Teaching a Machine to Feel Postoperative Pain: Combining High-Dimensional Clinical Data with Machine Learning Algorithms to Forecast Acute Postoperative Pain

Strengthening Discovery and Application of Artificial Intelligence in Anesthesiology: A Report from the Anesthesia Research Council

Interest in the potential applications of artificial intelligence in medicine, anesthesiology, and the world at large has never been higher. The Anesthesia Research Council steering committee formed an anesthesiologist artificial intelligence expert workgroup charged with evaluating the current state of artificial intelligence in anesthesiology, providing examples of future artificial intelligence applications and identifying barriers to artificial intelligence progress. The workgroup's findings are summarized here, starting with a brief introduction to artificial intelligence for clinicians, followed by overviews of current and anticipated artificial intelligence-focused research and applications in anesthesiology. Anesthesiology's progress in artificial intelligence is compared to that of other medical specialties, and barriers to artificial intelligence development and implementation in our specialty are discussed. The workgroup's recommendations address stakeholders in policymaking, research, development, implementation, training, and use of artificial intelligence-based tools for perioperative care.

Strengthening Discovery and Application of Artificial Intelligence in Anesthesiology: A Report from the Anesthesia Research Council

Interest in the potential applications of artificial intelligence in medicine, anesthesiology, and the world at large has never been higher. The Anesthesia Research Council steering committee formed an anesthesiologist artificial intelligence expert workgroup charged with evaluating the current state of artificial intelligence in anesthesiology, providing examples of future artificial intelligence applications and identifying barriers to artificial intelligence progress. The workgroup's findings are summarized here, starting with a brief introduction to artificial intelligence for clinicians, followed by overviews of current and anticipated artificial intelligence-focused research and applications in anesthesiology. Anesthesiology's progress in artificial intelligence is compared to that of other medical specialties, and barriers to artificial intelligence development and implementation in our specialty are discussed. The workgroup's recommendations address stakeholders in policymaking, research, development, implementation, training, and use of artificial intelligence-based tools for perioperative care.

Gender Differences in Predicting Metabolic Syndrome Among Hospital Employees Using Machine Learning Models: A Population-Based Study

BACKGROUND

Metabolic syndrome (MetS) is a complex condition that captures several markers of dysregulation, including obesity, elevated blood glucose levels, dyslipidemia and hypertension. Using an approach to early prediction of MetS risk in hospital employees that takes into account the differing effects of gender may be expected to improve cardiovascular disease-related health outcomes.

PURPOSE

In this study, machine learning techniques were applied to construct an optimized MetS prediction model for use on hospital employees.

METHODS

This population-based study survey included 3,537 participants aged 20 to 65 years old. Participant demographic, anthropometric data, medical history, lifestyle-related factor, and biochemical data were collected from the hospital's Health Management Information System from 2018 to 2020. MetS prediction and the investigation of gender differences were performed using six machine learning models based on the following algorithms: K-nearest neighbor, random forest, logistic regression, support vector machine, neural network, and Naïve Bayes. All analyses were performed by sequentially inputting the features in three steps according to their characteristics.

RESULTS

MetS was detected in 8.91% of the participants. Among the MetS prediction models, Naïve Bayes showed the best performance, with a sensitivity of 0.825, an accuracy of 0.859 and an area under the receiver operating characteristic curve of 0.936. Body mass index and alanine transaminase were identified as important predictive factors for MetS in participants of both genders. Age, uric acid, and aspartate transaminase were identified as important predictive factors in men, while chronic disease and phosphorous were identified as important predictive factors in women.

CONCLUSIONS

The results indicate Naïve Bayes model to be useful and accurate in identifying MetS in hospital employees independent of gender. The early prediction of MetS using a model that accounts for gender differences is an important part of routine health screening and requires a multidimensional approach, including self-administered questionnaires and anthropometric and biochemical measurements.

Moving towards the use of artificial intelligence in pain management

BACKGROUND AND OBJECTIVE

While the development of artificial intelligence (AI) technologies in medicine has been significant, their application to acute and chronic pain management has not been well characterized. This systematic review aims to provide an overview of the current state of AI in acute and chronic pain management.

DATABASES AND DATA TREATMENT

This review was registered with PROSPERO (ID# CRD42022307017), the international registry for systematic reviews. The search strategy was prepared by a librarian and run in four electronic databases (Embase, Medline, Central, and Web of Science). Collected articles were screened by two reviewers. Included studies described the use of AI for acute and chronic pain management.

RESULTS

From the 17,601 records identified in the initial search, 197 were included in this review. Identified applications of AI were described for treatment planning as well as treatment delivery. Described uses include prediction of pain, forecasting of individualized responses to treatment, treatment regimen tailoring, image-guidance for procedural interventions and self-management tools. Multiple domains of AI were used including machine learning, computer vision, fuzzy logic, natural language processing and expert systems.

CONCLUSION

There is growing literature regarding applications of AI for pain management, and their clinical use holds potential for improving patient outcomes. However, multiple barriers to their clinical integration remain including lack validation of such applications in diverse patient populations, missing infrastructure to support these tools and limited provider understanding of AI.

SIGNIFICANCE

This review characterizes current applications of AI for pain management and discusses barriers to their clinical integration. Our findings support continuing efforts directed towards establishing comprehensive systems that integrate AI throughout the patient care continuum.

Machine learning approach to predict postoperative pain after spinal morphine administration during caesarean delivery

Background

A major barrier to optimal pain management is the difficulty in predicting and assessing patients at high risk for significant pain across multiple locations within the institution in a timely manner. This is compounded by the fragmented display of clinical information on enterprise clinical platform, which further hinders delay the reviews and hence the increased risk of untreated pain. We evaluated and compared the predictive performance of six modelling techniques in predicting significant pain, defined as the maximum pain score of 3 or more on movement at the 13th to 24th hour after spinal morphine administration during caesarean delivery.

