On Wednesday, the U.K. Health Security Agency quietly released a review which found that the evidence base for the effectiveness of non-pharmaceutical interventions (NPIs) on Covid transmission was weak, writes Kevin Bardosh in UnHerd. These include measures such as lockdown, the ‘rule of 6’, test and isolation, face masks, border restrictions and more. Here’s an excerpt.
The review identified 151 studies conducted in the U.K. Two-thirds of the evidence identified was based on modelling studies, and there were only two randomised control trials included. Studies focused on measures to identify or isolate infectious people or reduce the number of human contacts were the most numerous. Only nine studies focused on measures to protect the most vulnerable.
The report concluded that “there is a lack of strong evidence on the effectiveness of NPIs to reduce COVID-19 transmission”:
“The body of evidence available on effectiveness of NPIs in the U.K. provides weak evidence in terms of study design, as it is mainly based on modelling studies, ecological studies, mixed-methods studies and qualitative studies.”
The review noted the many challenges with evaluating NPI effectiveness, which makes the current evidence-base prone to confounding and bias.
“In most cases, [current NPI effectiveness] would be graded as low or very low certainty in the Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) framework and rule out causal inference.”
This includes the way NPIs were implemented, levels of adherence and compliance, wider behavioural changes, the difficulties of isolating one NPI from a range (or ‘package’) of measures, endpoint measurements, poor study designs and study quality, and ethical considerations. The report went on to recommend a strengthening of evaluation research for NPI interventions in the future.
Dr. Bardosh notes that, as with the recent Royal Society Report, the new report once again fails to consider the wider social, economic, psychological and political impacts of interventions.
The report fails to mention that public health research should, in principle, weigh the impact of specific interventions on a virus against their full societal impact, including their social harms and unintended consequences.
This myopic definition of “effectiveness” continues to hamper an intelligent and holistic assessment of whether NPIs were, all said and done, really the best of many bad options.
Worth reading in full.