What SDV is really for? How it develops and to what form it evolves? What is the future role of a CRA?
This article depicts a role of SDV in a clinical monitoring process, shows the path of SDV evolution and helps to understand what we all need to be prepared in the future.
Source data and the source data verification (SVD) process remain an essential part of quality assurance in clinical research. For about 60 years, SDV has been the central and most time consuming method of quality assurance in clinical trials conducted by monitors (CRAs).
In ICH Good Clinical Practice (GCP), clinical monitoring is defined as
“the act of overseeing the progress of a clinical trial and of ensuring that it is conducted, recorded, and reported in accordance with the protocol,[standard operating procedures], GCP, and the applicable regulatory requirement(s).”
In traditional literature, the main activity of clinical monitoring—SDV—is described as a process that “ensures the integrity of the data and therefore forms a crucial part of the on-site QC process and from a regulatory perspective provides evidence that the data is true and accurate when submitting for a product license” .
From the above mentioned definition, we can identify two major assumptions of SDV:
- The source data are always true.
- Ensuring the congruency of source data and the case record form (CRF) improves the data quality and integrity.
Let us now analyze them in detail.
Assumption No 1: The source data are always true
Dr. M. Herschel, a former Director of Clinical Operations of GSK Germany, writes in his blog that SDV is only
“ensuring that two sets of data are congruent. This does not mean that both sets of data are true […] Conversely, some investigators use the CRF to record their findings and then transcribe the data onto the Patient Chart because this is more convenient for them” .
Therefore, even a CRF can even become a source document, and later comparison with an erroneous copy may even decrease the quality of clinical data.
Furthermore, a patient can provide wrong information (so-called professional patients) or a doctor can invent data (fraud) or wrongly document (sloppiness). In these cases, SDV cannot identify the discrepancies.
SDV focuses on and puts much efforts into controlling data transformation from source data to CRFs and avoiding other areas such as patient -> source data, or CRF -> data base, assuming that these transfers always work well. Is this always the case?
Figure 1. The isolated focus of SDV in the data acquisition and submission path
Assumption No 2: SDV improves data quality and integrity
In recent research involving over 1100 clinical trials in Phases I to IV, the authors questioned the added value of SVD compared to consumed time—e.g., the corrected amount of data as a direct result of SDV is around 1.1 percent, including both critical and noncritical data .
Figure 2. Percentage of electronic case report form corrections attributed to source data verification (SDV) versus other data correction methods (Source: Sheetz et al.)
So obviously, the SDV improves data quality and integrity, but its influence on general data quality improvement is quite negligible.
Does clinical monitoring just mean SDV?
So, why do pharmaceutical companies invest 22 percent of their clinical costs in clinical monitoring? In reality, the on-site visit of a monitor has the purpose not only of fulfilling SDV but also of helping a site to embrace regulatory requirements, understanding how a site treats a clinical trial, and so on. It influences the quality of a site’s output, and answers a lot of additional questions. Besides, a site feels support from the attention of a sponsor. This is particularly important to sites which are newcomers to clinical research.
Increase in complexity
Another big trend in clinical trials today is the increase in the amount of data and the increase in the complexity of clinical trials, discussed in ICH’s GCP addendum as one of the major trends and a reason for the GCP adjustment. These trends force sponsors and CROs to conduct trials more efficiently by utilizing new and better technological solutions .
Recently, new techniques for evaluating clinical data have appeared which are able to find certain anomalies and discrepancies. These automatic or semiautomatic procedures assume that the amount of effort necessary in SDV will decline over time, although the full chain of data acquisition and transformation will be under control.
The time of a CRA will be used more efficiently, as they will be empowered by new automatic algorithms which, with a high degree of efficiency, can detect fraud, bad performance and anomalies of other sorts. The advantage of this approach is the ability to cover bigger amounts of data in a shorter time while still remaining efficient.
Emergence of electronic CRFs
From another angle, there is a big trend for electronic CRFs (eCRFs), which will eliminate the need for the SDV completely. Taking into account that the SDV in 10–15 years will become history, we can forecast that on-site monitoring as a measure to build up personal contact with a site will remain. Still, we can forecast that the number of on-site visits will decline and become more targeted or triggered, as well as partially replaced by remote conferences.
A new definition of the CRA role
The role of a monitor will be newly defined. It is now evolving from an entry job to the role of a highly qualified expert who can interpret risk indicators, estimate triggers, and react to them accordingly, applying the right tool from the CRA tool-set. The CRA tool-set already includes remote contact with a site, visits on-site, SDV audit, reports about conduct, statistical research, review of the data from quality control programs, queries, corrective action plans, re-audits, etc.
Summing up, nowadays, both SDV as an activity and clinical monitoring as a process experience a great deal of pressure from industrial trends, such as the increase in data amount, increase in clinical trial complexity, and questioning of their efficiency compared to costs. The way out from this pressure is through the application of more sophisticated algorithms for data quality control which can oversee and control the whole data transformation chain. The role of a CRA, in this scenario, evolves to a highly-educated data expert, who can manage a clinical trial from a central place using the appropriate tool from the CRA tool-box.
ICH, “ICH Guideline: Good Clinical Practice.” [Online]. Available: http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Efficacy/E6/E6_R1_Guideline.pdf. [Accessed: 07-Mar-2016].
P. A. Carson, N. J. Dent, and R. S. of C. (Great Britain), Good Clinical, Laboratory and Manufacturing Practices: Techniques for the QA Professional. Royal Society of Chemistry, 2007.
M. Herschel, “Does Source Data Verification (SDV) need.” [Online]. Available: http://herschel-health.de/michaelherschelsheresies/aeltere-ausgaben/michaelherschelsourcedataverification/index.html. [Accessed: 23-Feb-2016].
Sheetz, Nicole, et al., “Evaluating source data verification as a quality control measure in clinical trials.,” Therapeutic Innovation & Regulatory Science, vol. 48.6 (2014), pp. 671–680.
ICH, “Integrated addendum to ICH E6(R1): Guideline for Good Clinical Practice E6(R2),” 11-Jun-2015. [Online]. Available: http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Efficacy/E6/E6_R2__Addendum_Step2.pdf. [Accessed: 16-Jun-2015].
R. Olsen, A. R. Bihlet, F. Kalakou, and J. R. Andersen, “The impact of clinical trial monitoring approaches on data integrity and cost-a review of current literature,” Eur. J. Clin. Pharmacol., Jan. 2016.
J. R. Andersen, I. Byrjalsen, A. Bihlet, F. Kalakou, H. C. Hoeck, G. Hansen, H. B. Hansen, M. A. Karsdal, and B. J. Riis, “Impact of source data verification on data quality in clinical trials: an empirical post hoc analysis of three phase 3 randomized clinical trials,” Br J Clin Pharmacol, vol. 79, no. 4, pp. 660–668, Apr. 2015.