Predicting common maternal postpartum complications: leveraging health administrative data and machine learning

Leveraging Machine Learning to Predict and Assess Disparities in Severe Maternal Morbidity in Maryland

BACKGROUND

Severe maternal morbidity (SMM) is increasing in the United States. The main objective of this study is to test the use of machine learning (ML) techniques to develop models for predicting SMM during delivery hospitalizations in Maryland. Secondarily, we examine disparities in SMM by key sociodemographic characteristics.

METHODS

We used the linked State Inpatient Database (SID) and the American Hospital Association (AHA) Annual Survey data from Maryland for 2016-2019 (N = 261,226 delivery hospitalizations). We first estimated relative risks for SMM across key sociodemographic factors (e.g., race, income, insurance, and primary language). Then, we fitted LASSO and, for comparison, Logit models with 75 and 18 features. The selection of SMM features was based on clinical expert opinion, a literature review, statistical significance, and computational resource constraints. Various model performance metrics, including the area under the receiver operating characteristic curve (AUC), accuracy, precision, and recall values were computed to compare predictive performance.

RESULTS

During 2016-2019, 76 per 10,000 deliveries (1976 of 261,226) were in patients who experienced an SMM event. The Logit model with a full list of 75 features achieved an AUC of 0.71 in the validation dataset, which marginally decreased to 0.69 in the reduced model with 18 features. The LASSO algorithm with the same 18 features demonstrated slightly superior predictive performance and an AUC of 0.80. We found significant disparities in SMM among patients living in low-income areas, with public insurance, and who were non-Hispanic Black or non-English speakers.

CONCLUSION

Our results demonstrate the feasibility of utilizing ML and administrative hospital discharge data for SMM prediction. The low recall score is a limitation across all models we compared, signifying that the algorithms struggle with identifying all SMM cases. This study identified substantial disparities in SMM across various sociodemographic factors. Addressing these disparities requires multifaceted interventions that include improving access to quality care, enhancing cultural competence among healthcare providers, and implementing policies that help mitigate social determinants of health.

Artificial Intelligence in Predicting Postpartum Hemorrhage in Twin Pregnancies Undergoing Cesarean Section

This study aimed to create a risk prediction model with artificial intelligence (AI) to identify patients at higher risk of postpartum hemorrhage using perinatal characteristics that may be associated with later postpartum hemorrhage (PPH) in twin pregnancies that underwent cesarean section. The study was planned as a retrospective cohort study at University Hospital. All twin cesarean deliveries were categorized into two groups: those with and without PPH. Using the perinatal characteristics of the cases, four different machine learning classifiers were created: Logistic regression (LR), support vector machine (SVM), random forest (RF), and multilayer perceptron (MLP). LR, RF, and SVM models were created a second time by including class weights to manage the underlying imbalances in the data. A total of 615 twin pregnancies were included in the study. There were 150 twin pregnancies with PPH and 465 without PPH. Dichorionity, PAS, and placenta previa were significantly higher in the PPH-positive group (p = .045, p = .004, p = .001 respectively). In our model, LR with class weight was the best model with the highest negative predictive value. The AUC in our LR with class weight model was %75.12 with an accuracy of 70.73%, a PPV of 47.92%, and an NPV of 85.33% in our data. Although the application of machine learning to create predictive models using clinical risk factors and our model's 70% accuracy rate are encouraging, it is not sufficient. Machine learning modeling needs further study and validation before being incorporated into clinical use.

Prediction of preterm birth in multiparous women using logistic regression and machine learning approaches

To predict preterm birth (PTB) in multiparous women, comparing machine learning approaches with traditional logistic regression. A population-based cohort study was conducted using data from the Ontario Better Outcomes Registry and Network (BORN). The cohort included all multiparous women who delivered a singleton birth at 20-42 weeks' gestation in an Ontario hospital between April 1, 2012 and March 31, 2014. The primary outcome was PTB < 37 weeks, with spontaneous PTB analyzed as a secondary outcome. Stepwise logistic regression and the Boruta machine learning were used to select the important variables during the first and second trimester. For building prediction models, the whole data set were divided for the two independent parts: two-third for training the classifiers (Logistic regression, random forests, decision trees, and artificial neural networks) and one-third for model validation. Then, the training data set were balanced by random over sampling technique. The best hyper parameters were obtained by the tenfold cross validation. The performance of all models was evaluated by sensitivity, specificity, positive predictive value, negative predictive value, and the area under the receiver operating characteristics (AUC). The cohort included 145,846 births, of which 8125 (5.57%) were preterm. In first-trimester models, the strongest predictors of PTB were previous PTB, preexisting diabetes, and abnormal pregnancy-associated plasma protein-A. In the testing data set, the highest predictive ability was seen for artificial neural networks, with an area under the receiver operating characteristic curve (AUC) of 68.8% (95% CI 67.6-70.1%). In second-trimester models, addition of infant sex, attendance at first-trimester appointment, medication exposure, and abnormal alpha-fetoprotein concentrations increased the AUC to 72.1% (95% CI 71.1-73.1%) with logistic regression. With the inclusion of the variable complications during pregnancy, the AUC increased to 80.5% (95% CI 79.6-81.5%) using logistic regression. For both overall and spontaneous PTB, during both the first and second trimesters, models yielded negative predictive values of 97%. Overall, machine learning and logistic regression produced similar performance for prediction of PTB. For overall and spontaneous PTB, both first- and second-trimester models provided negative predictive values of ~ 97%, higher than that of fetal fibronectin.

