Prediction of low Apgar score at five minutes following labor induction intervention in vaginal deliveries: machine learning approach for imbalanced data at a tertiary hospital in North Tanzania.

BACKGROUND: Prediction of low Apgar score for vaginal deliveries following labor induction intervention is critical for improving neonatal health outcomes. We set out to investigate important attributes and train popular machine learning (ML) algorithms to correctly classify neonates with a low Apgar scores from an imbalanced learning perspective. METHODS: We analyzed 7716 induced vaginal deliveries from the electronic birth registry of the Kilimanjaro Christian Medical Centre (KCMC). 733 (9.5%) of which constituted of low (< 7) Apgar score neonates. The 'extra-tree classifier' was used to assess features' importance. We used Area Under Curve (AUC), recall, precision, F-score, Matthews Correlation Coefficient (MCC), balanced accuracy (BA), bookmaker informedness (BM), and markedness (MK) to evaluate the performance of the selected six (6) machine learning classifiers. To address class imbalances, we examined three widely used resampling techniques: the Synthetic Minority Oversampling Technique (SMOTE) and Random Oversampling Examples (ROS) and Random undersampling techniques (RUS). We applied Decision Curve Analysis (DCA) to evaluate the net benefit of the selected classifiers. RESULTS: Birth weight, maternal age, and gestational age were found to be important predictors for the low Apgar score following induced vaginal delivery. SMOTE, ROS and and RUS techniques were more effective at improving "recalls" among other metrics in all the models under investigation. A slight improvement was observed in the F1 score, BA, and BM. DCA revealed potential benefits of applying Boosting method for predicting low Apgar scores among the tested models. CONCLUSION: There is an opportunity for more algorithms to be tested to come up with theoretical guidance on more effective rebalancing techniques suitable for this particular imbalanced ratio. Future research should prioritize a debate on which performance indicators to look up to when dealing with imbalanced or skewed data.

The Association Between Substance Use and Violence: Results from a Nationally Representative Sample of High School Students in the United States.

With over 52% of high school students reporting that they have tried alcohol or illicit drugs, 16% carrying a weapon, and 23% engaging in a physical fight, substance use and youth violence remain critical public health challenges in the United States. Using data from the 2017 Youth Risk Behavior Survey, study results revealed that youth who reported heavy use of either alcohol, marijuana, or illicit drugs were three to ten times more likely to report carrying a weapon or engaging in a physical fight. Similarly, youth with heavy substance use were one and half times to 14 times more likely to be a victim of violence or sexual or dating violence. The SEM analysis indicated that substance use had a significant effect on all aspects of violence. School-based behavioral health specialists and community-based pediatricians may need to develop targeted messages to address the potential for violence among youth who use alcohol and/or illicit drugs.

Self-Check-In Kiosks Utilization and Their Association With Wait Times in Emergency Departments in the United States.

BACKGROUND: Delayed care in emergency departments (EDs) is a serious problem in the United States. Patient wait time is considered a critical measure of delayed care in EDs. Several strategies have been employed by EDs to reduce wait time, including implementation of self-check-in kiosks. However, the effect of kiosks on wait time in EDs is understudied. OBJECTIVES: To assess the association between patient wait time and utilization of self-check-in kiosks in EDs. To investigate a series of other patient-, ED-, and hospital-level predictors of wait time in EDs. METHODS: Using data from the 2015 and 2016 National Hospital Ambulatory Medical Care Survey, we analyzed 40,528 ED visits by constructing a multivariable linear regression model of the log-transformed wait time data as an outcome, then computing percent changes in wait times. RESULTS: During the study period, about 9% of EDs in the United States implemented kiosks. In our linear regression model, the wait time in EDs with kiosk self-check-in services was 56.8% shorter (95% confidence interval ̶ 130% to ̶ 6.4%, p < 0.05) compared with EDs without kiosk services. In addition to kiosks, patients' day of visit, arrival time, triage assessment, arrival by ambulance, chronic medical conditions, ED boarding, hospitals' full-capacity protocol, and hospitals' location were significant predictors of wait time. CONCLUSIONS: Self-check-in kiosks are associated with shorter ED wait time in the United States. However, prolonged ED wait time continues to be a system-wide problem, and warrants multilayered interventions to address this challenge for those who are in acute need of immediate care.

