In a study published in the journal Discrete Dynamics in Nature and Society, Huami provided new clues for establishing a large-scale epidemic surveillance system, and helping improve the efficiency of public health monitoring and prediction. The research paper was titled “Learning from Large-Scale Wearable Device Data for Predicting Epidemics Trend of COVID-19.”
This study was supported by the Huami Corporation, a prediction model was established by using big data and artificial intelligence algorithms, which provides a new method for predicting epidemic trends for COVID-19, says a press release.
Researchers collected heart rate, physical activity, sleep, and other physiological data related to the above symptoms based on smart wearable devices. De-identified sensor data of about 1.3 million users who wore Huami devices from July 1, 2017, to April 8, 2020 were obtained according to appropriate security control. All the users were notified that their de-identified data could potentially be used for academic research.
Research found that, for every 1°C increase in human body temperature, heart rate increases by about 8.5 bpm. Based on this, the increase in heart rate caused by fevers related to COVID-19 or influenza-like diseases can be used as a starting point for a method to detect physiological abnormalities.
Huami researchers considered an individual’s resting heart rate at 1.5 standard deviations higher than the personal average for 5 consecutive days, and sleep duration not less than 0.5 standard deviations from the personal average as the criterion to determine an abnormality.
The prediction model’s analysis results show that in the listed cities of Wuhan, Beijing, Shenzhen, Hefei, and Nanjing, there was a clear outbreak period in the infection rate prediction curve for each city which corresponded to the epidemic’s outbreak in each city.
Taking Wuhan as an example, the infection rate predicted by the model peaked on January 28th, while the newly confirmed cases in Wuhan peaked at nearly 2,000 people on February 7. The predicted infection rate peak was 10 days earlier than the officially reported peak time.
Given the lag between COVID-19 infection and the emergence of symptoms and diagnosis, the model-derived results are also consistent with the results of a retrospective study on COVID-19 conducted by the Chinese Center for Disease Control.
Besides the academic research from Huami, Huami continued the efforts of Connect Health with Technology. The company has donated medical supplies and devices worth 11.5 million RMB during the coronavirus outbreak. Amazfit, a self-brand of Huami, started to working on a transparent N95 face mask called Amazfit AERI. The Innovative mask can clean itself and last for several weeks. For combating COVID-19, Huami also partnered with China National Clinical Research Center of Respiratory Disease (NCRCRD) and Guangdong Nanshan Medical Innovation Institute which led by Dr. Nanshan Zhong to build up a smart wearable joint laboratory.