Scientific misconduct and correcting the scientific literature

Academic Medicine

Volumn 74, Issue 3, 1999, Pages 221-230

  Parrish, Debra M ^a,b  

^a Titus and McConomy   (United States)

^b 20th floor Four Gateway Center   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

ETHICS; LEGAL ASPECT; MEDICAL LITERATURE; MEDICAL RESEARCH; PATENT; PRIORITY JOURNAL; PUBLICATION; REVIEW; SCIENTIFIC LITERATURE;

EID: 0033015963     PISSN: 10402446     EISSN: None     Source Type: Journal    
DOI: 10.1097/00001888-199903000-00009     Document Type: Review

References (59)

1. See, e.g., Anderson C. How not to publicize a misconduct finding. Science. March 25, 1994.
2. At least one other reported case reflected pressure on a prominent co-author to take corrective action. Dr. David Baltimore received pressure from the scientific community to correct a paper published in Cell, in which Dr. Thereza Imanishi-Kari was alleged to have fabricated data.
3. The study was funded by the National Cancer Institute of the National Institutes for Health of the Department of Health and Human Services.
4. April 24, 1994, memorandum from Dr. Kalt to ORI.
5. These letters explicitly referred to the St. Luc's data as "falsified data."
6. Angell M, Kassirer J. Setting the record straight. N Engl J Med. 1994; 330:1448-50 and 1458-62.
7. Dr. Fisher was an author of many, but not all, of these articles.
8. The annotations asserting "[scientific misconduct - data to be reanalyzed]" were removed and a new annotation "[prior annotation incorrect]" was inserted.
9. Importantly, the suit and claims regarding the improper tagging of the articles were dismissed before settlement. See Fisher v. National Institutes of Health, 934 F. Supp. 464 (D.D.C. 1996) and Fisher v. The University of Pittsburgh, No. 94-1160 D. Pa. filed July 8, 1994 (April 17, 1997 order granting dismissal of certain claims).
10. Because of the different levels of reliance, review, and correction accorded to textbooks and abstracts, they are excluded from the definition of publication as used herein.
11. 42 C.F.R. § 103(d)(3).
12. International Committee of Medical Journal Editors. Uniform requirements for manuscripts submitted to biomedical journals. JAMA. 1997; 277; 933.
13. ORI Newsletter. March 1997; 5, no. 2: 1.
14. In some cases, the negotiation of a settlement agreement begins before the agency makes a finding of misconduct.
15. Of the scientific misconduct cases handled by ORI that have been subjected to the full appeal process, the average amount of time from the issuance of the ORI finding to the conclusion of the appeal was 13 months.
16. After execution of a Voluntary Exclusion Agreement, the average delay in publication in the Federal Register is one month.
17. The ORI has found misconduct in nine cases involving unpublished manuscripts.
18. See, e.g., cases involving Cathy Q. Lee 61 Fed Reg. at 11010 (Mar. 18, 1996) and Yi Li 61 Fed Reg. at 63849 (Dec. 2, 1996).
19. The ORI policy states: The respondent is required to submit a letter within 30 days of notification of this action to [Journal] requesting correction of the article, [article], by [Description of correction]. This requirement will be noted in the ALERT System until the respondent has sent a copy of the letter of correction to ORI. The policy is similar for retractions. The explanatory note states that "ORI will notify the relevant journal and NLM [National Library of Medicine] of this action, and the respondent's name will remain on the ALERT system until he or she submits the correction [or retraction] request to the journal."
20. The ORI began this method of journal notification in 1994.
21. As discussed above, it typically takes the ORI ten to 12 months to confirm an institutional finding of misconduct, and then a month typically passes after the execution of a settlement agreement or conclusion of a hearing before notice is placed in the Federal Register.
22. Note that such a policy should have been inapplicable in the Fisher case because Dr. Poisson admitted to the misconduct.
23. This may be because the manuscripts or publications had been corrected or retracted before the NSF had made a finding of misconduct or because the NSF believes such notice would violate the confidentiality of the accused individuals.
24. It is also important to remember that the policies of many institutions do not provide for notification to journals unless there is a finding of misconduct.
