Prevalence and Determinants of Preterm Birth in Tehran, Iran: A Comparison between Logistic Regression and Decision Tree Methods

Associations Among Multimorbid Conditions in Hospitalized Middle-aged and Older Adults in China: Statistical Analysis of Medical Records

BACKGROUND

Multimorbidity has become a new challenge for medical systems and public health policy. Understanding the patterns of and associations among multimorbid conditions should be given priority. It may assist with the early detection of multimorbidity and thus improve quality of life in older adults.

OBJECTIVE

This study aims to comprehensively analyze and compare associations among multimorbid conditions by age and sex in a large number of middle-aged and older Chinese adults.

METHODS

Data from the home pages of inpatient medical records in the Shenzhen National Health Information Platform were evaluated. From January 1, 2017, to December 31, 2018, inpatients aged 50 years and older who had been diagnosed with at least one of 40 conditions were included in this study. Their demographic characteristics (age and sex) and inpatient diagnoses were extracted. Association rule mining, Chi-square tests, and decision tree analyses were combined to identify associations between multiple chronic conditions.

RESULTS

In total, 306,264 hospitalized cases with available information on related chronic conditions were included in this study. The prevalence of multimorbidity in the overall population was 76.46%. The combined results of the 3 analyses showed that, in patients aged 50 years to 64 years, lipoprotein metabolism disorder tended to be comorbid with multiple chronic conditions. Gout and lipoprotein metabolism disorder had the strongest association. Among patients aged 65 years or older, there were strong associations between cerebrovascular disease, heart disease, lipoprotein metabolism disorder, and peripheral vascular disease. The strongest associations were observed between senile cataract and glaucoma in men and women. In particular, the association between osteoporosis and malignant tumor was only observed in middle-aged and older men, while the association between anemia and chronic kidney disease was only observed in older women.

CONCLUSIONS

Multimorbidity was prevalent among middle-aged and older Chinese individuals. The results of this comprehensive analysis of 4 age-sex subgroups suggested that associations between particular conditions within the sex and age groups occurred more frequently than expected by random chance. This provides evidence for further research on disease clusters and for health care providers to develop different strategies based on age and sex to improve the early identification and treatment of multimorbidity.

Iranian midwives' awareness and performance of respectful maternity care during labor and childbirth

INTRODUCTION

Midwives' perceptions of Respectful Maternity Care (RMC) play an important role in promoting quality of care. This study aimed to explore the awareness and performance of Iranian midwives of RMC during childbirth.

METHODS

A cross-sectional study was carried out from November to December 2020 to evaluate 130 midwives' awareness and performance of RMC during childbirth at four public hospitals in Urmia province, Iran. Participants were midwives who were working in the labor unit and had at least one year of work experience. The Midwives' Knowledge and Practice Scale on Respectful Maternity Care was used to assess midwives' awareness and performance. The quality assessment of questionnaires was based on the mean for each item. A multivariate linear regression approach was developed to evaluate the relationship between midwives' age, academic education level plus occupational information and their awareness and performance of RMC.

RESULTS

This study revealed that Iranian midwives had good awareness but fair performance of RMC. The mean scores of the overall awareness and performance of RMC were 36.07±10.13 and 75.47±35.4, respectively. Midwives' performance on two domains was fair including 'Giving emotional support' and 'Providing safe care'. The results of multivariate linear regression analysis showed a significant negative relationship between job satisfaction and midwives' performance on RMC. Also work experience plus a Master's degree in midwifery had positive significant effects on midwives' awareness along with performance on RMC (p<0.05).

CONCLUSIONS

Promoting respectful maternity care requires essential interpersonal and communication skills and supportive attitudes from midwives.

Maternal predictive factors for preterm birth: A case-control study in Southern Iran

BACKGROUND

Preterm birth (PTB) is one of the most important factors that increase the risk of chronic diseases and postpartum death in infants. The aim of this study was to determine the maternal factors that affect the birth of preterm infants in the city of Bandar Abbas.

MATERIALS AND METHODS

This is a case-control study that was performed on 400 preterm infants. Sampling was done by a simple method, and information was gathered by interviewing the mothers and their medical records. Data were collected by SPSS software version 16. To compare risk factors in the two groups, conditional logistic regression was used, and P < 0.05 was considered statistically significant.

