- Research Article
- 10.1038/s44360-026-00098-x
Human MERS-CoV cases are falling but pose an ongoing pandemic threat
- Mar 19, 2026
- Nature Health
- Lorenzo Subissi + 36 more +36
Publications from 2021 to 2026
Showing 10 of 137 papers
Human MERS-CoV cases are falling but pose an ongoing pandemic threat
Elimination of visceral leishmaniasis as a public health problem in Bangladesh: Lessons learned and questions remaining.
In 2023, Bangladesh became the first country to achieve World Health Organization (WHO) validation of elimination of visceral leishmaniasis as a public health problem, defined as maintenance of annual kala-azar incidence at <1 case per 10,000 population at the subdistrict (upazila) level. The pillars of the programme are early diagnosis and effective treatment, indoor residual insecticide spraying, improved case detection, social mobilization and operational research, and effective disease surveillance. The Bangladesh National Kala-azar Elimination Programme was established in 2008, with introduction of rapid diagnostics and newer treatment modalities in health complexes at sub-district level in the endemic area in 2012-2015, initiation of blanket IRS in affected communities in 2012-2013 and adoption of a digital surveillance system in 2015. All subdistricts achieved and maintained the elimination threshold from 2017 onward. We present documentation of the course of KA elimination in Bangladesh and provide a perspective on the components necessary to maintain current success into the future.
Read moreEvaluation of Sustainable Agroecological Practices Based on the Application of Organic Substrates (Cocodust) in Vegetable Seedlings at the Nursery Level in the Producing Areas of Bangladesh
Seedling quality is a critical determinant of vegetable crop productivity, yet conventional soil-based nurseries in Bangladesh are frequently associated with high seedling mortality, soil-borne diseases, and excessive agrochemical use. This study assessed the agronomic performance and adoption potential of cocodust-based seedling nurseries compared with traditional soil-based practices across eight major vegetable crops in northern Bangladesh. Using a farm-level comparative mixed-methods design, data were collected from 80 vegetable farmers selected through stratified random sampling and grouped into cocodust users and traditional nursery users. Crop-wise differences in yield, harvesting time, and seedling mortality were analyzed using independent two-sample Welch’s t-tests. Cocodust application significantly increased yields of bitter gourd (17%), bottle gourd (18%), brinjal (23%), chili (23%), country bean (44%), and tomato (25%), while yield increases in cabbage (15%) and cauliflower (12%) were not statistically significant. Across all crops, seedling mortality was reduced by 68–82% under cocodust-based nurseries. Harvesting time was significantly shortened for all crops, with reductions of 8–15% in most cases and reaching 43–47% in chili and country bean. Cocodust use was associated with significantly reduced chemical fertilizer application in bitter gourd and marginally in tomato, and with reduced pesticide use in bitter gourd, country bean, and tomato, while organic fertilizer use remained high irrespective of nursery practice. Incidence of soil-borne diseases was significantly lower in bottle gourd and cauliflower under cocodust-based nurseries. Overall, cocodust-based seedling nurseries improved vegetable productivity, reduced early-stage seedling losses, shortened production cycles, and lowered reliance on chemical inputs in selected crops, highlighting cocodust as a low-cost and environmentally sustainable option for smallholder vegetable production systems.
Read moreNutrition-Sensitive Homestead Pond Polyculture Technology Empowers Women in Rural Bangladesh
Women’s empowerment in aquatic food systems plays a critical role in increasing production and ensuring food and nutrition security. This study used qualitative data from a mixed-method study to assess whether participation in a nutrition-sensitive homestead pond polyculture systems, targeting both women and men, strengthens women’s decision-making power in fish culture, harvesting, and household food purchase. The project promoted pond polyculture technology of nutrient-rich small and large fish species, which was adopted by the majority of project households (86.3%). Fish harvesting for family consumption and sale was more frequent in project households (24% twice/week; 54.9% once/year) compared to non-project households (13%; 47.8%). Joint decision-making by both women and men on the types of fish species to culture (80.39%) and harvesting time for sale (54.9%) was more common in project households, compared to non-project households (43.48% and 21.74%, respectively). More women in project households (41.2%) decided which fish species to harvest for family consumption, including harvested fish for family consumption (7.8%) and sale (3.9%), compared to non-project households (26.1%, 4.4% and none respectively). The findings reveal three interlinked empowerment pathways: the consumption pathway, via frequent fish harvesting, the income pathway, through fish sales, and the empowerment pathway, through enhanced decision making.
