The development and application of a prediction model for postpartum depression: optimizing risk assessment and prevention in the clinic

Construction of a depression risk prediction model for type 2 diabetes mellitus patients based on NHANES 2007-2014

BACKGROUND

Type 2 diabetes mellitus (T2DM) is a prevalent global health issue that has been linked to an increased risk of depression. The objective of this study was to construct a nomogram model for predicting depression in T2DM patients.

METHODS

A total of 4280 patients with T2DM were included in this study from the 2007-2014 NHANES. The entire dataset was split randomly into training set comprising 70 % of the data and a validation set comprising 30 % of the data. LASSO and multivariate logistic regression analyses identified predictors significantly associated with depression, and the nomogram was constructed with these predictors. The model was assessed by C-index, calibration curve, the hosmer-lemeshow test and decision curve analysis (DCA).

RESULTS

The nomogram model comprised of 9 predictors, namely age, gender, PIR, BMI, education attainment, smoking status, LDL-C, sleep duration and sleep disorder. The C-index of the training set was 0.780, while that of the validation set was 0.752, indicating favorable discrimination for the model. The model exhibited excellent clinical applicability and calibration in both the training and validation datasets. Moreover, the cut-off value of the nomogram is 223.

LIMITATIONS

This study has shortcomings in data collection, lack of external validation, and results non-extrapolation.

CONCLUSIONS

Our nomogram exhibits high clinical predictability, enabling clinicians to utilize this tool in identifying high-risk depressed patients with T2DM. It has the potential to decrease the incidence of depression and significantly improve the prognosis of patients with T2DM.

Predicting efficacy of sub-anesthetic ketamine/esketamine i.v. dose during course of cesarean section for PPD prevention, utilizing traditional logistic regression and machine learning models

OBJECTIVE

Increasing researches supported that intravenous ketamine/esketamine during the perioperative period of cesarean section could prevent postpartum depression(PPD). With the effective rate ranging from 87.2 % to 95.5 % in PPD, ketamine/esketamine's responsiveness was individualized. To optimize ketamine dose/form based on puerpera prenatal characteristics, reducing adverse events and improving the total efficacy rate, prediction models were developed to predict ketamine/esketamine's efficacy.

METHOD

Based on two randomized controlled trials, 12 prenatal features of 507 women administered the ketamine/esketamine intervention were collected. Traditional logistics regression, SVM, random forest, KNN and XGBoost prediction models were established with prenatal features and dosage regimen as predictors.

RESULTS

According to the logistic regression model (ain = 0.10, aout = 0.15, area under the receiver operating characteristic curve, AUC = 0.728), prenatal Edinburgh Postnatal Depression Scale (EPDS) score ≥ 10, thoughts of self-injury and bad mood during pregnancy were associated with poorer ketamine efficacy in PPD prevention, whilst a high dose of esketamine (0.25 mg/kg loading dose+2 mg/kg PCIA) was the most effective dosage regimen and esketamine was more recommended rather than ketamine in PPD. The AUCvalidation set of KNN and XGBoost model were 0.815 and 0.651, respectively.

CONCLUSION

Logistic regression and machine learning algorithm, especially the KNN model, could predict the effectiveness of ketamine/esketamine iv. during the course of cesarean section for PPD prevention. An individualized preventative strategy could be developed after entering puerpera clinical features into the model, possessing great clinical practice value in reducing PPD incidence.

Predictive models for predicting the risk of maternal postpartum depression: A systematic review and evaluation

OBJECTIVES

Clinical prediction models have been widely used to screen and diagnose postpartum depression (PPD). This study systematically reviews and evaluates the risk of bias and the applicability of PPD prediction models.

METHODS

A systematic search was performed in eight databases from inception to June 1, 2022. The literature was independently screened, and data were extracted by two investigators using the checklist for critical appraisal and data extraction for systematic reviews of prediction modeling studies (CHARMS). The risk of bias and applicability was assessed using the Prediction Model Risk of Bias Assessment Tool (PROBAST).

RESULTS

After the screening, 12 studies of PPD risk prediction models were included, with the area under the ROC curve of the models ranging from 0.611 to 0.937. The most-reported predictors of PPD included several aspects, including prenatal mood disorders, endocrine and hormonal influences, psychosocial aspects, the influence of family factors, and somatic illness factors. The applicability of all studies was good. However, there was some bias, mainly due to inadequate outcome events, missing data not appropriately handled, lack of model performance assessment, and overfitting of the models.

CONCLUSIONS

This systematic review and evaluation indicate that most present PPD prediction models have a high risk of bias during development and validation. Despite some models' predictive solid performance, the models' clinical practice rate is low. Therefore, future research should develop predictive models with excellent performance in all aspects and clinical applicability to better inform maternal medical decisions.