How to Detect, Manage, and Report Fraud and Fabricated Clinical Research Data

Clinical research associates (CRAs) must be aware of the many signs of possible fraud or fabricated data occurring at a clinical trial site. Whether indications of fraud appear as missing documentation, altered numbers, signatures that don’t match, or unavailable medical records, CRAs need to look for signs and understand the proper protocol to handle and report the misconduct during each monitoring visit. According to the NIH, different studies estimate rates of fraud between .01 and .40 percent. (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4340084/)

Before we get into the various signs of fraud and fabricated data, we must first take a moment to define data falsification/fabrication at a clinical trial site. This may include:

· Creating, altering, recording, or omitting data in such a way that the data does not represent what occurred during the trial. 

· Serious or repetitive failure to comply with Good Clinical Practice (GCP), regulatory requirements, or other applicable standards. This may result in false data due to the lack of understanding of regulatory consequences, inattention to detail, inadequate staffing, or lack of supervision.

Some of these actions at the trial site may be unintentional and not done to deliberately deceive, but can still jeopardize the safety of subjects and the reliability of data. Fraud, on the other hand, connotes these types of actions with the intent to deceive.

Commonly falsified documents include:

· EKGs/ECGs

· Blood pressure data

· Physical examinations

· Biological specimens

· Subject identities

· IP compliance records

There are three data falsification categories that we’ll look at in this article-omitted data, altered data, and manufactured data.

Omitted data refers to a conscious decision made to not reveal all data during the trial. Examples include not reporting adverse events and concomitant medications. If an investigator or designee does not make all subject medical records available for monitoring and inspection, and possibly has “shadow charts,” which are typically paper copies of original records retained apart from the primary custodial area; this could be a red flag.

Altered data is when a study team member consciously manipulates or changes data that was legitimately obtained, such as altering laboratory reports, study-specific test results or changing a test date or patient identifier. Unjustified changes to important data, late entries not fully explained, alterations to lab reports or other printouts from electronic systems should all be considered points of concern that data may have been altered.

Manufactured data occurs when information has been fabricated or results were created without actually performing the work. Examples of manufactured data include filling in values in the case report form (CRF) for which no data was obtained, photocopying data from one patient for another, and creating fictitious subjects. Clinical monitors should be on the lookout for inconsistent source documentation practices, suspicious signatures or handwriting, obliteration of subject identifiers, dates of birth, or visit/test dates. Additionally, dissimilar signatures or signatures that are too similar (i.e., copied, traced, or scanned from authentic signatures) can be signs of falsified documentation.

Detecting fraud and data fabrication

There are many steps a CRA can take to help ensure a trial site is providing authentic data or identify possible misconduct. First, determine the trial site’s customary record-keeping method for documenting the data points as required by the study protocol. Next, verify the existence of all original data and confirm that all CRF data are supported by the original source data.

If a study-specific subject chart (shadow chart) is maintained in addition to the customary records, such as medical or clinic charts and progress notes, then the CRA should access and monitor all records supporting the data collected in the study-specific chart. If a written procedure exists to ensure shadow charts contain all relevant information, the CRA may only need to spot-check occasionally to verify the process is effective and the procedure is being followed.

Possible signs of fraud or data fabrication happening at a trial site:

· Missing or incomplete source documents

· Forms of obliteration of data and/or frequent corrections to the source documentation

· Total obliteration of subjects’ names on medical records or tests

· Significantly more or fewer screen failures or early terminations than at other trial sites

· Fewer adverse events, concomitant medications, and physical exam findings than at other sites

· “Too-good-to-be-true” protocol and event schedule adherence

· Greater-than-average occurrence of laboratory sample issues (e.g., shipped beyond stability, lost, not done)

· Subject diary cards that have changes, appear to be completed all at once, or seem to be in pristine condition (e.g., free from normal wear and tear, same ink color, etc.)

· Too many subject visits on the same day

· Subject visits that occur on holidays, weekends, or dates on which the key study staff were known to be away from the clinic

· Similar handwriting on several consent forms, or varying signatures attributed to the same subject

If a CRA suspects misconduct at a trial site, and if repeated attempts to obtain missing records from responsible investigator site staff are unsuccessful, they must notify the clinical project manager or lead CRA immediately without alerting the suspected person(s). The lead CRA or supervisor should ensure the issue is escalated to the quality assurance team in accordance with applicable procedures. If the suspected person is an employee, the initial misconduct report may be made directly to the head of QA and/or the head of human resources.

All suspicions should be explained and supporting documentation should be provided. An accusation of research misconduct is very serious; therefore, an investigation should take place before the principal investigator (PI) gets involved. Doing otherwise could have serious legal implications for both the CRO managing the trial and the clinical monitor for the site.

It is important for a CRA to remember to not take “no” for an answer regarding access to source records. Also, CRAs should read and evaluate all source notes and test results for their legitimacy-rather than just taking inventory of the source documents. CRAs should question when standard record-keeping practices are not followed and when information is missing (e.g., dates, times, etc.), as well as when missing information is suddenly added. Lastly, CRAs should challenge questionable explanations, and point out inconsistencies in dates, handwriting, and signatures.

By being proactive and taking these steps, CRAs can ensure that possible fraud and fabricated data is detected during routine monitoring.

Lysa Triantafillou, Director of Quality Assurance, Rho