Use + Remix

Finding the right tool to predict when hospitals need resourcing

Many hospitals were overwhelmed at the peak of the pandemic. : Wikimedia Commons, Alberto Giuliani Many hospitals were overwhelmed at the peak of the pandemic: Alberto Giuliani, Wikimedia Commons

Knowing when the hospital system could be overwhelmed is key to managing a pandemic. Finding the right predictor can save lives.

As countries grappled with COVID-19, one piece of information remained key: when would the hospital systems be overwhelmed?

Using European Centre for Disease Control data, two public health researchers at the University of Malta found that of two commonly used predictive measures, one was able to accurately predict hospitalisation further in advance than the other. The work could help hospital systems still managing COVID and may help planning for future pandemics.

To estimate the coming caseload, hospitals tend to use either the ‘incidence rate’ – the officially reported number of COVID cases — or the ‘total positivity rate’ — the percentage of tests that were positive.

Both total positivity and incidence rates correlated with hospitalisation, intensive care admittance, and mortality. Both predicted a spike in intensive care admittance two weeks in advance, according to data from the European Centre for Disease Control (ECDC) for mid-March of 2020 and mid-April of 2021.

However, total positivity predicted a peak in COVID-19 hospitalisation and in COVID-19-associated mortality accurately one week further out than the incidence rate. This means total positivity rate is a better indicator for health administrators to maximise the time to prepare for a spike in demand for hospital beds for the care of COVID-19 patients. Total positivity rate was also correlated with COVID-related mortality and excess mortality two weeks later.

Excess mortality is used by epidemiologists to judge the impact of an emerging disease. It is the additional number of deaths over what is expected in any specific week of the year, based on the average number of deaths observed in previous years. The cause of death is not a factor in its calculation, as this is too dependent on the physicians accurately diagnosing and filling out death certificates – something that’s difficult when a new disease emerges. Total positivity rate for COVID correlated with excess deaths, further cementing the value of total positivity rate as the most useful predictor of the burden of a pandemic on the health of a population.

The research also evaluated introducing a minimum testing requirement of 300 tests per 100,000 people weekly across European Union member states. This greatly improved accuracy of the total positivity rate in predicting a spike in mortality, making it clear that these epidemiological indicators perform best when a minimum testing rate is in place. This finding raises concern, as it comes as several countries are restricting access to free testing as part of their ‘living with COVID’ strategies, rather than ensuring that a minimum level of testing is kept in place for surveillance purposes.

The incidence rate is more readily available than total positivity rate, but it is  limited by the amount of testing carried out in a population, especially with an infection which is known to cause little to no symptoms in many of those infected. The total positivity rate requires the incidence rate and the testing rate in a specific population or country, to be calculated.  The only outstanding variation would then be in testing strategy, which could also have some influence on comparability, but, alas, it can’t be adjusted for.

Over the last two years, epidemiologists have scrambled to predict upcoming epidemic waves. Researchers have not only looked into case numbers or related indicators, but also other sources such as social media data, with many holding huge potential for implementation by health policy makers. But few, if any, of these methods allow for a direct reliable comparison of pandemic burden in different populations, where testing capacity and strategy for a condition varies heavily. The task is made even harder by a disease that may present few or no symptoms and would therefore be undetectable if testing was not done.

Due to its built-in adjustment to testing rate, total positivity rate is among the best monitoring tools available for comparison of pandemic burden in different populations. It is only susceptible to variations in testing strategy. Its comparator, incidence rate, is susceptible to variations in testing strategy and to variation in testing numbers.

Total positivity rate holds great potential for large countries, or a bloc such as the European Union, to implement or ease lockdown measures and mobilise resources across regions or countries to respond to the threat in time. Such indicators can be used to forecast hospital beds, ICU beds and number of ventilators required up to two weeks in advance.

Public health responders can then be given advance notice to mobilise the workforce, increasing the number of case managers and contact tracers available to deal with cases and ultimately saving lives.

Professor Neville Calleja is head of the Department of Public Health, Faculty of Medicine & Surgery at the University of Malta.

The author declares no conflict of interest.

Originally published under Creative Commons by 360info™.

Editors Note: In the story “Pandemic warning systems” sent at: 20/03/2022 22:07.

This is a corrected repeat.

Authors
Neville Calleja, University of Malta
Editor
Sara Phillips, 360info
Special Report Articles
Monash University has established and is proud to host the global headquarters for 360info. Monash University is also the host of the Asia-Pacific Hub.
Supporting partners: