Code-switching is a ubiquitous phenomenon in multilingual communities. Natural language technologies that wish to communicate like humans must therefore adaptively incorporate code-switching techniques when they are deployed in multilingual settings. To this end, we propose a Hindi-English human-machine dialogue system that elicits code-switching conversations in a controlled setting. It uses different code-switching agent strategies to understand how users respond and accommodate to the agent's language choice. Through this system, we collect and release a new dataset COMMONDOST, comprising of 439 humanmachine multilingual conversations. We adapt pre-defined metrics to discover linguistic accommodation from users to agents. Finally, we compare these dialogues with Spanish-English dialogues collected in a similar setting, and analyze the impact of linguistic and socio-cultural factors on code-switching patterns across the two language pairs. 1

No abstract

The past decade has been notable for widespread dissemination of newborn resuscitation training in low-resource settings through simplified training programs including Helping Babies Breathe. Since 2020, implementation efforts have been impacted by restrictions on travel and in-person gatherings with the SARS-CoV-2 pandemic, prompting the development of alternative methods of training. While previous studies have demonstrated feasibility of remote neonatal resuscitation training, this perspective paper covers common barriers identified and key lessons learned developing a cadre of remote facilitators. Challenges of remote facilitation include mastering videoconferencing platforms, establishing personal connections, and providing effective oversight of skills practice. Training sessions can be used to support facilitators in acquiring comfort and competency in harnessing videoconferencing platforms for effective facilitation. Optimization of approaches and investment in capacity building of remote facilitators are imperative for effective implementation of remote neonatal resuscitation training.

Background Early Neonatal mortality (ENM) (< 7 days) remains a significant problem in low resource settings. Birth asphyxia (BA), prematurity and presumed infection contribute significantly to ENM. The study objectives were to determine: first, the overall ENM rate as well as yearly ENM rate (ENMR) from 2015 to 2019; second, the influence of decreasing GA (< 37 weeks) and BW (< 2500 g) on ENM; third, the contribution of intrapartum and delivery room factors and in particular fetal heart rate abnormalities (FHRT) to ENM; and fourth, the Fresh Still Birth Rates (FSB) rates over the same time period. Methods Retrospective cohort study undertaken in a zonal referral teaching hospital located in Northern Tanzania. Labor and delivery room data were obtained from 2015 to 2019 and included BW, GA, fetal heart rate (FHRT) abnormalities, bag mask ventilation (BMV) during resuscitation, initial temperature, and antenatal steroids use. Abnormal outcome was ENM < 7 days. Analysis included t tests, odds ratios (OR), and multivariate regression analysis. Results The overall early neonatal mortality rate (ENMR) was 18/1000 livebirths over the 5 years and did not change significantly comparing 2015 to 2019. Comparing year 2018 to 2019, the overall ENMR decreased significantly (OR 0.62; 95% confidence interval (CI) 0.45–0.85) as well as infants ≥37 weeks (OR 0.45) (CI 0.23–0.87) and infants < 37 weeks (OR 0.57) (CI 0.39–0.84). ENMR was significantly higher for newborns < 37 versus ≥37 weeks, OR 10.5 (p < 0.0001) and BW < 2500 versus ≥2500 g OR 9.9. For infants < 1000 g / < 28 weeks, the ENMR was ~ 588/1000 livebirths. Variables associated with ENM included BW - odds of death decreased by 0.55 for every 500 g increase in weight, by 0.89 for every week increase in GA, ENMR increased 6.8-fold with BMV, 2.6-fold with abnormal FHRT, 2.2-fold with no antenatal steroids (ANS), 2.6-fold with moderate hypothermia (all < 0.0001). The overall FSB rate was 14.7/1000 births and decreased significantly in 2019 when compared to 2015 i.e., 11.3 versus 17.3/1000 live births respectively (p = 0.02). Conclusion ENM rates were predominantly modulated by decreasing BW and GA, with smaller/ less mature newborns 10-fold more likely to die. ENM in term newborns was strongly associated with FHRT abnormalities and when coupled with respiratory depression and BMV suggests BA. In smaller newborns, lack of ACS exposure and moderate hypothermia were additional associated factors. A composite perinatal approach is essential to achieve a sustained reduction in ENMR.