Methods

We retrieved medical records from women who underwent caesarean delivery and received postoperative spinal morphine in a single specialist maternity hospital in Singapore between Aug 2019 and Aug 2022. We extracted 120 clinical variables from the medical records of eligible patients and further selected 23 to improve algorithm accuracies. The dataset was split randomly, with 80 % of patients (n = 5248) used for training the models, and 20 % (n = 1313) reserved for validation.

Results

The study cohort comprised 6561 patients with an incidence of significant postoperative pain of 7.9 %. Ridge regression demonstrated the best performance with both the full (AUC: 0.649) and selected (AUC: 0.719) feature sets. By reducing the number of features, Ridge regression, LASSO, Elastic net, and XGBoost showed similar in AUC (0.704-0.719), sensitivity (0.644-0.695), specificity (0.644-0.705), positive predictive value (0.155-0.179), and negative predictive value (0.949-0.955) in predicting significant postoperative pain. These were attributed to the top three variables, mainly the last recorded postoperative pain score (on movement) before the prediction point, mean and standard deviation of the hourly maximum postoperative pain score (at rest) at 0th to 12th hour.

Conclusions

Future research will aim to refine these models and explore their implementation in clinical settings to enhance real-time pain management and risk stratification for women after caesarean delivery.

Prognostic clinical prediction models for acute post-surgical pain in adults: a systematic review

BACKGROUND

Acute post-surgical pain is managed inadequately in many patients undergoing surgery. Several prognostic risk prediction models have been developed to identify patients at high risk of developing moderate to severe acute post-surgical pain. The aim of this systematic review was to describe and evaluate the methodological conduct of these prediction models.

METHODS

We searched MEDLINE, EMBASE and CINAHL for studies of prognostic risk prediction models for acute post-surgical pain using predetermined criteria. Prediction model performance was evaluated according to discrimination and calibration. Adherence to TRIPOD guidelines was assessed. Risk of bias and applicability was independently assessed by two reviewers using the prediction model risk of bias assessment tool.

RESULTS

We included 14 studies reporting on 17 prediction models. The most common predictors identified in final prediction models included age; surgery type; sex or gender; anxiety or fear of surgery; pre-operative pain intensity; pre-operative analgesic use; pain catastrophising; and expected surgical incision size. Discrimination, measured by the area under receiver operating characteristic curves or c-statistic, ranged from 0.61 to 0.83. Calibration was only reported for seven models. The median (IQR [range]) overall adherence rate to TRIPOD items was 62 (53-66 [47-72])%. All prediction models were at high risk of bias.

CONCLUSIONS

Effective prediction models could support the prevention and treatment of acute post-surgical pain; however, existing models are at high risk of bias which may affect their reliability to inform practice. Consideration should be given to the goals, timing of intended use and desired outcomes of a prediction model before development.

Postoperative Pain at Discharge From the Post-anesthesia Care Unit: A Case-Control Study

INTRODUCTION

Despite advancements in postoperative pain management, approximately 20% of patients still experience severe pain within the first 24 hours post-surgery. Previous studies utilizing machine learning have shown promise in predicting postoperative pain with various models. This study investigates postoperative pain predictors using a machine learning approach based on physiological indicators and demographic factors in a Mexican cohort.

METHODS

We conducted a retrospective case-control study to assess pain determinants at Post-anesthesia Care Unit (PACU) discharge at Hospital Ángeles Lomas in Mexico City. Data were collected from 550 patients discharged from the PACU, including 292 cases and 258 controls, covering a range of surgical procedures and illnesses. Machine learning techniques were employed to develop a predictive model for postoperative pain. Physiological responses, such as blood pressure, heart rate, respiratory rate, and anesthesia type, were recorded prior to PACU admission.

RESULTS

Significant differences were found between cases and controls, with factors such as sex, anesthesia type, and physiological responses influencing postoperative pain. Visual analog scale (VAS) scores at PACU admission were predictive of pain at discharge.

CONCLUSIONS

Our findings reinforce existing literature by highlighting sex-based disparities in pain experiences and the influence of anesthesia type on pain levels. The logistic regression model developed, incorporating physiological responses and sex, shows potential for refining pain management strategies. Limitations include the lack of detailed surgical data and psychological factors, and validation in a prospective cohort. Future research should focus on more comprehensive predictive models and longitudinal studies to further improve postoperative pain management.

Predicting Postoperative Pain and Opioid Use with Machine Learning Applied to Longitudinal Electronic Health Record and Wearable Data

BACKGROUND

 Managing acute postoperative pain and minimizing chronic opioid use are crucial for patient recovery and long-term well-being.

OBJECTIVES

 This study explored using preoperative electronic health record (EHR) and wearable device data for machine-learning models that predict postoperative acute pain and chronic opioid use.

METHODS

 The study cohort consisted of approximately 347 All of Us Research Program participants who underwent one of eight surgical procedures and shared EHR and wearable device data. We developed four machine learning models and used the Shapley additive explanations (SHAP) technique to identify the most relevant predictors of acute pain and chronic opioid use.

RESULTS

 The stacking ensemble model achieved the highest accuracy in predicting acute pain (0.68) and chronic opioid use (0.89). The area under the curve score for severe pain versus other pain was highest (0.88) when predicting acute postoperative pain. Values of logistic regression, random forest, extreme gradient boosting, and stacking ensemble ranged from 0.74 to 0.90 when predicting postoperative chronic opioid use. Variables from wearable devices played a prominent role in predicting both outcomes.

CONCLUSION

 SHAP detection of individual risk factors for severe pain can help health care providers tailor pain management plans. Accurate prediction of postoperative chronic opioid use before surgery can help mitigate the risk for the outcomes we studied. Prediction can also reduce the chances of opioid overuse and dependence. Such mitigation can promote safer and more effective pain control for patients during their recovery.

