Andersen K V, Helweg-Larsen K, Lange A P
Københavns Amts Sygehus i Glostrup, gynaekologisk/obstetrisk afdeling.
Ugeskr Laeger. 1991 May 27;153(22):1577-81.
The quality of registration of the causes of death in 273 perinatal and neonatal deaths was assessed. The present authors classified the causes of death according to WHO classification of disease. The authors classification was compared with the Danish Health Department's original classification of deaths and a classification undertaken by one of the staff of the Danish Health Department, the coding of which was based exclusively on information from the original death certificates (death certificate classification). Agreement between the respective forms of classification was assessed by the observed agreement for positive diagnosis and the kappa coefficients calculated by the three-figure level in the ICD diagnosis for categories and between categories of causes of perinatal and neonatal death, respectively. The investigation revealed that the deaths were classified with great accuracy by coding by the staff member when the death certificates were correctly completed. Comparison between the "authors classification" and the "Danish Health Department's original classification" and the "death certificate classification", respectively revealed that the observed agreement for positive diagnoses were 52% and 50% and that the kappa coefficients were between 0.80-0.07 and 0.81-0.04, respectively, on comparison with the three-figure levels in the ICD diagnoses and 0.93-0.32 and 0.85 and 0.32 on comparing the categories in between. It is concluded that considerable improvement in registration of perinatal and neonatal deaths may be obtained if the Danish Health Departments states precisely what information is required on the death certificate in connection with perinatal and neonatal deaths. Correspondingly, local efforts should be made to complete death certificates more correctly.
对273例围产期和新生儿死亡病例的死因登记质量进行了评估。本文作者根据世界卫生组织疾病分类对死因进行了分类。将作者的分类与丹麦卫生部最初的死亡分类以及丹麦卫生部一名工作人员进行的分类进行了比较，后者的编码完全基于原始死亡证明中的信息（死亡证明分类）。通过观察阳性诊断的一致性以及分别根据围产期和新生儿死亡原因类别在ICD诊断中三位数水平计算的kappa系数，评估了各种分类形式之间的一致性。调查显示，当死亡证明填写正确时，该工作人员通过编码对死亡进行了非常准确的分类。分别将“作者分类”与“丹麦卫生部最初分类”以及“死亡证明分类”进行比较，结果显示，与ICD诊断中的三位数水平相比，阳性诊断的观察一致性分别为52%和50%，kappa系数分别在0.80 - 0.07和0.81 - 0.04之间；在比较类别时，kappa系数分别为0.93 - 0.32、0.85和0.32。结论是，如果丹麦卫生部明确规定与围产期和新生儿死亡相关的死亡证明需要哪些信息，围产期和新生儿死亡登记可能会有显著改善。相应地，地方应努力更正确地填写死亡证明。