RESULTS

Results showed that factors such as type of delivery (odds ratio [OR] = 3.584, 95% confidence interval [CI]: 1.981-6.485), preeclampsia (OR = 2.688, 95% CI: 1.164-6.207), history of PTB (OR = 4.171, 95% CI: 1.483-11.728), premature rupture of membranes (OR = 3.273, 95% CI: 1.745-6.137), care during prenatal (OR = 0.334, 95% CI: 0.159-0.701), placental abruption (OR = 3.209, 95% CI: 1.209-8.519), placenta previa (OR = 9.333, 95% CI: 2.086-41.770), and cervical insufficiency (OR = 11, 95% CI: 1.381-87.641) were independent risk factors of preterm infant birth.

CONCLUSIONS

The PTB risk is higher for women with cervical insufficiency, history of placenta previa, and history of preterm. Early recognition and management of these high-risk conditions among pregnant women may lead to a reduction in PTB rates.

A diagnostic profile on the PartoSure test

Introduction: Preterm birth (PTB) is common, occurring in over 10% of all live births globally, and is increasing worldwide. The limitations of traditional biomarkers of PTB, such as fetal fibronectin (fFN) and phosphorylated insulin-like growth factor-binding protein-1 (phIGFBP-1) have been well demonstrated in the literature. Therefore, augmenting clinical assessment with newer biomarkers, such as placental alpha macroglobulin-1 (PAMG-1); PartoSure, has the potential to improve disease monitoring and the best interventions. Areas covered: The present expert opinion evaluates the utility and limitations of PAMG-1; PartoSure as a biomarker for PTB in light of the current literature. Expert opinion: Although fFN, phIGFBP-1 and PAMG-1; PartoSure test had similar negative predictive value (NPV) and negative likelihood ratio (LR-), the PAMG-1; PartoSure test had the highest specificity, positive predictive value (PPV), and positive likelihood ratio (LR+) across all at-risk pregnant women. Although findings of this review may be encouraging, the PartoSure test should not be interpreted as absolute evidence for prediction of PTB. The PartoSure test result should always be used in conjunction with information available from the clinical evaluation of the pregnant woman and other diagnostic procedures such as cervical examination, assessment of uterine activity, and evaluation of other risk factors.

Comparison of Multivariable Logistic Regression and Other Machine Learning Algorithms for Prognostic Prediction Studies in Pregnancy Care: Systematic Review and Meta-Analysis

Background

Predictions in pregnancy care are complex because of interactions among multiple factors. Hence, pregnancy outcomes are not easily predicted by a single predictor using only one algorithm or modeling method.

Objective

This study aims to review and compare the predictive performances between logistic regression (LR) and other machine learning algorithms for developing or validating a multivariable prognostic prediction model for pregnancy care to inform clinicians’ decision making.

Methods

Research articles from MEDLINE, Scopus, Web of Science, and Google Scholar were reviewed following several guidelines for a prognostic prediction study, including a risk of bias (ROB) assessment. We report the results based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Studies were primarily framed as PICOTS (population, index, comparator, outcomes, timing, and setting): Population: men or women in procreative management, pregnant women, and fetuses or newborns; Index: multivariable prognostic prediction models using non-LR algorithms for risk classification to inform clinicians’ decision making; Comparator: the models applying an LR; Outcomes: pregnancy-related outcomes of procreation or pregnancy outcomes for pregnant women and fetuses or newborns; Timing: pre-, inter-, and peripregnancy periods (predictors), at the pregnancy, delivery, and either puerperal or neonatal period (outcome), and either short- or long-term prognoses (time interval); and Setting: primary care or hospital. The results were synthesized by reporting study characteristics and ROBs and by random effects modeling of the difference of the logit area under the receiver operating characteristic curve of each non-LR model compared with the LR model for the same pregnancy outcomes. We also reported between-study heterogeneity by using τ² and I².

Results

Of the 2093 records, we included 142 studies for the systematic review and 62 studies for a meta-analysis. Most prediction models used LR (92/142, 64.8%) and artificial neural networks (20/142, 14.1%) among non-LR algorithms. Only 16.9% (24/142) of studies had a low ROB. A total of 2 non-LR algorithms from low ROB studies significantly outperformed LR. The first algorithm was a random forest for preterm delivery (logit AUROC 2.51, 95% CI 1.49-3.53; I²=86%; τ²=0.77) and pre-eclampsia (logit AUROC 1.2, 95% CI 0.72-1.67; I²=75%; τ²=0.09). The second algorithm was gradient boosting for cesarean section (logit AUROC 2.26, 95% CI 1.39-3.13; I²=75%; τ²=0.43) and gestational diabetes (logit AUROC 1.03, 95% CI 0.69-1.37; I²=83%; τ²=0.07).

