down on the number of psychiatry consults for delirium?
Main Outcome Measure: Number of psychiatry consults.
Size of Trial: I’m not sure but we get about 100 consultations a month and at least 20% are for delirium.
Description of Trial: A randomized trial comparing the number of psychiatry consult requests between general medicine units which either do or don’t use the Epic Delirium Order Set. I expect the number of consult requests would be lower from the unit which uses the order set.
Trial Instructions: Residents on the test unit will enter the Delirium Order Set upon either learning of a positive Delirium Observation Screening Scale (DOSS) score from nurses or diagnosing delirium themselves by the informal use of the Confusion Assessment Method (CAM) and the Mini-Cog. Residents on the control unit will not enter the Delirium Order Set under the same conditions.

We would need to extract an Epic report for baseline number of times the Epic Delirium order set is used. We could randomize internal medicine residents from one medical unit team to use Epic, and another team to not use it.”

And the outcome was probably predictable. The determination by the IRB Chair was as follows:

I have determined that the project described in the application does not meet the regulatory definition of human subjects research and does not require review by the IRB, because this project is aimed at increasing the use of an existing clinical tool and assessing its effectiveness; it is not likely to produce generalizable knowledge. Also it looks like this project can be done looking at aggregate data from EPIC with identifiers, if the data are extracted by UI IT.

Any readers who know their way around an IRB will immediately see how amateurish my proposal is. Funny thing, the computer gave me conflicting answers about whether this project was human subjects research or not. On the one hand, I got the comment by the IRB Chair. On the other, I also got a memo at the same time which identified it as human subjects research, a blatant contradiction. So I called the IRB office and they said it was their mistake or a computer glitch. They said my project was definitely not human subjects research and the IRB chair would be sending me an email confirmation of that.

I think my wooly-worded summary probably just blew their algorithm. In any case, now I need to decide whether to go ahead with the project or not. You know, all I’m really trying to do is connect with my colleagues in internal medicine. I’ve been trying to engage them in the fight against delirium for so long. I don’t know if I ever had their attention, or maybe I did at one time and just lost it.



  1. Sharon Van Fleet says:

    Dr. Amos, I like what you are trying to do here and I very much respect your willingness to share your vulnerability! I like your style, and your good heart!

    Off hand I don’t recall seeing much if any work looking at the impact of routine delirium screening/protocols on psychiatry consultation and to my view this is an important line of inquiry. I too have sat on an IRB and would like to reassure you that the need for approval isn’t always clear. I also agree with Dr. Dawson’s comment re randomization.To my view that aspect would tend to exceed the boundaries, if you will, of a quality improvement project….i.e., indeed making it research. (I would be able to track down the citation for a great article that really helps to clarify the difference between QI and research, from a nursing journal, if you are interested.) I’ve also understood that one is wise to at least check with the IRB, especially if one anticipates publishing. I would also wonder about the data collection also being a factor that might warrant an IRB opinion if not review.

    I think you have a worthwhile project, whether QI or research. Data trends regarding delirium incidence and psych consultation are important to help measure the impact of the screening and I would think you would want to be capturing that at minimum, whether or not you decide to proceed with randomizing your residents. Any positive trends that can be elucidated might be helpful in reigniting interest in the screening….and help support the continuation of the effort, which does require some nursing effort in the aggregate.

    Have you seen the CAM-S, recently published in Annals of Internal Medicine by Dr. Inouye and her colleagues, looking at delirium severity and outcomes? I do not have easy library access at present since I am “between jobs” and thus have not seen the article and tool, but I was able to see the first page of the editorial speaking compellingly to the need to attend to delirium which has the same mortality as MI. I thought of you when I saw the references to the article coming out. The Dutch group published in 2012 their early effort suggesting that the DOS Scale might serve as a delirium severity measure, which certainly fit with my experience in implementing it about a decade ago in one of my former settings and following patients from day-to-day in my role which was valued for case-by-case monitoring of sitter use and where the DOS became valued by other members of nursing leadership not for screening for delirium (for which there was little concern, despite my decade-long campaign) but for helping to determine the need for continued sitter use, because of its significant cost. I invested much in implementing the DOS and very much want to see your efforts succeed.

    Yes, keep trying!!!!

    Liked by 1 person

    • Great comments, Sharon; I can’t thank you enough for your encouragement and support! In fact, I did see the news about the CAM-S but didn’t get a chance to read the paper thoroughly. If you have time to get me the citation about distinguishing between QI and research, I would appreciate it.

      I hope all is well. Good luck!


      • Sharon Van Fleet says:

        Dr. Amos, Thank you for your reply.

        There are a number of sources, but I found these helpful in working with nurses trying to decide if they needed to submit proposals to the IRB:

        Ryan, M. & Rosario, R. (2012). Creating staff confidence in distinguishing between performance improvement and research studies: A user-friendly worksheet. Journal of Nursing Administration, 42 (7/8), 369-374.

        Newhouse, R.P. (2007). Diffusing confusion among evidence-based practice, quality improvement and research. Journal of Nursing Administration, 37 (10), 432-435.


      • Many thanks!


  2. Jim,

    As a former IRB Committee member – you could have fooled me about the need for approval. The issue to me would be the randomization rather than whether the clinical tool is in use or not.


  3. Sorry if my comment got you to jump through unnecessary hoops. I’m glad you’re good-to-go on the trial, if you decide to do it. One question I have is, Are the medical units equivalent? In the hospital where I did my residency, some floors had significantly better (or worse) nursing than others, and some floors were simply much busier than others, and I would think these would be confounders when assigning the delirium set. Certainly there are differences in the quality of residents, but since they rotate through units, that shouldn’t be a problem, provided the study lasts through more than one rotation.
    I don’t know if you’re inclined to go through with the trial, but I, for one, would be really interested in reading the resulting paper.

    Liked by 1 person

    • I’ll run it by other members of the Delirium Prevention Committee. Actually, the chair of the committee helped me with the IRB query.

      I would envision trying to run the trial on the units where nurses were piloting the delirium screening tool, the Delirium Observation Screening Scale (DOSS). The trouble is, I have to reason to believe there’s been some slippage in DOSS implementation lately. The committee is also looking to relight the fire under that initiative.

      The idea is to keep trying.


%d bloggers like this: