Anaesthetists’ prediction of postoperative hypotension in adult non-cardiac surgical patients: a single-centre prospective cohort study

Hypotension after surgery is common and associated with poor clinical outcomes. Prediction of hypotension is difficult. We aimed to characterise whether anaesthetists can predict hypotension in the first 48 h after non-cardiac surgery. We conducted a single-centre, prospective cohort study of anaesthetist-patient pairs during non-cardiac surgery. We asked anaesthetists to predict whether hypotension (systolic blood pressure <100 mm Hg) would occur in the first 48 h. We determined patients’ blood pressures over 48 h via the electronic medical record. We assessed anaesthetists’ predictions using sensitivity, specificity, positive and negative predictive values, positive and negative likelihood ratios, accuracy and area under the receiver operating characteristic curve (AUROC). We recruited 100 anaesthetist-patient pairs involving 49 individual anaesthetists. Hypotension occurred in 33 patients (33%). The true positive rate was 13%, the true negative rate was 46%, the false positive rate was 21% and the false negative rate was 20%, and accuracy was 59%. The AUROC was 0.54 (95% CI 0.44–0.64). The observed sensitivity was 39% (95% CI 23–58%), specificity was 69% (95% CI 56–79%), positive predictive value was 38% (95% CI 24–43%), negative predictive value was 70% (95% CI 26–52%), positive likelihood ratio was 1.3 (95%CI 0.7–2.2), and the negative likelihood ratio was 0.9 (95% CI 0.6–1.2). Adjusting for multiple predictions made by individual anaesthetists did not alter the interpretation of the results. Anaesthetists were unable to accurately predict postoperative hypotension. Better prediction tools are required to guide practice in high-risk patients.