According to several studies, 0.1–0.4% of all clinical research is conducted fraudulently. (Estimates of the frequency of negligence are not available, but they would certainly be greater.) As these numbers indicate, fraud is rare. But with more than 100 investigators operating under the cloud of an FDA reprimand, fraud and negligence are genuine research concerns.

Fraud
The motive for investigative fraud is almost always financial (although career advancement is a close second). The CFR requires that researchers disclose any proprietary interest they have in a project under their supervision. In addition, that disclosure must include reports of any “significant payments of other sorts,” such as funds provided for consultations, speaking engagements, advisory board memberships, and appointments to boards of directors. Financial connections are so common that The Journal of the American Medical Association estimated in its November 1, 2000, issue that 7.6% of academic researchers were tied to their sponsors through proprietary interests or significant payments of other sorts.

The fact that payments made to investigators for conducting clinical trials are based entirely on patient enrollment and retention is not addressed by the CFR. However, this is significant because if no patients are entered in a trial, the investigator collects no income from the project. Or if a patient discontinues early, the researcher is paid only for the portion of the study the participant has completed. It is therefore in the investigator’s financial interest to enroll large numbers of patients quickly and retain them as long as possible. Thus, taking nothing else into account, clinical investigators have an undisclosed monetary interest in their projects that could lead them to commit acts of fraud. The worst examples of fraud include the researcher who invents patients, purposely enters otherwise ineligible participants, and alters or lies about medical assessments that might force a patient’s early withdrawal from a study.

Besides the fraudulent actions just mentioned, several research practices are blatantly illegal, including faking medical records, deliberately changing patient evaluations, purposely omitting adverse events, knowingly providing excluded medication to study patients, forging informed consent documents, and hiding instances of negligence. Several of these violations, however, drop to the level of negligence when the adverbs “deliberately”, “purposely”, and “knowingly” are removed. (For example, a patient may inadvertently be given an excluded medication, and although this is a protocol violation, it is in no way fraudulent.)

Negligence
Many instances of negligence result from investigators’ laziness (in not properly reviewing research-related data and accompanying legal documents), sloppiness (in managing the trial or completing necessary paperwork), or ignorance (due to a lack of research training or regulatory awareness). Commonly, investigators and their staff simply make errors in study conduct, such as omitting a significant adverse event from the case report form document. When the errors are caught, investigative site personnel usually describe them as “protocol deviations”. All deviations are, in fact, violations that constitute negligence.

Investigators might argue that their actions did not harm any patients and therefore no real violation occurred, but the fact that patients were not harmed does not mean they couldn’t have been harmed. In addition, the integrity of data from single-site studies is necessarily compromised by such errors, as is data from multisite trials if errors are compounded across investigative centers or occur at high patient enrolling sites, that is, individual sites that enroll large numbers of patients in a single clinical study. Therefore, all violations are significant and constitute negligence.

This is not to say that the FDA would automatically discipline an investigator who accidentally provides an excluded study drug to a patient, such as marketed benzodiazepines to a patient enrolled in a trial of a novel anxiety medication. The study monitor will usually catch these errors early, ask the site to remedy them, follow up to confirm that action was taken, and document the violation and its remedy. The FDA would not generally consider the matter further. A serious problem emerges, however, when the violations are not caught, addressed, or rectified, and when the errors become serial in nature.

Getting caught
Negligence and fraud are detected by one of three mechanisms: monitoring, sponsor auditing, or FDA auditing. Monitors are usually the first lines of defense against fraud and negligence because they review every piece of data generated by a site before anyone else can. Site monitors also have the best vantage point from which to determine whether an investigator’s research is sloppy or willfully false (although this distinction is often extremely difficult to make).

For sites that enroll large numbers of patients, sponsoring agencies usually send their own auditors to ensure regulatory compliance and data integrity. Sponsor auditing is done in addition to regular monitoring site visits, and it serves as a kind of “backup review” of the collected research.

Finally, the FDA sends its own auditors to review research conduct. These audits generally occur least often; the FDA’s budget constraints restrict the number of sites visited each year. Sites that conduct “pivotal” trials are often chosen for audit, but those participating in Phase III trials are often among the most heavily targeted. Of these, centers that are preferentially selected for audit are the “high enrollers” and those that show significant screen failure rates, have large numbers of patients who discontinue early, and/or report significantly more adverse events than do other sites on the same trial.

Discovering and documenting protocol violations are relatively simple during a thorough data review. Determining whether fraud is the cause of these deviations, however, is often difficult. Warning signs often include

• case report forms that contain data for several patients over long time periods written using the same pen,
• patient visits that occur remarkably close to one another,
• different patients’ diary cards that are oddly similar,
• patient visit dates that occur on implausible days, such as holidays and weekends,
• similar patient handwriting that appears on different patients’ consent forms, and
• investigators who have signed review documents before their completion.

Penalties
The FDA routinely audits sites looking for instances of fraud and negligence. But charges are made with extreme caution. For example, an investigator who enrolls a patient who should have been excluded from a study because of a high blood pressure reading, would likely be scolded for a protocol violation that could have harmed the patient. If that investigator seems to make a habit of ignoring blood pressure readings, however, more serious actions related to a charge of negligence would be taken. Only when it becomes apparent that an otherwise conscientious researcher consistently overlooks significant vital sign readings would questions of fraud be considered. Even then, other evidence would have to be collected before a formal charge could be made.

It is generally up to the courts to address charges of fraud; the FDA simply follows up by disbarring convicted investigators. As stated at the beginning of this article, however, fraud charges are rare. Negligence, though, is a different matter. When FDA auditors discover issues of concern during site visits, they usually ask investigators to submit written responses to “inspectional observations” as noted on Form FDA 483. If the responses do not adequately address the concerns, a Warning Letter is sent to the researcher by certified mail. This letter demands a written response within 15 days of receipt.

Based on the FDA’s level of satisfaction with the response to the Warning Letter, one of three actions is taken. The first, known as an “assurance”, indicates that no punishment will be invoked because the investigator adequately assured the FDA that its concerns were addressed and that the problems will not occur again. The second action, known as a “restriction”, places limits on an investigator’s participation in research projects. This prohibition could disallow a physician to conduct trials for a specified number of years or to serve as principal investigator on any trial for some length of time. Finally, and less frequently, the FDA can outright disqualify investigators from participating in any clinical research project at any level.

Fraud and negligence are serious issues in clinical research. Given the number of trials conducted in the United States and the FDA’s limited budget, the vigilance of monitors to identify and resolve problems early is extremely important. The FDA has ways of protecting study participants from research harm, and safeguards to protect the general public from being treated with unsafe medications. But fraud still exists, and negligence, in the form of protocol violations, can significantly affect the health of research subjects and the quality of data. Thus, investigators should vigilantly adhere to the rules of their industry, and those charged with auditing research should be thorough in conducting their reviews.