News media coverage of COVID-19 public health and policy information

News media plays a crucial role during public health emergencies, communicating risks, shaping public perception, and framing policy debates. Effective pandemic response demands rapid, iterative actions under conditions of incomplete knowledge, requiring coordination among various agencies and sectors. News media serves as a primary information source, connecting health professionals, policymakers, and the public. The quality of news coverage significantly impacts public trust in science and accountability in decision-making. Past pandemics have demonstrated the risks of inadequate scientific quality in news coverage, potentially overstating or understating risks, misrepresenting the efficacy of protective measures, and hindering policy effectiveness. The COVID-19 pandemic provided an opportunity to assess the strengths and limitations of media coverage across different national contexts and political spectrums. Canada, the UK, and the US were selected for their varying governmental approaches to public health and pandemic response, offering a comparative analysis.

Prior research highlights the significant role of news media during health crises, emphasizing the influence of reporting volume, content, and tone on public perception and policy discussions. Studies on past pandemics, including H1N1, SARS, and Ebola, have revealed issues with the scientific quality and balance of media coverage. These issues include overemphasis on threats without sufficient detail on protective measures, the promotion of "false balance" by pairing credible scientific information with uninformed opinions, and a decline in coverage despite persistent public health risks. The authors cite previous work on media slant and the influence of newspapers on other media outlets.

This study employed a mixed-methods approach, combining public health data analysis with a content analysis of news media coverage. Public health data on COVID-19 cases, deaths, and testing were analyzed for Canada, the UK, and the US to contextualize media coverage. Print and online news articles were retrieved from the Factiva database for 12 newspapers representing the political spectrum of each country. A random sampling method was used to select one day of coverage per week over six 4-week periods, aiming for five eligible articles per outlet per day. Eligibility criteria included articles directly focused on COVID-19 public health implications in the newspaper's country of publication, excluding opinion pieces, editorials, etc. A validated coding tool was adapted to measure scientific quality (applicability, opinion vs. facts, validity, precision, context, global assessment) and sensationalism (exposing, speculating, generalizing, warning, extolling, global assessment) of the articles. Two independent coders evaluated each article, undergoing extensive training and calibration to ensure consistency. Interrater reliability was assessed using weighted Cohen's Kappa. Statistical analyses included Kruskal-Wallis one-way ANOVA and post-hoc pairwise comparisons.

From March to August 2020, the three countries differed substantially in COVID-19 outcomes and public health responses. The UK experienced the highest death rate, while the US had the highest case rate. Canada had the most effective outcomes. News outlets varied significantly in the amount of COVID-19 coverage. Overall, the scientific quality of reporting was moderate, but considerably lower for populist-right leaning newspapers. These outlets displayed weaknesses in validity, precision, and context, sometimes failing to distinguish between opinion and fact. Sensationalism was generally low across all outlets but lowest for populist-right outlets in Canada and the US. Left-leaning US newspapers provided more exposing and warning coverage, addressing policy failures and misinformation. Original reporting had significantly higher scientific quality than syndicated articles. Articles on healthcare topics had higher scientific quality than those on other topics. Sensationalism was highest for articles related to politics and foreign affairs. Coverage of public health policies varied over time, with consistent coverage of social distancing, testing and tracing, and protective equipment, and increasing coverage of mask mandates and reopening policies.

The findings underscore the complexity of communicating public health information during a pandemic. The study reveals significant variations in the scientific quality of COVID-19 news coverage across different political spectrums, highlighting the potential for partisan biases to influence how information is presented. The lower scientific quality of reporting in populist-right leaning newspapers raises concerns about the accuracy and reliability of information reaching the public. The generally low sensationalism, while often considered positive, could have paradoxically downplayed the urgency of public health threats and policy failures in some instances. The differences observed in US media coverage reflect and potentially contribute to the politicization of COVID-19 in the country, mirroring and amplifying existing political divisions. The study's focus on traditional national newspapers highlights their influence on broader media landscapes.

This research demonstrates that news media coverage of COVID-19 varied widely in scientific quality and sensationalism across the political spectrum. Populist-right outlets showed significantly lower scientific quality. While low sensationalism is generally positive, its combination with low scientific quality may have hindered effective public health communication. The study emphasizes the need for improved scientific accuracy in reporting and the critical role of media literacy in navigating complex health information during crises. Future research should explore the long-term impacts of biased reporting on public health outcomes and investigate strategies for improving the quality and accessibility of public health information.

The study's reliance on a random sample of days might not fully capture the nuances of daily news cycles. The focus on national newspapers might not fully represent the diversity of media outlets. The coding tool, while validated, is subject to interpretative variations. The analysis focuses on a specific period, limiting the scope of generalizability.