The team enrolled over 2,000 participants in a study that employed the algorithm to look for extended periods when heart rate is higher than normal — a sign that something might be amiss.