G. Nilsson et al., Evaluation of three Swedish ICD-10 primary care versions: Reliability and ease of use in diagnostic coding, METH INF M, 39(4-5), 2000, pp. 325-331
If computer-stored information is to be useful for purposes other than pati
ent care, reliability of the data is of utmost importance. In primary healt
hcare settings, however, it has been found to be poor. This paper presents
a study on the influence of coding tools on reliability and user acceptance
. Six general practitioners coded 152 medical problems each by means of thr
ee versions of ICD-10, one with a compositional structure. At code level th
e reliability was poor and was almost identical when the three versions wer
e compared. At aggregated level the reliability was good and somewhat bette
r in the compositional structure. Ideas for improved user acceptance arose,
a nd the study explored the need for several different tools to retrieve d
iagnostic codes.