Postoperative accurate pain assessment of children and artificial intelligence: A medical hypothesis and planned study

Although the pediatric perioperative pain management has been improved in recent years, the valid and reliable pain assessment tool in perioperative period of children remains a challenging task. Pediatric perioperative pain management is intractable not only because children cannot express their emotions accurately and objectively due to their inability to describe physiological characteristics of feeling which are different from those of adults, but also because there is a lack of effective and specific assessment tool for children. In addition, exposure to repeated painful stimuli early in life is known to have short and long-term adverse sequelae. The short-term sequelae can induce a series of neurological, endocrine, cardiovascular system stress related to psychological trauma, while long-term sequelae may alter brain maturation process, which can lead to impair neurodevelopmental, behavioral, and cognitive function. Children's facial expressions largely reflect the degree of pain, which has led to the developing of a number of pain scoring tools that will help improve the quality of pain management in children if they are continually studied in depth. The artificial intelligence (AI) technology represented by machine learning has reached an unprecedented level in image processing of deep facial models through deep convolutional neural networks, which can effectively identify and systematically analyze various subtle features of children's facial expressions. Based on the construction of a large database of images of facial expressions in children with perioperative pain, this study proposes to develop and apply automatic facial pain expression recognition software using AI technology. The study aims to improve the postoperative pain management for pediatric population and the short-term and long-term quality of life for pediatric patients after operational event.

How to Modulate Peripheral and Central Nervous System to Treat Acute Postoperative Pain and Prevent Pain Persistence

Chronic postoperative pain (CPSP) is a major issue after surgery, which may impact on patient's quality of life. Traditionally, CPSP is believed to rely on maladaptive hyperalgesia and risk factors have been identified that predispose to CPSP, including acute postoperative pain. Despite new models of prediction are emerging, acute pain is still a modifiable factor that can be challenged with perioperative analgesic strategies. In this review we present the issue of CPSP, focusing on molecular mechanism underlying the development of acute and chronic hyperalgesia. Also, we focus on how perioperative strategies can impact directly or indirectly (by reducing postoperative pain intensity) on the development of CPSP.

Parametric Modeling and Deep Learning for Enhancing Pain Assessment in Postanesthesia

OBJECTIVE

The problem of reliable and widely accepted measures of pain is still open. It follows the objective of this work as pain estimation through post-surgical trauma modeling and classification, to increase the needed reliability compared to measurements only.

METHODS

This article proposes (i) a recursive identification method to obtain the frequency response and parameterization using fractional-order impedance models (FOIM), and (ii) deep learning with convolutional neural networks (CNN) classification algorithms using time-frequency data and spectrograms. The skin impedance measurements were conducted on 12 patients throughout the postanesthesia care in a proof-of-concept clinical trial. Recursive least-squares system identification was performed using a genetic algorithm for initializing the parametric model. The online parameter estimates were compared to the self-reported level by the Numeric Rating Scale (NRS) for analysis and validation of the results. Alternatively, the inputs to CNNs were the spectrograms extracted from the time-frequency dataset, being pre-labeled in four intensities classes of pain during offline and online training with the NRS.

RESULTS

The tendency of nociception could be predicted by monitoring the changes in the FOIM parameters' values or by retraining online the network. Moreover, the tissue heterogeneity, assumed during nociception, could follow the NRS trends. The online predictions of retrained CNN have more specific trends to NRS than pain predicted by the offline population-trained CNN.

CONCLUSION

We propose tailored online identification and deep learning for artefact corrupted environment. The results indicate estimations with the potential to avoid over-dosing due to the objectivity of the information.

SIGNIFICANCE

Models and artificial intelligence (AI) allow objective and personalized nociception-antinociception prediction in the patient safety era for the design and evaluation of closed-loop analgesia controllers.

Validation of the Calgary Postoperative Pain after Spine Surgery Score for Poor Postoperative Pain Control after Spine Surgery

OBJECTIVE

The Calgary Postoperative Pain after Spine Surgery (CAPPS) score was developed to identify patients at risk of experiencing poorly controlled pain after spine surgery. The goal of this study was to independently validate the CAPPS score on a prospectively collected patient sample.

METHODS

Poor postoperative pain control was defined as a mean numeric rating scale (NRS) for pain >4 at rest in the first 24 hours after surgery. Baseline characteristics in this study (validation cohort) were compared to those of the development cohort used to create the CAPPS score. Predictive performance of the CAPPS score was assessed by the area under the curve (AUC) and percentage misclassification for discrimination. A graphical comparison between predicted probability vs. observed incidence of poorly controlled pain was performed for calibration.

RESULTS

Fifty-two percent of 201 patients experienced poorly controlled pain. The validation cohort exhibited lower depression scores and a higher proportion using daily opioid medications compared to the development cohort. The AUC was 0.74 [95%CI = 0.68-0.81] in the validation cohort compared to 0.73 [95%CI = 0.69-0.76] in the development cohort for the eight-tier CAPPS score. When stratified between the low- vs. extreme-risk and low- vs. high-risk groups, the percentage misclassification was 21.2% and 30.7% in the validation cohort, compared to 29.9% and 38.0% in the development cohort, respectively. The predicted probability closely mirrored the observed incidence of poor pain control across all scores.

CONCLUSIONS

The CAPPS score, based on seven easily obtained and reliable prognostic variables, was validated using a prospectively collected, independent sample of patients.

[Application possibilities of digital tools in postoperative pain therapy]

BACKGROUND

Recently, digital tools, such as smartphone-based applications and the use of artificial intelligence have increasingly found their way into pain medicine. This could enable new treatment approaches in postoperative pain management. Therefore, this article provides an overview of various digital tools and their potential application options in postoperative pain management.

MATERIAL AND METHODS

An orienting literature search was carried out in the MEDLINE and Web of Science databases, and a targeted selection of essential key publications was made in order to provide a structured presentation of different current possible applications and a discussion based on the most recent knowledge.

RESULTS

Today, possible applications of digital tools, even if they mostly have only a model character, include pain documentation and assessment, patient self-management and education, pain prediction, decision support for medical staff, and supportive pain therapy, for example in the form of virtual reality and videos. These tools offer advantages such as individualized treatment concepts, addressing specific patient groups, reduction of pain and analgesics, and the potential for early warning or detection of postoperative pain. Furthermore, the challenges of the technical implementation and appropriate user training are highlighted.