25. Journals have not taken the lead that they could have in establishing the necessary high ethical standards. For example, although the instructions to authors and criteria for authorship appear to set high standards, e.g., by requiring publishing researchers to make unique cell lines available, retaining the data for inspection, and including significant contributors in the acknowledgment, the journals have not retracted authorship or papers when those obligations are breached.
26. For example, note that although Dr. Fisher submitted the reanalysis that excluded the St. Luc's data on March 25, 1994, the NEJM declined to publish the reanalysis until an audit of the entire study was complete. The reanalysis of the most important portion of the study was not published until November 1995.
27. The first letter of correction was sent by several of the co-authors, including the accused, to the editors, and was published November 18, 1988 Cell. 1988; 55:541.
28. The second letter of correction was signed by several of the co-authors, including the accused. It was published on May 19, 1989. Cell. 1989; 57:515-6.
29. See the letter of retraction published May 17, 1991. Note that the accused author did not sign the letter of retraction. Cell. 1991; 65:536.
30. In the one case that went through a full appeal and in which the ORI finding of misconduct was upheld, the delay between from the time the institution asserted that it found scientific misconduct to publication in the Federal Register was approximately four years.
31. The obligation of co-authors to correct the literature based on the malfeasance of another co-author was also raised in the case of Dr. Thereza Imanishi-Kari.
32. Note that failing to correct the literature was not one of the allegations of scientific misconduct against Dr. Fisher.
33. In a few cases, the ORI has contacted journals before the conclusion of the investigation when it was necessary to preserve evidence that it was believed only the editor possessed, when an allegation was made that a reviewer of an article had committed plagiarism, and when a draft letter to the editor may have made false claims regarding the data.
34. See, e.g., Arch Gen Psychiatry. 1988; 45:685. See also the retraction issued by Dr. David Baltimore in the case involving Dr. Thereza Imanishi-Kari.
35. As noted above, the lawsuit and claims involving the tagging were dismissed.
36. The range of delays was seven weeks to 18 months.
37. See, e.g., correspondence regarding the correction of articles in the Farooq A. Siddiqui case.
38. The electronic databases were searched for retractions or corrections using a cite for the original article. Further, the editors of journals hosting the articles were contacted and asked about corrections and retractions for the articles.
39. Friedman PJ. Correcting the literature following fraudulent publication. JAMA. 1990; 263:1416-9.
40. However, while preparing this article, the author discovered several affected articles did not bear the tag despite notice in the NIH Guide to Grants and Contracts that the article was the subject of a misconduct finding.
41. Notice of a finding of the NIH Guide to Grants and Contracts is sent at approximately the same time that notice is sent to the Federal Register.
42. In researching this article, it became apparent that all the articles identified in the Federal Register as requiring correction or retraction had not been tagged by the NLM as having been the subject of an integrity review.
43. Ann Intern Med. 1988:108;304. See also: International Committee of Medical Journal Editors. Uniform requirements for manuscripts submitted to biomedical journals. JAMA. 1997; 277: 927-34.
44. Ann Intern Med. 1988:108;304. See also: International Committee of Medical Journal Editors. Uniform requirements for manuscripts submitted to biomedical journals. JAMA. 1997; 277: 927-34.
45. Most letters provide no reason for the withdrawal of the paper. Others state that the retraction stems from "methodological and procedural flaws" [Neurology. 1989;39:1261]; "invalid data" [Obstet Gynecol. 1995; 86: 728 and Fertility and Sterility. 1996; 65:211] or "arithmetical errors" [Lancet 1992;340:496]. But see the retraction in Cell [1992; 69:724] stating experimental evidence was "fabricated" by the accused scientist.
