Dataset Information

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.

ABSTRACT:

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.

SUBMITTER: Tarimo CS

PROVIDER: S-EPMC8647548 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundAcute kidney injury (AKI) is a recognized complication in critically ill patients. The epidemiology of AKI varies worldwide, depending on the diagnostic criteria used and the setting. The International Society of Nephrology has called for a reduction in preventable deaths from AKI to zero by the year 2025. It is suspected that the majority of AKI cases are in limited-resource countries, but the true burden of AKI in these settings remains unknown.ObjectiveWe aimed to determine, using standardized KDIGO (Kidney Disease Improving Global Outcomes) criteria, the prevalence of AKI, associated factors, and clinical characteristics of adult (≥18 years) patients admitted to intensive care units (ICUs) at a tertiary hospital in Tanzania.DesignProspective observational study from November 2017 to May 2018.MethodsIn all, 320 patients admitted to medical and surgical ICUs were consecutively enrolled. Baseline, clinical, and laboratory data were collected on admission and during their ICU stay. Serum creatinine and urine output were measured, and KDIGO criteria were used to determine AKI status.ResultsMore than half (55.3%) of ICU patients were diagnosed with AKI. Of these, 80% were diagnosed within 24 hours of admission. Acute kidney injury stage 3 accounted for 35% of patients with AKI. Patients with AKI were older, more likely to have cardiovascular comorbidities, and with higher baseline serum levels of creatinine, potassium, universal vital assessment admission scores, and total white cell count ≥12. Sepsis (odds ratio [OR] = 3.81; confidence interval [CI] = 1.21-11.99), diabetes (OR = 2.54; CI = 1.24-5.17), and use of vasopressors (OR = 3.78; CI = 1.36-10.54) were independently associated with AKI in multivariable logistic regression. Less than one-third of those who needed dialysis received it. There was 100% mortality in those who needed dialysis but did not receive (n = 19).LimitationsBeing based at a referral center, the findings do not represent the true burden of AKI in the community.ConclusionThe prevalence of AKI was very high in ICUs in Northern Tanzania. The majority of patients presented with AKI and were severely ill, suggesting late presentation, underscoring the importance of prioritizing prevention and early intervention. Further studies should explore locally suitable AKI risk scores that could be used to identify high-risk patients in the community health centers from where patients are referred.

Project description:ObjectiveWe aimed to determine the key predictors of perinatal deaths using machine learning models compared with the logistic regression model.DesignA 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.SettingThe 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.ParticipantsSingleton deliveries (n=42 319) with complete records from 2000 to 2015.Primary outcome measuresPerinatal death (composite of stillbirths and early neonatal deaths). These outcomes were only captured before mothers were discharged from the hospital.ResultsThe 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.ConclusionsIn 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.

Project description:BackgroundPrediction 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.MethodsWe 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.ResultsBirth 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.ConclusionThere 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.

Project description:IntroductionAdequate and timely access to pathology services is a key to scale up cancer control, however, there is an extremely shortage of pathologists in Tanzania. Telepathology (scanned images microscopy) has the potential to increase access to pathology services and it is increasingly being employed for primary diagnosis and consultation services. However, the experience with the use of telepathology in Tanzania is limited. We aimed to investigate the feasibility of using scanned images for primary diagnosis of pre-malignant and malignant cervical lesions by assessing its equivalency to conventional (glass slide) microscopy in Tanzania.MethodsIn this laboratory-based study, assessment of hematoxylin and eosin stained glass slides of 175 cervical biopsies were initially performed conventionally by three pathologists independently. The slides were scanned at x 40 and one to three months later, the scanned images were reviewed by the pathologists in blinded fashion. The agreement between initial and review diagnoses across participating pathologists was described and measured using Cohen's kappa coefficient (κ).ResultsThe overall concordance of diagnoses established on conventional microscopy compared to scanned images across three pathologists was 87.7%; κ = 0.54; CI (0.49-0.57).The overall agreement of diagnoses established by local pathologist on conventional microscopy compared to scanned images was 87.4%; κ = 0.73; CI (0.65-0.79). The concordance of diagnoses established by senior pathologist compared to local pathologist on conventional microscopy and scanned images was 96% and 97.7% respectively. The inter-observer agreement (κ) value were 0.93, CI (0.87-1.00) and 0.94, CI (0.88-1.00) for conventional microscopy and scanned images respectively.ConclusionsAll κ coefficients expressed good intra- and inter-observer agreement, suggesting that telepathology is sufficiently accurate for primary diagnosis in surgical pathology. The discrepancies in interpretation of pre-malignant lesions highlights the importance of p16 immunohistochemistry in definitive diagnosis in these lesions. Sustainability factors including hardware and internet connectivity are essential components to be considered before telepathology may be deemed suitable for widely use in Tanzania.

