Changes in the preferences of US physicians-in-training for medical interventions throughout medical education.

Context There exists a disparity between the views of physicians and the views of their patients on end-of-life decisions. However, the timing of when the end-of-life preferences of physicians and non-medically-trained individuals diverge is currently unknown. The objective of this paper is to characterise how preferences for medical interventions change throughout medical education and residency or fellowship training when confronted with scenarios of critical or terminal illness. Methods This is a single-centre cross-sectional study that enrolled medical students at Sidney Kimmel Medical College and residents and fellows at Thomas Jefferson University Hospital. Through an online survey we determined the preferences of medical trainees for specific interventions throughout medical training when presented with different clinical scenarios. Interventions were organised into three categories: standard, intermediate and aggressive. We analysed responses to questions regarding different scenarios in separate repeated measures logistic regression models. The probability of declining medical interventions was modelled, and significant predictors of refusal of interventions were identified. Results Years of training was a significant predictor of declining interventions for several scenarios. When faced with permanent physical disability, increased years of training led to a higher rate of refusal of intermediate ( OR = 1.14 [1.02-1.28], p = 0.02) and aggressive interventions ( OR = 1.15 [1.03-1.28], p = 0.01). For the scenario of terminal illness with associated physical disability, years of training significantly influenced refusal of intermediate ( OR = 1.14 [1.04-1.26], p = 0.006) and aggressive ( OR = 1.20 [1.08-1.34], p = 0.001) interventions. For the scenario of permanent cognitive impairment, increased years of training led to a higher rate of refusal of standard ( OR = 1.14 [1.01-1.29], p = 0.03), intermediate ( OR = 1.30 [1.13-1.50], p < 0.001) and aggressive ( OR = 1.38 [1.14-1.66], p = 0.001) interventions. Conclusion Changes in end-of-life preferences occur throughout medical training. Years of training influenced the likelihood of declining medical interventions when faced with scenarios of terminal illness and physical or cognitive disability. [ABSTRACT FROM AUTHOR]