Quality of Novel Coronavirus Related Health Information over the Internet: An Evaluation Study.

BACKGROUND: The novel coronavirus disease (COVID-19) has spread globally from its epicenter in Hubei, China, and was declared a pandemic by the World Health Organization (WHO) on March 11, 2020. The most popular search engine worldwide is Google, and since March 2020, COVID-19 has been a global trending search term. Misinformation related to COVID-19 from these searches is a problem, and hence, it is of high importance to assess the quality of health information over the internet related to COVID-19. The objective of our study is to examine the quality of COVID-19 related health information over the internet using the DISCERN tool. METHODS: The keywords included in assessment of COVID-19 related information using Google's search engine were "Coronavirus," "Coronavirus causes," "Coronavirus diagnosis," "Coronavirus prevention," and "Coronavirus management". The first 20 websites from each search term were gathered to generate a list of 100 URLs. Duplicate sites were excluded from this search, allowing analysis of unique sites only. Additional exclusion criteria included scientific journals, nonoperational links, nonfunctional websites (where the page was not loading, was not found, or was inactive), and websites in languages other than English. This resulted in a unique list of 48 websites. Four independent raters evaluated the websites using a 16-item DISCERN tool to assess the quality of novel coronavirus related information available on the internet. The interrater reliability agreement was calculated using the intracluster correlation coefficient. RESULTS: Results showed variation in how the raters assigned scores to different website categories. The .com websites received the lowest scores. Results showed that .edu and .org website category sites were excellent in communicating coronavirus related health information; however, they received lower scores for treatment effect and treatment choices. CONCLUSION: This study highlights the gaps in the quality of information that is available on the websites related to COVID-19 and study emphasizes the need for verified websites that provide evidence-based health information related to the novel coronavirus pandemic.

Mobile Health Application and e-Health Literacy: Opportunities and Concerns for Cancer Patients and Caregivers.

Health literacy is critical for cancer patients as they must understand complex procedures or treatment options. Caregivers' health literacy also plays a crucial role in caring for cancer patients. Low health literacy is associated with low adherence to medications, poor health status, and increased health care costs. There is a growing interest in the use of mobile health applications (apps) to improve health literacy. Mobile health apps can empower underserved cancer patients and their caregivers by providing features or functionalities to enhance interactive patient-provider communication and to understand medical information more readily. Despite the potentiality of improving health literacy through mobile health apps, there exist several related concerns: no equal access to mobile technology, no familiarity or knowledge of using mobile health apps, and privacy and security concerns. These elements should be taken into account for health policy making and mobile apps design and development. Importantly, mobile apps should be developed with the goal of achieving a high range of user access by considering all health literacy level and various cultural and linguistic needs.

Provision of Rehabilitation Services in Residential Care Facilities: Evidence From a National Survey.

OBJECTIVE: To examine the association between organizational factors and provision of rehabilitation services that include physical therapy (PT) and occupational therapy (OT) in residential care facilities (RCFs) in the United States. DESIGN: A cross-sectional, observational study conducted using a national sample from the 2010 National Survey of Residential Care Facilities conducted by the Centers for Disease Control and Prevention's National Center for Health Statistics. SETTINGS: U.S. RCFs. PARTICIPANTS: RCFs (N=2302; weighted sample, 31,134 RCFs). INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: The association between characteristics of the facilities, director and staff, and residents, and provision of PT and OT services was assessed using multivariate logistic regression analyses. RESULTS: Among all RCFs in the United States, 43.9% provided PT and 40.0% provided OT. Medicaid-certified RCFs, larger-sized RCFs, RCFs with a licensed director, RCFs that used volunteers, and RCFs with higher personal care aide hours per patient per day were more likely to provide both PT and OT, while private, for-profit RCFs were less likely to provide PT and OT. RCFs with a higher percentage of white residents were more likely to provide PT, while RCFs with chain affiliation were more likely to provide OT. CONCLUSIONS: Less than half of the RCFs in the United States provide PT and OT, and this provision of therapy services is associated with organizational characteristics of the facilities. Future research should explore the effectiveness of rehabilitation services in RCFs on residents' health outcomes.

