Artificial intelligence (AI) used in machine learning models can predict which patients will be at greatest risk of severe pain after surgery and help determine who would benefit most from personalized pain management plans that are not opioid alternatives use annual ANESTHESIOLOGY 2020.
Some patients experience more pain after surgery and need higher doses of opioids for longer periods of time, which increases their risk of opioid abuse disorder. By knowing which patients are at higher risk for severe post-operative pain, anesthesiologists can create an anesthesia plan that uses non-opioid alternatives such as nerve blocks, epididymis, and other drugs to more effectively manage pain and reduce the need for opioids .
Currently, doctors use time-consuming questionnaires to identify patients at higher risk for severe post-operative pain and ask about their history of anxiety, sleep quality, and depression. In this study, the researchers looked for a faster, more effective method using machine learning, where a system learns and evolves based on the data provided. They created three machine learning models that analyzed the electronic health records. It found that younger age, higher body mass index, female gender, pre-existing pain, and previous opioid use were the most predictive factors for postoperative pain.
"We plan to incorporate the models into our electronic health records to enable predictions of post-operative pain for each patient," said Dr. Mieke A. Soens, lead study author and anesthetist at Brigham and Women & # 39; s Hospital and anesthesiology teacher at Harvard Medical School, Boston. "If the patient is found to be at high risk for severe post-operative pain, the anesthetist can adjust the patient's anesthesia schedule to maximize opioid-free pain management strategies, which would reduce the need for opioids after surgery."
In the two-part study, researchers examined data from 5,944 patients who had a variety of surgeries, including gallbladder removal, hysterectomy, hip replacement, and prostate surgery. Of these, 1,287 (22%) had consumed 90 morphine milligram equivalents (MME) in the first 24 hours after surgery, which is considered a high dose. In the first part of the study, they used 163 potential factors to predict severe pain after surgery, based on a literature review and consultation with experts. From there, they created three machine learning algorithm models (logistic regression, random forest, and artificial neural networks) that were used to analyze patients 'medical records and reduce the 163 predictor factors to those that tracked patients' pain severity and potential opioid needs most accurately predicted the operation.
In the second part, they compared the models' predictions of actual opioid use in the same patients. They found that all three models showed similar predictive accuracy overall: 81% for logistic regression and random forest methods and 80% for artificial neural networks. This means that the models precisely identified who were more likely to experience severe pain and who needed higher doses of opioids in about 80% of the cases.
"Electronic health records are a valuable and underutilized source of patient information and can be used effectively to improve patient lives," said Dr. Soens. "The selective identification of patients who typically require high doses of opioids after surgery is important to reduce opioid abuse."
The study shows that opioid prescribing guidelines need to be balanced after surgery
American Society of Anesthetists
AI predicts patients at highest risk for severe pain and increased opioid use after surgery (2020, October 4).
accessed on October 4, 2020
This document is subject to copyright. Apart from fair treatment for the purpose of private study or research, no
Part may be reproduced without written permission. The content is provided for informational purposes only.