Read moreImmediate or early skin‐to‐skin contact for mothers and their healthy newborn infants
Review Update (2016) https://doi.org/10.1002/14651858.CD003519.pub4 Review Update (2012) https://doi.org/10.1002/14651858.CD003519.pub3 Review Update (2007) https://doi.org/10.1002/14651858.CD003519.pub2 Original review (2003) https://doi.org/10.1002/14651858.CD003519 Protocol (2002) DOI unavailable.
Read morePublic Mindset Towards Bangladeshi Celebrities: A Dual-analysis of Ensemble and Non-Ensemble Models of NLP on Facebook Comments
The paper proposes sentiment analysis model for assessing the Bangladeshi celebrity sentiments in Facebook Bangla comments involving state-of-the-art machine learning and deep learning models, and also an ensemble multi-channel system of machine learning and deep learning models. We labeled 6,596 gender balanced comments (5,817 female, 493 male) with sentiment V3 (three labels, positivity, negativity, neutrality) and seven reaction categories. The utilization of a combination of feature extraction options (TF-IDF, unigram, bigram, trigram, Word2Vec) addresses the linguistic complexity of the Bangla language. The suggested system utilizes eight machine learning and deep learning algorithms (Logistic Regression, Random Forest, K-Nearest Neighbors, Linear SVM, XGBoost, AdaBoost, Decision Tree, Artificial Neural Network) and standalone multi-channel models (one ensemble of CNN, LSTM, GloVe and another combination of LR, RF and SVM). This paper will produce the following contributions: (1) Gender-disaggregated study of celebrity perception in Bangladesh social networks (2) Comparative analysis among individual machine learning and deep learning models with balanced dataset using resampling technique and imbalanced dataset, and ensemble multi-channel systems with 80-10-10 split to comprehend the public mindset in Bangla, (3) Logistic Regression, KNN, and ANN achieved the top accuracies respectively, with Linear SVM showing 100% recall across all n-grams for both female and Male dataset: LR 75.93%, KNN 76.91%, ANN 77.57%, SVM 100% recall; CNN-LSTM-GloVe 71.88%, LR-RF-SVM 73%. (4) An ensemble using CNN, LSTM, and GloVe reached 72% accuracy, while a LR–RF–SVM combination with 5-fold cross-validation achieved 73% for the female dataset and the CNN-LSTM-GloVe ensemble achieved 56% accuracy, while the LR-RF-SVM ensemble reached 64% for the male dataset. (5) After resampling, SVM performed top-notch for the male dataset, and the hybrid model, which includes LR, RF, SVM, CNN and LSTM, also shows significant performance for the female dataset.
Read moreserocalculator, an R package for estimating seroincidence from cross-sectional serological data
Motivation:Seroincidence—the rate of new infections in a population—is a key measure for understanding pathogen transmission dynamics and informing public health action. Estimating seroincidence from cross-sectional data is complicated by antibody waning, cross-reactivity, and individual heterogeneity in antibody responses.Implementation:serocalculator is an open-source R package that uses a likelihood-based framework incorporating antibody decay models, biological variability, and measurement noise to estimate seroincidence rates from cross-sectional serological data.General features:The package supports overall and stratified seroincidence estimation using single or multiple biomarkers. It requires three inputs: (1) a pre-estimated seroresponse model characterizing post-infection antibody waning; (2) noise parameters capturing biological and assay-related variability; and (3) quantitative antibody responses from a cross-sectional survey. It is computationally efficient, well-documented, and includes a point-and-click R Shiny interface. These features promote usability across research and public health.Availability:The package serocalculator is freely available on CRAN, with development versions on GitHub.
