EUVIMEDEUVIMED entdecken
ÜbersichtsarbeitOpen Access · CC BY-NC-ND 4.0

Validity considerations in programmatic assessment of clinical reasoning

Christopher Runyon iD, Laura Zwaan iD, Monica M Cuddy, Andrew P J Olson iD, Verity Schaye iD

Medical Education · 11.09.2026

DOI: 10.1111/medu.70308 ↗

Kurz gefasst

Übersichtsarbeit in der Zeitschrift Medical Education, online erschienen am 11.09.2026. Verfasst von Christopher Runyon und 4 weiteren Autor:innen. Beteiligte Einrichtung: Growth and Innovation, NBME, Philadelphia, Pennsylvania, USA. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 60 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Problem Clinical reasoning is a core competency in health professions education, but its assessment is challenging. Clinical reasoning is multifaceted, content‐ and context‐dependent, and largely unobservable, and no single assessment can support a robust characterisation of a learner's clinical reasoning. Programmatic assessment has gained traction as a response to these challenges, but a validity argument explaining why a programmatic approach addresses the specific vulnerabilities of clinical reasoning assessment has received limited attention. Approach We examine validity issues in the programmatic assessment of clinical reasoning using Kane's argument‐based framework, organised around five inferences: performance, scoring, generalisation, extrapolation and consequences. For each inference, we identify the central validity question, describe issues specific to clinical reasoning at the level of the individual assessment and the programme, discuss how a programmatic approach mitigates weaknesses of single assessments and describe residual or new validity threats. A fictional learner followed across undergraduate medical education makes these issues concrete. We then consider generative artificial intelligence (GAI) as a cross‐cutting validity challenge affecting each inference. Key Points Programmatic assessment strengthens each inference in the validity chain, but it does not resolve every weakness and introduces new assumptions of its own, particularly around how conflicting evidence is synthesised. GAI compounds these challenges: it can shift what is actually being measured, complicate scoring and generalisation, and blur the line between assessment and target performance, while also offering new opportunities to make clinical reasoning more observable when incorporated deliberately. Implications We conclude with a three‐stage developmental framework—coherence and coverage, intentional depth, and multimodal longitudinal integration—to help institutions integrate existing assessments into a programme of clinical reasoning assessment. This framework reflects validity as an ongoing institutional commitment to evidence building for better learning, decisions and patient care.

Autor:innen und Institutionen

  • Growth and Innovation, NBME, Philadelphia, Pennsylvania, USA.
  • Institute of Medical Education Research Rotterdam, Erasmus Medical Center Rotterdam, Rotterdam, The Netherlands.
  • Department of Medicine, University of Minnesota Twin Cities, Minneapolis, Minnesota, USA.
  • Department of Medicine, NYU Grossman School of Medicine, New York, New York, USA.

ORCID-Profile: Christopher Runyon, Laura Zwaan, Andrew P J Olson, Verity Schaye

Zitieren

APA

Runyon, C., Zwaan, L., Cuddy, M. M., Olson, A. P. J., & Schaye, V. (2026). Validity considerations in programmatic assessment of clinical reasoning. Medical Education. https://doi.org/10.1111/medu.70308

Vancouver

Runyon C, Zwaan L, Cuddy MM, Olson APJ, Schaye V. Validity considerations in programmatic assessment of clinical reasoning. Medical Education. 2026. doi:10.1111/medu.70308

BibTeX
@article{Runyon2026,
  author = {Runyon, Christopher and Zwaan, Laura and Cuddy, Monica M and Olson, Andrew P J and Schaye, Verity},
  title = {Validity considerations in programmatic assessment of clinical reasoning},
  journal = {Medical Education},
  year = {2026},
  doi = {10.1111/medu.70308}
}
Kostenlos starten →