Errors in death certificates in a rural area of Greece.

Introduction: A death certificate is the formal document in which a physician records the time, cause and circumstances under which the death of an individual has occurred. Morbidity and mortality statistics are mainly based on the analysis of these certificates, and inaccuracies in the detail may lead to biased estimation in several epidemiological parameters. The aim of this study was to examine the quality of cause of death in death certificates in a rural area of Greece, and to identify factors that may be associated with inaccuracies in the completion of these death certificates.

Methods: All death certificates archived in the municipality of Tritaia during the period 1999-2006 were examined. Statistical analysis was performed by comparing the proportions of the unpaired case. The state of independence among the various variables was investigated by considering the class of discrete graphical models.

Results: In total, 516 death certificates were examined; 5.6% (29/516) were excluded because of insufficient demographic data. The remaining 487 death certificates were analyzed with the following findings: 51.5% were for males and 48.5% females (median age 82 years, range 5-103 years; and 83 years, range 0-104, respectively); and 39.4% (192/487) were correctly completed. In 168 the mechanism of death was given; in 72 multiple causal sequences were given; in 22 a single/not precise cause was given; and in 33 a single causal sequence with incorrect order was given. In all, 20.1% were completed by a physician of the regional health centre. Gender was not associated with the presence of error (p = 0.352). Errors were present in 63.8% (270/423) of the death certificates in deceased individuals > or =60 years and in 39.1% (25/64) of the death certificates in individuals < or =59 years (p<0.001). In 19.7% of the erroneously completed death certificates, the certifier was a physician working in primary health care. The presence of errors in death certificates decreased from 74.6% in 1999 to 51.8% in 2006 (p = 0.004).

Conclusions: Giving the mechanism instead of the cause of death was the most frequent type of error. A statistically significant increase in the presence of errors was observed as the age of the descendent increased. During these 8 years, there has been a statistically significant decrease in errors. However, efforts should be made by trainers and physicians in order to improve the accuracy of the information in death certificates. If this is accomplished, cause-of-death statistics will be more accurate and so enable better health planning.