Dataset Information

Prediction of perinatal death using machine learning models: a birth registry-based cohort study in northern Tanzania.

ABSTRACT:

Objective

We aimed to determine the key predictors of perinatal deaths using machine learning models compared with the logistic regression model.

Design

A secondary data analysis using the Kilimanjaro Christian Medical Centre (KCMC) Medical Birth Registry cohort from 2000 to 2015. We assessed the discriminative ability of models using the area under the receiver operating characteristics curve (AUC) and the net benefit using decision curve analysis.

Setting

The KCMC is a zonal referral hospital located in Moshi Municipality, Kilimanjaro region, Northern Tanzania. The Medical Birth Registry is within the hospital grounds at the Reproductive and Child Health Centre.

Participants

Singleton deliveries (n=42 319) with complete records from 2000 to 2015.

Primary outcome measures

Perinatal death (composite of stillbirths and early neonatal deaths). These outcomes were only captured before mothers were discharged from the hospital.

Results

The proportion of perinatal deaths was 3.7%. There were no statistically significant differences in the predictive performance of four machine learning models except for bagging, which had a significantly lower performance (AUC 0.76, 95% CI 0.74 to 0.79, p=0.006) compared with the logistic regression model (AUC 0.78, 95% CI 0.76 to 0.81). However, in the decision curve analysis, the machine learning models had a higher net benefit (ie, the correct classification of perinatal deaths considering a trade-off between false-negatives and false-positives)-over the logistic regression model across a range of threshold probability values.

Conclusions

In this cohort, there was no significant difference in the prediction of perinatal deaths between machine learning and logistic regression models, except for bagging. The machine learning models had a higher net benefit, as its predictive ability of perinatal death was considerably superior over the logistic regression model. The machine learning models, as demonstrated by our study, can be used to improve the prediction of perinatal deaths and triage for women at risk.

SUBMITTER: Mboya IB

PROVIDER: S-EPMC7574940 | biostudies-literature | 2020 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Prediction of perinatal death using machine learning models: a birth registry-based cohort study in northern Tanzania.

Mboya Innocent B IB Mahande Michael J MJ Mohammed Mohanad M Obure Joseph J Mwambi Henry G HG

BMJ open 20201019 10

<h4>Objective</h4>We aimed to determine the key predictors of perinatal deaths using machine learning models compared with the logistic regression model.<h4>Design</h4>A secondary data analysis using the Kilimanjaro Christian Medical Centre (KCMC) Medical Birth Registry cohort from 2000 to 2015. We assessed the discriminative ability of models using the area under the receiver operating characteristics curve (AUC) and the net benefit using decision curve analysis.<h4>Setting</h4>The KCMC is a zo ...[more]

PMID: 33077570

Similar Datasets

Project description:BACKGROUND:More than five million perinatal deaths occur each year globally. Despite efforts put forward during the millennium development goals era, perinatal deaths continue to increase relative to under-five deaths, especially in low- and middle-income countries. This study aimed to determine predictors of perinatal death in the presence of missing data using birth registry data from Kilimanjaro Christian Medical Center (KCMC), between 2000-2015. METHODS:This was a retrospective cohort study from the medical birth registry at KCMC referral hospital located in Moshi Municipality, Kilimanjaro region, northern Tanzania. Data were analyzed using Stata version 15.1. Multiple imputation by fully conditional specification (FCS) was used to impute missing values. Generalized estimating equations (GEE) were used to determine the marginal effects of covariates on perinatal death using a log link mean model with robust standard errors. An exchangeable correlation structure was used to account for the dependence of observations within mothers. RESULTS:Among 50,487 deliveries recorded in the KCMC medical birth registry between 2000-2015, 4.2% (95%CI 4.0%, 4.3%) ended in perinatal death (equivalent to a perinatal mortality rate (PMR) of 41.6 (95%CI 39.9, 43.3) deaths per 1,000 births). After the imputation of missing values, the proportion of perinatal death remained relatively the same. The risk of perinatal death was significantly higher among deliveries from mothers who resided in rural compared to urban areas (RR = 1.241, 95%CI 1.137, 1.355), with primary education level (RR = 1.201, 95%CI 1.083, 1.332) compared to higher education level, with <4 compared to ≥4 antenatal care (ANC) visits (RR = 1.250, 95%CI 1.146, 1.365), with postpartum hemorrhage (PPH) (RR = 2.638, 95%CI 1.997, 3.486), abruption placenta (RR = 4.218, 95%CI 3.438, 5.175), delivered a low birth weight baby (LBW) (RR = 4.210, 95%CI 3.788, 4.679), male child (RR = 1.090, 95%CI 1.007, 1.181), and were referred for delivery (RR = 2.108, 95%CI 1.919, 2.317). On the other hand, deliveries from mothers who experienced premature rupture of the membranes (PROM) (RR = 0.411, 95%CI 0.283, 0.598) and delivered through cesarean section (CS) (RR = 0.662, 95%CI 0.604, 0.724) had a lower risk of perinatal death. CONCLUSIONS:Perinatal mortality in this cohort is higher than the national estimate. Higher risk of perinatal death was associated with low maternal education level, rural residence, <4 ANC visits, PPH, abruption placenta, LBW delivery, child's sex, and being referred for delivery. Ignoring missing values in the analysis of adverse pregnancy outcomes produces biased covariate coefficients and standard errors. Close clinical follow-up of women at high risk of experiencing perinatal death, particularly during ANC visits and delivery, is of high importance to increase perinatal survival.