Quantitative prediction of postpartum hemorrhage in cesarean section on machine learning

BACKGROUND

Cesarean section-induced postpartum hemorrhage (PPH) potentially causes anemia and hypovolemic shock in pregnant women. Hence, it is helpful for obstetricians and anesthesiologists to prepare pre-emptive prevention when predicting PPH occurrence in advance. However, current works on PPH prediction focus on whether PPH occurs rather than assessing PPH amount. To this end, this work studies quantitative PPH prediction with machine learning (ML).

METHODS

The study cohort in this paper was selected from individuals with PPH who were hospitalized at Shijiazhuang Obstetrics and Gynecology Hospital from 2020 to 2022. In this study cohort, we built a dataset with 6,144 subjects covering clinical parameters, anesthesia operation records, laboratory examination results, and other information in the electronic medical record system. Based on our built dataset, we exploit six different ML models, including logistic regression, linear regression, gradient boosting, XGBoost, multilayer perceptron, and random forest, to automatically predict the amount of bleeding during cesarean section. Eighty percent of the dataset was used as model training, and 20 % was used for verification. Those ML models are constantly verified and improved by root mean squared error(RMSE) and mean absolute error(MAE). Moreover, we also leverage the importance of permutation and partial dependence plot (PDP) to discuss their feasibility.

RESULT

The experiment results show that random forest obtains the highest accuracy for PPH amount prediction compared to other ML methods. Random forest reaches the mean absolute error of 21.7, less than 5.4 % prediction error. It also gains the root mean squared error of 33.75, less than 9.3 % prediction error. On the other hand, the experimental results also disclose indicators that contributed most to PPH prediction, including Ca, hemoglobin, white blood cells, platelets, Na, and K.

CONCLUSION

It effectively predicts the amount of PPH during a cesarean section by ML methods, especially random forest. With the above insight, ML predicting PPH amounts provides early warning for clinicians, thus reducing complications and improving cesarean sections' safety. Furthermore, the importance of ML and permutation, complemented by incorporating PDP, promises to provide clinicians with a transparent indication of individual risk prediction.

Development of a prediction model of postpartum hospital use using an equity-focused approach

BACKGROUND

Racial inequities in maternal morbidity and mortality persist into the postpartum period, leading to a higher rate of postpartum hospital use among Black and Hispanic people. Delivery hospitalizations provide an opportunity to screen and identify people at high risk to prevent adverse postpartum outcomes. Current models do not adequately incorporate social and structural determinants of health, and some include race, which may result in biased risk stratification.

OBJECTIVE

This study aimed to develop a risk prediction model of postpartum hospital use while incorporating social and structural determinants of health and using an equity approach.

STUDY DESIGN

We conducted a retrospective cohort study using 2016-2018 linked birth certificate and hospital discharge data for live-born infants in New York City. We included deliveries from 2016 to 2017 in model development, randomly assigning 70%/30% of deliveries as training/test data. We used deliveries in 2018 for temporal model validation. We defined "Composite postpartum hospital use" as at least 1 readmission or emergency department visit within 30 days of the delivery discharge. We categorized diagnosis at first hospital use into 14 categories based on International Classification of Diseases-Tenth Revision diagnosis codes. We tested 72 candidate variables, including social determinants of health, demographics, comorbidities, obstetrical complications, and severe maternal morbidity. Structural determinants of health were the Index of Concentration at the Extremes, which is an indicator of racial-economic segregation at the zip code level, and publicly available indices of the neighborhood built/natural and social/economic environment of the Child Opportunity Index. We used 4 statistical and machine learning algorithms to predict "Composite postpartum hospital use", and an ensemble approach to predict "Cause-specific postpartum hospital use". We simulated the impact of each risk stratification method paired with an effective intervention on race-ethnic equity in postpartum hospital use.

RESULTS

The overall incidence of postpartum hospital use was 5.7%; the incidences among Black, Hispanic, and White people were 8.8%, 7.4%, and 3.3%, respectively. The most common diagnoses for hospital use were general perinatal complications (17.5%), hypertension/eclampsia (12.0%), nongynecologic infections (10.7%), and wound infections (8.4%). Logistic regression with least absolute shrinkage and selection operator selection retained 22 predictor variables and achieved an area under the receiver operating curve of 0.69 in the training, 0.69 in test, and 0.69 in validation data. Other machine learning algorithms performed similarly. Selected social and structural determinants of health features included the Index of Concentration at the Extremes, insurance payor, depressive symptoms, and trimester entering prenatal care. The "Cause-specific postpartum hospital use" model selected 6 of the 14 outcome diagnoses (acute cardiovascular disease, gastrointestinal disease, hypertension/eclampsia, psychiatric disease, sepsis, and wound infection), achieving an area under the receiver operating curve of 0.75 in training, 0.77 in test, and 0.75 in validation data using a cross-validation approach. Models had slightly lower performance in Black and Hispanic subgroups. When simulating use of the risk stratification models with a postpartum intervention, identifying high-risk individuals with the "Composite postpartum hospital use" model resulted in the greatest reduction in racial-ethnic disparities in postpartum hospital use, compared with the "Cause-specific postpartum hospital use" model or a standard approach to identifying high-risk individuals with common pregnancy complications.

CONCLUSION

The "Composite postpartum hospital use" prediction model incorporating social and structural determinants of health can be used at delivery discharge to identify persons at risk for postpartum hospital use.

