Arzu Kanik on clinician biostatistics

Prof. Arzu Kanik used a September 8 post on X to make a narrow argument about how clinical research should be done: biostatistics, she wrote, has to be understandable to clinicians and reproducible in practice. She tied that argument to collaboration, not substitution, saying clinicians, biostatisticians and AI teams should work together rather than be framed in opposition. She also pointed readers to EPICOSAI, a research platform she has presented publicly, as a way to move from a clinical question through study design, analysis and reporting. ### What exactly was Kanik arguing? Kanik’s post centered on a familiar problem in medical research: statistical methods can become detached from the clinical decisions they are supposed to inform. Her public work describes clinical biostatistics as something that should begin with the research question, population, outcome and study design, rather than with a preselected test or model. (designer.epicosai.com) The EPICOSAI book and product materials use similar language. Volume I of *Clinical Biostatistics with EPICOSAI* says clinical evidence begins before a p-value is calculated, when uncertainty is turned into a research question and a design determines what conclusions the data can support. The series says the platform is used as an “interactive clinical-biostatistics laboratory,” not as a black box. (amazon.nl) ### Why does that matter for clinicians? Clinical researchers often face a practical gap between statistical output and bedside relevance. Kanik’s published descriptions of her framework say good biostatistical practice should ask whether evidence is valid, reliable and clinically meaningful, not only whether it reaches statistical significance. That framing tracks with how EPICOSAI presents itself publicly. (amazon.nl) The study-design assistant says it offers recommendations for randomized trials, cohort studies, systematic reviews and more than 170 study designs, while the documentation says the system selects reporting checklists after identifying the study type. ### What is EPICOSAI supposed to do? EPICOSAI’s public pages describe a platform built around the full research workflow rather than a single statistical function. (amazon.nl) Its documentation says it is designed for researchers in clinical and academic settings, and the reviewer module says it can automatically identify study design and recommend a reporting guideline. The design assistant page says users can get guidance for study design before analysis begins. (designer.epicosai.com) Product listings and book descriptions linked to the platform say the workflow runs from clinical questions and evidence design to statistical analysis, interpretation and methodological judgment. ### Where does Kanik fit into the project? ESC 365, the European Society of Cardiology’s content platform, lists Kanik as a professor at MedicReS AI in Houston and says she led development of E-PICOS while a visiting professor at Harvard University. (docs.epicosai.com) The same profile says she has worked on biostatistics departments, clinical research commissions and researcher education. Google Scholar lists Kanik at the University of Mersin’s Department of Biostatistics and MedicReS, with research interests including clinical research, biostatistics and meta-analysis. (designer.epicosai.com) ESC 365 also hosts a 2025 presentation titled “EPICOSAI” under her name. ### Is this a case for AI replacing statisticians? Kanik’s public materials point the other way. The recurring claim across the post context, the book descriptions and the platform pages is that AI should support question formulation, design choice, analysis workflow and reporting discipline, while statistical judgment remains explicit and inspectable. (esc365.escardio.org) That is why her emphasis on reproducibility matters. The platform’s documentation ties study-type selection to reporting checklists, and the books repeatedly describe critical reading of output, challenging assumptions and reporting evidence transparently. (scholar.google.com) EPICOSAI’s public tools and documentation remain live as of early September 2026, and Kanik’s ESC 365 profile and presentation pages remain accessible for readers who want the primary material behind the post. (amazon.com.au) (designer.epicosai.com) (docs.epicosai.com)