Machine learning for maternal health: Predicting delivery location in a community health worker program in Zanzibar.

Background: Maternal and neonatal health outcomes in low- and middle-income countries (LMICs) have improved over the last two decades. However, many pregnant women still deliver at home, which increases the health risks for both the mother and the child. Community health worker programs have been broadly employed in LMICs to connect women to antenatal care and delivery locations. More recently, employment of digital tools in maternal health programs have resulted in better care delivery and served as a routine mode of data collection. Despite the availability of rich, patient-level data within these digital tools, there has been limited utilization of this type of data to inform program delivery in LMICs. Methods: We use program data from 38,787 women enrolled in Safer Deliveries, a community health worker program in Zanzibar, to build a generalizable prediction model that accurately predicts whether a newly enrolled pregnant woman will deliver in a health facility. We use information collected during the enrollment visit, including demographic data, health characteristics and current pregnancy information. We apply four machine learning methods: logistic regression, LASSO regularized logistic regression, random forest and an artificial neural network; and three sampling techniques to address the imbalanced data: undersampling of facility deliveries, oversampling of home deliveries and addition of synthetic home deliveries using SMOTE. Results: Our models correctly predicted the delivery location for 68%-77% of the women in the test set, with slightly higher accuracy when predicting facility delivery versus home delivery. A random forest model with a balanced training set created using undersampling of existing facility deliveries accurately identified 74.4% of women delivering at home. Conclusions: This model can provide a "real-time" prediction of the delivery location for new maternal health program enrollees and may enable early provision of extra support for individuals at risk of not delivering in a health facility, which has potential to improve health outcomes for both mothers and their newborns. The framework presented here is applicable in other contexts and the selection of input features can easily be adapted to match data availability and other outcomes, both within and beyond maternal health.

Characterizing Extracellular Vesicles and Particles Derived from Skeletal Muscle Myoblasts and Myotubes and the Effect of Acute Contractile Activity.

Extracellular vesicles (EVs), released from all cells, are essential to cellular communication and contain biomolecular cargo that can affect recipient cell function. Studies on the effects of contractile activity (exercise) on EVs usually rely on plasma/serum-based assessments, which contain EVs from many different cells. To specifically characterize skeletal muscle−derived vesicles and the effect of acute contractile activity, we used an in vitro model where C2C12 mouse myoblasts were differentiated to form myotubes. EVs were isolated from conditioned media from muscle cells at pre-differentiation (myoblasts) and post-differentiation (myotubes) and also from acutely stimulated myotubes (1 h @ 14 V, C-Pace EM, IonOptix, Westwood, MA, USA) using total exosome isolation reagent (TEI, ThermoFisher (Waltham, MA, USA), referred to as extracellular particles [EPs]) and differential ultracentrifugation (dUC; EVs). Myotube-EPs (~98 nm) were 41% smaller than myoblast-EPs (~167 nm, p < 0.001, n = 8−10). Two-way ANOVA showed a significant main effect for the size distribution of myotube vs. myoblast-EPs (p < 0.01, n = 10−13). In comparison, myoblast-EPs displayed a bimodal size distribution profile with peaks at <200 nm and 400−600, whereas myotube-Eps were largely 50−300 nm in size. Total protein yield from myotube-EPs was nearly 15-fold higher than from the myoblast-EPs, (p < 0.001 n = 6−9). Similar biophysical characteristics were observed when EVs were isolated using dUC: myotube-EVs (~195 nm) remained 41% smaller in average size than myoblast-EVs (~330 nm, p = 0.07, n = 4−6) and had comparable size distribution profiles to EPs isolated via TEI. Myotube-EVs also had 4.7-fold higher protein yield vs. myoblast EVs (p < 0.05, n = 4−6). Myotube-EPs exhibited significantly decreased expression of exosomal marker proteins TSG101, CD63, ALIX and CD81 compared with myoblast-EPs (p < 0.05, n = 7−12). Conversely, microvesicle marker ARF6 and lipoprotein marker APO-A1 were only found in the myotube-EPs (p < 0.05, n = 4−12). There was no effect of acute stimulation on myotube-EP biophysical characteristics (n = 7) or on the expression of TSG101, ARF6 or CD81 (n = 5−6). Myoblasts treated with control or acute stimulation−derived EPs (13 µg/well) for 48 h and 72 h showed no changes in mitochondrial mass (MitoTracker Red, ThermoFisher, Waltham, MA, USA), cell viability or cell count (n = 3−4). Myoblasts treated with EP-depleted media (72 h) exhibited ~90% lower cell counts (p < 0.01, n = 3). Our data show that EVs differed in size, distribution, protein yield and expression of subtype markers pre vs. post skeletal muscle−differentiation into myotubes. There was no effect of acute stimulation on biophysical profile or protein markers in EPs. Acute stimulation−derived EPs did not alter mitochondrial mass or cell count/viability. Further investigation into the effects of chronic contractile activity on the biophysical characteristics and cargo of skeletal muscle−specific EVs are warranted.