Project description:BackgroundPreterm birth is a significant contributor of under-five and newborn deaths globally. Recent estimates indicated that, Tanzania ranks the tenth country with the highest preterm birth rates in the world, and shares 2.2% of the global proportion of all preterm births. Previous studies applied binary regression models to determine predictors of preterm birth by collapsing gestational age at birth to <37 weeks. For targeted interventions, this study aimed to determine predictors of preterm birth using multinomial regression models accounting for missing data.MethodsWe carried out a secondary analysis of cohort data from the KCMC zonal referral hospital Medical Birth Registry for 44,117 women who gave birth to singletons between 2000-2015. KCMC is located in the Moshi Municipality, Kilimanjaro region, northern Tanzania. Data analysis was performed using Stata version 15.1. Assuming a nonmonotone pattern of missingness, data were imputed using a fully conditional specification (FCS) technique under the missing at random (MAR) assumption. Multinomial regression models with robust standard errors were used to determine predictors of moderately to late ([32,37) weeks of gestation) and very/extreme (<32 weeks of gestation) preterm birth.ResultsThe overall proportion of preterm births among singleton births was 11.7%. The trends of preterm birth were significantly rising between the years 2000-2015 by 22.2% (95%CI 12.2%, 32.1%, p<0.001) for moderately to late preterm and 4.6% (95%CI 2.2%, 7.0%, p = 0.001) for very/extremely preterm birth category. After imputation of missing values, higher odds of moderately to late preterm delivery were among adolescent mothers (OR = 1.23, 95%CI 1.09, 1.39), with primary education level (OR = 1.28, 95%CI 1.18, 1.39), referred for delivery (OR = 1.19, 95%CI 1.09, 1.29), with pre-eclampsia/eclampsia (OR = 1.77, 95%CI 1.54, 2.02), inadequate (<4) antenatal care (ANC) visits (OR = 2.55, 95%CI 2.37, 2.74), PROM (OR = 1.80, 95%CI 1.50, 2.17), abruption placenta (OR = 2.05, 95%CI 1.32, 3.18), placenta previa (OR = 4.35, 95%CI 2.58, 7.33), delivery through CS (OR = 1.16, 95%CI 1.08, 1.25), delivered LBW baby (OR = 8.08, 95%CI 7.46, 8.76), experienced perinatal death (OR = 2.09, 95%CI 1.83, 2.40), and delivered male children (OR = 1.11, 95%CI 1.04, 1.20). Maternal age, education level, abruption placenta, and CS delivery showed no statistically significant association with very/extremely preterm birth. The effect of (<4) ANC visits, placenta previa, LBW, and perinatal death were more pronounced on the very/extremely preterm compared to the moderately to late preterm birth. Notably, extremely higher odds of very/extreme preterm birth were among the LBW babies (OR = 38.34, 95%CI 31.87, 46.11).ConclusionsThe trends of preterm birth have increased over time in northern Tanzania. Policy decisions should intensify efforts to improve maternal and child care throughout the course of pregnancy and childbirth towards preterm birth prevention. For a positive pregnancy outcome, interventions to increase uptake and quality of ANC services should also be strengthened in Tanzania at all levels of care, where several interventions can easily be delivered to pregnant women, especially those at high-risk of experiencing adverse pregnancy outcomes.

