Potential Biomarkers for Predicting Depression in Diabetes Mellitus

Deep learning based prediction of depression and anxiety in patients with type 2 diabetes mellitus using regional electronic health records

BACKGROUND

Depression and anxiety are prevalent mental health conditions among individuals with type 2 diabetes mellitus (T2DM), who exhibit unique vulnerabilities and etiologies. However, existing approaches fail to fully utilize regional heterogeneous electronic health record (EHR) data. Integrating this data can provide a more comprehensive understanding of depression and anxiety in T2DM patients, leading to more personalized treatment strategies.

OBJECTIVE

This study aims to develop and validate a deep learning model, the Regional EHR for Depression and Anxiety Prediction Model (REDAPM), using regional EHR data to predict depression and anxiety in patients with T2DM.

METHODS

A case-control development and validation study was conducted using regional EHR data from the Nanjing Health Information Center (NHIC). Two retrospective, matched (1:3) datasets were constructed from the full cohort for the model's internal and external validation. These two datasets were selected from the NHIC data of 2020 and 2022, respectively. The REDAPM incorporates both structured and unstructured EHR data, capturing the temporal dependency of clinical events. The performance of REDAPM was compared to a set of baseline models, evaluated using the area under the receiver operating characteristic curve (ROC-AUC) and the area under the precision-recall curve (PR-AUC). Subgroup, ablation, and interpretation analyses were conducted to identify relevant clinical features available from EHRs.

RESULTS

The internal and external validation datasets comprised 24,724 and 34,340 patients, respectively. The REDAPM outperformed baseline models in both datasets, achieving ROC-AUC scores of 0.9029±0.008 and 0.7360±0.005, and PR-AUC scores of 0.8124±0.011 and 0.5504±0.009. Ablation and subgroup experiments confirmed the significant contribution of patients' medical history text to the model's performance. Integrated gradient score analysis identified the predictive importance of other mental disorders.

CONCLUSION

The REDAPM effectively leverages the heterogeneous characteristics of regional EHR data, demonstrating strong predictive performance for depression onset in diabetic patients. It also shows potential for discovering significant clinical features, indicating considerable promise for clinical utility.

Utilizing machine learning algorithms for predicting Anxiety-Depression Comorbidity Syndrome in Gastroenterology Inpatients (ADCS-GI)

BACKGROUND

Accurately diagnosing Anxiety-Depression Comorbidity Syndrome in Gastroenterology Inpatients (ADCS-GI) shows significant challenges as traditional diagnostic methods fail to meet expectations due to patient hesitance and non-psychiatric healthcare professionals' limitations. Therefore, the need for objective diagnostics highlights the potential of machine learning in identifying and treating ADCS-GI.

METHODS

A total of 1186 ADCS patients were recruited for this study. We conducted extensive studies for the dataset, including data quantification, equilibrium, and correlation analysis. Eight machine learning models, including Gaussian Naive Bayes (NB), Support Vector Classifier (SVC), K-Neighbors Classifier, RandomForest, XGB, CatBoost, Cascade Forest, and Decision Tree, were utilized to compare prediction efficacy, with an effort to minimize the dependency on subjective questionnaires.

RESULTS

Among eight machine learning algorithms, the Decision Tree and K-nearest neighbors models demonstrated an accuracy exceeding 81% and a sensitivity in the same range for detecting ADCS in patients. Notably, when identifying moderate and severe cases, the models achieved an accuracy above 88% and a sensitivity of 90%. Furthermore, the models trained without reliance on subjective questionnaires showed promising performance, indicating the feasibility of developing questionnaire-free early detection applications.

CONCLUSION

Machine learning algorithms can be used to identify ADCS among gastroenterology patients. This can help facilitate the early detection and intervention of psychological disorders in gastroenterology patients' care.

The curvilinear relationship between Framingham Steatosis Index and depression: insights from a nationwide study

Background

The Framingham Steatosis Index (FSI) serves as a diagnostic metric for fatty liver. While research has established a link between depression and fatty liver, the association with the Framingham Steatosis Index (FSI) remains undocumented. The aim of this study is to explore the potential correlation between FSI and depression, addressing this research void.

Methods

Our data originates from the National Health and Nutrition Examination Survey (NHANES) database. We employed the PHQ-9 questionnaire for the evaluation of depressive symptoms. We investigated the association between FSI and depression using a weighted multiple logistic regression model and stratified analysis. Non-linear associations were explored using fitted smooth curves. A recursive method was employed to identify inflection points. Subgroup analyses were conducted to examine differences in the association between FSI and depression within subgroups.

Results

Our research encompassed a total of 19,697 participants. Multivariate logistic regression analysis, adjusted for potential confounding factors, demonstrated a significant positive association between FSI and depression, with OR of 1.14 (95% CI: 1.10, 1.18). Stratified analysis indicated that a significant positive correlation exists between FSI and depression among all groups except those with BMI below 30. The non-linear relationship was further confirmed by the restricted cubic splines analysis, which revealed an inflection point at an FSI value of 29.72. Below this threshold, there was no significant correlation, while above it, a positive correlation was observed. Subgroup analysis revealed statistically significant interactions between FSI and depression within the educational attainment groups.

Conclusion

Our study's discovery is the curvilinear relationship between FSI and depression. Factors such as inflammation, hormonal levels, and metabolic disruptions could be the underlying mechanisms driving this relationship. This finding offers valuable insights that could inform the development of comprehensive intervention strategies for managing depression in clinical settings.

Progress and trends of research on mineral elements for depression

Objective

To explore the research progress and trends on mineral elements and depression.