Seeking Medical Information Using Mobile Apps and the Internet: Are Family Caregivers Different from the General Public?

Family caregivers play an important role to care cancer patients since they exchange medical information with health care providers. However, relatively little is known about how family caregivers seek medical information using mobile apps and the Internet. We examined factors associated with medical information seeking by using mobile apps and the Internet among family caregivers and the general public using data from the 2014 Health Information National Trends Survey 4 Cycle 1. The study sample consisted of 2425 family caregivers and 1252 non-family caregivers (the general public). Guided by Comprehensive Model of Information Seeking (CMIS), we examined related factors' impact on two outcome variables for medical information seeking: mobile apps use and Internet use with multivariate logistic regression analyses. We found that online medical information seeking is different between family caregivers and the general public. Overall, the use of the Internet for medical information seeking is more common among family caregivers, while the use of mobile apps is less common among family caregivers compared with the general public. Married family caregivers were less likely to use mobile apps, while family caregivers who would trust cancer information were more likely to use the Internet for medical information seeking as compared to the general public. Medical information seeking behavior among family caregivers can be an important predictor of both their health and the health of their cancer patients. Future research should explore the low usage of mobile health applications among family caregiver population.

Promoting colorectal cancer screening through a new model of delivering rural primary care in the USA: a qualitative study.

INTRODUCTION: Despite the known benefits of colorectal cancer (CRC) screening, rural areas have consistently reported lower screening rates than their urban counterparts. Alternative healthcare delivery models, such as accountable care organizations (ACOs), have the potential to increase CRC rates through collaboration among healthcare providers with the aim of improving quality and decreasing cost. However, researchers have not sufficiently explored how this innovative model could influence the promotion of cancer screening. The purpose of the study was to explore the mechanism of how CRC screening can be promoted in ACO-participating rural primary care clinics. METHODS: The study collected qualitative data from in-depth interviews with 21 healthcare professionals employed in ACO-participating primary care clinics in rural Nebraska. Participants were asked about their views on opportunities and challenges to promote CRC screening in an ACO context. Data were analyzed using a grounded theory approach. RESULTS: The study found that the new healthcare delivery model can offer opportunities to promote cancer screening in rural areas through enhanced electronic health record use, information sharing and collaborative learning within ACO networks, use of standardized quality measures and performance feedback, a shift to preventive/comprehensive care, adoption of team-based care, and empowered care coordinators. The perceived challenges were found in financial instability, increased staff workload, lack of provider training/education, and lack of resources in rural areas. CONCLUSIONS: This study found that the innovative care delivery model, ACO, could provide a well-designed platform for promoting CRC screening in rural areas, if sustainable resources (eg finance, health providers, and education) are provided. This study provides 'practical' information to identify effective and sustainable intervention programs to promote preventive screening. Further efforts are needed to facilitate delivery system reforms in rural primary care, such as improving performance evaluation measures and methods.

Paid sick leave is associated with fewer ED visits among US private sector working adults.

