Post-surgical (bacterial) infections are the most frequent post-surgical complications, including deep or superficial wound infections, urinary tract infections, pneumonia, and even sepsis. Approximately 6.5-25% of all surgical patients will develop any type of bacterial infection. To personalize surgical infection management, (Artificial Intelligence) models are in the making to predict which patients are at high or low risk of developing a post-surgical infection. In order to benchmark these prediction models to the predictive capabilities of surgeons, the investigators aim to investigate the performance of surgeons in predicting the risk of a patient developing (any type) of post-surgical infection within 30 days.
A prospective non-interventional study is performed to collect surgeons' predictions on the risk of a patient developing a postoperative infection within 30 days of surgery. Surgeons are asked to fill in a short questionnaire asking about the estimated infection risk. The actual outcome (infection \< 30 days of surgery) of a patient will be collected retrospectively after completion of the study. This study will have no effect on standard care: surgical interventions and postoperative care will be carried out according to standard clinical practice. Besides a one-time estimate of the surgeon, immediately after the surgical procedure, no other interventions will be performed and surgical specialists will carry out their normal post-surgical care, including screening and treating (if necessary) their patients for postoperative infections.
Study Type
OBSERVATIONAL
Enrollment
594
Surgeons will be asked to fill in a short questionnaire after surgery on risk of postoperative infection
Sesmu Arbous
Leiden, South Holland, Netherlands
The discriminative predictive performance of surgeons with respect to estimating the risk of developing (any type) bacterial post-surgical within 30 days of surgery
The primary outcome measure of discrimination are area under the receiving operating characteristic curve (AUROC). Predictions are compared to the occurrence of a postoperative infection requiring treatment, surgical intervention or registration within 30 days of surgery.
Time frame: 30 days
The calibration properties of surgeons with respect to estimating the risk of developing (any type) bacterial post-surgical within 30 days of surgery
Calibration plots with slope and intercept
Time frame: 30 days
Relationship between the certainty in estimate and the predictive performance of surgeons
Surgeons are questioned on their certainty in the provided estimate
Time frame: 30 days
Relationship between patient factors and predicted risk
Surgeons are questioned to indicate for a list of patient factors whether they were of impact to the decision
Time frame: 30 days
Predictive performance per surgeons and patients subgroups
Surgeons subgroups are based on specialty, years of experience, level of experience, sex. Patient subgroups include, surgical specialty, age groups, type of surgical procedure, planned or emergency intervention.
Time frame: 30 days
Relationship between predicted risk of surgeons and if they perform additional actions
Surgeons are asked to indicate whether they performed additional actions for this patient in the questionnaire.
Time frame: 30 days
Comparison between the predicted risk of surgeons and an artificial intelligence algorithm
The performance of the physicians is compared to that of the artificial intelligence algorithm by means of AUROC
Time frame: 30 days
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.