AI is Smart. Is it Wise? Quantifying the Effect of Patient Choice (β) on Physical Outcomes

AI is Smart. Is it Wise? Quantifying the Effect of Patient-Choice (β) on Physical Outcomes* | medRxiv Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search AI is Smart. Is it Wise? Quantifying the Effect of Patient-Choice (β) on Physical Outcomes * View ORCID Profile Ogan Gurel , View ORCID Profile Matthew F. Rasmussen , View ORCID Profile Venkatesh Veginati , View ORCID Profile James N. Weinstein doi: https://doi.org/10.64898/2026.03.10.26348069 Ogan Gurel 1 Health Careers Institute, Division of Enterprise Development & Department of Bioengineering, College of Engineering, The University of Texas at Arlington , Arlington, TX, USA MD Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Ogan Gurel For correspondence: ogan.gurel{at}uta.edu Matthew F. Rasmussen 2 Student & Lead Peer Mentor, Health Professions Office, and Department of Chemistry & Biochemistry, College of Science, The University of Texas at Arlington , Arlington, TX, USA MPhil Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Matthew F. Rasmussen Venkatesh Veginati 3 ASDS MS Graduate, The University of Texas at Arlington, Division of Data Science, College of Science, The University of Texas at Arlington , Arlington, TX, USA MS Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Venkatesh Veginati James N. Weinstein 4 Microsoft Research , NH, USA MS, DO Roles: Senior Vice President Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for James N. Weinstein Abstract Full Text Info/History Metrics Data/Code Preview PDF Abstract Large language models (LLMs) increasingly guide clinical decisions through population-level evidence, yet they cannot encode individual patient preferences. When treatments yield comparable outcomes, patient choice may drive decisions, though its effect remains unquantified. The Spine Patient Outcomes Research Trial (SPORT)—marked by similar surgical and nonoperative results and substantial crossover—provided a natural experiment to apply causal-inference methods to estimate unbiased treatment effects and quantify the contribution of patient choice to outcomes. Using only published aggregate results from SPORT, we conducted two-stage least squares instrumental-variable analysis with randomized treatment assignment as the instrument, with Complier Average Causal Effects (CACE) and E-values assessing sensitivity to unmeasured confounding. Primary outcomes were SF-36 Bodily Pain, SF-36 Physical Function, and the Oswestry Disability Index. We decomposed treatment effects into α , the biological treatment mechanism, and β , the preference-driven selection component reflecting the downstream effect of patient choice. Aggregate estimates revealed α G = 15.7 (0.5) and β G = 7.4 (3.4), with a negligible net difference between surgical and nonoperative treatment effects (Δ α ≈ 0.65). This analysis demonstrates that patient choice ( β ) is a measurable and clinically meaningful contributor to physical outcomes. When treatment effects are comparable, β —a dimension inaccessible to current LLMs trained primarily on α -biased population-level evidence—emerges as the dominant driver of decision-making. These findings provide empirical grounding for informed choice, clarify the limits of present-day AI in preference-sensitive care, and identify a structural constraint in AI-driven clinical decision support. Key messages The effect of patient choice ( β ) on physical outcomes is real, measurable, and clinically meaningful. β becomes the dominant driver of outcomes when biological treatment differences (Δ α ) are small. LLMs cannot encode β because they are trained on β-biased population-level evidence. These findings provide the empirical foundation for informed choice—not just informed consent. Competing Interest Statement The authors have declared no competing interest. Clinical Trial NCT00000410 (SPORT trial) Funding Statement This study did not receive any funding Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: We used only aggregate, published statistics from the JAMA papers, and no patient‑level data-identified or de‑identified-were accessed for this work. I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes Footnotes This revision tightens and clarifies the manuscript without altering analyses, results, or conclusions. Medical literacy is more clearly linked to beta as an enabling substrate for informed choice, improving conceptual coherence. References 35 and 36 clarified. Minor formatting and presentation changes were applied, primarily to improve table readability. Data Availability This study used no human-subject, clinical, or patient-level data. All information used is conceptual, theoretical, or derived from previously published sources. All computational spreadsheets and formulas used to generate illustrative estimates are available upon reasonable request. Copyright The copyright holder for this preprint is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . Back to top Previous Next Posted March 18, 2026. Download PDF Data/Code Email Thank you for your interest in spreading the word about medRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following AI is Smart. Is it Wise? Quantifying the Effect of Patient-Choice (β) on Physical Outcomes* Message Subject (Your Name) has forwarded a page to you from medRxiv Message Body (Your Name) thought you would like to see this page from the medRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share AI is Smart. Is it Wise? Quantifying the Effect of Patient-Choice (β) on Physical Outcomes * Ogan Gurel , Matthew F. Rasmussen , Venkatesh Veginati , James N. Weinstein medRxiv 2026.03.10.26348069; doi: https://doi.org/10.64898/2026.03.10.26348069 Share This Article: Copy Citation Tools AI is Smart. Is it Wise? Quantifying the Effect of Patient-Choice (β) on Physical Outcomes * Ogan Gurel , Matthew F. Rasmussen , Venkatesh Veginati , James N. Weinstein medRxiv 2026.03.10.26348069; doi: https://doi.org/10.64898/2026.03.10.26348069 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Orthopedics Subject Areas All Articles Addiction Medicine (637) Allergy and Immunology (917) Anesthesia (348) Cardiovascular Medicine (4974) Dentistry and Oral Medicine (493) Dermatology (426) Emergency Medicine (674) Endocrinology (including Diabetes Mellitus and Metabolic Disease) (1682) Epidemiology (16207) Forensic Medicine (36) Gastroenterology (1238) Genetic and Genomic Medicine (7250) Geriatric Medicine (751) Health Economics (1092) Health Informatics (5262) Health Policy (1463) Health Systems and Quality Improvement (1825) Hematology (599) HIV/AIDS (1375) Infectious Diseases (except HIV/AIDS) (16466) Intensive Care and Critical Care Medicine (1199) Medical Education (696) Medical Ethics (156) Nephrology (747) Neurology (7552) Nursing (377) Nutrition (1120) Obstetrics and Gynecology (1293) Occupational and Environmental Health (1038) Oncology (3730) Ophthalmology (1087) Orthopedics (406) Otolaryngology (463) Pain Medicine (499) Palliative Medicine (143) Pathology (720) Pediatrics (1852) Pharmacology and Therapeutics (752) Primary Care Research (784) Psychiatry and Clinical Psychology (6099) Public and Global Health (9957) Radiology and Imaging (2515) Rehabilitation Medicine and Physical Therapy (1506) Respiratory Medicine (1264) Rheumatology (657) Sexual and Reproductive Health (799) Sports Medicine (595) Surgery (799) Toxicology (109) Transplantation (318) Urology (307)

Related documents

• Patient Preference Study in Type-2 Diabetes Mellitus: Discrete Choice Experiment (DCE) to Support Novo Nordisk's Long-Acting Insulin#1033517
• Does Choice Improve the Pain Relief Derived From a Brief Intervention#1094196
• On/Off-Discrepancies in Medical Decision-Making: Utilising the Reversibility of Deep Brain Stimulation to Strengthen Patient Autonomy#111349
• The Self Match Study: A Study of Informed Choice in the Treatment of Addiction#1095753
• Identifying and measuring important outcomes for evaluating the impact of pharmacist prescribing in Ireland: A modified Delphi study#5268