Risk of heat- or cold-related illnesses across a broad population of Ontario workers

In Brief

  • It's not just extreme temperatures that put workers at risk of health hazards related to heat or cold exposure. According to emergency department records, hazards exist even at moderate temperatures.
  • Above 22 degrees Celsius, the risk of heat stress and heat illnesses increases by 75 per cent with every one-degree increase in temperature.
  • Below zero degrees Celsius, the risk of hypothermia, frostbite and other cold-related illnesses increases by 15 per cent with every one-degree decrease in temperature.

Why was this study done?

It's long been known that extreme temperatures can pose health hazards, from heat stress disorders to hypothermia and frost-bite. With climate change expected to bring about more extreme weather events, there is growing concern about possible health effects on workers. Research done to date on the work-related hazards of hot or cold weather has focused on specific industries, occupations or worksites. This is the first known study to look at the hazards of hot or cold weather across an entire working population.

How was the study done?

Using the National Ambulatory Care Reporting System, which documents all emergency department visits in Canada, the researchers focused on visits in southwestern Ontario between 2004 and 2010. They looked for visits involving heat- or cold-related illnesses that were deemed by the medical staff to be work-related. (The study looked at temperature-related illnesses such as heat stroke, frostbite and hypothermia, not weather-related injuries such as falls on ice.) For daily weather and pollution records, researchers used data from Environment Canada and the Ontario Ministry of the Environment. They then compared illness rates on days with different temperatures, while controlling for day of the week, monthly workforce size, seasonal trends and other factors.

What did the researchers find?

In southwestern Ontario, the risk of heat illness is 5.5 emergency department visits per 100,000 full-time equivalent workers each year. Above 22 degrees Celsius, every one-degree rise in the temperature sees the risk of workers seeking treatment at the ER rise by 75 per cent.

When it comes to the cold, the risk of illness such as hypothermia is 0.93 visit per 100,000 full time worker per year. Each one-degree fall in the temperature below zero-degree Celsius comes with a 15-per-cent increase in the risk of workers seeking help at the ER. Wind speed also plays a part. Each one-kilometre increase in wind speed brings a six-per-cent increase in the rate of ER visits by workers experiencing cold-related illnesses.

What are the implications of the study’s findings?

A strong relationship exists between outdoor temperature and illness among the working population. And, as this study shows, it's not just in extreme temperatures that workers are at risk. If workers complain of symptoms or show signs of heat stress or hypothermia, their situation shouldn’t be dismissed just because the temperature outside isn't extreme.

What are some strengths and weaknesses of the study?

The use of emergency department records was a strength of this study, as it allowed the researchers to capture work-related illnesses that are not covered by workers' compensation. Another strength is the study's scope, encompassing the entire working population in a large area across six years. As a result, the study was able to identify risks in non-extreme conditions.

A limitation of this study is its reliance on data from central monitoring sites as a stand-in for the actual temperature to which workers were exposed. Another limitation is a potential misclassification of where patients got sick, due to the use of residential postal code information from emergency department records. As a result, people who commuted long distances may have been exposed to different weather conditions than what was assigned to them in the data analysis.

Publication Information



Fortune M, Mustard C, Brown P


Environmental Research, 2014: vol. 132, pp 449-456