A U.S. government watchdog for scientific misconduct has floated the possibility of revising some of its regulations, and it wants your thoughts on what should change.
The Office of Research Integrity recently issued a Request for Information – essentially an email inbox open for suggestions – to help shape its potential revision of the 2005 Public Health Service Policies on Research Misconduct, 42 C.F.R. Part 93 in the federal code.
These regulations define what “research misconduct” means for work funded by the U.S. Public Health Service – the oft-quoted “falsification, fabrication, and plagiarism” – and establish how the government and research institutions respond to these issues.
The current regulations replaced rules issued in 1989, the same year the Office of Scientific Integrity in the National Institutes of Health and the Office of Scientific Integrity Review in the U.S. Department of Health and Human Services were created. These two offices were merged into the Office of Research Integrity in 1992.
Here’s the meat of the request:
ORI is not seeking specific regulatory language at this time, only the identification of potential topic(s), issue(s), or area(s) that stakeholders and other members of the general public see as being important to consider when revising the 2005 ORI regulation at 42 CFR part 93. Responders may find it helpful to consider the following questions when preparing responses (the order of the questions below should not be taken to imply importance, priority, or precedence):
(1) Which section(s) should be changed or augmented when revising 42 CFR part 93? Why? How should the section(s) be changed or augmented?
(2) Which section(s) should be retained as it currently is in 42 CFR part 93? Why?
(3) Which section(s) should be considered for removal when revising 42 CFR part 93? Why?
ORI views this RFI as a brainstorming process. Short responses, limited to just a few words on a given topic, issue, or area will facilitate the organization and categorization of responses. If an idea specifically relates to a part of the current regulation, citing that section (e.g., § 314.3) would be helpful.
If you’ve sent in comments to ORI, or plan to, please also paste them in our comments section below for discussion. ORI wants responses “no later than 5:00 p.m. ET on October 31, 2022.”
Hat tip: Theresa Defino
Like Retraction Watch? You can make a tax-deductible contribution to support our work, follow us on Twitter, like us on Facebook, add us to your RSS reader, or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].
Admittedly, these are not carefully thought through, but some of these might be nice.
Here are some ideas for revising the ORI regulation:
More clear-cut guidance is needed from the Office of Research Integrity about ethics in publishing. Perhaps adopt a “code of ethics” derived from this: https://www.nature.com/nature-portfolio/editorial-policies. While this code may only be enforceable on NIH-funded papers, having this as US policy might nudge US institutions (at least) to follow suit.
Require research to be published in open-access journals; public pays for it, public should get access (something like this may already be in the works per White House announcement)
Require data retention for 10 years with a data custodian (can be institution or for-profit data warehouse) pre-identified in grant requests and published paper and extend ORI investigation window to 10 years
Require data to be publicly available – with anonymization for patient data and PII
Require funded researchers to agree to submit to independent audit or ORI examination upon request, with automatic grant money claw-back if they do not agree when being investigated (like if they lawyer-up).
Require that institutional whistleblower protections are in place before grants awarded
Require funded institutions to have a publicly identified “research integrity officer”
Require institutions (instead of authors) to correct published research when research misconduct investigations are concluded.
Not only does the data need to be retained, any commercial software must be documented (including how it was used) and any software written must be included with the data. Processing of the data can be as important to the accuracy of the results as the data is.
My suggestion (email in progress) would be to change the steps required for misconduct investigations. As it stands now, institutes are required to investigate themselves, but that creates a wild conflict of interest (feel free to reply with your favourite example), and then take it to the ORI.
Instead ORI could perhaps force institutes to bring in third parties. For example, if a stem cell researcher at Harvard were implicated in data fabrication, Harvard would alert ORI and bring in stem cell experts from outside of Harvard, e.g. MIT, Yale, Oxford, wherever, who would investigate. There is a procedure for perpetrators of sexual assault to be reported to potential employers in academia (Title IX, I think it’s called), something similar could also be implemented.
ORI might also offer some protection for institutes when publishing misconduct. As it stands, I suspect a lot of misconduct investigations end when threats of legal action start rising. (This is probably outside the power of ORI, now that I think of it). Even when the conclude, there are cases where institutes announce that some misconduct has occurred, but sometimes don’t even mention which papers are effected.
Related to Cheshire’s post “Require institutions (instead of authors) to correct published research when research misconduct investigations are concluded.” is an excellent point. Explicitly requiring institutes to at least contact journals (or some similar requirement) seems like a no-brainer.
I would be curious if anyone has any feedback on these (admittedly rough) suggestions.