Use of Machine Learning for Early Detection of Maternal Cardiovascular Conditions: Retrospective Study Using Electronic Health Record Data

BACKGROUND

Cardiovascular conditions (eg, cardiac and coronary conditions, hypertensive disorders of pregnancy, and cardiomyopathies) were the leading cause of maternal mortality between 2017 and 2019. The United States has the highest maternal mortality rate of any high-income nation, disproportionately impacting those who identify as non-Hispanic Black or Hispanic. Novel clinical approaches to the detection and diagnosis of cardiovascular conditions are therefore imperative. Emerging research is demonstrating that machine learning (ML) is a promising tool for detecting patients at increased risk for hypertensive disorders during pregnancy. However, additional studies are required to determine how integrating ML and big data, such as electronic health records (EHRs), can improve the identification of obstetric patients at higher risk of cardiovascular conditions.

OBJECTIVE

This study aimed to evaluate the capability and timing of a proprietary ML algorithm, Healthy Outcomes for all Pregnancy Experiences-Cardiovascular-Risk Assessment Technology (HOPE-CAT), to detect maternal-related cardiovascular conditions and outcomes.

METHODS

Retrospective data from the EHRs of a large health care system were investigated by HOPE-CAT in a virtual server environment. Deidentification of EHR data and standardization enabled HOPE-CAT to analyze data without pre-existing biases. The ML algorithm assessed risk factors selected by clinical experts in cardio-obstetrics, and the algorithm was iteratively trained using relevant literature and current standards of risk identification. After refinement of the algorithm's learned risk factors, risk profiles were generated for every patient including a designation of standard versus high risk. The profiles were individually paired with clinical outcomes pertaining to cardiovascular pregnancy conditions and complications, wherein a delta was calculated between the date of the risk profile and the actual diagnosis or intervention in the EHR.

RESULTS

In total, 604 pregnancies resulting in birth had records or diagnoses that could be compared against the risk profile; the majority of patients identified as Black (n=482, 79.8%) and aged between 21 and 34 years (n=509, 84.4%). Preeclampsia (n=547, 90.6%) was the most common condition, followed by thromboembolism (n=16, 2.7%) and acute kidney disease or failure (n=13, 2.2%). The average delta was 56.8 (SD 69.7) days between the identification of risk factors by HOPE-CAT and the first date of diagnosis or intervention of a related condition reported in the EHR. HOPE-CAT showed the strongest performance in early risk detection of myocardial infarction at a delta of 65.7 (SD 81.4) days.

CONCLUSIONS

This study provides additional evidence to support ML in obstetrical patients to enhance the early detection of cardiovascular conditions during pregnancy. ML can synthesize multiday patient presentations to enhance provider decision-making and potentially reduce maternal health disparities.

A Prediction Model for Severe Maternal Morbidity and Mortality After Delivery Hospitalization

OBJECTIVE

To develop a risk stratification model for severe maternal morbidity (SMM) or mortality after the delivery hospitalization based on information available at the time of hospital discharge.

METHODS

This population-based cohort study included all pregnancies among Ohio residents with Medicaid insurance from 2012 to 2017. Pregnant individuals were identified using linked live birth and fetal death records and Medicaid claims data. Inclusion was restricted to those with continuous postpartum Medicaid enrollment and delivery at 20 or more weeks of gestation. The primary outcome of the study was SMM or mortality after the delivery hospitalization and was assessed up to 42 days postpartum and up to 1 year postpartum separately. Variables considered for the model included patient-, obstetric health care professional-, and system-level data available in vital records or Medicaid claims data. Parsimonious models were created with logistic regression and were internally validated. Receiver operating characteristic curves were used to evaluate model performance, and model calibration was assessed.

RESULTS

There were 343,842 pregnant individuals who met inclusion criteria with continuous Medicaid enrollment through 42 days postpartum and 287,513 with continuous enrollment through 1 year. After delivery hospitalization discharge, the incidence of SMM or mortality was 140.5 per 10,000 pregnancies through 42 days of delivery and 330.7 per 10,000 pregnancies through 1 year postpartum. The final model predicting SMM or mortality through 42 days postpartum included maternal prepregnancy body mass index, age, gestational age at delivery, mode of delivery, chorioamnionitis, and maternal diagnosis of cardiac disease, preeclampsia or gestational hypertension, or a mental health condition. Similar variables were included in the model predicting SMM or mortality through 365 days with chronic hypertension, pregestational diabetes, and illicit substance use added and chorioamnionitis removed. Both models demonstrated moderate prediction (area under the curve [AUC] 0.77, 95% CI 0.76-0.78 for 42-day model; AUC 0.72, 95% CI 0.71-0.73 for the 1-year model) and good calibration.

CONCLUSION

A prediction model for SMM or mortality up to 1 year postpartum was created and internally validated with information available to health care professionals at the time of hospital discharge. The utility of this model for patient counseling and strategies to optimize postpartum care for high-risk individuals will require further evaluation.

Machine Learning Models of Postoperative Atrial Fibrillation Prediction After Cardiac Surgery

OBJECTIVES

This study aimed to use machine learning algorithms to build an efficient forecasting model of atrial fibrillation after cardiac surgery, and to compare the predictive performance of machine learning to traditional logistic regression.

DESIGN

A retrospective study.

SETTING

Second Affiliated Hospital of Zhejiang University School of Medicine.