CONTEXT: The United States (US) is the only developed country that does not guarantee short-term or longer-term paid sick leave. OBJECTIVE: This study used a multiyear nationally representative database to examine the association between availability of paid sick leave and frequency of emergency department (ED) use among US private sector employees. STUDY SAMPLE: We used the National Health Interview Survey data (2012-2014). The final study sample consists of 42,460 US adults between 18 and 64years of age and working in nongovernmental private sector. RESULTS: Our results suggest that availability of paid sick leave is significantly associated with lower likelihood of ED use, for both moderate (1-3 times/year) and repeated users (4 or more times/year). After controlling for confounding factors, respondents with paid sick leave are 14% less likely to be moderate ED users (adjusted odds ratio, 0.86; 95% CI, 0.79-0.93) and 32% less likely to be repeated ED users (adjusted odds ratio, 0.68; 95% CI, 0.50-0.91). DISCUSSION: Although expansion of health insurance coverage under the Affordable Care Act has not been shown to reduce utilization of high cost health care services such as the ED, our study suggests other factors such as the availability of paid sick leave may do so, by allowing patients to seek care through other more cost-effective mechanisms (eg, primary care providers). To reduce ED utilization, health policymakers should consider alternative reforms including paid sick leave.

Use of Mobile Health Applications for Health-Seeking Behavior Among US Adults.

This study explores the use of mobile health applications (mHealth apps) on smartphones or tablets for health-seeking behavior among US adults. Data was obtained from cycle 4 of the 4th edition of the Health Information National Trends Survey (HINTS 4). Weighted multivariate logistic regression models examined predictors of 1) having mHealth apps, 2) usefulness of mHealth apps in achieving health behavior goals, 3) helpfulness in medical care decision-making, and 4) asking a physician new questions or seeking a second opinion. Using the Andersen Model of health services utilization, independent variables of interest were grouped under predisposing factors (age, gender, race, ethnicity, and marital status), enabling factors (education, employment, income, regular provider, health insurance, and rural/urban location of residence), and need factors (general health, confidence in their ability to take care of health, Body Mass Index, smoking status, and number of comorbidities). In a national sample of adults who had smartphones or tablets, 36 % had mHealth apps on their devices. Among those with apps, 60 % reported the usefulness of mHealth apps in achieving health behavior goals, 35 % reported their helpfulness for medical care decision-making, and 38 % reported their usefulness in asking their physicians new questions or seeking a second opinion. The multivariate models revealed that respondents were more likely to have mHealth apps if they had more education, health insurance, were confident in their ability to take good care of themselves, or had comorbidities, and were less likely to have them if they were older, had higher income, or lived in rural areas. In terms of usefulness of mHealth apps, those who were older and had higher income were less likely to report their usefulness in achieving health behavior goals. Those who were older, African American, and had confidence in their ability to take care of their health were more likely to respond that the mHealth apps were helpful in making a medical care decision and asking their physicians new questions or for a second opinion. Potentially, mHealth apps may reduce the burden on primary care, reduce costs, and improve the quality of care. However, several personal-level factors were associated with having mHealth apps and their perceived helpfulness among their users, indicating a multidimensional digital divide in the population of US adults.

Use of information technology for medication management in residential care facilities: correlates of facility characteristics.

The effectiveness of information technology in resolving medication problems has been well documented. Long-term care settings such as residential care facilities (RCFs) may see the benefits of using such technologies in addressing the problem of medication errors among their resident population, who are usually older and have numerous chronic conditions. The aim of this study was two-fold: to examine the extent of use of Electronic Medication Management (EMM) in RCFs and to analyze the organizational factors associated with the use of EMM functionalities in RCFs. Data on RCFs were obtained from the 2010 National Survey of Residential Care Facilities. The association between facility, director and staff, and resident characteristics of RCFs and adoption of four EMM functionalities was assessed through multivariate logistic regression. The four EMM functionalities included were maintaining lists of medications, ordering for prescriptions, maintaining active medication allergy lists, and warning of drug interactions or contraindications. About 12% of the RCFs adopted all four EMM functionalities. Additionally, maintaining lists of medications had the highest adoption rate (34.5%), followed by maintaining active medication allergy lists (31.6%), ordering for prescriptions (19.7%), and warning of drug interactions or contraindications (17.9%). Facility size and ownership status were significantly associated with adoption of all four EMM functionalities. Medicaid certification status, facility director's age, education and license status, and the use of personal care aides in the RCF were significantly associated with the adoption of some of the EMM functionalities. EMM is expected to improve the quality of care and patient safety in long-term care facilities including RCFs. The extent of adoption of the four EMM functionalities is relatively low in RCFs. Some RCFs may strategize to use these functionalities to cater to the increasing demands from the market and also to provide better quality of care.