Read morePopulation pharmacokinetics of artemether–lumefantrine plus amodiaquine in patients with uncomplicated Plasmodium falciparum malaria
AimsResistance to the artemisinins and the artemisinin‐based combination therapy (ACT) partner drugs has developed in Southeast Asia, and artemisinin resistance has also emerged in eastern Africa. Triple ACTs (triple artemisinin‐based combination therapies, TACT), consisting of two partner drugs with different mechanisms of action and similar pharmacokinetic profiles, combined with an artemisinin derivative can help to delay or prevent artemisinin resistance and prolong the useful lifetime of the partner drugs. This study aims to characterize the pharmacokinetic properties of a recommended TACT, artemether‐lumefantrine plus amodiaquine, using data from two large clinical trials.MethodsWe analysed data from two randomized, controlled intervention trials conducted between 2015 and 2020 in one African country and two Southeast Asian countries, in which artemether‐lumefantrine was administered alone (n = 443) or together with amodiaquine (n = 442) to patients with uncomplicated P. falciparum malaria. Both studies included a sub‐cohort with dense pharmacokinetic sampling, combined with sparse data in the other patients. Concentration–time data of artemether, dihydroartemisinin, lumefantrine, desbutyllumefantrine, amodiaquine and desethylamodiaquine were analysed using nonlinear mixed‐effects modelling.ResultsPharmacokinetic models were developed for all drugs and demonstrated good predictive performance and goodness‐of‐fit diagnostics. Coadministered amodiaquine was not a significant covariate on pharmacokinetic properties of artemether‐lumefantrine. Model‐predicted Cmax and AUC (median [95% confidence interval, CI]) for artemether were 256 (159–407) ng/mL and 2850 (1820‐4920) h·ng/mL for artemether‐lumefantrine alone, and 230 (123–391) ng/mL and 2800 (1570‐4570) h·ng/mL for artemether‐lumefantrine‐amodiaquine. For dihydroartemisinin, values were 135 (54.5–214) ng/mL and 1870 (813–3015) h·ng/mL for artemether‐lumefantrine alone, and 116 (40.8–186) ng/mL and 1580 (547–2680) h·ng/mL for artemether‐lumefantrine‐amodiaquine. For lumefantrine, values were 15.2 (2.90–31.3) μg/mL and 600 (275–1230) h·μg/mL for artemether‐lumefantrine alone, and 14.1 (2.72–31.4) μg/mL and 586 (269–1070) h·μg/mL for artemether‐lumefantrine‐amodiaquine. Day 7 concentrations of lumefantrine were 452 (215–1240) and 438 (204–1030) μg/mL for artemether‐lumefantrine alone and artemether‐lumefantrine‐amodiaquine, respectively. All geometric mean ratios (GMRs) for the drug–drug interaction (DDI) effect on key pharmacokinetic parameters of artemether, dihydroartemisinin and lumefantrine fell within the 0.80–1.25 range, with the majority of the corresponding 90% CI also contained within this range. This indicates no clinically relevant DDIs between artemether‐lumefantrine and amodiaquine.ConclusionsThe DDI effect of amodiaquine on the pharmacokinetics of artemether‐lumefantrine is expected to be minimal, the based on the current analysis. However, further large‐scale clinical trials are needed to confirm this finding.