PARTICIPANTS

The study comprised 1,400 patients who underwent valve and/or coronary artery bypass grafting surgery with cardiopulmonary bypass from September 1, 2013 to December 31, 2018.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Two machine learning approaches (gradient-boosting decision tree and support-vector machine) and logistic regression were used to build predictive models. The performance was compared by the area under the curve (AUC). The clinical practicability was assessed using decision curve analysis. Postoperative atrial fibrillation occurred in 519 patients (37.1%). The AUCs of the support-vector machine, logistic regression, and gradient boosting decision tree were 0.777 (95% CI: 0.772-0.781), 0.767 (95% CI: 0.762-0.772), and 0.765 (95% CI: 0.761-0.770), respectively. As decision curve analysis manifested, these models had achieved appropriate net benefit.

CONCLUSION

In the authors' study, the support-vector machine model was the best predictor; it may be an effective tool for predicting atrial fibrillation after cardiac surgery.

Evaluation of clinical versus non-clinical continuing education in terms of preferences and value for oral healthcare workers

BACKGROUND

Continuing professional development (clinical) and continuing education (non-clinical) is fundamental to education and self-improvement of all categories of staff within a large healthcare facility.

AIM

This study sought to examine the attendance preferences and perceived value of clinical and non-clinical oral healthcare workers towards clinical continuing professional development (CPD) and non-clinical, continuing education (CE) activities.

METHODS

A retrospective cross-sectional survey design was used capturing 8640 self-reported evaluations collected across 8 successive years and 160 CPD and CE activities in a large dental hospital. Analysis was performed using descriptive statistics including mean scores, independent t-test and cross tabulations using chi-square.

RESULTS

A strongly significant association (p < 0.001) was found between attendee position type (clinical or non-clinical) and attendance preference to either clinical or non-clinical education. Dental assistants, compared to Dentist/Specialist (p < 0.001) found the programs more accurate, relevant, improved their knowledge, would use what was learned and rated the sessions higher overall. Clinical CPD was deemed more relevant (p = 0.025) and improved knowledge (p = 0.01) while non-clinical CE had higher presenter quality (p < 0.001) and overall mean scores (p = 0.015).

CONCLUSION

There was a preference towards attending clinical CPD over non-clinical CE, by not only clinical, but also non-clinical oral healthcare workers. Non-clinical CE was scored higher by both clinical and non-clinical participants and should therefore be considered for inclusion in CPD education programs with similar settings.

Prediction of Maternal Hemorrhage Using Machine Learning: Retrospective Cohort Study

Background

Postpartum hemorrhage remains one of the largest causes of maternal morbidity and mortality in the United States.

Objective

The aim of this paper is to use machine learning techniques to identify patients at risk for postpartum hemorrhage at obstetric delivery.

Methods

Women aged 18 to 55 years delivering at a major academic center from July 2013 to October 2018 were included for analysis (N=30,867). A total of 497 variables were collected from the electronic medical record including the following: demographic information; obstetric, medical, surgical, and family history; vital signs; laboratory results; labor medication exposures; and delivery outcomes. Postpartum hemorrhage was defined as a blood loss of ≥1000 mL at the time of delivery, regardless of delivery method, with 2179 (7.1%) positive cases observed. Supervised learning with regression-, tree-, and kernel-based machine learning methods was used to create classification models based upon training (21,606/30,867, 70%) and validation (4630/30,867, 15%) cohorts. Models were tuned using feature selection algorithms and domain knowledge. An independent test cohort (4631/30,867, 15%) determined final performance by assessing for accuracy, area under the receiver operating curve (AUROC), and sensitivity for proper classification of postpartum hemorrhage. Separate models were created using all collected data versus models limited to data available prior to the second stage of labor or at the time of decision to proceed with cesarean delivery. Additional models examined patients by mode of delivery.

Results

Gradient boosted decision trees achieved the best discrimination in the overall model. The model including all data mildly outperformed the second stage model (AUROC 0.979, 95% CI 0.971-0.986 vs AUROC 0.955, 95% CI 0.939-0.970). Optimal model accuracy was 98.1% with a sensitivity of 0.763 for positive prediction of postpartum hemorrhage. The second stage model achieved an accuracy of 98.0% with a sensitivity of 0.737. Other selected algorithms returned models that performed with decreased discrimination. Models stratified by mode of delivery achieved good to excellent discrimination but lacked the sensitivity necessary for clinical applicability.

Conclusions

Machine learning methods can be used to identify women at risk for postpartum hemorrhage who may benefit from individualized preventative measures. Models limited to data available prior to delivery perform nearly as well as those with more complete data sets, supporting their potential utility in the clinical setting. Further work is necessary to create successful models based upon mode of delivery and to validate the findings of this study. An unbiased approach to hemorrhage risk prediction may be superior to human risk assessment and represents an area for future research.

A novel oppositional binary crow search algorithm with optimal machine learning based postpartum hemorrhage prediction model

Postpartum hemorrhage (PPH) is an obstetric emergency instigated by excessive blood loss which occurs frequently after the delivery. The PPH can result in volume depletion, hypovolemic shock, and anemia. This is particular condition is considered a major cause of maternal deaths around the globe. Presently, physicians utilize visual examination for calculating blood and fluid loss during delivery. Since the classical methods depend on expert knowledge and are inaccurate, automated machine learning based PPH diagnosis models are essential. In regard to this aspect, this study introduces an efficient oppositional binary crow search algorithm (OBCSA) with an optimal stacked auto encoder (OSAE) model, called OBCSA-OSAE for PPH prediction. The goal of the proposed OBCSA-OSAE technique is to detect and classify the presence or absence of PPH. The OBCSA-OSAE technique involves the design of OBCSA based feature selection (FS) methods to elect an optimum feature subset. Additionally, the OSAE based classification model is developed to include an effective parameter adjustment process utilizing Equilibrium Optimizer (EO). The performance validation of the OBCSA-OSAE technique is performed using the benchmark dataset. The experimental values pointed out the benefits of the OBCSA-OSAE approach in recent methods.

