Clinical trials play important role in development of innovative medical treatments. Effective clinical trials relies on a number of factors, including the pro-active methods and tools, protection of patient safety, data quality, etc.
Data quality can be delivered only with support of efficacious tools and innovative approaches to clinical trial monitoring, among of them is the statistical monitoring. It is specific data analysis, which implemented in parallel with the on-site and the off-site monitoring activities. This method helps with prompt identification of systematic mistakes or non-compliance in clinical trial (which could significantly influence the data integrity and jeopardize the clinical trial conduct). Statistical monitoring it is one of the way to ensure the quality of clinical trials and its accuracy.
There is no holistic and consolidated industry approach in handling data integrity and management of non-compliance yet. In order to investigate the effectiveness on statistical monitoring TransCelerate BioPharma conducted an experiment checking the possibility of implementation of statistical methods for detecting fabricated data and other signals of non-compliance1, 3, 4.
TransCelerate tested statistical monitoring on a data set from a chronic obstructive pulmonary disease (COPD) clinical study with 178 sites and 1554 subjects. A TransCelerate member company provided deidentified clinical database from a randomized, double-blind, placebo-controlled, study in chronic obstructive pulmonary disease (COPD) with already locked database. Fabricated data were selectively incorporated in 7 sites and 43 subjects. Analyses of vital signs, spirometry, visit dates, and adverse events included distributions of standard deviations, correlations, repeated values and digit preference1, 2.
There were 4 simulated studies with different number of subjects enrolled:
- Study 2A (61 sites, 338 subjects)
- Study 2 (61 sites, 627 subjects)
- Study 1A (178 sites, 824 subjects)
- Study 1 (178 sites, 1554 subjects)
An interpretation team, including clinicians, statisticians, site monitoring, and data management, reviewed the results and created an algorithm to flag sites for fabricated data. The algorithm identified 11 sites (19%), 19 sites (31%), 28 sites (16%), and 45 sites (25%) as having potentially fabricated data for studies 2A, 2, 1A, and 1, respectively. For study 2A, 3 of 7 sites with fabricated data were detected, 5 of 7 were detected for studies 2 and 1A, and 6 of 7 for study one1, 2. Only for study 2A the approach was not effective, for the rest 3 studies the algorithm had good sensitivity and specificity (>70%) for identifying sites with fabricated data.
Listed above experiment showed high effectiveness of statistical monitoring in clinical trials and TransCelerate recommended a cross-functional approach to statistical monitoring that can be adopted to the study design and data source1. Definitely, you can find more detailed information about this experiment with detailed explanation of the algorithm and all methods.
From my point of view, statistical monitoring has high potential for further development. It is visible that implementation of this approach to clinical trial oversight would increase data quality and efficiency of clinical trials conduct while assuring subject safety.
- Develop, Innovate, Advance (DIA)
- TransCelerate BioPharma Inc. (2013) Position paper: risk-based monitoring methodology
- USA Department of Health and Human Services, Food and Drug Administration (2011) Guidance for industry: oversight of clinical investigations—a risk-based approach to monitoring
- European Medicines Agency (2011) Reflection paper on risk based quality management in clinical trials. EMA/INS/GCP/394194/2011