Applying machine learning methods to psychosocial screening data to improve identification of prenatal depression: Implications for clinical practice and research.

We utilized machine learning (ML) methods on data from the PROMOTE, a novel psychosocial screening tool, to quantify risk for prenatal depression for individual patients and identify contributing factors that impart greater risk for depression. Random forest algorithms were used to predict likelihood for being at high risk for prenatal depression (Edinburgh Postnatal Depression Scale; EPDS ≥ 13 and/or positive self-injury item) using data from 1715 patients who completed the PROMOTE. Performance matrices were calculated to assess the ability of the PROMOTE to accurately classify patients. Probability for depression was calculated for individual patients. Finally, recursive feature elimination was used to evaluate the importance of each PROMOTE item in the classification of depression risk. PROMOTE data were successfully used to predict depression with acceptable performance matrices (accuracy = 0.80; sensitivity = 0.75; specificity = 0.81; positive predictive value = 0.79; negative predictive value = 0.97). Perceived stress, emotional problems, family support, age, major life events, partner support, unplanned pregnancy, current employment, lifetime abuse, and financial state were the most important PROMOTE items in the classification of depression risk. Results affirm the value of the PROMOTE as a psychosocial screening tool for prenatal depression and the benefit of using it in conjunction with ML methods. Using such methods can help detect underreported outcomes and identify what in patients' lives makes them more vulnerable, thus paving the way for effective individually tailored precision medicine.

The effect of a scheduled regimen of acetaminophen and ibuprofen on opioid use following cesarean delivery.

Objective The primary objective was to evaluate if the administration of ibuprofen and acetaminophen at regularly scheduled intervals impacts pain scores and total opioid consumption, when compared to administration based on patient demand. Methods A retrospective chart review was performed comparing scheduled vs. as-needed acetaminophen and ibuprofen regimens, with 100 women included in each arm. Demographics and delivery characteristics were collected in addition to pain scores and total ibuprofen, acetaminophen and oxycodone use at 24, 48 and 72 h postoperatively. Results The scheduled dosing group was found to have a statistically significant decrease in pain scores at all time intervals. Acetaminophen and ibuprofen usage were also noted to be higher in this group while narcotic use was reduced by 64%. Conclusion Scheduled dosing of non-narcotic pain medications can substantially decrease opioid usage after cesarean delivery and improve post-operative pain.

The effect of an oxytocin washout period on blood loss at cesarean delivery.

Objectives Prolonged oxytocin exposure may result in increased blood loss during delivery. Our objective was to determine whether an oxytocin rest period before cesarean delivery had an impact on blood loss. Methods We performed a retrospective cohort study of women who underwent primary cesarean delivery after oxytocin augmentation. The primary outcome was change between pre- and postoperative hematocrit (Hct) in women with less than 60-min oxytocin rest period (<60 min) and greater than 60-min rest period (>60 min). Results There was no difference in demographic characteristics (age, BMI, or gestational age at delivery) between the two groups. Women in the >60 min group had a higher cumulative dose and longer duration of oxytocin administration. There was no significant difference in change in Hct between the two groups when controlling for these factors. Conclusions We did not find a significant correlation between the duration of the oxytocin rest period and blood loss. Oxytocin washout periods of greater than 60 min may not result in decreased blood loss at cesarean delivery, and thus, women may not benefit from such oxytocin washout periods.

The utility of clinical findings to predict laboratory values in hypertensive disorders of pregnancy.

OBJECTIVE: Preeclampsia is the 2nd leading cause of maternal mortality in the United States. Women with new-onset or worsening hypertension are commonly evaluated for laboratory abnormalities. We aim to investigate whether demographic and/or clinical findings correlate with abnormal laboratory values. STUDY DESIGN: A retrospective chart review of women who presented for evaluation of hypertension in pregnancy during 2010. Demographic information, medical history, symptoms, vital signs, and laboratory results were collected. Bivariate analysis was used to investigate associations between predictors and the outcome. RESULT: Of the 481 women in the sample, 22 were identified as having abnormal laboratory test results (4.6%). Women who reported right upper quadrant pain or tenderness had significantly increased likelihood of having laboratory abnormalities compared to those without the complaint. CONCLUSION: Only a small percentage of women evaluated were determined to have abnormal laboratory findings, predominantly among women with severe preeclampsia. Right upper quadrant pain or tenderness was positively correlated with laboratory abnormalities. The restriction of laboratory analysis in women with clinical evidence of severe disease may be warranted - a broader study should, however, first be used to confirm our findings.

Does a video module improve patient adherence to recommended gestational weight gain guidelines?

