The effect of clinical experience, judgment task difficulty and time pressure on nurses’ confidence calibration in a high fidelity clinical simulation

BACKGROUND: Misplaced or poorly calibrated confidence in healthcare professionals’ judgments compromises the quality of health care. Using higher fidelity clinical simulations to elicit clinicians’ confidence 'calibration' (i.e. overconfidence or underconfidence) in more realistic settings is a promising but underutilized tactic. In this study we examine nurses’ calibration of confidence with judgment accuracy for critical event risk assessment judgments in a high fidelity simulated clinical environment. The study also explores the effects of clinical experience, task difficulty and time pressure on the relationship between confidence and accuracy. METHODS: 63 student and 34 experienced nurses made dichotomous risk assessments on 25 scenarios simulated in a high fidelity clinical environment. Each nurse also assigned a score (0–100) reflecting the level of confidence in their judgments. Scenarios were derived from real patient cases and classified as easy or difficult judgment tasks. Nurses made half of their judgments under time pressure. Confidence calibration statistics were calculated and calibration curves generated. RESULTS: Nurse students were underconfident (mean over/underconfidence score −1.05) and experienced nurses overconfident (mean over/underconfidence score 6.56), P = 0.01. No significant differences in calibration and resolution were found between the two groups (P = 0.80 and P = 0.51, respectively). There was a significant interaction between time pressure and task difficulty on confidence (P = 0.008); time pressure increased confidence in easy cases but reduced confidence in difficult cases. Time pressure had no effect on confidence or accuracy. Judgment task difficulty impacted significantly on nurses’ judgmental accuracy and confidence. A 'hard-easy' effect was observed: nurses were overconfident in difficult judgments and underconfident in easy judgments. CONCLUSION: Nurses were poorly calibrated when making risk assessment judgments in a high fidelity simulated setting. Nurses with more experience tended toward overconfidence. Whilst time pressure had little effect on calibration, nurses’ over/underconfidence varied significantly with the degree of task difficulty. More research is required to identify strategies to minimize such cognitive biases.