Noise and neglect: Social-media signals expose attention gaps for dengue, chikungunya, lymphatic filariasis and kala-azar in India’s vector-borne NTDs

TL;DR

The study shows that public attention on social media for certain tropical diseases in India doesn't match their health impact, suggesting a need for better communication strategies.

Contribution

A novel digital surveillance approach combining sentiment analysis and topic modeling to assess public attention gaps for neglected tropical diseases in India.

Findings

Dengue received over half of online mentions despite other diseases having higher epidemiological burdens.

Kala-azar had minimal online visibility despite being endemic.

Public sentiment was neutral-to-positive, focusing on prevention, treatment, and vaccine news.

Abstract

Neglected tropical and vector-borne diseases, including dengue, chikungunya, lymphatic filariasis, and kala-azar, pose substantial public health burdens in India. Despite WHO recommendations for enhanced disease surveillance and targeted communication strategies, little is known about public perceptions and discussions of these diseases across digital platforms. Understanding these perceptions can guide evidence-based policy making and public health messaging.

We conducted an in silico analysis of publicly accessible social and news media data related to dengue, chikungunya, filariasis, and kala-azar in India from January 2019 to December 2023. YouTube comments and Google News headlines were systematically retrieved, pre-processed, and analysed through sentiment analysis (VADER lexicon) and Latent Dirichlet Allocation (LDA) topic modelling. Facebook and Twitter data were not included due to Application programming interface (API) restrictions and their current subscription-based models, limiting free access even for research purposes. We visualized disease-specific digital attention in comparison to epidemiological burden and created chord, Sankey, and network diagrams to elucidate thematic and sentiment-based interactions.

Across keyword-matched items (n = 330), dengue accounted for 173 (~52%) and also had the highest mean annual reported burden (163,679 cases/year; 2019–2023). Lymphatic filariasis showed disproportionately high attention (106 items/mentions vs 3,060 reported cases/year), while kala-azar had minimal visibility (5 items; none on YouTube). Sentiment was overall neutral-to-positive, with Google News more neutral and YouTube more positive. Topics emphasized outbreak alerts, public-health campaigns, and prevention/treatment, with recurring vaccine/innovation themes.

Our study presents a novel approach combining digital surveillance, sentiment analysis, and topic modelling to provide insights into public perceptions of NTDs in India. The observed mismatch between epidemiological burden and online attention underscores the need for strategic public health messaging, aligning with WHO recommendations for community engagement and tailored disease-awareness campaigns. This research provides a valuable tool for policymakers to enhance the effectiveness of communication strategies and improve targeted intervention planning for neglected tropical diseases in India.

Neglected tropical diseases (NTDs) including dengue, chikungunya, lymphatic filariasis, and kala-azar continue to afflict millions across India, yet public attention and conversation around them remain inconsistent. We examined more than 45,000 YouTube comments and 270 Google News reports posted between January 2019 and December 2023 to see how these four NTDs are discussed online. After automated text cleaning, VADER sentiment scoring and Latent Dirichlet Allocation topic modelling, we overlaid the resulting tone-and-topic maps on official disease-burden data. Dengue dominated the discussion accounting for well over half of all references, whereas kala-azar, though still endemic, drew scarcely any notice. Overall sentiment skewed neutral-to-positive and focused largely on prevention, treatment and vaccine news. Interactive bubble maps, Sankey flows and chord diagrams vividly exposed the gap between epidemiological need and digital attention. We could not analyse Facebook or Twitter because their new, pay-walled APIs make large-scale data collection prohibitively expensive for researchers, underscoring a growing obstacle for digital epidemiology. Our reproducible, low-cost workflow highlights which NTDs are being overlooked online, providing Indian health authorities with actionable evidence and supporting the World Health Organization’s call for stronger community engagement in the fight against NTDs.

Figures