If You Build Them, Data Will Come: Making the Most of CRFs
To successfully meet strategic clinical development goals, a significant amount of time and attention goes into the design and execution of study protocols. In contrast, case report form (CRF) design, with its sole purpose of accurately capturing critical data, often does not get adequate scrutiny. First and foremost, CRFs need to include data fields, which directly or indirectly inform every study endpoint. This connection may seem obvious, but we continue to be surprised at how often study teams find data deficiencies well after the study is underway. Alternatively, some CRFs capture too much data, resulting in the wasteful expenditure of study team or vendor resources for clinical review, reconciliation, and source data verification of what are ultimately uninformative or redundant data.
While the vast majority of CRFs are standard and uniformly required across all studies, the careful inclusion of specialized data fields or entire CRF pages could be a key to a drug’s successful development and approval. Recognizing the essential data and how to capture them is vital to success. Here are a few examples:
Adverse events of special interest (AESIs); especially when multiple adverse event (AE) verbatim/coded preferred terms may inform an AESI category (eg, opportunistic infections, immune-related AEs, infusion reactions, hepatotoxicity). While the basic AESI data characteristics (start/stop dates, grade, causality, outcome, action on dosing, etc) are recorded on standard AE CRFs, a separate AESI CRF allows the Sponsor to collect additional information that will facilitate accurate summarization at the time of data presentation:
a. CRF will require site to specify the protocol-defined AESI category for each AE reported as an AESI. This will facilitate Sponsor team review and confirmation that the AE meets AESI criteria and expedite necessary queries for correction or removal.
b. CRF can capture specific management of interest, eg, steroid dosing for immune-related AESIs (or confirm not used), rescue medications for infusion reactions, antimicrobial medications for systemic infections, procedures, diagnostic testing.
Concomitant medications of clinical interest. Some drug classes or specific indications have proposed prophylaxis requirements, and a separate CRF makes it easy to account for compliance to the protocol or local best practices, which can be easily summarized for protocol deviation and site remediation practices or to support regulatory information requests or filings.
Important inclusion/exclusion criteria that define a protocol patient population. If patients require prior treatment with a specific therapy, it would be valuable to include a field on the Prior Therapy CRF to confirm and capture all necessary treatment details, which may be critical to support the intended drug indication.
CRFs that auto-derive calculated data, eg, RECIST efficacy data (sum of longest diameter, % change from baseline/nadir) significantly save site data entry time, reduce data entry error, and standardize response derivations.
Finally, while not always factored into development timelines, it is well worth the time to create strategic edit checks and well-defined CRF completion instructions (with detailed examples) right from the start. These additional timesaving, adjunctive tools will minimize the likelihood of needing expensive database updates and their consequential delays.
Virginia Kelly, MD
Acumen CMO and Founder