"Urban-Satellite" estimates in the ABCD Study: Linking Neuroimaging and Mental Health to Satellite Imagery Measurements of Macro Environmental Factors.

While numerous studies over the last decade have highlighted the important influence of environmental factors on mental health, globally applicable data on physical surroundings are still limited. Access to such data and the possibility to link them to epidemiological studies is critical to unlocking the relationship of environment, brain and behaviour and promoting positive future mental health outcomes. The Adolescent Brain Cognitive Development (ABCD) Study is the largest ongoing longitudinal and observational study exploring brain development and child health among children from 21 sites across the United States. Here we describe the linking of the ABCD study data with satellite-based "Urban-Satellite" (UrbanSat) variables consisting of 11 satellite-data derived environmental indicators associated with each subject's residential address at their baseline visit, including land cover and land use, nighttime lights, and population characteristics. We present these UrbanSat variables and provide a review of the current literature that links environmental indicators with mental health, as well as key aspects that must be considered when using satellite data for mental health research. We also highlight and discuss significant links of the satellite data variables to the default mode network clustering coefficient and cognition. This comprehensive dataset provides the foundation for large-scale environmental epidemiology research.

Association of death or illness from COVID-19 among family and friends on vaccine uptake within four months of the Emergency Use Authorization. Findings from a national survey in the United States.

OBJECTIVE: To examine the relationship between knowing that a friend or family member became ill with, or died from, COVID-19 and receiving a vaccine dose within four months of the FDA's Emergency Use Authorization. METHODS: A national sample of 1,517 respondents were surveyed from April 7 to April 12, 2021, 1,193 of whom were eligible for the vaccine when the data were collected. RESULTS: Respondents who knew someone who became ill with COVID-19 (AOR = 2.32, 95 % CI 1.74-3.09) or knew someone who died (AOR = 2.29, 95 % CI 1.32-3.99) from COVID-19 were more likely to receive at least a single COVID-19 vaccine dose. CONCLUSION: Encouraging people to share their COVID-19 illness and bereavement experiences with their local network such as friends, families, social-networks and via social media might help increase vaccine uptake.

A systematic review for the impacts of global approaches to regulating electronic nicotine products.

Background: The rapid increase in electronic nicotine product (ENP) use among young people has been a global public health challenge, given the potential harm of ENPs and nicotine dependence. Many countries have recently introduced legislations to regulate ENPs, but the impacts of these policies are poorly understood. This systematic review aims to critically synthesise empirical studies on the effects of global regulations regarding ENPs on the prevalence of use, health outcomes and their determinants, using the 4A marketing mix framework (acceptability, affordability, accessibility and awareness). Methods: Following the PRISMA guideline, we searched PubMed, Embase, Scopus, Web of Science, Academic Search Complete, Business Source Complete, and APA PsycINFO databases from inception until June 14, 2022 and performed citation searches on the included studies. Reviewed literature was restricted to peer-reviewed, English-language articles. We included all pre-post and quasi-experimental studies that evaluated the impacts of e-cigarette policies on the prevalence of ENP use and other health outcomes. A modified Joanna Briggs Institute (JBI) Critical Appraisal checklist for quasi-experimental studies was used for quality assessment. Due to heterogeneity of the included studies, we conducted a narrative synthesis of evidence. Results: Of 3991 unduplicated records screened, 48 (1.2%) met the inclusion criteria, most were from high-income countries in North America and Europe and 26 studies measured self-reported ENPs use. Flavour restrictions significantly decreased youth ENP use and taxation reduced adult use; mixed results were found for the impacts of age restrictions. Indoor vaping restrictions and the European Tobacco Products Directive (TPD) did not seem to reduce ENP use based on existing studies. Changes in determinants such as sales and perceptions corroborated our conclusions. Few studies assessed the impacts of other regulations such as advertising restrictions and retail licensing requirements. Conclusions: Flavour restrictions and taxes have the strongest evidence to support effective control of ENPs, while others need powerful enforcement and meaningful penalties to ensure their effectiveness. Future research should focus on under-examined policies and differential impacts across sociodemographic characteristics and countries. Registration: PROSPERO CRD42022337361.

