From Critique to Analysis, and Ultimately to Deep Understanding: Teaching Students to Read Clinical Research Articles in the AI Era

Abstract

Teaching students to read clinical research has often been organised around journal critique and the identification of methodological weaknesses. Although critical appraisal remains essential, a fault-finding approach can reduce reading to a checklist of errors. This working paper traces an approximately twenty-year evolution in my own teaching through three overlapping phases: critique, in which students looked for flaws; analysis, in which they identified study designs and separated papers into methodological components; and deep understanding, in which they reconstruct the main idea, trace how the design and evidence support it, define the boundary of the claim, and judge its clinical meaning. The current Clinical Research Design Appraisal approach therefore places critique after reconstruction rather than using criticism as the starting point. The shift has become more important in the era of generative artificial intelligence, when summaries, PICO statements and plausible lists of limitations can be produced without demonstrating learner understanding. It has also been informed by a practice-based observation that a paper may be cited for a peripheral detail while its central contribution and evidential conditions remain unexamined. Ten question cards structure the seminar, and responses follow an Answer–Evidence–Meaning format. Students read the paper before class, work through every card in small groups, may be randomly selected to present any card, and complete an individual exit reflection. Preliminary observations from the first seminar group suggest movement beyond naming study designs towards reasoning about bias control, external validity, statistical versus clinical significance, and possible redesign. The framework is presented as a developing pedagogical model, with evaluation planned across three seminar groups.