Methods

After querying the MeSH database and referring to the search rules, the search terms were selected and optimized to obtain the target literature collection. We analyzed the general characteristics of the literature, conducted network clustering and co-occurrence analysis, and carried out a narrative review of crucial literature.

Results

Bipolar disorder was a dominant topic in the retrieved literature, which saw a significant increase in 2010 and 2019-2020. Most studies focused on mineral elements, including lithium, calcium, magnesium, zinc, and copper. The majority of journals and disciplines were in the fields of psychiatry, neuropsychology, neuropharmacology, nutrition, medical informatics, chemistry, and public health. The United States had the highest proportion in terms of paper sources, most-cited articles, high-frequency citations, frontier citations, and high centrality citation. Regarding the influence of academic institutions, the top five were King's College London, the Chinese Academy of Sciences, University of Barcelona, INSERM, and Heidelberg University. Frontier keywords included bipolar disorder, drinking water, (neuro)inflammation, gut microbiota, and systematic analysis. Research on lithium response, magnesium supplementation, and treatment-resistant unipolar depression increased significantly after 2013.

Conclusion

Global adverse events may have indirectly driven the progress in related research. Although the literature from the United States represents an absolute majority, its influence on academic institutions is relatively weaker. Multiple pieces of evidence support the efficacy of lithium in treating bipolar disorder (BD). A series of key discoveries have led to a paradigm shift in research, leading to increasingly detailed studies on the role of magnesium, calcium, zinc, and copper in the treatment of depression. Most studies on mineral elements remain diverse and inconclusive. The potential toxicity and side effects of some elements warrant careful attention.

Bilirubin and postpartum depression: an observational and Mendelian randomization study

Background

Postpartum depression (PPD) is one of the most common complications of delivery and is usually disregarded. Several risk factors of PPD have been identified, but its pathogenesis has not been completely understood. Serum bilirubin has been found to be a predictor of depression, whose relationship with PPD has not been investigated.

Methods

Observational research was performed followed by a two-sample Mendelian randomization (MR) analysis. From 2017 to 2020, the clinical data of pregnant women were retrospectively extracted. Logistic regression and random forest algorithm were employed to assess the risk factors of PPD, including the serum levels of total bilirubin and direct bilirubin. To further explore their potential causality, univariable and multivariable Mendelian randomization (MVMR) were conducted. Sensitivity analyses for MR were performed to test the robustness of causal inference.

Results

A total of 1,810 patients were included in the PPD cohort, of which 631 (34.87%) were diagnosed with PPD. Compared with the control group, PPD patients had a significantly lower level of total bilirubin (9.2 μmol/L, IQR 7.7, 11.0 in PPD; 9.7 μmol/L, IQR 8.0, 12.0 in control, P < 0.001) and direct bilirubin (2.0 μmol/L, IQR 1.6, 2.6 in PPD; 2.2 μmol/L, IQR 1.7, 2.9 in control, P < 0.003). The prediction model identified eight independent predictive factors of PPD, in which elevated total bilirubin served as a protective factor (OR = 0.94, 95% CI 0.90-0.99, P = 0.024). In the MR analyses, genetically predicted total bilirubin was associated with decreased risk of PPD (IVW: OR = 0.86, 95% CI 0.76-0.97, P = 0.006), which remained consistent after adjusting educational attainment, income, and gestational diabetes mellitus. Conversely, there is a lack of solid evidence to support the causal relationship between PPD and bilirubin.

Conclusion

Our results suggested that decreased total bilirubin was associated with the incidence of PPD. Future studies are warranted to investigate its potential mechanisms and illuminate the pathogenesis of PPD.

Applying contrastive pre-training for depression and anxiety risk prediction in type 2 diabetes patients based on heterogeneous electronic health records: a primary healthcare case study

OBJECTIVE

Due to heterogeneity and limited medical data in primary healthcare services (PHS), assessing the psychological risk of type 2 diabetes mellitus (T2DM) patients in PHS is difficult. Using unsupervised contrastive pre-training, we proposed a deep learning framework named depression and anxiety prediction (DAP) to predict depression and anxiety in T2DM patients.

MATERIALS AND METHODS

The DAP model consists of two sub-models. Firstly, the pre-trained model of DAP used unlabeled discharge records of 85 085 T2DM patients from the First Affiliated Hospital of Nanjing Medical University for unsupervised contrastive learning on heterogeneous electronic health records (EHRs). Secondly, the fine-tuned model of DAP used case-control cohorts (17 491 patients) selected from 149 596 T2DM patients' EHRs in the Nanjing Health Information Platform (NHIP). The DAP model was validated in 1028 patients from PHS in NHIP. Evaluation included receiver operating characteristic area under the curve (ROC-AUC) and precision-recall area under the curve (PR-AUC), and decision curve analysis (DCA).

RESULTS

The pre-training step allowed the DAP model to converge at a faster rate. The fine-tuned DAP model significantly outperformed the baseline models (logistic regression, extreme gradient boosting, and random forest) with ROC-AUC of 0.91±0.028 and PR-AUC of 0.80±0.067 in 10-fold internal validation, and with ROC-AUC of 0.75 ± 0.045 and PR-AUC of 0.47 ± 0.081 in external validation. The DCA indicate the clinical potential of the DAP model.

CONCLUSION

The DAP model effectively predicted post-discharge depression and anxiety in T2DM patients from PHS, reducing data fragmentation and limitations. This study highlights the DAP model's potential for early detection and intervention in depression and anxiety, improving outcomes for diabetes patients.