Contributions of artificial intelligence reported in Obstetrics and Gynecology journals: a systematic review

Background

The applications of artificial intelligence (AI) processes have grown significantly in all medical disciplines during the last decades. Two main types of AI have been applied in medicine: symbolic AI (eg, knowledge base and ontologies) and nonsymbolic AI (eg, machine learning and artificial neural networks). Consequently, AI has also been applied across most obstetrics and gynecology (OB/GYN) domains, including general obstetrics, gynecology surgery, fetal ultrasound, and assisted reproductive medicine, among others.

Objective

The aim of this study was to provide a systematic review to establish the actual contributions of AI reported in OB/GYN discipline journals.

Methods

The PubMed database was searched for citations indexed with “artificial intelligence” and at least one of the following medical subject heading (MeSH) terms between January 1, 2000, and April 30, 2020: “obstetrics”; “gynecology”; “reproductive techniques, assisted”; or “pregnancy.” All publications in OB/GYN core disciplines journals were considered. The selection of journals was based on disciplines defined in Web of Science. The publications were excluded if no AI process was used in the study. Review, editorial, and commentary articles were also excluded. The study analysis comprised (1) classification of publications into OB/GYN domains, (2) description of AI methods, (3) description of AI algorithms, (4) description of data sets, (5) description of AI contributions, and (6) description of the validation of the AI process.

Results

The PubMed search retrieved 579 citations and 66 publications met the selection criteria. All OB/GYN subdomains were covered: obstetrics (41%, 27/66), gynecology (3%, 2/66), assisted reproductive medicine (33%, 22/66), early pregnancy (2%, 1/66), and fetal medicine (21%, 14/66). Both machine learning methods (39/66) and knowledge base methods (25/66) were represented. Machine learning used imaging, numerical, and clinical data sets. Knowledge base methods used mostly omics data sets. The actual contributions of AI were method/algorithm development (53%, 35/66), hypothesis generation (42%, 28/66), or software development (3%, 2/66). Validation was performed on one data set (86%, 57/66) and no external validation was reported. We observed a general rising trend in publications related to AI in OB/GYN over the last two decades. Most of these publications (82%, 54/66) remain out of the scope of the usual OB/GYN journals.

Conclusions

In OB/GYN discipline journals, mostly preliminary work (eg, proof-of-concept algorithm or method) in AI applied to this discipline is reported and clinical validation remains an unmet prerequisite. Improvement driven by new AI research guidelines is expected. However, these guidelines are covering only a part of AI approaches (nonsymbolic) reported in this review; hence, updates need to be considered.

Application of logistic regression and convolutional neural network in prediction and diagnosis of high-risk populations of lung cancer

OBJECTIVES

The early detection, early diagnosis, and early treatment of lung cancer are the best strategies to improve the 5-year survival rate. Logistic regression analysis can be a helpful tool in the early detection of high-risk groups of lung cancer. Convolutional neural network (CNN) could distinguish benign from malignant pulmonary nodules, which is critical for early precise diagnosis and treatment. Here, we developed a risk assessment model of lung cancer and a high-precision classification diagnostic model using these technologies so as to provide a basis for early screening of lung cancer and for intelligent differential diagnosis.

METHODS

A total of 355 lung cancer patients, 444 patients with benign lung disease and 472 healthy people from The First Affiliated Hospital of Zhengzhou University were included in this study. Moreover, the dataset of 607 lung computed tomography images was collected from the above patients. The logistic regression method was employed to screen the high-risk groups of lung cancer, and the CNN model was designed to classify pulmonary nodules into benign or malignant nodules.

RESULTS

The area under the curve of the lung cancer risk assessment model in the training set and the testing set were 0.823 and 0.710, respectively. After finely optimizing the settings of the CNN, the area under the curve could reach 0.984.

CONCLUSIONS

This performance demonstrated that the lung cancer risk assessment model could be used to screen for high-risk individuals with lung cancer and the CNN framework was suitable for the differential diagnosis of pulmonary nodules.

Machine learning-based prediction of postpartum hemorrhage after vaginal delivery: combining bleeding high risk factors and uterine contraction curve

PURPOSE

This work used a machine learning model to improve the accuracy of predicting postpartum hemorrhage in vaginal delivery.

METHODS

Among the 25,098 deliveries in the obstetrics department of the First Hospital of Jinan University recorded from 2016 to 2020, 10,520 were vaginal deliveries with complete study data. Further review selected 850 cases of postpartum hemorrhage (amount of bleeding > 500 mL) and 54 cases of severe postpartum hemorrhage (amount of bleeding > 1000 mL). Indicators of clinical risk factors for postpartum hemorrhage were retrieved from electronic medical records. Features of the uterine contraction curve were extracted 2 h prior to vaginal delivery and modeled using a 49-variable machine learning with 90% of study cases used in the training set and 10% of study cases used in the test set. Accuracy was compared among the assessment table, classical statistical models, and machine learning models used to predict postpartum hemorrhage to assess their clinical use. The assessment table contained 16 high-risk factor scores to predict postpartum hemorrhage. The classical statistical model used was Logistic Regression (LR). The machine learning models were Random Forest (RF), K Nearest Neighbor (KNN), and the one integrated with Lightgbm (LGB) and LR. The effect of model prediction was evaluated by area under the receiver operating characteristic curve (AUC), namely, C-static, calibration curve Brier score, decision curve, F-measure, sensitivity (SE), and specificity (SP).

