Login     Sign Up     |      Home      Regions      Verify a Certification     EBA Portal    

Types of Data Fraud and Detection in Clinical Trial data

Data fraud is of serious concern to the sponsors of clinical trials and the regulatory agencies responsible for monitoring these trials. This could be due to the rise of Risk-Based Monitoring as an effective strategy in term of cost and efficiency when conducting trials.
As long as clinical data is involved, it is impossible for the clinical trial database to be 100% error free. One of the most important requirements of the ICH-GCP is that the data gotten from clinical trials are correct, comprehensive and from reliable sources. This begs the question, at what point do we differentiate from an error from a fraud?
The perception is that Data Fraud in clinical trials is close to non-existent. Estimated figures put this at less than 1% of trial cases recorded. The reason for this is that the process of identifying and documenting fraud could be considered time-consuming and costly with the potential to cause irreparable damage to the reputation of the company carrying out the research.
Are we underestimating the importance of fraud detection in clinical trials? Should we concern ourselves with looking for fraud?
  • The lives of patients and overall health could be at risk
  • Minimize the number of failed outcomes
  • Maintain and increase the reputation of clinical research

Types of Clinical Data Fraud
The following mistakes could have very serious effect on the outcome of clinical trials if they lead to increase treatment effects :
  • Plagiarism
  • Manufactured Data
  • Manipulate Data in order to achieve the desired objective: for instance, making patients eligible, showing a treatment effect.
Fraud Detection in Clinical Trials
A typical way to detect fraud in clinical trials would typically involve a visit to medical centers to verify the source data and also make further investigation where necessary. This approach is both costly and time-consuming.
Alternatively, you could depend on statistical methods to help detect irregular data patterns. These patterns could be converted into graphical checks to help detect fraudulent data. Data prone to fraud include; eligibility criteria, repeated measurements, adverse events, assessment of medical compliance, assessment dates and patient diaries.
If suspected fraud is confirmed to be true, then steps need to be taken to notify the monitors and relevant onsite audit and proper investigation should be carried out.