Source Data Verification (SDV)

Source data verification (SDV) is the process, performed as part of clinical-trial monitoring, of comparing data recorded on a case report form (CRF or eCRF) against the source document it was originally transcribed from — confirming the transcription is accurate, complete, and consistent with the underlying source data. It answers one specific question for a given data point: does what’s on the CRF actually match what the source document says? SDV is a monitoring activity, not a software platform — it is commonly confused with, but distinct from, the Electronic Data Capture (EDC) system that hosts the eCRF being checked. An EDC platform can support SDV (e.g., by presenting the eCRF for side-by-side comparison, or flagging fields not yet verified), but the verification itself is a human or process-driven quality-control step layered on top of the EDC, not a function the software performs by itself.

How SDV fits into the broader monitoring workflow

SDV is one specific technique within the larger discipline of clinical trial monitoring, most often performed by a Clinical Research Associate (CRA) during a site monitoring visit, whether conducted on-site or remotely. The Case Report Form (CRF) entry is the item being checked; the source document (a clinic chart, lab printout, imaging read, or diary) is what it’s checked against. When SDV surfaces a discrepancy — a mismatched date, an incorrectly transcribed lab value, a missing field — it is documented as a query and routed back to site staff for correction or a documented explanation before that data is considered clean. Any correction made after initial CRF entry must preserve a visible audit trail rather than silently overwrite the original value, consistent with the ALCOA+ data-integrity attributes (attributable, legible, contemporaneous, original, accurate, plus complete, consistent, enduring, available) that clinical trial data is generally expected to meet.

100% SDV vs. targeted, risk-based SDV

For years, many sponsors defaulted to verifying essentially every CRF field against source for every participant — ‘100% SDV’ — as an unwritten industry norm. Neither ICH E6(R2) nor FDA guidance has ever actually mandated this. FDA’s 2013 guidance for industry, Oversight of Clinical Investigations — A Risk-Based Approach to Monitoring (elaborated by a follow-up FDA Q&A guidance finalized in April 2023), explicitly encourages sponsors to focus monitoring resources — including SDV — on the data and processes most likely to affect participant safety and the reliability of trial results, rather than verifying every field by default. ICH E6(R2)’s Section 5.18 (Monitoring) and Section 5.0 (Quality Management) formalized this shift for Good Clinical Practice generally, and ICH E6(R3), which reached ICH Step 4 on 6 January 2025, carries the same risk-proportionate principle forward with even more explicit emphasis on quality management as a governing approach rather than a monitoring-technique afterthought.

In practice, this means a modern monitoring plan typically specifies targeted SDV: verifying a defined set of critical data (commonly primary-endpoint data, serious adverse events, and informed-consent documentation) at or near 100%, while sampling a smaller percentage of lower-risk secondary data — rather than applying the same verification intensity uniformly across every field. The rationale, and its evidence base, is genuinely debated in the literature: a 2024 scoping review found no absolute, settled measures of SDV accuracy and noted that manual SDV is itself susceptible to human error in ways similar to the record abstraction it’s meant to check — meaning the shift toward reduced/targeted SDV proceeded without a fully settled evidence base proving 100% SDV produces meaningfully more error-free data. Industry estimates (reported by vendors such as Medidata, among others) have put SDV at roughly a quarter to 40% of on-site monitoring cost and a substantial share of on-site monitoring time historically, which is the practical reason sponsors have had strong incentive to target it rather than apply it uniformly.

SDV within risk-based monitoring (RBM)

SDV is one tool within the broader Risk-Based Monitoring (RBM) strategy, not a synonym for it. RBM is the overall approach — combining centralized (remote, statistical/analytics-based) monitoring, on-site visits, and quality tolerance limits — that a sponsor’s monitoring plan uses to concentrate oversight where risk to participant safety or data reliability is greatest; SDV is specifically the source-document-comparison technique deployed within that plan. ICH E6(R2) Section 5.18.3 explicitly permits centralized monitoring — ‘a remote evaluation of accumulating data, performed in a timely manner, supported by appropriately qualified and trained persons’ — alongside or instead of exclusively on-site verification, with the chosen mix documented and justified. This has driven growth in remote SDV (sometimes rSDV): CRAs verifying CRF entries against source documents made available electronically (e.g., through a site’s electronic health record portal or a secure document-sharing system) without traveling to the site, rather than only through in-person on-site visits.

SDV vs. source data review (SDR)

Industry usage sometimes distinguishes SDV from a related but narrower activity, source data review (SDR): reviewing source documents for participant safety and protocol-compliance signals (e.g., is an adverse event apparent in the chart that hasn’t been reported yet?) without necessarily performing a field-by-field comparison against every corresponding CRF entry. This SDV/SDR distinction is a common industry convention rather than a term formally defined in ICH E6 itself, and organizations vary in how strictly they separate the two activities in practice — but the underlying difference is useful to keep straight: SDV confirms transcription accuracy for specific already-entered data points, while SDR is a broader safety/compliance-oriented read of the source record.

Related CASRAI content

• Electronic Data Capture (EDC) — the software platform that typically hosts the eCRF being verified; SDV is a monitoring activity performed against data an EDC system stores, not a function the EDC performs itself.
• Case Report Form (CRF) — the instrument whose entries SDV checks against source documents; covers CRF design, protocol traceability, and the query process in more depth.
• eSource (Electronic Source Data) — when source data itself originates electronically, the ‘source document’ SDV verifies against is the electronic record (or a certified copy), not a paper original.
• Risk-Based Monitoring (RBM) — the broader monitoring strategy that targeted/reduced SDV is typically deployed within.
• Clinical Research Associate (CRA) — the role most commonly responsible for performing SDV during monitoring visits.
• Clinical Data Management Tools: How EDC, CDMS, CTMS, eTMF, RTSM, and RBM Fit Together — how SDV and the systems that support it fit into the broader clinical data management technology stack.
• Data Management Plan for Clinical Trials — how CRF flow, verification, and database lock fit into a trial’s overall data management plan.