Read moreEffect of maternal infection on stillbirths and early neonatal deaths: nested case-control studies in pregnancy cohorts in East Africa
Summary Background Reducing perinatal deaths is a priority, but data on maternal infectious causes are sparse in low- and middle-income countries where the burden is highest. We aimed to describe maternal infections at delivery and their association with perinatal death in Kilifi County Hospital (KCH), Kenya and Hiwot Fana Comprehensive Specialised Hospital (HFCSH), Ethiopia. Methods We investigated 642 mothers delivering stillbirths/newborns dying in the first 24h after birth (cases) and 855 mothers with newborns surviving > 24h (controls), from well-characterised pregnancy cohorts in a nested case-control design in KCH (2011-17) retrospectively and HFCSH (2019-20) prospectively. We tested maternal blood for infection at delivery using molecular methods with 60 PCR targets (TaqMan Array Cards, TAC). In HFCSH, vagino-rectal swabs (VRS) and oropharyngeal swabs (OPS) were also tested, with 28 and 42 PCR targets respectively, along with conventional microbiological testing. We tested associations between maternal infection and perinatal death for each site, separately, and combined, and adjusted for potential confounders. We did a sensitivity analysis using only controls with good pregnancy outcomes in KCH. Where appropriate, we calculated the population attributable fraction (PAF). Results In HFCSH, maternal bacteraemia was associated with perinatal death (adjusted odds ratio aOR3.7 [1.5-9.2]). In KCH, bacterial detection in maternal blood was associated with perinatal death (aOR2.7 [1.2-6.0]), but only in sensitivity analysis. Escherichia coli /Shigella was associated with perinatal death when cultured/detected in blood (aOR2.6 [1.1-6.3]) across sites. Though infrequent, Bordetella sp . was associated with perinatal death (OR4.9[1.1-23.0]) on OPS in HFCSH. No other individual infections were associated with perinatal death. The PAF for perinatal deaths among hospital deliveries was 6.1% (4.0%-8.2%) for maternal bacteraemia in HFCSH and 4.9% (2.6%-7.2%) for bacterial detection in KCH. Conclusions Maternal bacterial infection is associated with perinatal death in high-burden settings, and in our study accounted for around 5% of perinatal deaths in hospital deliveries. The study was underpowered to detect species-specific infections associated with perinatal death. Funding Wellcome (205184)
Read moreSocial and healthcare-seeking experiences of people affected with lymphedema in Bangladesh.
Neglected tropical diseases (NTDs) such as lymphatic filariasis (LF) are a significant concern in developing countries like Bangladesh. Understanding the health and social needs of individuals with LF is essential for improving their healthcare-seeking experiences and advancing the healthsystem'scapacity. Therefore, this qualitative study aimed to explore the social and care-seeking experiences of persons with LF in Bangladesh. Semi-structured, face-to-face, in-depth interviews were conducted with people with LF in two highly endemic districts in northern Bangladesh. Online key informant interviews were also conducted among stakeholders associated with NTD care. Recorded interviews were analysed using thematic analysis. 28 participants (20 with LF and 8 stakeholders) were interviewed, and five major themes emerged after analysis. In terms of disease-related knowledge, the participants perceived lymphedema as a condition characterised by swelling, pain, and fever, which they believed was caused by eating stale food, being infected by others, or being punished by God. Overall, females in particular shared their experiences of negative attitudes from the community. LF adversely affected their daily life, including mental health and well-being. Most respondents sought support from conventional healthcare services; however, their perception of incurable disease led to low medication adherence and dissatisfaction. Lack of knowledge, inaccessibility of healthcare services, financial challenges, and physical disability were major barriers to seeking care. Participants emphasised the importance of financial assistance, community awareness, enhancing the accessibility and quality of care, and occupational rehabilitation scopes with governmental aid. Our findings highlighted the importance of ensuring an accessible and affordable healthcare infrastructure for people with LF. Additionally, the involvement of government and related stakeholders is essential to improve service users' experiences and attain high standards, combined with the need for inclusive well-being-related services. Concentrated efforts should be made to design culturally acceptable interventions to raise awareness and reduce stigma.
Read more