46. Most letters provide no reason for the withdrawal of the paper. Others state that the retraction stems from "methodological and procedural flaws" [Neurology. 1989;39:1261]; "invalid data" [Obstet Gynecol. 1995; 86: 728 and Fertility and Sterility. 1996; 65:211] or "arithmetical errors" [Lancet 1992;340:496]. But see the retraction in Cell [1992; 69:724] stating experimental evidence was "fabricated" by the accused scientist.
47. Most letters provide no reason for the withdrawal of the paper. Others state that the retraction stems from "methodological and procedural flaws" [Neurology. 1989;39:1261]; "invalid data" [Obstet Gynecol. 1995; 86: 728 and Fertility and Sterility. 1996; 65:211] or "arithmetical errors" [Lancet 1992;340:496]. But see the retraction in Cell [1992; 69:724] stating experimental evidence was "fabricated" by the accused scientist.
48. Most letters provide no reason for the withdrawal of the paper. Others state that the retraction stems from "methodological and procedural flaws" [Neurology. 1989;39:1261]; "invalid data" [Obstet Gynecol. 1995; 86: 728 and Fertility and Sterility. 1996; 65:211] or "arithmetical errors" [Lancet 1992;340:496]. But see the retraction in Cell [1992; 69:724] stating experimental evidence was "fabricated" by the accused scientist.
49. Most letters provide no reason for the withdrawal of the paper. Others state that the retraction stems from "methodological and procedural flaws" [Neurology. 1989;39:1261]; "invalid data" [Obstet Gynecol. 1995; 86: 728 and Fertility and Sterility. 1996; 65:211] or "arithmetical errors" [Lancet 1992;340:496]. But see the retraction in Cell [1992; 69:724] stating experimental evidence was "fabricated" by the accused scientist.
50. When a letter is sent to editors after the ORI has made a finding of scientific misconduct or is settling a case with such a finding, the contents of letters of correction or retraction may be negotiated by the ORI with the accused and non-accused authors.
51. See Retraction: Oct-3 is a maternal factor required for the first mouse embryonic division. Cell 1992; 69: 724.
52. Inhibition of Rev-mediated HIV-1 expression: retraction. Science. 1991; 264:492; Retraction. N Engl J Med. 1991; 325:1487; Retraction: effects of interleukin-1 on platelet counts. Lancet. 1992; 340:496.
53. Inhibition of Rev-mediated HIV-1 expression: retraction. Science. 1991; 264:492; Retraction. N Engl J Med. 1991; 325:1487; Retraction: effects of interleukin-1 on platelet counts. Lancet. 1992; 340:496.
54. Inhibition of Rev-mediated HIV-1 expression: retraction. Science. 1991; 264:492; Retraction. N Engl J Med. 1991; 325:1487; Retraction: effects of interleukin-1 on platelet counts. Lancet. 1992; 340:496.
55. Retraction. N Engl J Med. 1991; 325: 1487 (signed by the non-accused authors and an institutional official).
56. See, e.g., Cell. 1992; 69: 724 (stating "Recent investigations have revealed that the experimental evidence supporting the conclusions of our paper II has been fabricated by [the accused scientist] without any knowledge by the others. We therefore retract this paper in its entirety").
57. Journal editors often are subjected to pressure from institutions, co-authors, and funding agencies regarding the wording of a letter of correction or retraction and find themselves embroiled in the dispute with threats of litigation. One journal, The Journal of Histochemistry and Cytochemistry, published the accusation of plagiarism, the rebuttal, and an editorial, leaving it to the readers to make their own determinations regarding whether plagiarism had occurred. J. Histochem Cytochem. 1990 and 1991; 38:267-73 and 39:379.
58. See, e.g., Obstet Gynecol. 1995; 86:728. In the ORI case involving James H. Abbs, PhD, as a condition of a Voluntary Exclusion Agreement, Dr. Abbs was required to write to the editor and notify him of the ORI's desire that an article be retracted although Dr. Abbs believed that the article should not be retracted. Based on that letter, the editor issued a retraction that stated: The retraction of this paper is based on the Department of Health and Human Services' Office of Research Integrity (ORI) Investigation Report finding that James H. Abbs, PhD, engaged in scientific misconduct by deliberately falsifying and fabricating certain figures and research results that were published in the . . . paper. The ORI believes it is necessary to retract the paper to correct the scientific literature.
59. See Neurology. 1996:47:340.