Extracellular Vesicles as Predictors of Individual Response to Exercise Training in Youth Living with Obesity.

BACKGROUND: Exercise is associated with health benefits, including the prevention and management of obesity. However, heterogeneity in the adaptive response to exercise training exists. Our objective was to evaluate if changes in extracellular vesicles (EVs) after acute aerobic exercise were associated with the responder phenotype following 6-weeks of resistance training (RT). METHODS: This is a secondary analysis of plasma samples from the EXIT trial (clinical trial#02204670). Eleven sedentary youth with obesity (15.7 ± 0.5 yrs, BMI ≥95th percentile) underwent acute exercise (60% HRR, 45 min). Blood was collected at baseline [AT0 min], during [AT15-45 min], and 75 min post-recovery [AT120], and EVs purified using size exclusion chromatography from extracted plasma. Afterward, youth participated in 6-weeks RT and were categorized into responders or non-responders based on changes in insulin sensitivity. RESULTS: We assessed EV biophysical profile (size, zeta potential, protein yield, and EV subtype protein expression) in a single-blind fashion. Overall, there was a general increase in EV production in both groups. Average EV size was larger in responders (~147 nm) vs. non-responders (~124 nm; p < 0.05). EV size was positively associated with absolute change in Matsuda index (insulin sensitivity) following RT (r = 0.44, p = 0.08). EV size distribution revealed responders predominantly expressed EVs sized 150-300 nm, whereas non-responders expressed EVs sized 50-150 nm (p < 0.05). At baseline, responders had ~25% lower TSG101, ~85% higher MMP2 levels. EV protein yield was higher in responders than non-responders at AT15 (p < 0.05). CONCLUSIONS: Our data suggest that youth with obesity that respond to RT produce larger EVs that are TSG101+ and CD63+, with increased EV protein yield during acute exercise.

Improving health facility delivery rates in Zanzibar, Tanzania through a large-scale digital community health volunteer programme: a process evaluation.

The utilization of community health worker (CHW) programmes to improve maternal and neonatal health outcomes has become widely applied in low- and middle-income countries. While current research has focused on discerning the effect of these interventions, documenting the process of implementing, scaling and sustaining these programmes has been largely ignored. Here, we focused on the implementation of the Safer Deliveries CHW programme in Zanzibar, a programme designed to address high rates of maternal and neonatal mortality by increasing rates of health facility delivery and postnatal care visits. The programme was implemented and brought to scale in 10 of 11 districts in Zanzibar over the course of 3 years by D-tree International and the Zanzibar Ministry of Health. As the programme utilized a mobile app to support CHWs during their visits, a rich data resource comprised of 133 481 pregnancy and postpartum home visits from 41 653 women and 436 CHWs was collected, enabling the evaluation of numerous measures related to intervention fidelity and health outcomes. Utilizing the framework of Steckler et al., we completed a formal process evaluation of the primary intervention, CHW home visits to women during their pregnancy and postpartum period. Our in-depth analysis and discussion will serve as a model for process evaluations of similar CHW programmes and will hopefully encourage future implementers to report analogous measures of programme performance.

Does intrauterine saline infusion by intrauterine insemination (IUI) catheter as endometrial injury during IVF cycles improve pregnancy outcomes among patients with recurrent implantation failure?: An RCT.