With respect to having institutions bring in third parties (i.e., the wild CoI; a much needed change to the current system), I note that in Japan it had been proposed that half of the members of the team investigating research misconduct cases should come from outside of the institution (see, http://retractionwatch.com/wp-content/uploads/2019/01/APRIN-Checklist.pdf). I don’t know whether THAT approach is how research misconduct cases are currently investigated there. But, if it is being used, we should take a serious look at that system, see if it works, and, if so, make whatever adjustments to adapt it here in the USA and elsewhere, and make it even more effective.
Yes, we agree with the conflict of interest concern you have, and also that third parties would be a good approach to try to resolve those. As we published a few years ago:
“There are a number of federal audit guidelines for addressing the aforementioned threats to independence and conflict of interest. Some of those involve the important concept of performance audits.
“2.10 Performance audits are defined as audits that provide findings or conclusions based on an evaluation of sufficient, appropriate evidence against criteria.”
Performance audits are also more broadly of value in determining whether an oversight system is actually functioning adequately.”
Loikith and Bauchwitz, Sci Eng Ethics. 2016 Aug;22(4):1027-1049.
My submitted comment:
I believe that SELF-PLAGIARISM needs to be added to the list of offenses that constitutes misconduct. Under the current regulations (42 CFR part 93) self-plagiarism is not considered misconduct. This situation allows individuals to publish the same data in multiple venues, thus gaming the system of metrics by which scientists are judged for promotions, grants, awards, etc. Self-plagiarism is double-dipping, and is not a practice that should be encouraged by being beyond the reach of ORI.
I agree! Covert forms of data duplication, augmentation, segmentation (i.e., certain forms of salami publication) in which readers are intentionally misled (e.g., by the absence or ambiguous placement of a key citation to earlier data) about the true nature and/or provenance of the data (whether they are new, previously published or a mixture of new and old) and that ends up distorting the scientific record or potentially does so should be considered research misconduct.
The following are comments made concerning suggested modifications to the Code of Federal Regulations of the USA (42 CFR part 93). The recommendations were made on October 31, 2022, in response to a Request for Information (RFI) by the federal Department of Health and Human Services’ Office of the Secretary. For the full text of the actual recommendation made, see https://amerares.com/blogs/frinvblog/recommended-modifications-to-the-2005-public-health-service-policies-on-research-misconduct-42-c-f-r-part-93/.
(1) Re recommended modifications to 42 C.F.R. § 93.100 (General Policy) part (b):
HHS takes ultimate responsibility for oversight. However, doing so effectively will require making many more of ORI’s functions mandatory, rather than discretionary, and most importantly, instituting independent audit procedures of research misconduct program.
(2) Section 93.307 (Institutional inquiry) part (a)(3), (b), and (c)
(a) Auditability of the inquiry initiation process would encompass the handling of allegations. But what is “sufficient” about credibility, and how “specific” must evidence be? The lack of definitiion of these terms is an example of why federally compliant evidentiary audit standards should be instituted within these regulations. (See also Sci Eng Ethics. 2016 Aug;22(4):1027-1049.) It is strongly recommended that this RFI be presented to the federal CIGIE for comment, along with other U.S. and international audit societies, such as the Institute of Internal Auditors.
(b) Recommendation is made to strike out language permitting any institutional discretion, e.g. by their own interpretations of “reasonable and practical”, to not take timely custody of all relevant records at the earliest moment. Ensuring the acquisition of all relevant evidence would promote more effective assessment and auditability of allegations. If comprehensive acquisition and archiving of research records really cannot be accomplished, the institution can document why not, and such claims can themselves be audited.
(3) Section 93.402 (ORI allegation assessments) parts (a) and (b)
Amending this section to make ORI’s performance more mandatory could provide a means of comparing the interest of informants in contacting ORI directly, rather than working solely within their institutions. The latter is known to be a very highly risky prospect to the informant. By contrast, ORI might be less conflicted than the affected institution in honoring confidentiality and anti-retaliation requirements needed to protect the confidential informant/complainant. (Anonymous reporting options could also be specified in these regulations). The addition of an expert panel option to assist ORI could potentially allow auditors to compare such outcomes with those produced by using a fully institutional route.
Robert Bauchwitz notes:
Technical problems are affecting the original link in the comment above.
The same information can also be found at https://figshare.com/articles/presentation/Recommended_Modifications_to_the_2005_Public_Health_Service_Policies_on_Research_Misconduct_42_C_F_R_Part_93/21938282