Correspondence (letter to the editor): Possible Overcoding

As commendable as secondary analyses of routine hospital data are, one should ponder the results in particular if they conflict substantially with everyday experience in the extramural sector —even if these empirical data appear more imprecise. Such a result as presented in the article with an incidence rate of 365 sepsis patients per 100 000 persons in the general population represents just such a discrepancy.

Approximately 1000 patients visit a general medical practice per quarter and around 1700 patients receive ongoing care in the background (since not all patients attend every quarter). As such, a practice should expect 6.2 sepsis cases per year. According to the data published by Fleischmann et al. (1), a quarter of these patients would die—amounting to almost 1.5 patients/year. This does not even include those individuals—an additional 12% according to Fleischmann et al.—who die of the disease after a sometimes rapid fatal course at home or in nursing institutions.

However, according to recollection, the number of patients with sepsis in my own practice—which is in line with data gathered from colleagues—is between five and 10 times lower: between one and three cases in the last 2 years—albeit, as mentioned, only according to recollection. However, a real sepsis is one of those cases that one tends not to forget easily – even if one only learns about it in a medical report from a hospital.

Although Fleischmann et al. point out that some cases could have been coded incorrectly, they attribute no relevance to this. But those responsible for coding, however, have a conflict of interests that I consider to be relevant, given that they are acting on behalf of a hospital geared to revenue. I therefore assume that considerable over-coding takes place. Moreover, the people performing the coding are not deterred in any way from over-coding, since for them this “small sin” is not resulting in any obvious harm.