The pandemic has sharpened our attention to the disparities in access to health care and medical research between developed countries and low- and middle-income countries (LMICs). Fortunately, the medical research community often comes together in times of need, and building local skills and knowledge in neighboring LRICs is a valuable way to improve health globally.
Epidemiological modeler Associate Professor James Trauer has recently taken this route, lending its expertise to help Malaysia, Vietnam and the Philippines produce effective and localized COVID mitigation policies, and build their internal capacity to continue independent modeling.
He recently edited a publication in Lancet Regional Health – Western Pacific which describes both the technical process involved and the vital establishment of trust and communication channels necessary for success.
Epidemiological modeling allows us both to assess the effectiveness of past public health interventions and to predict the future effects of these interventions under a range of different scenarios. This requires a mix of public health and medical knowledge, statistical knowledge, and coding skills.
A/Prof Trauer is a trained respiratory and sleep physician, but gradually became interested in this niche area. “My late father encouraged me to do my own statistical analysis on small clinical projects that I ran, rather than relying on statisticians. Then I was fortunate to be supervised during my PhD by Professor Emma McBryde of James Cook University, who is also a major collaborator on a number of my projects today. She has great quantitative skills, and that’s how my interest in modeling grew.
A/Prof Trauer led the development of the open access modeling software AuTuMN in 2016, with Dr Romain Ragonnet, Dr Angus Hughes and David Shipman (also of Monash), Professor McBryde and other researchers. Designed to help in the fight against tuberculosis in the Pacific region, it is easily adaptable to other infectious agents.
AuTuMN’s application to develop a COVID policy for Malaysia, Vietnam and the Philippines formed the basis of one of the World Health Organization’s first Western Pacific Innovation Challenge awards. Given the magnitude of the task, the team divided the workload between them, each contributing their part.
The Western Pacific Region includes demographically, socioeconomically and culturally diverse countries, which has resulted in major differences in the timing, scale and duration of COVID-19 waves. The Philippines and Vietnam are currently classified as low-income countries, while Malaysia is middle-income.
Successful epidemiological modeling must incorporate local conditions. Immunization coverage rates, cultural acceptance of public health interventions, ability to maintain closed borders, and access to personal protective equipment are just a few of the many variables that must be entered into projections.
Acting Professor Trauer and his team worked directly with the relevant local health authorities and research institutes in each country. These in turn served as intermediaries to local policy makers, identifying the burning questions the modellers needed to answer and collecting the health data needed to feed their analyses. Once the questions and data were finalized, the researchers connected them to the platform and sent the results back to their local collaborators for dissemination. These most often took the form of interactive online dashboards that policy makers could interrogate.
“The questions we can answer are things like, given the decline in immunity over time, what hospitalization rates can we expect if the national vaccination rate is maintained at 100,000 per day, versus 200,000 per day, versus 300,000 per day? This allows them to set vaccination goals that should minimize the burden on the hospital system. he says.
“Or they might ask something like, what changes would we see in hospitalizations and death rates if we reopen schools in a month or reopen international borders? What if we did both at the same time?
The technical processes were part of the project; building relationships and trust with local contacts was another big challenge according to A/Prof Trauer.
“We needed to make them feel comfortable sharing national data with us, and also build their confidence in the accuracy of the models we were providing. We in turn had to trust them on the quality of the local data provided.
“A good example of where this worked well was in the Philippines at the start of the pandemic, we noticed unusual patterns in reported COVID-19 mortality. After questioning this, local public health surveillance experts were able to identify the difficulties in classifying COVID-19 as a cause of death and the resulting delays in cause of death reporting.
“Knowing that this area is still fairly new, many countries have very limited staff who can immediately understand the strengths and limitations of the models, and what should and should not be inferred from them. We have worked hard to build their capacity for local critical appraisal, with frequent discussions around the results. »
The project was successful in many ways.
First, the models have clearly been successful in helping these countries determine the most effective course of action. They accurately predicted the timing and magnitude of epidemic peaks in the Philippines, and identified a probable epidemic resurgence due to the emergence of a new variant of concern. They also helped Malaysia estimate the effects of vaccination rates on health system resources. The models were often mentioned in public health press releases issued in Vietnam and the Philippines.
Second, working relationships with established individuals and organizations across the region, along with increased local skills and knowledge, will enable a faster response to future waves or other outbreaks.
Third, the results contributed to general knowledge about the spread of COVID and the effectiveness of mitigation strategies in LMICs.
And finally, the project improved the AuTuMN platform itself. A/Prof Trauer says: “Exposing this still new platform to so much data has given it a real workout, and we are even more confident that it is fit for purpose. Every time you project a scenario and then come back to see how accurate it was, you create a cycle of continuous improvement.
“I am a strong proponent of sharing knowledge with others, especially in scenarios like this where sharing a little expert knowledge can be of such benefit to millions of people in our regional neighborhood. I’m incredibly proud of the platform, the relationships we’ve built, and the entire team here in Australia and across the region who have made this possible.
Click here for more news from the School of Public Health and Preventive Medicine