In this study, we used a novel approach to distill information from the cumulative number of diagnosed cases of COVID-19 infection. Among various types of surveillance data, this data often reported the earliest and on a continuous basis with high completeness and are most widely available. In addition, patients with a diagnosed infection are those with high likelihoods to spread the virus to others. Findings from this study provided useful information in a real time manner to monitor, evaluate and forecast the COVID-19 epidemic in China. The methods used in this study although somewhat mathematical, are easy to follow while information extracted from the commonly used data with the methods are highly useful and more sensitive than the daily new and cumulative cases.