Abstract
The Division of Investigative Oversight within the U.S. Office of Research Integrity (ORI) is responsible for conducting oversight review of institutional inquiries and investigations of possible research misconduct. It is also responsible for determining whether Public Health Service findings of research misconduct are warranted. Although ORI findings rely primarily on the scope and quality of the institution’s analyses and determinations, ORI often has been able to strengthen the original findings by employing a variety of analytical methods, often computer based. Although ORI does not conduct inquiries or investigations, it has broad authority to provide assistance to institutions at all stages of their reviews of allegations. This assistance can range from providing advice on best practices, to legal assistance, to suggestions for how best to investigate specific allegations. When asked, ORI can also conduct certain forensic analyses, such as a statistical examination of questioned digits or a simple examination of a questioned figure in Photoshop. ORI will not provide opinions or render judgment on such analyses while the institution is still conducting its investigation. Such analyses can be done without knowing much else about the case.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Because of the extensive publicity in this case, in this section the authors do not adhere to ORI’s usual policy of not identifying respondents except when announcing Public Health Service (PHS) findings after closing a case.
The data were accumulated in an Excel© spreadsheet, but the terms ‘spreadsheet’ and ‘database’ are used interchangeably throughout the record of the case and in this paper.
Formulas can be removed from a spreadsheet by copying its contents to a new sheet with the Paste Special command, and choosing to paste values only.
Poehlman presumably knew that program project grant (PO1) applications are reviewed by NIH institutes while research grants (R01) applications are reviewed centrally by the Center for Scientific Review, reducing the likelihood that the same individuals would see the preliminary data in R01 and P01 applications, for example. On the other hand, if falsifications had been made to preliminary data in consecutive R01 applications, the reviewers would have received both the previous and current applications.
ORI noted that Poehlman’s colleagues did not have access to his grant applications except to help with editing of selected portions; they were involved in the preparation and review of manuscripts for publication.
ELISA stands for Enzyme Linked Immuno-Sorbent Assay. A reagent such as a purified protein or antibody is bound to the surface of small plastic trays typically containing multiple wells (such as in an 8 × 12 grid), and control and test substances are examined for their ability to bind to that substance. The extent of this binding is measured by an instrument called an ELISA reader after a chemical is added to each well which generates color proportional to the degree of binding.
ORI’s oversight review of these eight experiments established on other grounds that these results were also falsified, and the respondent eventually agreed that all of the data in the original paper had been falsified, and that he had in large part fabricated the results. Thus, the data provided to the institution a year later as the raw data were in fact either fabricated, as in the three experiments without counter printouts, or falsified, as in the eight assays with printouts, which actually were obtained from printouts from irrelevant sources which the respondent crafted into very creative protocols to fit the scintillation counter tapes that otherwise made no logical sense in terms of how one would normally design the assay.
This was documented during the University’s investigation and copies of the documentation and correspondence were sent to ORI with the investigation report.
The multiple samples in each assay are plotted next to each other to avoid being obscured by overlapping each other. Therefore the chronological order is only in terms of consecutive experiments and not between consecutive data points within an assay.
It is not possible in the space available to provide much of the additional information that helped ORI to make a determination in this case. However, additional evidence developed during ORI’s oversight review established that the respondent was notably sloppy in conducting his experiments (inconsistent with his remarkably clean results), and in the two experiments observed by the laboratory it was determined that the respondent had not added radioactive chromium to the assay plates designed to determine the degree of chromium release due to the interaction of effector and target cells. For example, residue in tubes from which the samples were purportedly taken for scintillation counting was shown to lack radioactivity.
The Public Health Service is a part of the Department of Health and Human Services (HHS) and includes the Office of Public Health and Science and the following Operating Divisions: Agency for Healthcare Research and Quality, Agency for Toxic Substances and Disease Registry, Centers for Disease Control and Prevention, Food and Drug Administration, Health Resources and Services Administration, Indian Health Service, National Institutes of Health, the Substance Abuse and Mental Health Services Administration, and the offices of the Regional Health Administrators.
References
Dahlberg, J. E., & Mahler, C. C. (2006). The Poehlman case: Running away from the truth. Science and Engineering Ethics, 12, 157–173.
Martinson, B. C., Anderson, M. S., & de Vries, R. (2005). Scientists behaving badly. Nature, 435(7043), 737–738.
Mosimann, J. E., Dahlberg, J. E., Davidian, N. M., & Krueger, J. W. (2002). Terminal digits and the examination of questioned data. Accountability in Research, 9, 75–92.
Mosimann, J. E., & Ratnaparkhi, M. V. (1996). Uniform occurrence of digits for folded and mixture distributions at finite intervals. Communications in Statistics, 25(2), 481–506.
Mosimann, J. E., Wiseman, C. V., & Edelman, R. E. (1995). Data fabrication: Can people generate random digits? Accountability in Research, 4, 31–55.
Poehlman, E. T., Toth, M. J., & Gardner, A. W. (1995). Changes in energy balance and body composition at menopause: A controlled longitudinal study. Annals of Internal Medicine, 123(9), 673–675. Retraction in: Sox, H.C. (2003). Annals of Internal Medicine, 139(8), 702.
Sox, H. C., & Rennie, D. (2006). Research misconduct, retraction, and cleansing the medical literature: Lessons from the Poehlman case. Annals of Internal Medicine, 144(8), 609–613.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Dahlberg, J.E., Davidian, N.M. Scientific Forensics: How the Office of Research Integrity can Assist Institutional Investigations of Research Misconduct During Oversight Review. Sci Eng Ethics 16, 713–735 (2010). https://doi.org/10.1007/s11948-010-9208-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11948-010-9208-4