Use of Assistive Technology for Persons with Psychosocial Disability: Systematic Review.

BACKGROUND: Assistive technology (AT) refers to assistive products (AP) and associated systems and services that are relevant for function, independence, well-being, and quality of life for individuals with disabilities. There is a high unmet need for AT for persons with disabilities and this is worse for persons with cognitive and mental or psychosocial disabilities (PDs). Further, information and knowledge on AT for PDs is limited. OBJECTIVE: The aim of this review was to explore the pattern of AT use among persons with PDs and its associated socioeconomic and health benefits. METHODS: The review was reported according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses), and we conducted systematic searches in the 4 databases: PubMed, Embase.com, APA PsycInfo (Ebsco), and Web of Science (Core Collection) with the following index terms: "Assistive Technology," "Self-Help Devices," "Quality of Life," "Activities of Daily Living," "Mental Disorders." We included only AT individuals with PDs can independently use without reliance on a provider. Identified papers were exported to EndNote (Clarivate) and we undertook a narrative synthesis of the included studies. RESULTS: In total, 5 studies were included in the review which reported use of different AT for schizophrenia, bipolar disorder, depression and anxiety disorders. The APs described in the included studies are Palm tungsten T3 handheld computer, MOBUS, personal digital assistant, automated pill cap, weighted chain blankets, and smartphone function. All the AT products identified in the studies were found to be easily usable by individuals with PDs. The APs reported in the included studies have broad impact and influence on social function, productivity, and treatment or management. The studies were heterogeneous and were all conducted in high-income countries. CONCLUSIONS: Our study contributes to and strengthens existing evidence on the relevance of AT for PDs and its potential to support socioeconomic participation and health. Although AT has the potential to improve function and participation for individuals with PDs; this review highlights that research on the subject is limited. Further research and health policy changes are needed to improve research and AT service provision for individuals with PDs especially in low-income settings. TRIAL REGISTRATION: PROSPERO CRD42022343735; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=343735.

Combining Resampling Strategies and Ensemble Machine Learning Methods to Enhance Prediction of Neonates with a Low Apgar Score After Induction of Labor in Northern Tanzania.

OBJECTIVE: The goal of this study was to establish the most efficient boosting method in predicting neonatal low Apgar scores following labor induction intervention and to assess whether resampling strategies would improve the predictive performance of the selected boosting algorithms. METHODS: A total of 7716 singleton births delivered from 2000 to 2015 were analyzed. Cesarean deliveries following labor induction, deliveries with abnormal presentation, and deliveries with missing Apgar score or delivery mode information were excluded. We examined the effect of resampling approaches or data preprocessing on predicting low Apgar scores, specifically the synthetic minority oversampling technique (SMOTE), borderline-SMOTE, and the random undersampling (RUS) technique. Sensitivity, specificity, precision, area under receiver operating curve (AUROC), F-score, positive predicted values (PPV), negative predicted values (NPV) and accuracy of the three (3) boosting-based ensemble methods were used to evaluate their discriminative ability. The ensemble learning models tested include adoptive boosting (AdaBoost), gradient boosting (GB) and extreme gradient boosting method (XGBoost). RESULTS: The prevalence of low (<7) Apgar scores was 9.5% (n = 733). The prediction models performed nearly similar in their baseline mode. Following the application of resampling techniques, borderline-SMOTE significantly improved the predictive performance of all the boosting-based ensemble methods under observation in terms of sensitivity, F1-score, AUROC and PPV. CONCLUSION: Policymakers, healthcare informaticians and neonatologists should consider implementing data preprocessing strategies when predicting a neonatal outcome with imbalanced data to enhance efficiency. The process may be more effective when borderline-SMOTE technique is deployed on the selected ensemble classifiers. However, future research may focus on testing additional resampling techniques, performing feature engineering, variable selection and optimizing further the ensemble learning hyperparameters.

