Structured&

Comments on the SDC Use Case Document (version dated 4/26/2013):

Lines 92 and 496 reference an assumption that the data pre-populated from the EHR will be the patient's "most recent" data. While using only the "most recent" data simplifies the process and could be feasible in some settings for some studies, it is not always feasible:
 * This is not the current practice for clinical research (research data is most often captured some time after the actual visit) and even if that practice changes, the busy demands of health care could interfere with timely research documentation.
 * In a large medical center setting, a patient could have multiple encounters with different clinicians all over the span of a few hours and all documented in the same system. If the research data capture process is not immediate, the "most recent" data could come from an encounter that occurred after the "research visit".
 * In some settings (e.g. inpatient, emergency dept, critical care, etc.), the patient encounter could span over several days and could include the same assessments documented numerous times over the time frame. The end user, may need the ability to select the data gathered specific to a research visit based on timeframe or other factors such as the "study nurse" documenting the data vs. the "floor nurse". Example: vital signs are regularly documented in the EHR during an inpatient stay but the vital signs collected 4 hours after a specific procedure are the only ones of interest to the research study. The site research personnel need the ability to select the specific set of vital signs (4 hrs. post procedure).

Line 207: with auto pre-population, some data elements need to be linked/grouped to ensure they stay together. For instance, systolic and diastolic BP are captured as discrete data elements but they should be "linked together" when they are captured and should remain linked during any edits. Not sure if this is in-scope or out of scope

Line 334: Suggest the addition of a use case assumption:
 * In some use cases, data needs to be captured and continually documented as a running record (e.g. concommitant medications and adverse events). Data captured on subsequent visits may need to be reconciled with the previously collected data to ensure accuracy of the continuing record. Example: on visit 1, lisinopril 10 mg was captured; on visit 2, lisinopril 10mg and Furosemide 40 mg were captured. Is the Lisiniopril 10mg the same Rx or is it a 2nd Rx (20mg total)?

Line 497: The pre-populated data in the eCRF should reflect the data sourced from the EHR and therefore pre-populated data should not be edited within the eCRF. Edits should be made within the EHR (source system) so the pre-populated data is accurate. For clinical research, auditors must be able to discern the clear sequenece of events from the audit trail. Therefore the archive should include an audit trail documenting: EHR source system; date/time/creator of the data within the EHR system; date/time/end-user who pre-populated and saved the eCRF; date/time/end-user for any updates. For data captured de-novo, i.e. not sourced from the EHR, the audit trail needs to capture date/time/creator and date/time/end-user for any updates.