Surveillance data can also be summarized by when the disease or health event of interest occurred to see if temporal clustering occurs. Time intervals used may be years, months, weeks, or whatever unit bears particular relevance. The date used may be the date of onset of illness, diagnosis, report to the surveillance system, or other factor that will help determine the best means of designing programs and policies that reduce or prevent the adverse health condition from occurring. As in previous examples, a graphic summary of the data will allow us to easily recognize trends over time.

In our surveillance data for bicycle accidents, we observed that the greatest number occurred during the months of June, July, and August. To understand why this occurred, we carried out additional analyses and cross classified our cases by age. This data summary helped us realize that the majority of cases reported to our surveillance system occurred among school-aged children during their three-month summer vacation. This suggested to us that, a school-based educational program emphasizing proper bicycle riding and use of helmets and pads needed to be developed and implemented during the month of May, one month before children left school to begin their summer vacation.

An important factor to remember when summarizing surveillance data by time is that reporting delays do occur. Thus, monthly reports summarized at the end of the year may differ from those summarized at the end of the each month earlier in the year. For example, between June 1 and June 30, only 30 bicycle accidents were reported to our surveillance system; however, by the end of the year, an additional 8 cases had been reported as having occurred in June.

Accounting for delays in reporting is especially important when summarizing case data for diseases or health outcomes which occur at the end of a calendar year. That is, surveillance data for the first several months of the year following that of interest should be reviewed to find any cases which are reported late. The magnitude of these delays can be determined by tracking the interval between the occurrence of the health event of interest and the date of the initial report of its occurrence to your surveillance system. For example, if our interest was in bicycle accidents occurring during school vacations in December rather than the summer months, and we knew that reporting delays were as much as three months long, we would examine surveillance data for January, February, and March 2002 (in addition to that for December 2001) to identify any bicycle accidents which occurred during December 2001.