Project description:BackgroundCaesarean delivery (CD) is the commonest obstetric surgery and surgical intervention to save lives of the mother and/or the new-borns. Despite been accepted as safe procedure, caesarean delivery has an increased risk of adverse maternal and fetal outcomes. The rising rate of caesarean delivery has been a major public health concern worldwide and the consequences that come along with it urgently need to be assessed, especially in resource limited settings. We aimed to examine the relationship between first birth caesarean delivery and adverse maternal and perinatal outcomes in the second pregnancy among women who delivered at a tertiary hospital in Northern Tanzania.MethodsA retrospective cohort study was conducted using maternally-linked data from Kilimanjaro Christian Medical Centre. All women who had singleton second delivery between the years 2011 to 2015 were studied. A total of 5,984 women with singleton second delivery were analysed. Multivariable log-binomial regression was used to determine the association between first caesarean delivery and maternal-perinatal outcomes in the second pregnancy.ResultsCaesarean delivery in the first birth was associated with an increased risk of adverse maternal and perinatal outcomes in the second pregnancy. These included repeated CD (ARR 1.19; 95% CI: 1.05-1.34), pre/eclampsia (ARR 1.38; 95% CI: 1.06-1.78), gestational diabetes mellitus (ARR 2.80; 95% CI: 1.07-7.36), uterine rupture (ARR 1.56; CI: 1.05-2.32), peri-partum hysterectomy (ARR 2.28; CI: 1.04-5.02) and preterm birth (ARR 1.21; CI: 1.05-1.38).ConclusionCaesarean delivery in their first pregnancy had an increased risk of repeated caesarean delivery and other adverse maternal-perinatal outcomes in the following pregnancy. Findings from this study highlight the importance of devising regional specific measures to mitigate unnecessary primary caesarean delivery. Additionally, these findings may help both clinicians and women in deciding against or for trial of labor after previous caesarean delivery in an event of absent direct obstetric indication.

Project description:IntroductionMaternal HIV infection is associated with increased risk of having a preterm delivery, low birth weight baby, small for gestational age baby and stillbirth. Maternal use of combination antiretroviral treatment is also associated with preterm delivery and low birth weight, although the effects vary by the type of drugs and timing of initiation.ObjectiveTo examine time trends in adverse perinatal outcomes among HIV-positive compared with HIV-negative women.DesignRegistry-based cohort study.SettingNorthern Tanzania, 2000-2018.Study sampleMother-baby pairs of singleton deliveries (n = 41 156).MethodsPerinatal outcomes of HIV-positive women were compared with HIV-negative women during time periods representing shifts in prevention of mother-to-child transmission guidelines. Monotherapy was used as first-line therapy before 2007 while combination antiretroviral treatment was routinely used from 2007. Log binomial and quantile regression were used to analyze the data.Main outcome measuresPreterm delivery, low birth weight, perinatal death, stillbirth, low Apgar score, transfer to neonatal care unit and small for gestational age.ResultsOverall, maternal HIV infection was associated with a higher risk of low birth weight and small for gestational age. Moreover, this pattern became more pronounced over time for low birth weight, the last time period being an exception. For other outcomes we found none or only a small overall association with maternal HIV infection, although a trend towards higher risk over time in HIV-positive compared with HIV-negative women was observed for preterm delivery and perinatal death. Quantile regression showed an increase in birth weight in babies born to HIV-negative women over time and a corresponding decline in birth weight in babies born to HIV-positive women.ConclusionUnfavourable trends in some of the selected perinatal outcomes were seen for HIV-positive compared with HIV-negative women. Potential side-effects of combination antiretroviral treatment in pregnancy should be further explored.

Project description:Chagas disease (CD) is recognized by the World Health Organization as one of the thirteen most neglected tropical diseases. More than 80% of people affected by CD will not have access to diagnosis and continued treatment, which partly supports the high morbidity and mortality rate. Machine Learning (ML) can identify patterns in data that can be used to increase our understanding of a specific problem or make predictions about the future. Thus, the aim of this study was to evaluate different models of ML to predict death in two years of patients with CD. ML models were developed using different techniques and configurations. The techniques used were: Random Forests, Adaptive Boosting, Decision Tree, Support Vector Machine, and Artificial Neural Networks. The adopted settings considered only interview variables, only complementary exam variables, and finally, both mixed. Data from a cohort study with CD patients called SaMi-Trop were analyzed. The predictor variables came from the baseline; and the outcome, which was death, came from the first follow-up. All models were evaluated in terms of Sensitivity, Specificity and G-mean. Among the 1694 individuals with CD considered, 134 (7.9%) died within two years of follow-up. Using only the predictor variables from the interview, the different techniques achieved a maximum G-mean of 0.64 in predicting death. Using only the variables from complementary exams, the G-mean was up to 0.77. In this configuration, the protagonism of NT-proBNP was evident, where it was possible to observe that an ML model using only this single variable reached G-mean of 0.76. The configuration that mixed interview variables and complementary exams achieved G-mean of 0.75. ML can be used as a useful tool with the potential to contribute to the management of patients with CD, by identifying patients with the highest probability of death. Trial Registration: This trial is registered with ClinicalTrials.gov, Trial ID: NCT02646943.

