Winbo I G, Serenius F H, Dahlquist G G, Källén B A
Department of Pediatrics, Umeå University, Sweden.
Int J Epidemiol. 1998 Jun;27(3):499-504. doi: 10.1093/ije/27.3.499.
Stillbirths and neonatal deaths are often the result of a complicated chain of events. For epidemiological purposes a classification into single cause of death groups is essential. For large-scale studies, a method is needed which enables such grouping based on available register data.
A cause of death classification system called NICE is presented. It is hierarchical and is aetiologically orientated. A computerized method is adapted which makes use of data in four central Swedish registries. A validation of the computer method has been made from the medical records on a 10% sample of all stillbirths and neonatally dead infants in Sweden from 1983 to 1990.
The specificity of the computer method is high, sensitivity is less satisfactory for some subgroups. A time trend analysis illustrates the usefulness of the classification system and shows a decline with time for two groups: placental abruption and obstetric complications.
The NICE classification system fulfils the criteria of an aetiologically orientated classification system which can be used in a computerized environment.
死产和新生儿死亡往往是一系列复杂事件的结果。出于流行病学目的，将死因分类为单一死因组至关重要。对于大规模研究，需要一种能够基于现有登记数据进行此类分组的方法。
提出了一种名为NICE的死因分类系统。它是分层的且以病因学为导向。采用了一种计算机化方法，该方法利用瑞典四个中央登记处的数据。对1983年至1990年瑞典所有死产和新生儿死亡婴儿10%样本的病历进行了计算机方法的验证。
计算机方法的特异性较高，某些亚组的敏感性不太令人满意。时间趋势分析说明了分类系统的有用性，并显示两组随时间下降：胎盘早剥和产科并发症。
NICE分类系统符合可在计算机环境中使用的病因学导向分类系统的标准。
