Machine learning model combines timing and weather data.
A branch of artificial intelligence (AI) called machine learning can accurately predict the risk of out-of-hospital cardiac arrest – when the heart suddenly stops beating – using a combination of timing and weather data, according to research online published in the journal Heart.
Machine learning is the study of computer algorithms, and based on the idea that systems can learn from data and identify patterns to make decisions with minimal intervention.
The risk of cardiac arrest was highest on Sundays, Mondays, holidays, and when the temperature dropped sharply within or between days, the findings show.
This information could be used as an early warning system for citizens, to reduce their risk and improve their chances of survival, and to improve the preparedness of emergency medical services, the researchers propose.
Out-of-hospital cardiac arrest is common all over the world, but is generally associated with low survival rates. The risk is influenced by the prevailing weather conditions.
But meteorological data is extensive and complex, and machine learning has the potential to detect associations not identified by conventional one-dimensional statistical approaches, the Japanese researchers say.
To further investigate this, they assessed the ability of machine learning to predict daily out-of-hospital cardiac arrest, using daily weather (temperature, relative humidity, rainfall, snowfall, cloud cover, wind speed, and atmospheric pressure readings) and timing (year, season). , day of the week, time of day and holidays).
Of the 1,299,784 cases that occurred between 2005 and 2013, machine learning was applied to 525,374, using weather or timing data, or both (training data set).
The results were then compared to 135,678 cases that occurred in 2014-15 to test the accuracy of the model for predicting the number of daily cardiac arrests in other years (test data set).
And to see just how accurate the local-level approach could be, the researchers conducted a “ heatmap analysis, ” using a different data set from the location of cardiac arrests outside the hospital in Kobe City between January 2016 and December 2018.
The combination of weather and timing data most accurately predicted an out-of-hospital cardiac arrest in both the training and test data sets. It predicted that Sundays, Mondays, holidays, winter, low temperatures, and sharp drops in temperature within and between days were more closely associated with cardiac arrest than just the weather or timing data.
The researchers acknowledge that they did not have detailed information on the location of cardiac arrest except in Kobe City, nor did they have data on pre-existing medical conditions, both of which may have influenced the results.
But they suggest, “Our predictive model for the daily incidence of [out-of-hospital cardiac arrest] is generally generalizable to the general population in developed countries, because this study had a large sample size and used extensive meteorological data.
” They add, “The methods developed in this study serve as an example of a new predictive analysis model that can be applied to other relevant clinical outcomes associated with life-threatening acute cardiovascular disease.”
And they conclude, “This predictive model could be useful in preventing [out-of-hospital cardiac arrest] and improving patient prognosis … through an alert system for citizens and [emergency medical services] on risk days in the future.
” In a linked editorial, Dr. David Foster Gaieski of Sidney Kimmel Medical College at Thomas Jefferson University agrees.
Knowing what the weather is likely to be in the coming week can generate ‘cardiovascular emergency alerts’ for those at risk – by notifying the elderly and others of upcoming periods of heightened danger, similar to how weather data is used to inform people of upcoming dangerous road conditions during winter storms, ”he explains.
cardiac catheterization know about and be prepared for the number of [cases] expected in the coming days. , “he adds. References:
” Machine learning model for predicting out-of-hospital cardiac arrest using meteorological and chronological data “May 17, 2021, Heart. DOI: 10.1136 / heartjnl-2020-318726” The weather forecast for next week : cloudy, cold, with risk of cardiac arrest “May 17, 2021, Heart. DOI: 10.1136 / heartjnl-2021-318950 Funding:
Environmental Restoration and Conservation Agency of Japan; Japan Society for the Promotion of Science; Intramural Research Fund for Cardiovascular Diseases of the national cerebral and cardiovascular center