RESULTS

1: Among the tested tools, the machine learning model LGB + LR has the best performance in predicting postpartum hemorrhage. Its Brier, AUC, and F-measure scores are better than those of other models in each group, and its SE and SP reach 0.694 and 0.800, respectively. The predictive performance of the classical statistical model LR is AUC: 0.729, 95%CI [0.702-0.756]). 2: Verification on the testing set reveals that the features of uterine contraction contribute to the improved accuracy of the model prediction. 3: LGB + LR model suggested that among the 49 indicators for predicting severe postpartum hemorrhage, the importance of the first 10 characteristics in descending order is as follows: hematocrit (%), shock index, frequency of contractions (min-1), white blood cell count, gestational hypertension, neonatal weight (kg), time of second labor (min), mean area of contractions (mmHg s), total amniotic fluid (mL), and body mass index (BMI). The prediction effect is close to that of the model after training with all 49 features. The predictive effect was close to that of the model after training using all 49 features. 4: Contraction frequency and intensity Mean_Area (representing effective contractions) have a high predictive value for severe postpartum hemorrhage. 5: Blood loss amount within 2 h has a high warning effect on postpartum hemorrhage, and the increase in AUC to 0.95 indicates that postpartum bleeding mostly occurs within 2 h after delivery.

CONCLUSION

Machine learning models incorporated with uterine contraction features can further improve the accuracy of postpartum hemorrhage prediction in vaginal delivery and provide a reference for clinicians to intervene early and reduce adverse pregnancy outcomes.

Artificial intelligence in obstetrics

This study reviews recent advances on the application of artificial intelligence for the early diagnosis of various maternal-fetal conditions such as preterm birth and abnormal fetal growth. It is found in this study that various machine learning methods have been successfully employed for different kinds of data capture with regard to early diagnosis of maternal-fetal conditions. With the more popular use of artificial intelligence, ethical issues should also be considered accordingly.

The process of sourcing and preparing EHR data to implement a machine-learning algorithm for early identification of maternal cardiovascular risk (Preprint)

Background

Health care data are fragmenting as patients seek care from diverse sources. Consequently, patient care is negatively impacted by disparate health records. Machine learning (ML) offers a disruptive force in its ability to inform and improve patient care and outcomes. However, the differences that exist in each individual’s health records, combined with the lack of health data standards, in addition to systemic issues that render the data unreliable and that fail to create a single view of each patient, create challenges for ML. Although these problems exist throughout health care, they are especially prevalent within maternal health and exacerbate the maternal morbidity and mortality crisis in the United States.

Objective

This study aims to demonstrate that patient records extracted from the electronic health records (EHRs) of a large tertiary health care system can be made actionable for the goal of effectively using ML to identify maternal cardiovascular risk before evidence of diagnosis or intervention within the patient’s record. Maternal patient records were extracted from the EHRs of a large tertiary health care system and made into patient-specific, complete data sets through a systematic method.

Methods

We outline the effort that was required to define the specifications of the computational systems, the data set, and access to relevant systems, while ensuring that data security, privacy laws, and policies were met. Data acquisition included the concatenation, anonymization, and normalization of health data across multiple EHRs in preparation for their use by a proprietary risk stratification algorithm designed to establish patient-specific baselines to identify and establish cardiovascular risk based on deviations from the patient’s baselines to inform early interventions.

Results

Patient records can be made actionable for the goal of effectively using ML, specifically to identify cardiovascular risk in pregnant patients.

Conclusions

Upon acquiring data, including their concatenation, anonymization, and normalization across multiple EHRs, the use of an ML-based tool can provide early identification of cardiovascular risk in pregnant patients.

Application of Artificial Intelligence in Community-Based Primary Health Care: Systematic Scoping Review and Critical Appraisal

BACKGROUND

Research on the integration of artificial intelligence (AI) into community-based primary health care (CBPHC) has highlighted several advantages and disadvantages in practice regarding, for example, facilitating diagnosis and disease management, as well as doubts concerning the unintended harmful effects of this integration. However, there is a lack of evidence about a comprehensive knowledge synthesis that could shed light on AI systems tested or implemented in CBPHC.

OBJECTIVE

We intended to identify and evaluate published studies that have tested or implemented AI in CBPHC settings.

METHODS

We conducted a systematic scoping review informed by an earlier study and the Joanna Briggs Institute (JBI) scoping review framework and reported the findings according to PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analysis-Scoping Reviews) reporting guidelines. An information specialist performed a comprehensive search from the date of inception until February 2020, in seven bibliographic databases: Cochrane Library, MEDLINE, EMBASE, Web of Science, Cumulative Index to Nursing and Allied Health Literature (CINAHL), ScienceDirect, and IEEE Xplore. The selected studies considered all populations who provide and receive care in CBPHC settings, AI interventions that had been implemented, tested, or both, and assessed outcomes related to patients, health care providers, or CBPHC systems. Risk of bias was assessed using the Prediction Model Risk of Bias Assessment Tool (PROBAST). Two authors independently screened the titles and abstracts of the identified records, read the selected full texts, and extracted data from the included studies using a validated extraction form. Disagreements were resolved by consensus, and if this was not possible, the opinion of a third reviewer was sought. A third reviewer also validated all the extracted data.