CONCLUSION

The use of digital tools, although so far integrated in clinical routine in a relatively selective and exemplary manner, promises to be an innovative approach for personalized postoperative pain therapy in the future. Future studies and projects should help to integrate the promising research approaches into everyday clinical practice.

Artificial intelligence algorithms predict the efficacy of analgesic cocktails prescribed after orthopedic surgery

BACKGROUND

Mixtures ('cocktails') of various analgesics are more effective in controlling post-operative pain because of potential synergetic effects. Few studies have investigated such effects in large combinations of analgesics and no studies have determined the probabilities of effectiveness.

METHODS

We used one-hot encoding of the categorical variables reported pain levels and the administered cocktails (from a total of eight analgesics) and then applied an unsupervised neural network and then the unsupervised DBSCAN algorithm to detect clusters of cocktails. We used Bayesian statistics to classify the effectiveness of these cocktails.

RESULTS

Of the 61 different cocktails administered to 750 patients, we found that four combinations of three to four analgesics were by far the most effective. All these cocktails contained Metamizole and Paracetamol; three contained Hydromorphone and two contained Diclofenac and one Diclofenac-Orphenadrine. The ML probability that these cocktails decreased pain levels ranged from 0.965 to 0.981. Choice of a most effective cocktail involves choosing the optimum in a 4-dimensional parameter space: maximum probability of efficacy, confidence interval about maximum probability, fraction of patients with increase in pain levels, relative number of patients with successful pain level decrease.

CONCLUSIONS

We observed that administering one analgesic or at most two is not effective. We found no statistical indicators that interactions between analgesics in the most effective cocktails decreased their effectiveness. Pairs of most effective cocktails differed by the addition of only one analgesic (Diclofenac-Orphenadrine for one pair and Hydromorphone for the other). We conclude that the listed cocktails are to be recommended.

Fairness in the prediction of acute postoperative pain using machine learning models

Introduction

Overall performance of machine learning-based prediction models is promising; however, their generalizability and fairness must be vigorously investigated to ensure they perform sufficiently well for all patients.

Objective

This study aimed to evaluate prediction bias in machine learning models used for predicting acute postoperative pain.

Method

We conducted a retrospective review of electronic health records for patients undergoing orthopedic surgery from June 1, 2011, to June 30, 2019, at the University of Florida Health system/Shands Hospital. CatBoost machine learning models were trained for predicting the binary outcome of low (≤4) and high pain (>4). Model biases were assessed against seven protected attributes of age, sex, race, area deprivation index (ADI), speaking language, health literacy, and insurance type. Reweighing of protected attributes was investigated for reducing model bias compared with base models. Fairness metrics of equal opportunity, predictive parity, predictive equality, statistical parity, and overall accuracy equality were examined.

Results

The final dataset included 14,263 patients [age: 60.72 (16.03) years, 53.87% female, 39.13% low acute postoperative pain]. The machine learning model (area under the curve, 0.71) was biased in terms of age, race, ADI, and insurance type, but not in terms of sex, language, and health literacy. Despite promising overall performance in predicting acute postoperative pain, machine learning-based prediction models may be biased with respect to protected attributes.

Conclusion

These findings show the need to evaluate fairness in machine learning models involved in perioperative pain before they are implemented as clinical decision support tools.

Role of Artificial Intelligence and Machine Learning in the prediction of the pain: A scoping systematic review

Artificial Intelligence in healthcare is growing quickly in diagnostics and treatment management. Despite the quantity and variety of studies its role in clinical care is not clear. To identify the evidence gaps and characteristics of the Artificial Intelligence and Machine Learning techniques in predicting/diagnosing the pain? Pubmed/Embase were searched from the inception to October 2021 for articles without any language restrictions specifically addressing the following: the accuracy of AI in pain considering Brain Imaging, Patient-reported measures, and Electrophysiology, the ability of AI to differentiate stratify severity/types of pain, the ability of AI to predict pain and lastly the most accurate AI technique for given inputs. All the included studies were on humans. Eight hundred forty abstracts were reviewed, and 23 articles were finally included. Identified records were independently screened and relevant data was extracted. We performed conceptual synthesis by grouping the studies using available concepts of AL/ML techniques in diagnosing pain. Then we summarized the number of features/physiological measurements. Structured tabulation synthesis was used to show patterns predictions along with a narrative commentary. A total of 23 articles, published between 2015 and 2020 from 12 countries were included. Most studies were experimental in design. The most common design was cross sectional. Chronic or acute pains were predicted more often. Compared to control, the pain prediction was in the range of 57%-96% by AI techniques. Support Vector Machine and deep learning showed higher accuracy for classifying pain. From this study, it can be inferred that AI/ML can be used to differentiate healthy controls from patients. It can also facilitate categorizing them into new and different clinical subgroups. Lastly, it can predict future pain. The limitations are with respect to studies done after the search period. AL/ ML has a supportive role in pain diagnostics.

Improving Pain Assessment Using Vital Signs and Pain Medication for Patients With Sickle Cell Disease: Retrospective Study

BACKGROUND

Sickle cell disease (SCD) is the most common inherited blood disorder affecting millions of people worldwide. Most patients with SCD experience repeated, unpredictable episodes of severe pain. These pain episodes are the leading cause of emergency department visits among patients with SCD and may last for several weeks. Arguably, the most challenging aspect of treating pain episodes in SCD is assessing and interpreting a patient's pain intensity level.

OBJECTIVE

This study aims to learn deep feature representations of subjective pain trajectories using objective physiological signals collected from electronic health records.

METHODS

This study used electronic health record data collected from 496 Duke University Medical Center participants over 5 consecutive years. Each record contained measures for 6 vital signs and the patient's self-reported pain score, with an ordinal range from 0 (no pain) to 10 (severe and unbearable pain). We also extracted 3 features related to medication: medication type, medication status (given or applied, or missed or removed or due), and total medication dosage (mg/mL). We used variational autoencoders for representation learning and designed machine learning classification algorithms to build pain prediction models. We evaluated our results using an accuracy and confusion matrix and visualized the qualitative data representations.