Project description:BackgroundPreterm birth can lead to impaired language development. This study aimed to predict language outcomes at 2 years corrected gestational age (CGA) for children born preterm.MethodsWe analysed data from 89 preterm neonates (median GA 29 weeks) who underwent diffusion MRI (dMRI) at term-equivalent age and language assessment at 2 years CGA using the Bayley-III. Feature selection and a random forests classifier were used to differentiate typical versus delayed (Bayley-III language composite score <85) language development.ResultsThe model achieved balanced accuracy: 91%, sensitivity: 86%, and specificity: 96%. The probability of language delay at 2 years CGA is increased with: increasing values of peak width of skeletonized fractional anisotropy (PSFA), radial diffusivity (PSRD), and axial diffusivity (PSAD) derived from dMRI; among twins; and after an incomplete course of, or no exposure to, antenatal corticosteroids. Female sex and breastfeeding during the neonatal period reduced the risk of language delay.ConclusionsThe combination of perinatal clinical information and MRI features leads to accurate prediction of preterm infants who are likely to develop language deficits in early childhood. This model could potentially enable stratification of preterm children at risk of language dysfunction who may benefit from targeted early interventions.ImpactA combination of clinical perinatal factors and neonatal DTI measures of white matter microstructure leads to accurate prediction of language outcome at 2 years corrected gestational age following preterm birth. A model that comprises clinical and MRI features that has potential to be scalable across centres. It offers a basis for enhancing the power and generalizability of diagnostic and prognostic studies of neurodevelopmental disorders associated with language impairment. Early identification of infants who are at risk of language delay, facilitating targeted early interventions and support services, which could improve the quality of life for children born preterm.

Project description:BackgroundSchizophrenia (SC) is one of the most common mental illnesses. However, the underlying genes that cause it and its effective treatments are unknown. Programmed cell death (PCD) is associated with many immune diseases and plays an important role in schizophrenia, which may be a diagnostic indicator of the disease.MethodsTwo groups as training and validation groups were chosen for schizophrenia datasets from the Gene Expression Omnibus Database (GEO). Furthermore, the PCD-related genes of the 12 patterns were extracted from databases such as KEGG. Limma analysis was performed for differentially expressed genes (DEG) identification and functional enrichment analysis. Machine learning was employed to identify minimum absolute contractions and select operator (LASSO) regression to determine candidate immune-related center genes, construct protein-protein interaction networks (PPI), establish artificial neural networks (ANN), and validate with consensus clustering (CC) analysis, then Receiver operating characteristic curve (ROC curve) was drawn for diagnosis of schizophrenia. Immune cell infiltration was developed to investigate immune cell dysregulation in schizophrenia, and finally, related drugs with candidate genes were collected via the Network analyst online platform.ResultsIn schizophrenia, 263 genes were crossed between DEG and PCD-related genes, and machine learning was used to select 42 candidate genes. Ten genes with the most significant differences were selected to establish a diagnostic prediction model by differential expression profiling. It was validated using artificial neural networks (ANN) and consensus clustering (CC), while ROC curves were plotted to assess diagnostic value. According to the findings, the predictive model had a high diagnostic value. Immune infiltration analysis revealed significant differences in Cytotoxic and NK cells in schizophrenia patients. Six candidate gene-related drugs were collected from the Network analyst online platform.ConclusionOur study systematically discovered 10 candidate hub genes (DPF2, ATG7, GSK3A, TFDP2, ACVR1, CX3CR1, AP4M1, DEPDC5, NR4A2, and IKBKB). A good diagnostic prediction model was obtained through comprehensive analysis in the training (AUC 0.91, CI 0.95-0.86) and validation group (AUC 0.94, CI 1.00-0.85). Furthermore, drugs that may be useful in the treatment of schizophrenia have been obtained (Valproic Acid, Epigallocatechin gallate).

Dataset Information

Prediction of perinatal death using machine learning models: a birth registry-based cohort study in northern Tanzania.

Objective

Design

Setting

Participants

Primary outcome measures

Results

Conclusions

Publications

Prediction of perinatal death using machine learning models: a birth registry-based cohort study in northern Tanzania.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets