Stylised example of chikungunya case count data availability for a given week.

This matrix provides a stylised example of the chikungunya case count data available when nowcasting cases for a given week. In this example, we hold data from week 1 onwards and are currently in week 7. For ease of illustration, we assume here that the maximum delay in entering a case into the surveillance system is five weeks. Each row represents a previous week (t) of entered cases, and the column represents the entry delay (d) in weeks. For example, we can see that there were initially 15 cases entered into the system in week 2, 8 further cases after a delay of 1 week, 10 cases after a delay of 2 weeks, and so on. Case data is incomplete, not only for week 7 but also weeks 3 through 6. The incompleteness is usually worse the closer the week to the current period, so there is a running “triangle” of unknown case counts and associated delays to be estimated for previous weeks. Estimating each cell in the last row of this triangle yields a nowcast of the total case count for week 7. The method introduced by Bastos et al. [21] provides an approach for generating these estimates.