RESULTS

We designed a classification model using raw data and deep representational learning to predict subjective pain scores with average accuracies of 82.8%, 70.6%, 49.3%, and 47.4% for 2-point, 4-point, 6-point, and 11-point pain ratings, respectively. We observed that random forest classification models trained on deep represented features outperformed models trained on unrepresented data for all pain rating scales. We observed that at varying Likert scales, our models performed better when provided with medication data along with vital signs data. We visualized the data representations to understand the underlying latent representations, indicating neighboring representations for similar pain scores with a higher resolution of pain ratings.

CONCLUSIONS

Our results demonstrate that medication information (the type of medication, total medication dosage, and whether the medication was given or missed) can significantly improve subjective pain prediction modeling compared with modeling with only vital signs. This study shows promise in data-driven estimated pain scores that will help clinicians with additional information about the patient's condition, in addition to the patient's self-reported pain scores.

Artificial intelligence and telemedicine in anesthesia: potential and problems

INTRODUCTION

The application of novel technologies like Artificial Intelligence (AI), Machine Learning (ML) and telemedicine in anesthesiology could play a role in transforming the future of health care. In the present review we discuss the current applications of AI and telemedicine in anesthesiology and perioperative care, exploring their potential influence and the possible hurdles.

EVIDENCE ACQUISITION

AI technologies have the potential to deeply impact all phases of perioperative care from accurate risk prediction to operating room organization, leading to increased cost-effective care quality and better outcomes. Telemedicine is reported as a successful mean within the anaesthetic pathway, including preoperative evaluation, remote patient monitoring, and postoperative care.

EVIDENCE SYNTHESIS

The utilization of AI and telemedicine is promising encouraging results in perioperative management, nevertheless several hurdles remain to be overcome before these tools could be integrated in our daily practice.

CONCLUSIONS

AI models and telemedicine can significantly influence all phases of perioperative care, helping physicians in the development of precision medicine.

Machine Learning in Pain Medicine: An Up-To-Date Systematic Review

INTRODUCTION

Pain is the unpleasant sensation and emotional experience that leads to poor quality of life for millions of people worldwide. Considering the complexity in understanding the principles of pain and its significant impact on individuals and society, research focuses to deliver innovative pain relief methods and techniques. This review explores the clinical uses of machine learning (ML) for the diagnosis, classification, and management of pain.

METHODS

A systematic review of the current literature was conducted using the PubMed database library.

RESULTS

Twenty-six papers related to pain and ML research were included. Most of the studies used ML for effectively classifying the patients' level of pain, followed by use of ML for the prediction of manifestation of pain and for pain management. A less common reason for performing ML analysis was for the diagnosis of pain. The different approaches are thoroughly discussed.

CONCLUSION

ML is increasingly used in pain medicine and appears to be more effective compared to traditional statistical approaches in the diagnosis, classification, and management of pain.

Development of a random forest model for hypotension prediction after anesthesia induction for cardiac surgery

BACKGROUND

Hypotension after the induction of anesthesia is known to be associated with various adverse events. The involvement of a series of factors makes the prediction of hypotension during anesthesia quite challenging.

AIM

To explore the ability and effectiveness of a random forest (RF) model in the prediction of post-induction hypotension (PIH) in patients undergoing cardiac surgery.

METHODS

Patient information was obtained from the electronic health records of the Second Affiliated Hospital of Hainan Medical University. The study included patients, ≥ 18 years of age, who underwent cardiac surgery from December 2007 to January 2018. An RF algorithm, which is a supervised machine learning technique, was employed to predict PIH. Model performance was assessed by the area under the curve (AUC) of the receiver operating characteristic. Mean decrease in the Gini index was used to rank various features based on their importance.

RESULTS

Of the 3030 patients included in the study, 1578 (52.1%) experienced hypotension after the induction of anesthesia. The RF model performed effectively, with an AUC of 0.843 (0.808-0.877) and identified mean blood pressure as the most important predictor of PIH after anesthesia. Age and body mass index also had a significant impact.

CONCLUSION

The generated RF model had high discrimination ability for the identification of individuals at high risk for a hypotensive event during cardiac surgery. The study results highlighted that machine learning tools confer unique advantages for the prediction of adverse post-anesthesia events.

Prioritizing Pain-Associated Targets with Machine Learning

While hundreds of genes have been associated with pain, much of the molecular mechanisms of pain remain unknown. As a result, current analgesics are limited to few clinically validated targets. Here, we trained a machine learning (ML) ensemble model to predict new targets for 17 categories of pain. The model utilizes features from transcriptomics, proteomics, and gene ontology to prioritize targets for modulating pain. We focused on identifying novel G-protein-coupled receptors (GPCRs), ion channels, and protein kinases because these proteins represent the most successful drug target families. The performance of the model to predict novel pain targets is 0.839 on average based on AUROC, while the predictions for arthritis had the highest accuracy (AUROC = 0.929). The model predicts hundreds of novel targets for pain; for example, GPR132 and GPR109B are highly ranked GPCRs for rheumatoid arthritis. Overall, gene-pain association predictions cluster into three groups that are enriched for cytokine, calcium, and GABA-related cell signaling pathways. These predictions can serve as a foundation for future experimental exploration to advance the development of safer and more effective analgesics.

Machine Learning and Surgical Outcomes Prediction: A Systematic Review

BACKGROUND

Machine learning (ML) has garnered increasing attention as a means to quantitatively analyze the growing and complex medical data to improve individualized patient care. We herein aim to critically examine the current state of ML in predicting surgical outcomes, evaluate the quality of currently available research, and propose areas of improvement for future uses of ML in surgery.

METHODS

A systematic review was conducted in accordance with the Preferred Reporting Items for a Systematic Review and Meta-Analysis (PRISMA) checklist. PubMed, MEDLINE, and Embase databases were reviewed under search syntax "machine learning" and "surgery" for papers published between 2015 and 2020.

