https://iborz.org/tags/disease-progression-modeling/Disease Progression ModelingHugoBlox Kit (https://hugoblox.com)en-usThu, 18 Jun 2026 00:00:00 +0000https://iborz.org/media/logo.svghttps://iborz.org/tags/disease-progression-modeling/https://iborz.org/blog/tpe/Thu, 18 Jun 2026 00:00:00 +0000https://iborz.org/blog/tpe/<p>A disease does not progress in a straight line. It drifts, then turns, and the turn is the part that matters.</p> <p>A biomarker can changes gradually until it passes a critical point and then accelerates. The trajectory bends, and from one patient to the next that bend lands in a different place, pushed around by age, genetics, and everything else that makes a person who they are. Average the curves together and the turn smears into nothing.</p> <p>I built the transition point estimator (TPE) to find that turn directly. It uses machine learning to model the nonlinear link between a biomarker and an outcome, keeps the confounders in check, and reads off the point where each marker&rsquo;s behavior shifts. That point becomes its cutpoint.</p> <p>A cutpoint set this way is important in how the disease actually moves. Find the turn, and you find the moment worth acting on.</p>