A large portion of source documents are electronic and may be available to the sponsor remotely such as laboratory and radiology reports or source documents submitted by the CI for other purposes (e.g. documenting serious adverse events or adjudicated events). Consistent with ICH E6 and ISO 14155:2011, original observations can be entered directly into the eCRF or transmitted to the eCRF from various locations, devices, or instruments.
The FDA recognizes that sponsors may not have remote access to all electronic health records. Particularly records maintained by hospitals, universities, and other institutions because of data privacy, security concerns, and/or technological challenges.
As discussed in this guidance, a variety of centralized monitoring techniques can be used to replace, supplement, or target on-site monitoring activities. The majority of these techniques can be performed regardless of the extent of use of electronic records in the study. These techniques include site checks for completeness of data, frequency of protocol violations, and screen failure rates.
For example, the majority of these techniques can be performed using CRF data collected either using electronic data capture systems or entered into a database from a paper CRF collected by the sponsor.
A recent publication discusses statistical techniques for identifying various types of data errors. The statistical techniques described in this guidance may not be routinely used by all sponsors and may not be appropriate for every trial, but they are included in this guidance as examples of monitoring techniques that may be considered by sponsors.
Additional monitoring techniques, such as routine review of data as they are submitted, are possible for studies that use electronic CRFs. Although not a monitoring technique, another method of ensuring data quality routinely implemented in eCRFs is the use of electronic prompts in the eCRF. This method minimizes errors and omissions at the time of data entry, particularly if data are entered directly into the eCRF.
Source Data Verification and Corroboration:
The sponsor should consider the quantity and types of source data that need to be verified against CRFs or corroborated against other records during the sponsor’s identification of critical data and processes, risk assessment, or both. For example, review of medical record to corroborate a subject’s response of “no hospitalizations” since the previous visit on a CRF.
Also to note, a description of the quantity and types of source records should be included to verify or corroborate in the monitoring plan. Source records that provide the most meaningful information about a subject’s participation and the CI’s conduct and oversight should also be considered.
For example, for a particular study, there may be minimal benefit in comparing 100% of the source data for each subject to the CRFs for each study visit. Rather, it may be sufficient to compare the most critical data points for a sample of subjects and study visits as an indicator of data accuracy. Similarly, for a particular study, all concomitant medications, body temperature, and body weight are required by the protocol, to be collected, documented, and transcribed to a CRF. However, they may not be identified by the sponsor as critical data, because a small error rate in those variables would not affect the outcome of the trial.