AI sleep data could give one-week early warning of flu and COVID-19

AI-powered sleep data could help detect rising levels of flu and COVID-19 around a week earlier, according to a new study by the UK Health Security Agency (UKHSA) and sleep technology company Sleep Cycle

The research examined whether cough data collected passively through a popular sleep app could provide an early indication of respiratory illness spreading in communities across England.

The findings suggest that monitoring coughs recorded during sleep complements existing public health surveillance systems and gives experts a faster picture of changing respiratory illness levels.

Cough data tracked through sleep app

Sleep Cycle is a smartphone app that uses artificial intelligence to analyse sounds during sleep. As people use the app overnight, its technology can detect and record coughing without requiring users to report symptoms actively.

Researchers studied this passively collected data to assess whether changes in coughing were linked to wider patterns of respiratory illness.

The study found that cough levels recorded through the app closely followed respiratory illness trends reported through NHS 111. In several instances, increases in coughing appeared approximately one week before rises in influenza and COVID-19 activity were detected.

This suggests that sleep-based cough monitoring could provide an earlier warning of changes in respiratory disease levels.

A faster view of respiratory illness

Traditional disease surveillance depends heavily on people seeking healthcare and reporting symptoms. This can create delays and may vary depending on factors such as public awareness, access to healthcare and differences between communities.

Laboratory testing and reporting processes can also take time, meaning that official figures may not always provide an immediate picture of what is happening.

The Sleep Cycle system, by contrast, generates a cough signal automatically during normal sleep. The data is updated daily, potentially giving public health teams a near real-time indicator of respiratory illness activity.

Supporting existing surveillance systems

The researchers stressed that digital health data would not replace established surveillance methods. Instead, it could provide an additional information source alongside systems already used by public health authorities.

Combining different sources could help experts identify seasonal changes and emerging increases in respiratory illnesses sooner. Earlier signals could also give authorities more time to assess potential outbreaks and plan their response.

The study also highlighted the potential of consumer-generated health data to contribute to population-level health monitoring while using privacy-preserving approaches.

Potential for future public health monitoring

The findings point to a growing role for everyday technology in disease surveillance. Smartphones and other connected devices can generate large amounts of information without requiring people to complete surveys or visit healthcare services.

If further research confirms the results, passive cough monitoring could become another tool for tracking respiratory illness and improving situational awareness.