EUVIMEDEUVIMED entdecken
CrossrefInfokarte

Monitoring Canada’s Health Care Polycrisis Through News Headline Analysis: Infodemiology Study

John Weirstrass Muteba Mwamba, Vijay Mago

JMIR Infodemiology · 2026-10-29 · 6 · e99655

DOI: 10.2196/99655 ↗
AbstractBibliografische Angaben

Abstract

Background The health care polycrisis in Canada comprises several concurrent crises, namely workforce shortages; food insecurity; housing stress; and health inequities affecting Indigenous peoples, low-income families, and racialized communities. A scalable infoveillance system is needed to monitor these related but distinct crises concurrently using natural language processing (NLP) and statistical forecasting. Objective This study applied a multimethod NLP pipeline to characterize discourse dynamics, causal interdependencies among crisis themes, and the responsiveness of policy-making to those dynamics within Canada’s health care system from 2020 to 2025. Methods We used an NLP pipeline to examine the language and sentiment of 6757 online news headlines regarding Canada’s health care system over approximately 5 years (2020-2025), using 2 methods for sentiment analysis: a fine-tuned BERT (Bidirectional Encoder Representations from Transformers) model (test accuracy of 0.97) and VADER (Valence Aware Dictionary and Sentiment Reasoner) for valence and subjectivity. We used latent Dirichlet allocation to identify topics within the headlines, then applied time series analysis, using both ARIMA (autoregressive integrated moving average) and ETS (error, trend, and seasonality) methods, to predict trends in each topic. Finally, we used Granger causality testing to determine whether past values of one theme could predict future values of another. Results Seven major discourse topics emerged, of which 5 were retained for the surveillance analysis. The health care system crisis theme was the most prevalent (1200/6757, 17.8% of all headlines), followed by economic determinants of health (1017/6757, 15.1%), housing as a health determinant (958/6757, 14.2%), food insecurity health impact (929/6757, 13.7%), and social inequality health effects (699/6757, 10.3%). Strong statistical evidence showed that the volume of discourse around food insecurity health impact was temporally predictive of subsequent discourse volume around housing as a health determinant (F1,54=17.361; false discovery rate–adjusted P=.002). Sentiment analysis showed no statistically significant change after any policy intervention (Cohen d

Wissen, das zu Ihrer Arbeit passt.

Vom kostenlosen Einstieg über die persönliche Recherche bis zur vollständigen Research Suite.

Tarife ansehen →

Sind alle Verlagsvolltexte enthalten?

Nein. Bibliografische Nachweise, Abstracts und Volltexte sind unterschiedliche Inhalte. Ob ein Volltext verfügbar ist, hängt von der Quelle, Open Access und Ihren vorhandenen Lizenzen ab. Ein EUVIMED-Tarif ersetzt keine zusätzlichen Verlagsrechte.

Tarife ansehen →