Machine Learning Models for Prediction of Maternal Hemorrhage and Transfusion: Model Development Study.

BACKGROUND: Current postpartum hemorrhage (PPH) risk stratification is based on traditional statistical models or expert opinion. Machine learning could optimize PPH prediction by allowing for more complex modeling. OBJECTIVE: We sought to improve PPH prediction and compare machine learning and traditional statistical methods. METHODS: We developed models using the Consortium for Safe Labor data set (2002-2008) from 12 US hospitals. The primary outcome was a transfusion of blood products or PPH (estimated blood loss of ≥1000 mL). The secondary outcome was a transfusion of any blood product. Fifty antepartum and intrapartum characteristics and hospital characteristics were included. Logistic regression, support vector machines, multilayer perceptron, random forest, and gradient boosting (GB) were used to generate prediction models. The area under the receiver operating characteristic curve (ROC-AUC) and area under the precision/recall curve (PR-AUC) were used to compare performance. RESULTS: Among 228,438 births, 5760 (3.1%) women had a postpartum hemorrhage, 5170 (2.8%) had a transfusion, and 10,344 (5.6%) met the criteria for the transfusion-PPH composite. Models predicting the transfusion-PPH composite using antepartum and intrapartum features had the best positive predictive values, with the GB machine learning model performing best overall (ROC-AUC=0.833, 95% CI 0.828-0.838; PR-AUC=0.210, 95% CI 0.201-0.220). The most predictive features in the GB model predicting the transfusion-PPH composite were the mode of delivery, oxytocin incremental dose for labor (mU/minute), intrapartum tocolytic use, presence of anesthesia nurse, and hospital type. CONCLUSIONS: Machine learning offers higher discriminability than logistic regression in predicting PPH. The Consortium for Safe Labor data set may not be optimal for analyzing risk due to strong subgroup effects, which decreases accuracy and limits generalizability.

Who's Posting What? A Qualitative, Cross-Sectional Study Comparing Twitter and Instagram Patterns Between Health Care Professionals and Patients.

IMPORTANCE: There is a paucity of evidence-based, physician-authored content available on social media. Data are lacking on physicians use of social media, including intended audience and content. OBJECTIVE: The aim of this study was to explore the patterns of Twitter and Instagram use for popular urogynecology hashtags between physicians, patients, and allied health professionals (AHPs). STUDY DESIGN: Twelve hashtags derived from the Urogynecology Tag Ontology project were used as search terms to select Twitter and Instagram posts. Up to 5 top posts per hashtag per author type (physician, patient, or AHP) were included. Posts were analyzed using Dedoose qualitative analytic software by author, hashtag, intended audience, and themes. RESULTS: On Twitter, 109 posts met inclusion criteria: 41% written by physicians, 40% patients, and 18.3% AHPs. For Instagram, 72 posts were included: 50% written by patients, 39% AHPs, and 11% physicians. Twitter physician posts were mainly intended for health professionals (64%) with only 18% for patients. Patients posted to the general public (57%) and patients (36%). Instagram physician posts were intended for health professionals (49%), whereas 62% of AHPs posted to patients. Most patient posts were directed to other patients (90%). Physicians posted about academic peer discussions, medical education, and advocacy. Patients posted about personal experiences, treatments, or dissatisfaction. CONCLUSIONS: Physicians are more likely to post on Twitter than Instagram, with content focused primarily on their peer group, and physicians/patients are unlikely to engage with each other. There is an opportunity to improve social media interactions between physicians and the public while increasing high-quality patient education.

High Levels of Serum IgG for Opisthorchis viverrini and CD44 Expression Predict Worse Prognosis for Cholangiocarcinoma Patients after Curative Resection.

BACKGROUND: Opisthorchis viverrini (OV)-associated cholangiocarcinoma (CCA) has a high immune response with chronic inflammation and oxidative stress. CD44 and Nestin, two cancer stem cell (CSC) markers, play major roles in cancer cell survival. Effects of immune response and expression CSC markers on survival of patients with CCA remain unclear. OBJECTIVE: To investigate the effects of level of OV IgG together with CSC marker expression and also the combination of these markers on survival of CCA patients after curative resection. METHODS: All serum specimens from CCA patients who underwent curative surgery from 2005 to 2015 were examined for IgG for OV antigen by ELISA. Tissue specimens were studied for CD44 and Nestin expression. Survival analysis by Cox proportional hazard model was used for estimating hazard ratio (HR) with a 95% confidence interval (CI). RESULTS: In this study, 122 (69.3%) of 176 were positive for OV IgG, and 35 (19.9%) were considered to have high-positive OV IgG. CD44s positive expression was found in 54 (40%), CD44v6 high expression in 96 (69.6%), CD44v8-10 high expression in 87 (63.5%) and Nestin high expression in 21 (16.1%). Multivariate survival analysis found that high-positive OV IgG and late stage tumor were independent prognostic factors with the adjusted HR of 2.24 (95% CI 1.27-3.93) and 2.78 (95% CI 1.46-5.29), respectively. Subgroup analysis in early and late stage CCA showed that a combined positive OV IgG and CD44s expression with the high expression of CD44v8-10 had the significantly poorest prognosis with HR of 3.75 (95% CI 1.61-8.72) and HR of 1.76 (95% CI 1.02-3.03), respectively. CONCLUSION: A high level of OV IgG as well as a high level of CSC markers resulted in an aggressive CCA. OV IgG level together with CSC markers can be used as the prognostic markers for CCA patients' survival. The study of the CD44 pathway is promising for adjuvant treatment.

GABPα and CREB1 Binding to Double Nucleotide Polymorphisms of Their Consensus Motifs and Cooperative Binding to the Composite ETS ⇔ CRE Motif (ACCGGAAGTGACGTCA).

Previously, cooperative binding of the bZIP domain of CREB1 and the ETS domain of GABPα was observed for the composite DNA ETS ⇔ CRE motif (A 0 C 1 C 2 G 3 G 4 A 5 A 6 G 7 T 8 G 9 A 10 C 11 G 12 T 13 C 14 A 15 ). Single nucleotide polymorphisms (SNPs) at the beginning and end of the ETS motif (ACCGGAAGT) increased cooperative binding. Here, we use an Agilent microarray of 60-mers containing all double nucleotide polymorphisms (DNPs) of the ETS ⇔ CRE motif to explore GABPα and CREB1 binding to their individual motifs and their cooperative binding. For GABPα, all DNPs were bound as if each SNP acted independently. In contrast, CREB1 binding to some DNPs was stronger or weaker than expected, depending on the locations of each SNP. CREB1 binding to DNPs where both SNPs were in the same half site, T 8 G 9 A 10 or T 13 C 14 A 15 , was greater than expected, indicating that an additional SNP cannot destroy binding as much as expected, suggesting that an individual SNP is enough to abolish sequence-specific DNA binding of a single bZIP monomer. If a DNP contains SNPs in each half site, binding is weaker than expected. Similar results were observed for additional ETS and bZIP family members. Cooperative binding between GABPα and CREB1 to the ETS ⇔ CRE motif was weaker than expected except for DNPs containing A 7 and SNPs at the beginning of the ETS motif.