No abstract

Background Infants are at highest risk of pneumococcal disease. Their added protection through herd effects is a key part in the considerations on optimal pneumococcal vaccination strategies. Yet, little is currently known about the main transmission pathways to this vulnerable age group. Hence, this study investigates pneumococcal transmission routes to infants in the coastal city of Nha Trang, Vietnam. Methods and findings In October 2018, we conducted a nested cross-sectional contact and pneumococcal carriage survey in randomly selected 4- to 11-month-old infants across all 27 communes of Nha Trang. Bayesian logistic regression models were used to estimate age specific carriage prevalence in the population, a proxy for the probability that a contact of a given age could lead to pneumococcal exposure for the infant. We used another Bayesian logistic regression model to estimate the correlation between infant carriage and the probability that at least one of their reported contacts carried pneumococci, controlling for age and locality. In total, 1,583 infants between 4 and 13 months old participated, with 7,428 contacts reported. Few infants (5%, or 86 infants) attended day care, and carriage prevalence was 22% (353 infants). Most infants (61%, or 966 infants) had less than a 25% probability to have had close contact with a pneumococcal carrier on the surveyed day. Pneumococcal infection risk and contact behaviour were highly correlated: If adjusted for age and locality, the odds of an infant’s carriage increased by 22% (95% confidence interval (CI): 15 to 29) per 10 percentage points increase in the probability to have had close contact with at least 1 pneumococcal carrier. Moreover, 2- to 6-year-old children contributed 51% (95% CI: 39 to 63) to the total direct pneumococcal exposure risks to infants in this setting. The main limitation of this study is that exposure risk was assessed indirectly by the age-dependent propensity for carriage of a contact and not by assessing carriage of such contacts directly. Conclusions In this study, we observed that cross-sectional contact and infection studies could help identify pneumococcal transmission routes and that preschool-age children may be the largest reservoir for pneumococcal transmission to infants in Nha Trang, Vietnam.

Background: Infants are at highest risk of pneumococcal disease. Their added protection through herd effects is a key part in the considerations on optimal pneumococcal vaccination strategies. Yet, little is currently known about the main transmission pathways to this vulnerable age group. Methods and findings: We conducted a nested cross-sectional contact and nasopharyngeal swabbing survey in randomly selected infants across all 27 communes of Nha Trang, Vietnam. Bayesian logistic regression models were used to estimate age specific carriage prevalence in the population, a proxy for the probability that a contact of a given age could lead to pneumococcal exposure for the infant. We used another Bayesian logistic regression model to estimate the correlation between infant carriage and the probability that at least one of their reported contacts carried pneumococci, controlling for age and locality. In total 1583 infants between 4 and 13 months old participated, with 7428 contacts reported. Few infants (5%) attended day care and carriage prevalence was 22%. Most infants (61%) had less than a 25% probability to have had close contact with a pneumococcal carrier on the surveyed day. Pneumococcal infection risk and contact behaviour were highly correlated: if adjusted for age and locality the odds of an infant's carriage increased by 22% (95%CI:15-29) per 10 percentage points increase in the probability to have had close contact with at least one pneumococcal carrier. Two to six year old children contributed 51% (95%CI: 39-63) to the total pneumococcal exposure risks to infants in this setting. Conclusions: Cross-sectional contact and infection studies can help identify pneumococcal transmission routes. In Nha Trang, preschool age children are the largest reservoir for pneumococcal transmission to infants.

Smartphones and wearable sensors are increasingly used for personalised prediction and management in healthcare contexts. Personalisation requires tuning/learning a model of the user. However, traditional machine learning approaches for personalised modelling typically require the availability of sufficient personal data of a suitable nature for training, which can be a challenge in such contexts. We propose a parametric transfer learning approach based on the Fisher divergence to address this challenge. This makes it possible to create patientspecific models and make predictions of self-reported well-being scores, when training is performed incrementally on sparse data becoming slowly available over time. This approach allows us to make informed predictions even in the early stages of data collection, by leveraging external information coming from other patients, in the form of a prior used within a Markov-Chain Monte Carlo process. Our approach performs favourably against competing models and standard baselines, particularly when long-term forecasts are required but training data cover only a short period.

No abstract