Strategies to accelerate evidence generation and application in colorectal cancer care

Randomized control trials (RCTs) inform clinical practice recommendations for screening and the management of colorectal cancer (CRC), while also supporting regulatory decisions of novel therapies. However, RCTs require significant time investment that delays patient access to novel interventions. Strategies to expedite evidence generation are needed to improve RCT utility within clinical and regulatory frameworks. This thesis explored computerized simulations of CRC screening, self-directed and asynchronous electronic consenting (e-consenting) and surrogate endpoint use in RCTs as acceleratory strategies. First, OncoSim-CRC was used to support timely re-evaluation of national screening guidance due to rising rates of early-onset CRC by modelling the economic impacts of earlier screening. In a simulated cohort of 10 million Canadians with age-adjusted early-onset rates, earlier fecal immunochemical testing saved lives while increasing colonoscopy requirements 3.9–6.6%, while colonoscopy-first strategies also reduced mortality but increased colonoscopy requirements 89–116.7%. All strategies were cost-effective compared to current screening recommendations. Second, a self-directed e-consent was evaluated in a prospective oncology study to demonstrate patient acceptance of a new enrolment strategy aimed at improving recruitment. Fifty-one patients with cancer initially self-enrolled. After a follow-up call, 47 (92%) provided full consent and 41 (87%) did so asynchronously. Thirty-eight participants (93%) expressed comfortability enrolling without interacting with a research coordinator, suggesting strong acceptance of self-directed approaches. Third, a novel framework was developed to evaluate the validity of surrogate endpoints used to grant earlier access to promising therapies. A literature search was conducted, identifying 39 analyses of 12 surrogate endpoints used in CRC trials for framework evaluation. Framework-derived evidence scores were assigned based on seven domains of the surrogate relationship. Only disease-free survival was a strong surrogate for adjuvant trials (≥A18, median R² = 0.92). All endpoints in advanced-stage trials were unreliable, including progression-free survival (B6–A15, median R² = 0.55) and response rate (B6–A13, median R² = 0.33). These works advance the existing strategies that accelerate or complement RCT evidence. Microsimulations will support timely guideline revisions, asynchronous e-consent offers a novel enrolment strategy and the surrogate endpoint framework helps validate surrogates used for earlier clinical trial read out and gaining expedited regulatory approvals.