RESULTS

We retrieved 22,113 documents. After the removal of duplicates, 16,870 documents were screened, and 90 peer-reviewed publications met our inclusion criteria. Machine learning (ML) (41/90, 45%), natural language processing (NLP) (24/90, 27%), and expert systems (17/90, 19%) were the most commonly studied AI interventions. These were primarily implemented for diagnosis, detection, or surveillance purposes. Neural networks (ie, convolutional neural networks and abductive networks) demonstrated the highest accuracy, considering the given database for the given clinical task. The risk of bias in diagnosis or prognosis studies was the lowest in the participant category (4/49, 4%) and the highest in the outcome category (22/49, 45%).

CONCLUSIONS

We observed variabilities in reporting the participants, types of AI methods, analyses, and outcomes, and highlighted the large gap in the effective development and implementation of AI in CBPHC. Further studies are needed to efficiently guide the development and implementation of AI interventions in CBPHC settings.

The Potential for Health Information Technology Tools to Reduce Racial Disparities in Maternal Morbidity and Mortality

Health information technology (health IT) potentially is a promising vital lever to address racial and ethnic, socioeconomic, and geographic disparities in maternal morbidity and mortality (MMM). This is especially relevant given that approximately 60% of maternal deaths are considered preventable.1-36 Interventions that leverage health IT tools to target the underlying drivers of disparities at the patient, clinician, and health care system levels potentially could reduce disparities in quality of care throughout the continuum (antepartum, intrapartum, and postpartum) of maternity care. This article presents an overview of the research (and gaps) on the potential of health IT tools to document SDoH and community-level geocoded data in EHR-based CDS systems, minimize implicit bias, and improve adherence to clinical guidelines and coordinated care to inform multilevel (patient, clinician, system) interventions throughout the continuum of maternity care for health disparity populations impacted by MMM. Telemedicine models for improving access in rural areas and new technologies for risk assessment and disease management (e.g., regarding preeclampsia) also are discussed.

Predicting neonatal respiratory distress syndrome and hypoglycaemia prior to discharge: Leveraging health administrative data and machine learning

OBJECTIVES

A major challenge for hospitals and clinicians is the early identification of neonates at risk of developing adverse conditions. We develop a model based on routinely collected administrative data, which accurately predicts two common disorders among early term and preterm (<39 weeks) neonates prior to discharge.

STUDY DESIGN

The data included all inpatient live births born prior to 39 weeks (n = 154,755) occurring in the Australian state of Queensland between January 2009 and December 2015. Predictor variables included all maternal data captured in administrative records from the beginning of gestation up to, and including, the delivery, as well as neonatal data recorded at the delivery. Gradient boosted trees were used to predict neonatal respiratory distress syndrome and hypoglycaemia prior to discharge, with model performance benchmarked against a logistic regression models.

RESULTS

The gradient boosted trees model achieved very high discrimination for respiratory distress syndrome [AUC = 0.923, 95% CI (0.917, 0.928)] and good discrimination for hypoglycaemia [AUC = 0.832, 95% CI (0.827, 0.837)] in the validation data, as well as outperforming the logistic regression models.

CONCLUSION

Our study suggests that routinely collected health data have the potential to play an important role in assisting clinicians to identify neonates at risk of developing selected disorders shortly after birth. Despite achieving high levels of discrimination, many issues remain before such models can be implemented in practice, which we discuss in relation to our findings.

An Expanded Obstetric Comorbidity Scoring System for Predicting Severe Maternal Morbidity

OBJECTIVE

To develop and validate an expanded obstetric comorbidity score for predicting severe maternal morbidity that can be applied consistently across contemporary U.S. patient discharge data sets.

METHODS

Discharge data from birth hospitalizations in California during 2016-2017 were used to develop the score. The outcomes were severe maternal morbidity, defined using the Centers for Disease Control and Prevention index, and nontransfusion severe maternal morbidity (excluding cases where transfusion was the only indicator of severe maternal morbidity). We selected 27 potential patient-level risk factors for severe maternal morbidity, identified using International Classification of Diseases, Tenth Revision, Clinical Modification diagnosis codes. We used a targeted causal inference approach integrated with machine learning to rank the risk factors based on adjusted risk ratios (aRRs). We used these results to assign scores to each comorbidity, which sum to a single numeric score. We validated the score in California and national data sets and compared the performance to that of a previously developed obstetric comorbidity index.

RESULTS

Among 919,546 births, the rates of severe maternal morbidity and nontransfusion severe maternal morbidity were 168 and 74 per 10,000 births, respectively. The highest risk comorbidity was placenta accreta spectrum (aRR of 30.5 for severe maternal morbidity and 54.7 for nontransfusion severe maternal morbidity) and the lowest was gestational diabetes mellitus (aRR of 1.06 for severe maternal morbidity and 1.12 for nontransfusion severe maternal morbidity). Normalized scores based on the aRR were developed for each comorbidity, which ranged from 1 to 59 points for severe maternal morbidity and from 1 to 36 points for nontransfusion severe maternal morbidity. The overall performance of the expanded comorbidity scores was good (C-statistics were 0.78 for severe maternal morbidity and 0.84 for nontransfusion severe maternal morbidity in California data and 0.82 and 0.87, respectively, in national data) and improved on prior comorbidity indices developed for obstetric populations. Calibration plots showed good concordance between predicted and actual risks of the outcomes.

