Source documents are the original records in which clinical research data are first captured. They may include hospital files, clinic notes, laboratory reports, pharmacy records, imaging reports, participant diaries, consent forms, vital signs charts, electronic health records, device outputs, field notebooks, or study-specific worksheets. Source data are the values and observations recorded in these documents that are necessary for reconstructing and evaluating the
study.
The relationship between source documents and the research database is central to data integrity. A database value should be traceable to an appropriate source unless the eCRF itself is defined as the source. For example, if a nurse measures a participant’s temperature and enters it directly into REDCap during the encounter, the REDCap entry may function as the source record if the protocol and data management plan define it that way. If the nurse first writes the value on a clinic worksheet and a data clerk later enters it into REDCap, the worksheet is the source document and the REDCap entry is a transcription of that source.
Source Data Verification, often abbreviated as SDV, is the process of comparing database values against source documents to confirm accuracy, completeness, and consistency. In traditional monitoring models, monitors may verify a high proportion of key variables, especially primary outcomes, informed consent, eligibility criteria, and safety events. In risk-based approaches, SDV may focus on critical data and processes rather than every field. Regardless of the model, thestudyteammustknowwhichsourcedocumentssupporteachdatabasevariable.
Source documentation is important because it allows reconstruction of the study. If a regulator, sponsor, monitor, auditor, or investigator asks how a value entered the final dataset, the team should be able to trace it back to the original observation or certified copy. Without source traceability, the credibility of the data is weakened. In addition, source documents protect participants and study teams by demonstrating that procedures occurred as required by the protocol.
The data manager should document source expectations in the data management plan or source document agreement. This is especially important when data come from multiple sources. A laboratory value may come from a laboratory information system, a printed report, or an uploaded file. A follow-up outcome may come from a clinic visit record or a phone follow up form. A medication exposure may come from pharmacy dispensing logs or clinician notes.
The source hierarchy should be clear.