RESULTS

Of the initial 2677 studies, 45 papers met inclusion and exclusion criteria. Fourteen different subspecialties were represented with neurosurgery being most common. The most frequently used ML algorithms were random forest (n = 19), artificial neural network (n = 17), and logistic regression (n = 17). Common outcomes included postoperative mortality, complications, patient reported quality of life and pain improvement. All studies which compared ML algorithms to conventional studies which used area under the curve (AUC) to measure accuracy found improved outcome prediction with ML models.

CONCLUSIONS

While still in its early stages, ML models offer surgeons an opportunity to capitalize on the myriad of clinical data available and improve individualized patient care. Limitations included heterogeneous outcome and imperfect quality of some of the papers. We therefore urge future research to agree upon methods of outcome reporting and require basic quality standards.

Big Data in Studying Acute Pain and Regional Anesthesia

The digital transformation of healthcare is advancing, leading to an increasing availability of clinical data for research. Perioperative big data initiatives were established to monitor treatment quality and benchmark outcomes. However, big data analyses have long exceeded the status of pure quality surveillance instruments. Large retrospective studies nowadays often represent the first approach to new questions in clinical research and pave the way for more expensive and resource intensive prospective trials. As a consequence, the utilization of big data in acute pain and regional anesthesia research has considerably increased over the last decade. Multicentric clinical registries and administrative databases (e.g., healthcare claims databases) have collected millions of cases until today, on which basis several important research questions were approached. In acute pain research, big data was used to assess postoperative pain outcomes, opioid utilization, and the efficiency of multimodal pain management strategies. In regional anesthesia, adverse events and potential benefits of regional anesthesia on postoperative morbidity and mortality were evaluated. This article provides a narrative review on the growing importance of big data for research in acute postoperative pain and regional anesthesia.

Novel Analgesic Index for Postoperative Pain Assessment Based on a Photoplethysmographic Spectrogram and Convolutional Neural Network: Observational Study

BACKGROUND

Although commercially available analgesic indices based on biosignal processing have been used to quantify nociception during general anesthesia, their performance is low in conscious patients. Therefore, there is a need to develop a new analgesic index with improved performance to quantify postoperative pain in conscious patients.

OBJECTIVE

This study aimed to develop a new analgesic index using photoplethysmogram (PPG) spectrograms and a convolutional neural network (CNN) to objectively assess pain in conscious patients.

METHODS

PPGs were obtained from a group of surgical patients for 6 minutes both in the absence (preoperatively) and in the presence (postoperatively) of pain. Then, the PPG data of the latter 5 minutes were used for analysis. Based on the PPGs and a CNN, we developed a spectrogram-CNN index for pain assessment. The area under the curve (AUC) of the receiver-operating characteristic curve was measured to evaluate the performance of the 2 indices.

RESULTS

PPGs from 100 patients were used to develop the spectrogram-CNN index. When there was pain, the mean (95% CI) spectrogram-CNN index value increased significantly-baseline: 28.5 (24.2-30.7) versus recovery area: 65.7 (60.5-68.3); P<.01. The AUC and balanced accuracy were 0.76 and 71.4%, respectively. The spectrogram-CNN index cutoff value for detecting pain was 48, with a sensitivity of 68.3% and specificity of 73.8%.

CONCLUSIONS

Although there were limitations to the study design, we confirmed that the spectrogram-CNN index can efficiently detect postoperative pain in conscious patients. Further studies are required to assess the spectrogram-CNN index's feasibility and prevent overfitting to various populations, including patients under general anesthesia.

TRIAL REGISTRATION

Clinical Research Information Service KCT0002080; https://cris.nih.go.kr/cris/search/search_result_st01.jsp?seq=6638.

Intersections of machine learning and epidemiological methods for health services research

The field of health services research is broad and seeks to answer questions about the health care system. It is inherently interdisciplinary, and epidemiologists have made crucial contributions. Parametric regression techniques remain standard practice in health services research with machine learning techniques currently having low penetrance in comparison. However, studies in several prominent areas, including health care spending, outcomes and quality, have begun deploying machine learning tools for these applications. Nevertheless, major advances in epidemiological methods are also as yet underleveraged in health services research. This article summarizes the current state of machine learning in key areas of health services research, and discusses important future directions at the intersection of machine learning and epidemiological methods for health services research.

Profiling Patients by Intensity of Nursing Care: An Operative Approach Using Machine Learning

Physical function is a patient-oriented indicator and can be considered a proxy for the assignment of healthcare personnel. The study aims to create an algorithm that classifies patients into homogeneous groups according to physical function. A two-step machine-learning algorithm was applied to administrative data recorded between 2015 and 2018 at the University Hospital of Padova. A clustering-large-applications (CLARA) algorithm was used to partition patients into homogeneous groups. Then, machine learning technique (MLT) classifiers were used to categorize the doubtful records. Based on the results of the CLARA algorithm, records were divided into three groups according to the Barthel index: <45, >65, ≥45 and ≤65. The support vector machine was the MLT showing the best performance among doubtful records, reaching an accuracy of 66%. The two-step algorithm, since it splits patients into low and high resource consumption, could be a useful tool for organizing healthcare personnel allocation according to the patients' assistance needs.

The Future Directions of Research in Cardiac Anesthesiology

This article provides an overview of knowledge gaps that need to be addressed in cardiac anesthesia, including mitigating the inflammatory effects of cardiopulmonary bypass, defining myocardial infarction after cardiac surgery, improving perioperative neurologic outcomes, and the optimal management of patients undergoing valve replacement. In addition, emerging approaches to research conduct are discussed, including the use of new analytical techniques like machine learning, pragmatic trials, and adaptive designs.

Finding NEEMO: towards organizing smart digital solutions in orthopaedic trauma surgery

There are many digital solutions which assist the orthopaedic trauma surgeon. This already broad field is rapidly expanding, making a complete overview of the existing solutions difficult.The AO Foundation has established a task force to address the need for an overview of digital solutions in the field of orthopaedic trauma surgery.Areas of new technology which will help the surgeon gain a greater understanding of these possible solutions are reviewed.We propose a categorization of the current needs in orthopaedic trauma surgery matched with available or potential digital solutions, and provide a narrative overview of this broad topic, including the needs, solutions and basic rules to ensure adequate use in orthopaedic trauma surgery. We seek to make this field more accessible, allowing for technological solutions to be clearly matched to trauma surgeons' needs. Cite this article: EFORT Open Rev 2020;5:408-420. DOI: 10.1302/2058-5241.5.200021.