Active commuting and the risk of obesity, hypertension and diabetes: a systematic review and meta-analysis of observational studies.

Active commuting may hold a potential for preventing adverse health outcomes. However, evidence of the association of active commuting and the risk of health outcomes remains debatable. The current study systematically and quantitatively summarised research findings on the association between active commuting and the risk of the mentioned health outcomes. We comprehensively searched four databases (PubMed, EMBASE, Web of Science and Open Grey) from inception to 2 August 2020 for observational studies investigating the associations among adult population. Summary relative risks (RRs) and 95% CIs were estimated for the association. Heterogeneity was investigated using Cochran's Q test and the I 2 statistic. Restricted cubic splines were used to evaluate linear and nonlinear relations. The search yielded 7581 initial references. We included 28 articles in the meta-analysis. Compared with inactive commuting, active commuting reduced the risk of obesity (RR=0.88, 95% CI 0.83 to 0.94, I2=69.1%), hypertension (RR=0.95, 95% CI 0.87 to 1.04, I2=82.2%) and diabetes (RR=0.82, 95% CI 0.76 to 0.90, I2=44.5%). Restricted cubic splines showed linear associations between active commuting and obesity, hypertension and diabetes (Pnonlinearity=0.640; Pnonlinearity=0.886; Pnonlinearity=0.099). As compared with the lowest active commuting group, the risk of obesity, hypertension and diabetes in the highest active commuting group were reduced by 13% (95% CI 0.82 to 0.93, I2=65.2%); 6% (95% CI 0.86 to 1.02, I2=75.2%) and 19% (95% CI 0.73 to 0.91, I2=49.8%) respectively. Active commuting seemed to be associated with lower risk of obesity, hypertension and diabetes. However, the results should be interpreted cautiously because this meta-analysis was based solely on observational studies.PROSPERO registration numberCRD42020202723.

Validating machine learning models for the prediction of labour induction intervention using routine data: a registry-based retrospective cohort study at a tertiary hospital in northern Tanzania.

OBJECTIVES: We aimed at identifying the important variables for labour induction intervention and assessing the predictive performance of machine learning algorithms. SETTING: We analysed the birth registry data from a referral hospital in northern Tanzania. Since July 2000, every birth at this facility has been recorded in a specific database. PARTICIPANTS: 21 578 deliveries between 2000 and 2015 were included. Deliveries that lacked information regarding the labour induction status were excluded. PRIMARY OUTCOME: Deliveries involving labour induction intervention. RESULTS: Parity, maternal age, body mass index, gestational age and birth weight were all found to be important predictors of labour induction. Boosting method demonstrated the best discriminative performance (area under curve, AUC=0.75: 95% CI (0.73 to 0.76)) while logistic regression presented the least (AUC=0.71: 95% CI (0.70 to 0.73)). Random forest and boosting algorithms showed the highest net-benefits as per the decision curve analysis. CONCLUSION: All of the machine learning algorithms performed well in predicting the likelihood of labour induction intervention. Further optimisation of these classifiers through hyperparameter tuning may result in an improved performance. Extensive research into the performance of other classifier algorithms is warranted.