Modelling the lagged registers of COVID-19 confirmed deaths

On a daily basis, the Mexican government publishes a data base that includes the records of COVID-19 confirmed death cases. This data base is cumulative, in the sense that each day new records are included. As a consequence of the process with which the health officials collect and test the suspected COVID-19 cases that are included in the data bases, there is a lag between the date of occurrence of the death and the date in which it is registered in the data base. The new confirmed deaths that are included between two dates t1 and t2 (t1 < t2) do not correspond necessarily to the death cases that occurred between these two dates, but they also correspond to deaths that occurred before, i.e., at dates t <= t1. An implication is that the shape of the curve of cumulative registered death counts at a given date does not correspond to the "real"" count, which may lead to wrong interpretations of the state of the pandemic in the country. Here, a model is proposed in order to estimate the lagged death cases and thus provide a better approximation of the state of the pandemic.