The Montefiore Einstein Department of Anesthesiology is leading efforts to improve how physicians prepare patients for surgical procedures. In 2023, the department developed machine learning–derived ASA and case cancellation risk scores that automatically pull data from the electronic medical record to identify patients at risk for perioperative medical complications or surgery cancellation.

These clinical decision support tools are embedded directly in the electronic medical record and are used by Montefiore Einstein’s surgical and anesthesiology teams to determine which patients require additional preoperative evaluation prior to surgery.

Some patients, however, do not yet have a comprehensive medical record, particularly those who are new to Montefiore Einstein. To address this gap, the Department of Anesthesiology created an Anesthesia Health Questionnaire, which is used during the preoperative evaluation process to determine whether a patient requires further assessment before surgery. This effort is led by Dr. Carina Himes, director of Preoperative Evaluation in the Department of Anesthesiology.

The questionnaire was designed to identify patients at the highest risk, allowing clinicians to determine when additional information, testing, or medical optimization may be necessary prior to surgery.

The Anesthesia Health Questionnaire is integrated into MyChart and has been used successfully in collaboration with the Department of Otolaryngology over the past several years. When combined with the machine learning–derived ASA score, the questionnaire helps triage patients appropriately for further evaluation while preventing unnecessary medical visits for lower-risk patients.

This approach has reduced surgery delays and improved convenience for patients. Approximately 40 percent of otolaryngology patients at the Hutchinson Ambulatory Surgery Center receive anesthesia telehealth visits prior to surgery, while about 60 percent of inpatients at Moses and Weiler typically qualify for telehealth visits. Use of the questionnaire is now expanding to patients undergoing urologic and gastroenterologic procedures.

For higher-risk patients, the preoperative evaluation process balances medical complexity with surgical urgency. Clinicians assess how a patient’s condition affects quality of life, consider patient preferences, and often involve colleagues from other medical subspecialties. Early identification allows time for thoughtful discussion and medical optimization before surgery.

In collaboration with Dr. Himes, the Anesthesiology Digital Health Lab advanced this work further by developing the Anesthesia Risk Assessment Score (ARAS), a tool designed to accurately predict which patients face the highest risk of adverse outcomes. The candidate predictors used to create ARAS were drawn directly from responses to the anesthesia questionnaire.

The risk stratification model was developed by analyzing electronic medical record data from more than 59,000 adult patients who underwent inpatient surgery at Montefiore Einstein between 2016 and 2023. Questionnaire responses were evaluated to determine predictors for the final ARAS model.

The six independent predictors associated with the highest risk of 30-day postoperative mortality include a history of stroke within one year, seizure within one year, heart failure or pacemaker or defibrillator implantation, liver failure, history of blood or bleeding disorders, and metabolic equivalents less than or equal to four.

The ARAS also accurately predicts the risk of adverse discharge, defined as discharge to a skilled nursing facility after surgery among patients who previously lived at home. This represents an important patient-centered outcome. Notably, ARAS performs comparably to the American Society of Anesthesiologists Physical Status Classification System in identifying high-risk patients.

The ARAS score is calculated entirely from patient-reported questionnaire responses and does not require direct input from an anesthesiologist. It generates individualized risk percentages for 30-day postoperative mortality and adverse discharge, providing the preoperative team with valuable data early in the triage process to determine whether additional evaluation is necessary.

Unlike traditional risk assessment tools that rely on subjective clinician interpretation or large data infrastructures, ARAS offers a standardized and efficient approach that can be applied broadly across clinical settings.

The integration of the Anesthesia Risk Assessment Score into Epic preoperative assessment templates enables anesthesia providers to make more informed decisions, including which patients can safely undergo surgery at freestanding ambulatory care centers and which patients require additional diagnostic testing or perioperative resources.

By leveraging these tools, anesthesia teams can develop more comprehensive and individualized treatment plans, ultimately improving patient safety, efficiency, and surgical outcomes.