Conclusions

Prediction models with the best performances across studies were not necessarily those that used LR but also used random forest and gradient boosting that also performed well. We recommend a reanalysis of existing LR models for several pregnancy outcomes by comparing them with those algorithms that apply standard guidelines.

Trial Registration

PROSPERO (International Prospective Register of Systematic Reviews) CRD42019136106; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=136106

A systematic review shows no performance benefit of machine learning over logistic regression for clinical prediction models

OBJECTIVES

The objective of this study was to compare performance of logistic regression (LR) with machine learning (ML) for clinical prediction modeling in the literature.

STUDY DESIGN AND SETTING

We conducted a Medline literature search (1/2016 to 8/2017) and extracted comparisons between LR and ML models for binary outcomes.

RESULTS

We included 71 of 927 studies. The median sample size was 1,250 (range 72-3,994,872), with 19 predictors considered (range 5-563) and eight events per predictor (range 0.3-6,697). The most common ML methods were classification trees, random forests, artificial neural networks, and support vector machines. In 48 (68%) studies, we observed potential bias in the validation procedures. Sixty-four (90%) studies used the area under the receiver operating characteristic curve (AUC) to assess discrimination. Calibration was not addressed in 56 (79%) studies. We identified 282 comparisons between an LR and ML model (AUC range, 0.52-0.99). For 145 comparisons at low risk of bias, the difference in logit(AUC) between LR and ML was 0.00 (95% confidence interval, -0.18 to 0.18). For 137 comparisons at high risk of bias, logit(AUC) was 0.34 (0.20-0.47) higher for ML.

CONCLUSION

We found no evidence of superior performance of ML over LR. Improvements in methodology and reporting are needed for studies that compare modeling algorithms.

Algorithm on age partitioning for estimation of reference intervals using clinical laboratory database exemplified with plasma creatinine

BACKGROUND

We describe an algorithm to determine age-partitioned reference intervals (RIs) exemplified for creatinine using data collection from the clinical laboratory database.

METHODS

The data were acquired from the test results of creatinine of 164,710 outpatients aged <18 years in Beijing Children's Hospital laboratories' databases between January 2016 and December 2016. The tendency of serum creatinine with age was examined visually using box plot by gender first. The age subgroup was divided automatically by the decision tree method. Subsequently, the statistical tests of the difference between subgroups were performed by Harris-Boyd and Lahti methods.

RESULTS

A total of 136,546 samples after data cleaning were analyzed to explore the partition of age group for serum creatinine from birth to 17 years old. The suggested age partitioning of RIs for creatinine by the decision tree method were for eight subgroups. The difference between age subgroups was demonstrated to be statistically significant by Harris-Boyd and Lahti methods. In addition, the results of age partitioning for RIs estimation were similar to the suggested age partitioning by the Canadian Laboratory Initiative in Pediatric Reference Intervals study. Lastly, a suggested algorithm was developed to provide potential methodological considerations on age partitioning for RIs estimation.

CONCLUSIONS

Appropriate age partitioning is very important for establishing more accurate RIs. The procedure to explore the age partitioning using clinical laboratory data was developed and evaluated in this study, and will provide more opinions for designing research on establishment of RIs.

Comparison of three data mining models for prediction of advanced schistosomiasis prognosis in the Hubei province

BACKGROUND

In order to better assist medical professionals, this study aimed to develop and compare the performance of three models-a multivariate logistic regression (LR) model, an artificial neural network (ANN) model, and a decision tree (DT) model-to predict the prognosis of patients with advanced schistosomiasis residing in the Hubei province.

METHODOLOGY/PRINCIPAL FINDINGS

Schistosomiasis surveillance data were collected from a previous study based on a Hubei population sample including 4136 advanced schistosomiasis cases. The predictive models use LR, ANN, and DT methods. From each of the three groups, 70% of the cases (2896 cases) were used as training data for the predictive models. The remaining 30% of the cases (1240 cases) were used as validation groups for performance comparisons between the three models. Prediction performance was evaluated using area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and accuracy. Univariate analysis indicated that 16 risk factors were significantly associated with a patient's outcome of prognosis. In the training group, the mean AUC was 0.8276 for LR, 0.9267 for ANN, and 0.8229 for DT. In the validation group, the mean AUC was 0.8349 for LR, 0.8318 for ANN, and 0.8148 for DT. The three models yielded similar results in terms of accuracy, sensitivity, and specificity.

CONCLUSIONS/SIGNIFICANCE

Predictive models for advanced schistosomiasis prognosis, respectively using LR, ANN and DT models were proved to be effective approaches based on our dataset. The ANN model outperformed the LR and DT models in terms of AUC.