Project description:BackgroundPediatric injuries are a leading cause of morbidity and mortality in low- and middle-income countries (LMICs). It is important that injured children get quality care in order to improve their outcomes. Injured children are nearly always accompanied by family member caregivers invested in their outcome, and who will be responsible for their recovery and rehabilitation after discharge.ObjectiveThe purpose of this study was to identify family member caregiver perspectives on strengths and challenges in pediatric injury care throughout hospitalization at a tertiary hospital in Northern Tanzania.MethodsThis study was conducted at a zonal referral hospital in Northern Tanzania. Qualitative semi-structured in-depth interviews (IDIs) were conducted by trained interviewers who were fluent in English and Swahili in order to examine the strengths and challenges in pediatric injury care. IDIs were completed from November 2020 to October 2021 with 30 family member caregivers of admitted pediatric injured patients. De-identified transcripts were synthesized in memos and analyzed through a team-based, thematic approach informed by applied thematic analysis.ResultsStrengths and challenges were identified throughout the hospital experience, including emergency medicine department (EMD) care, inpatient wards care, and discharge. Across the three phases, strengths were identified such as how quickly patients were evaluated and treated, professionalism and communication between healthcare providers, attentive nursing care, frequent re-evaluation of a patient's condition, and open discussion with caregivers about readiness for discharge. Challenges identified related to lack of communication with caregivers, perceived inability of caregivers to ask questions, healthcare providers speaking in English during rounds with lack of interpretation into the caregivers' preferred language, and being sent home without instructions for rehabilitation, ongoing care, or guidance for follow-up.ConclusionCaregiver perspectives highlighted strengths and challenges throughout the hospital experience that could lead to interventions to improve the care of pediatric injury patients in Northern Tanzania. These interventions include prioritizing communication with caregivers about patient status and care plan, ensuring all direct communication is in the caregivers' preferred language, and standardizing instructions regarding discharge and follow-up.

Project description:Paediatric rheumatic disorders are common in children and result in significant impairment in quality of life, morbidity and mortality. There is limited information on the burden of these disorders in lower income countries especially in sub-Saharan Africa. Few case reports have documented presence of paediatric rheumatic disorders in Tanzania. This study was conducted to determine the spectrum of rheumatic disorders among children at Muhimbili National Hospital (MNH). This was a retrospective study conducted among children who were attended at MNH between January 2012 and August 2019. Paediatric patients seen in the out-patient clinics and those admitted in the wards were eligible. All patients with diagnosis of rheumatic disorders were identified from admission books and outpatient clinic logbooks, and later data were collected from their case notes and were recorded in clinical research forms. Collected information included age, sex, clinical features and laboratory tests results. A total of 52 children with mean age of 9.5?±?4.3?years, 12 (40.4%) participants were aged above 10?years and 32 (61.5%) were females. Frequently reported clinical presentations were joint pain 44 (84.6%), joint swelling 34 (65.4%), fever 24 (46.2%) and skin rashes 21(40.4%). Juvenile idiopathic arthritis (JIA) was the predominant diagnosis reported in 28 (53.8%) participants followed by juvenile systemic lupus erythematosus 8 (15.4%), mixed connective tissue diseases 4 (7.7%) and juvenile dermatomyositis 4 (7.7%). Antinuclear antibody test was performed in 16 participants it was positive in 9 (56.2%). Nine participants were tested for anti-double stranded DNA test and 5 (55.6%) were positive for this test. C-reactive protein was tested in 46 participants out of which 32 (69.6%) had elevated levels. HIV was tested in 24 (46.2%) participants and results were negative. Thirty-five out of 52 (67.3%) participants had anaemia. Predominant drugs used for treatment of JIA include prednisolone and methotrexate. Paediatric rheumatic disorders are not uncommon in Tanzania-and were noted to affect more female children in this study. Predominant conditions included juvenile idiopathic arthritis (JIA), juvenile systemic lupus erythematosus (JSLE) and juvenile dermatomyositis (JDM).