OBJECTIVES: We aim to assess if implementation of an educational video module can improve patient adherence to recommended weight gain guidelines. Secondarily, we investigated if patients' knowledge about gestational weight gain was improved with use of the video, as well as if there was a difference in maternal and neonatal outcomes, and patient satisfaction. METHODS: This was an IRB-approved, prospective cohort study conducted from February 2019 to October 2019. Patients were recruited from a large academic practice during their first trimester of pregnancy. Patients in the control cohort received routine care. Patients in the video cohort watched a 5-min educational video module about gestational weight gain. Pre-pregnancy weight and baseline demographics were recorded. All patients took a baseline questionnaire assessing gestational weight gain knowledge upon enrollment, and again 4 weeks later. Pre and post score differences were calculated. On admission to the hospital for delivery, all patients' gestational weight gain was calculated, and the overall gestational weight gain differences between the two groups were calculated. Maternal and neonatal delivery outcomes were also collected. T-tests, Mann-Whitney U tests, and Chi-square analyses were used to compare groups, and a p-value of <.05 was deemed statistically significant. RESULTS: During the study period, 155 patients were recruited, with 79 in control cohort and 76 in video cohort, respectively. There was no significant difference in the percentage of patients who gained the appropriate amount of weight between the two groups; 25% (18/74) of patients in the control vs. 25% (17/68) of patients in video cohort (p = .926). There was no difference in the improvement of the pre and post assessment scores when compared between the two cohorts; the average score improvement was 1.72 ± 15.09% for the control, vs. 6.20 ± 12.51% for video cohort (p = .129). There was no difference in maternal or neonatal outcomes between the two groups. Patients were overall satisfied with the video module, with 67.6% (n = 45) reporting the video to be very educational. CONCLUSIONS: Use of a video module did not improve GWG outcomes or knowledge in our study. Future work can focus on use of a recurring intervention throughout pregnancy, either with app-based technology or multiple videos.

Improving the selection of recommended prophylactic antibiotics using an electronic medical record.

OBJECTIVE: To estimate the effect of reference text in an electronic medical record (EMR) on the recommended use of preoperative antibiotics for cesarean deliveries. METHODS: A retrospective cohort study of all cesarean deliveries at Stony Brook Hospital between May 2009 and June 2011, was performed. Reference text was added to the EMR in May 2010. The cesarean deliveries performed before (group 1) and after (group 2) the addition of the reference text were compared for the type and dose of antibiotic ordered preoperatively, as well as the patient's body mass index (BMI), drug allergies, and diagnosis of chorioamnionitis. Those with chorioamnionitis were excluded. The χ and t tests were performed to study differences between the groups. RESULTS: Of the 2,273 deliveries identified, 172 were excluded for chorioamnionitis or incomplete medical records. The women in group 1 had a higher BMI (33.3 ± 7.3 kg/m²) compared with the women in group 2 (32.6 ± 6.3 kg/m²; P=.005). Otherwise, there were no differences between the groups regarding maternal weight, penicillin allergy, or other antibiotics. There was a significant increase in the number of patients receiving the adequate (92.6% compared with 85.7%; P<.005) and recommended (58.1% compared with 43.4%; P<.005) antibiotics in group 2 when compared with group 1. CONCLUSION: Reference text included within the EMR that prompts physicians to select the recommended antibiotics significantly improves preoperative antibiotic use.

Control of simian/human immunodeficiency virus viremia and disease progression after IL-2-augmented DNA-modified vaccinia virus Ankara nasal vaccination in nonhuman primates.

A successful HIV vaccine may need to stimulate antiviral immunity in mucosal and systemic immune compartments, because HIV transmission occurs predominantly at mucosal sites. We report here the results of a combined DNA-modified vaccinia virus Ankara (MVA) vaccine approach that stimulated simian/human immunodeficiency virus (SHIV)-specific immune responses by vaccination at the nasal mucosa. Fifteen male rhesus macaques, divided into three groups, received three nasal vaccinations on day 1, wk 9, and wk 25 with a SHIV DNA plasmid producing noninfectious viral particles (group 1), or SHIV DNA plus IL-2/Ig DNA (group 2), or SHIV DNA plus IL-12 DNA (group 3). On wk 33, all macaques were boosted with rMVA expressing SIV Gag-Pol and HIV Env 89.6P, administered nasally. Humoral responses were evaluated by measuring SHIV-specific IgG and neutralizing Abs in plasma, and SHIV-specific IgA in rectal secretions. Cellular responses were monitored by evaluating blood-derived virus-specific IFN-gamma-secreting cells and TNF-alpha-expressing CD8+ T cells, and blood- and rectally derived p11C tetramer-positive T cells. Many of the vaccinated animals developed both mucosal and systemic humoral and cell-mediated anti-SHIV immune responses, although the responses were not homogenous among animals in the different groups. After rectal challenge of vaccinated and naive animals with SHIV89.6P, all animals became infected. However a subset, including all group 2 animals, were protected from CD4+ T cell loss and AIDS development. Taken together, these data indicate that nasal vaccination with SHIV-DNA plus IL-2/Ig DNA and rMVA can provide significant protection from disease progression.