BACKGROUND: Recurrent implantation failure is one of the most issues in IVF cycles. Some researchers found that beneficial effects of endometrial Scratching in women with recurrent implantation failure, while some authors demonstrated contrary results. OBJECTIVE: The present study aimed to investigate the effect of intrauterine. Saline infusion as a form of endometrial injury, during fresh in vitro fertilization-embryo transfer cycle, among patients with recurrent implantation failure. MATERIALS AND METHODS: In this clinical trial study 63 women undergoing assisted reproductive technology were divided into two groups either local endometrial injury by intrauterine saline infusion during day 3-5 of the ongoing controlled ovarian stimulation cycle, or IVF protocol performed without any other intervention in Taleghani Hospital, Tehran, Iran. The main outcome measure was clinical pregnancy rates. RESULTS: Patients who received intra uterine saline infusion (n=20), had significantly lower clinical pregnancy numbers (1 vs. 9, p<0.05) and implantation rates (4.7% vs. 41.6%, p<0.05), compared to controls (n=39). However, there was no significant difference in miscarriage rates (9.4% vs. 8.7%, p>0.05) and multiple pregnancy numbers (1 vs. 3, p>0.05) between groups. CONCLUSION: When intrauterine saline infusion as a form of endometrial injury is performed during the ongoing IVF cycles it has negative effect on reproductive outcomes among patients with recurrent implantation failure.

A comparative study of dydrogesterone and micronized progesterone for luteal phase support during in vitro fertilization (IVF) cycles.

The aim of the present study was to compare the efficacy, tolerability and patients' satisfaction after the use of oral dydrogesterone with vaginal micronized progesterone for luteal-phase support (LPS) among infertile women undergoing in vitro fertilization (IVF). A total of 210 women (aged 20-40 years old) with a history of infertility, who underwent controlled ovarian stimulation for fresh intra-cytoplasmic sperm injection-embryo transfer cycles, were included in the study. Consequently, they were randomized to receive LPS with dydrogesterone 20 mg twice daily (n = 96) or micronized progesterone 400 mg twice daily at the day of oocyte retrieval (n = 114). The clinical success rate (31% versus 33%; p = 0.888), miscarriage rate (5.0% versus 3.0%; p = 0.721), ongoing pregnancy rate (30.0% versus 30.0%; p = 1.000), implantation (22.0% versus 24.0%; p = 0.254) and multiple pregnancy rate (5.30% versus 7.20%; p = 0.394) were comparable among the two groups. Serum progesterone levels were significantly lower among the patients receiving dydrogesterone than the control group (13.62 ± 13.83 ng/ml versus 20.66 ± 18.09 ng/ml; p = 0.001). However, there was no statistically significant difference regarding the patients' satisfaction (p = 0.825) and tolerability (0.790) between the two groups. Our results showed that oral dydrogesterone (40 mg/day) is as effective as vaginal micronized progesterone considering its clinical outcomes and patients' satisfaction and tolerability, for LPS among women undergoing IVF.

EGFP reporter protein: its immunogenicity in Leishmania-infected BALB/c mice.

Optical reporter genes such as green fluorescent protein (GFP) and luciferase are efficiently and widely used in monitoring and studying the protective/therapeutic potential of candidate agents in leishmaniasis. But several observations and controversial reports have generated a main concern, whether enhanced GFP (EGFP) affects immune response. To address this issue, we studied the immunogenicity of EGFP in vivo by two lines of stably transfected parasites (Leishmania major (EGFP) or L. major (EGFP-LUC)) in BALB/c model and/or as a recombinant protein (rEGFP) produced in vitro by bacteria in parallel. Disease progression was followed by footpad swelling measurements and parasite burden in draining lymph nodes using microtitration assay and real-time PCR, and immune responses were also evaluated in spleen. EGFP-expressing parasites generated larger swellings in comparison with wild-type (L. major) while mice immunized with rEGFP and challenged with wild-type parasite were quite comparable in footpad swelling with control group without significant difference. However, both conventional and molecular approaches revealed no significant difference in parasite load between different groups. More importantly, no significant inflammatory responses were detected in groups with higher swelling size measured by interferon-γ (IFN-γ), interleukin (IL)-10, IL-5, and nitric oxide against frozen and thawed lysate of parasite as stimulator. Altogether, these results clearly revealed that EGFP protein expressed in prokaryotic and eukaryotic hosts is not an immunological reactive molecule and acts as a neutral protein without any side effects in mice. So, EGFP expressing Leishmania could be a safe and reliable substitution for wild-types that simplifies in situ follow-up and eliminates the animal scarification wherever needed during the study.