CONCLUSION

We developed and validated an expanded obstetric comorbidity score to improve comparisons of severe maternal morbidity rates across patient populations with different comorbidity case mixes.

Predicting postpartum psychiatric admission using a machine learning approach

AIMS

The accurate identification of mothers at risk of postpartum psychiatric admission would allow for preventive intervention or more timely admission. We developed a prediction model to identify women at risk of postpartum psychiatric admission.

METHODS

Data included administrative health data of all inpatient live births in the Australian state of Queensland between January 2009 and October 2014. Analyses were restricted to mothers with one or more indicator of mental health problems during pregnancy (n = 75,054 births). The predictors included all maternal data up to and including the delivery, and neonatal data recorded at delivery. We used multiple machine learning methods to predict hospital admission in the 12 months following delivery in which the primary diagnosis was recorded as an ICD-10 psychotic, bipolar or depressive disorders.

RESULTS

The boosted trees algorithm produced the best performing model, predicting postpartum psychiatric admission in the validation data with good discrimination [AUC = 0.80; 95% CI = (0.76, 0.83)] and achieving good calibration. This model outperformed benchmark logistic regression model and an elastic net model. In addition to indicators of maternal metal health history, maternal and neonatal anthropometric measures and social/lifestyle factors were strong predictors.

CONCLUSION

Our results indicate the potential of a big data approach when aiming to identify mothers at risk of postpartum psychiatric admission. Mothers at risk could be followed-up and supported after neonatal discharge to either remove the need for admission or facilitate more timely admission.

Machine Learning and Statistical Models to Predict Postpartum Hemorrhage

OBJECTIVE

To predict a woman's risk of postpartum hemorrhage at labor admission using machine learning and statistical models.

METHODS

Predictive models were constructed and compared using data from 10 of 12 sites in the U.S. Consortium for Safe Labor Study (2002-2008) that consistently reported estimated blood loss at delivery. The outcome was postpartum hemorrhage, defined as an estimated blood loss at least 1,000 mL. Fifty-five candidate risk factors routinely available on labor admission were considered. We used logistic regression with and without lasso regularization (lasso regression) as the two statistical models, and random forest and extreme gradient boosting as the two machine learning models to predict postpartum hemorrhage. Model performance was measured by C statistics (ie, concordance index), calibration, and decision curves. Models were constructed from the first phase (2002-2006) and externally validated (ie, temporally) in the second phase (2007-2008). Further validation was performed combining both temporal and site-specific validation.

RESULTS

Of the 152,279 assessed births, 7,279 (4.8%, 95% CI 4.7-4.9) had postpartum hemorrhage. All models had good-to-excellent discrimination. The extreme gradient boosting model had the best discriminative ability to predict postpartum hemorrhage (C statistic: 0.93; 95% CI 0.92-0.93), followed by random forest (C statistic: 0.92; 95% CI 0.91-0.92). The lasso regression model (C statistic: 0.87; 95% CI 0.86-0.88) and logistic regression (C statistic: 0.87; 95% CI 0.86-0.87) had lower-but-good discriminative ability. The above results held with validation across both time and sites. Decision curve analysis demonstrated that, although all models provided superior net benefit when clinical decision thresholds were between 0% and 80% predicted risk, the extreme gradient boosting model provided the greatest net benefit.

CONCLUSION

Postpartum hemorrhage on labor admission can be predicted with excellent discriminative ability using machine learning and statistical models. Further clinical application is needed, which may assist health care providers to be prepared and triage at-risk women.

Machine learning in infection management using routine electronic health records: tools, techniques, and reporting of future technologies

BACKGROUND

Machine learning (ML) is increasingly being used in many areas of health care. Its use in infection management is catching up as identified in a recent review in this journal. We present here a complementary review to this work.

OBJECTIVES

To support clinicians and researchers in navigating through the methodological aspects of ML approaches in the field of infection management.

SOURCES

A Medline search was performed with the keywords artificial intelligence, machine learning, infection∗, and infectious disease∗ for the years 2014-2019. Studies using routinely available electronic hospital record data from an inpatient setting with a focus on bacterial and fungal infections were included.

CONTENT

Fifty-two studies were included and divided into six groups based on their focus. These studies covered detection/prediction of sepsis (n = 19), hospital-acquired infections (n = 11), surgical site infections and other postoperative infections (n = 11), microbiological test results (n = 4), infections in general (n = 2), musculoskeletal infections (n = 2), and other topics (urinary tract infections, deep fungal infections, antimicrobial prescriptions; n = 1 each). In total, 35 different ML techniques were used. Logistic regression was applied in 18 studies followed by random forest, support vector machines, and artificial neural networks in 18, 12, and seven studies, respectively. Overall, the studies were very heterogeneous in their approach and their reporting. Detailed information on data handling and software code was often missing. Validation on new datasets and/or in other institutions was rarely done. Clinical studies on the impact of ML in infection management were lacking.

IMPLICATIONS

Promising approaches for ML use in infectious diseases were identified. But building trust in these new technologies will require improved reporting. Explainability and interpretability of the models used were rarely addressed and should be further explored. Independent model validation and clinical studies evaluating the added value of ML approaches are needed.