Data science and machine learning in anesthesiology

Machine learning (ML) is revolutionizing anesthesiology research. Unlike classical research methods that are largely inference-based, ML is geared more towards making accurate predictions. ML is a field of artificial intelligence concerned with developing algorithms and models to perform prediction tasks in the absence of explicit instructions. Most ML applications, despite being highly variable in the topics that they deal with, generally follow a common workflow. For classification tasks, a researcher typically tests various ML models and compares the predictive performance with the reference logistic regression model. The main advantage of ML lies in its ability to deal with many features with complex interactions and its specific focus on maximizing predictive performance. However, emphasis on data-driven prediction can sometimes neglect mechanistic understanding. This article mainly focuses on the application of supervised ML to electronic health record (EHR) data. The main limitation of EHR-based studies is in the difficulty of establishing causal relationships. However, the associated low cost and rich information content provide great potential to uncover hitherto unknown correlations. In this review, the basic concepts of ML are introduced along with important terms that any ML researcher should know. Practical tips regarding the choice of software and computing devices are also provided. Towards the end, several examples of successful ML applications in anesthesiology are discussed. The goal of this article is to provide a basic roadmap to novice ML researchers working in the field of anesthesiology.

Artificial Intelligence and Machine Learning in Anesthesiology

Commercial applications of artificial intelligence and machine learning have made remarkable progress recently, particularly in areas such as image recognition, natural speech processing, language translation, textual analysis, and self-learning. Progress had historically languished in these areas, such that these skills had come to seem ineffably bound to intelligence. However, these commercial advances have performed best at single-task applications in which imperfect outputs and occasional frank errors can be tolerated.The practice of anesthesiology is different. It embodies a requirement for high reliability, and a pressured cycle of interpretation, physical action, and response rather than any single cognitive act. This review covers the basics of what is meant by artificial intelligence and machine learning for the practicing anesthesiologist, describing how decision-making behaviors can emerge from simple equations. Relevant clinical questions are introduced to illustrate how machine learning might help solve them-perhaps bringing anesthesiology into an era of machine-assisted discovery.

Primer on machine learning: utilization of large data set analyses to individualize pain management

PURPOSE OF REVIEW

Pain researchers and clinicians increasingly encounter machine learning algorithms in both research methods and clinical practice. This review provides a summary of key machine learning principles, as well as applications to both structured and unstructured datasets.

RECENT FINDINGS

Aside from increasing use in the analysis of electronic health record data, machine and deep learning algorithms are now key tools in the analyses of neuroimaging and facial expression recognition data used in pain research.

SUMMARY

In the coming years, machine learning is likely to become a key component of evidence-based medicine, yet will require additional skills and perspectives for its successful and ethical use in research and clinical settings.

Preoperative Assessment and Management of Patients with Pain and Anxiety Disorders

Purpose of Review

This review summarizes selected recent evidence on issues important for preoperative pain evaluation.

Recent Findings

Opioids, though a mainstay of postoperative pain management, are associated with both short and increasingly recognized long-term risks, including persistent opioid use. Risk factors for high levels of acute postoperative pain as well as chronic postsurgical pain may overlap, including psychological factors such as depression, anxiety, and catastrophizing. Tools to predict those at risk for poor postoperative pain outcomes are being studied.

Summary

Preoperative pain and psychological factors can affect postoperative pain outcomes. More work is needed in the future to develop practical interventions in the preoperative period to address these factors.

Use of Mobile Health Apps and Wearable Technology to Assess Changes and Predict Pain During Treatment of Acute Pain in Sickle Cell Disease: Feasibility Study

BACKGROUND

Sickle cell disease (SCD) is an inherited red blood cell disorder affecting millions worldwide, and it results in many potential medical complications throughout the life course. The hallmark of SCD is pain. Many patients experience daily chronic pain as well as intermittent, unpredictable acute vaso-occlusive painful episodes called pain crises. These pain crises often require acute medical care through the day hospital or emergency department. Following presentation, a number of these patients are subsequently admitted with continued efforts of treatment focused on palliative pain control and hydration for management. Mitigating pain crises is challenging for both the patients and their providers, given the perceived unpredictability and subjective nature of pain.

OBJECTIVE

The objective of this study was to show the feasibility of using objective, physiologic measurements obtained from a wearable device during an acute pain crisis to predict patient-reported pain scores (in an app and to nursing staff) using machine learning techniques.

METHODS

For this feasibility study, we enrolled 27 adult patients presenting to the day hospital with acute pain. At the beginning of pain treatment, each participant was given a wearable device (Microsoft Band 2) that collected physiologic measurements. Pain scores from our mobile app, Technology Resources to Understand Pain Assessment in Patients with Pain, and those obtained by nursing staff were both used with wearable signals to complete time stamp matching and feature extraction and selection. Following this, we constructed regression and classification machine learning algorithms to build between-subject pain prediction models.

RESULTS

Patients were monitored for an average of 3.79 (SD 2.23) hours, with an average of 5826 (SD 2667) objective data values per patient. As expected, we found that pain scores and heart rate decreased for most patients during the course of their stay. Using the wearable sensor data and pain scores, we were able to create a regression model to predict subjective pain scores with a root mean square error of 1.430 and correlation between observations and predictions of 0.706. Furthermore, we verified the hypothesis that the regression model outperformed the classification model by comparing the performances of the support vector machines (SVM) and the SVM for regression.

CONCLUSIONS

The Microsoft Band 2 allowed easy collection of objective, physiologic markers during an acute pain crisis in adults with SCD. Features can be extracted from these data signals and matched with pain scores. Machine learning models can then use these features to feasibly predict patient pain scores.

The Future Directions of Research in Cardiac Anesthesiology

This article provides an overview of knowledge gaps that need to be addressed in cardiac anesthesia, including mitigating the inflammatory effects of cardiopulmonary bypass, defining myocardial infarction after cardiac surgery, improving perioperative neurologic outcomes, and the optimal management of patients undergoing valve replacement. In addition, emerging approaches to research conduct are discussed, including the use of new analytical techniques like machine learning, pragmatic trials, and adaptive designs.

Supervised Machine-learning Predictive Analytics for Prediction of Postinduction Hypotension

Editor’s Perspective

What We Already Know about This Topic

What This Article Tells Us That Is New

Background

Hypotension is a risk factor for adverse perioperative outcomes. Machine-learning methods allow large amounts of data for development of robust predictive analytics. The authors hypothesized that machine-learning methods can provide prediction for the risk of postinduction hypotension.

Methods

Data was extracted from the electronic health record of a single quaternary care center from November 2015 to May 2016 for patients over age 12 that underwent general anesthesia, without procedure exclusions. Multiple supervised machine-learning classification techniques were attempted, with postinduction hypotension (mean arterial pressure less than 55 mmHg within 10 min of induction by any measurement) as primary outcome, and preoperative medications, medical comorbidities, induction medications, and intraoperative vital signs as features. Discrimination was assessed using cross-validated area under the receiver operating characteristic curve. The best performing model was tuned and final performance assessed using split-set validation.

Results

Out of 13,323 cases, 1,185 (8.9%) experienced postinduction hypotension. Area under the receiver operating characteristic curve using logistic regression was 0.71 (95% CI, 0.70 to 0.72), support vector machines was 0.63 (95% CI, 0.58 to 0.60), naive Bayes was 0.69 (95% CI, 0.67 to 0.69), k-nearest neighbor was 0.64 (95% CI, 0.63 to 0.65), linear discriminant analysis was 0.72 (95% CI, 0.71 to 0.73), random forest was 0.74 (95% CI, 0.73 to 0.75), neural nets 0.71 (95% CI, 0.69 to 0.71), and gradient boosting machine 0.76 (95% CI, 0.75 to 0.77). Test set area for the gradient boosting machine was 0.74 (95% CI, 0.72 to 0.77).

Conclusions

The success of this technique in predicting postinduction hypotension demonstrates feasibility of machine-learning models for predictive analytics in the field of anesthesiology, with performance dependent on model selection and appropriate tuning.

Using a Guided Machine Learning Ensemble Model to Predict Discharge Disposition following Meningioma Resection

Objective  Machine learning (ML) algorithms are powerful tools for predicting patient outcomes. This study pilots a novel approach to algorithm selection and model creation using prediction of discharge disposition following meningioma resection as a proof of concept. Materials and Methods  A diversity of ML algorithms were trained on a single-institution database of meningioma patients to predict discharge disposition. Algorithms were ranked by predictive power and top performers were combined to create an ensemble model. The final ensemble was internally validated on never-before-seen data to demonstrate generalizability. The predictive power of the ensemble was compared with a logistic regression. Further analyses were performed to identify how important variables impact the ensemble. Results  Our ensemble model predicted disposition significantly better than a logistic regression (area under the curve of 0.78 and 0.71, respectively, p  = 0.01). Tumor size, presentation at the emergency department, body mass index, convexity location, and preoperative motor deficit most strongly influence the model, though the independent impact of individual variables is nuanced. Conclusion  Using a novel ML technique, we built a guided ML ensemble model that predicts discharge destination following meningioma resection with greater predictive power than a logistic regression, and that provides greater clinical insight than a univariate analysis. These techniques can be extended to predict many other patient outcomes of interest.

Advancing Symptom Science Through Symptom Cluster Research: Expert Panel Proceedings and Recommendations

An overview of proceedings, findings, and recommendations from the workshop on "Advancing Symptom Science Through Symptom Cluster Research" sponsored by the National Institute of Nursing Research (NINR) and the Office of Rare Diseases Research, National Center for Advancing Translational Sciences, is presented. This workshop engaged an expert panel in an evidenced-based discussion regarding the state of the science of symptom clusters in chronic conditions including cancer and other rare diseases. An interdisciplinary working group from the extramural research community representing nursing, medicine, oncology, psychology, and bioinformatics was convened at the National Institutes of Health. Based on expertise, members were divided into teams to address key areas: defining characteristics of symptom clusters, priority symptom clusters and underlying mechanisms, measurement issues, targeted interventions, and new analytic strategies. For each area, the evidence was synthesized, limitations and gaps identified, and recommendations for future research delineated. The majority of findings in each area were from studies of oncology patients. However, increasing evidence suggests that symptom clusters occur in patients with other chronic conditions (eg, pulmonary, cardiac, and end-stage renal disease). Nonetheless, symptom cluster research is extremely limited and scientists are just beginning to understand how to investigate symptom clusters by developing frameworks and new methods and approaches. With a focus on personalized care, an understanding of individual susceptibility to symptoms and whether a "driving" symptom exists that triggers other symptoms in the cluster is needed. Also, research aimed at identifying the mechanisms that underlie symptom clusters is essential to developing targeted interventions.

Prediction of Patient-Controlled Analgesic Consumption: A Multimodel Regression Tree Approach

Several factors contribute to individual variability in postoperative pain, therefore, individuals consume postoperative analgesics at different rates. Although many statistical studies have analyzed postoperative pain and analgesic consumption, most have identified only the correlation and have not subjected the statistical model to further tests in order to evaluate its predictive accuracy. In this study involving 3052 patients, a multistrategy computational approach was developed for analgesic consumption prediction. This approach uses data on patient-controlled analgesia demand behavior over time and combines clustering, classification, and regression to mitigate the limitations of current statistical models. Cross-validation results indicated that the proposed approach significantly outperforms various existing regression methods. Moreover, a comparison between the predictions by anesthesiologists and medical specialists and those of the computational approach for an independent test data set of 60 patients further evidenced the superiority of the computational approach in predicting analgesic consumption because it produced markedly lower root mean squared errors.