Limitations of the death certificate only index as a measure of incompleteness of cancer registration

Abstract

Summary The death certificate only (DCO) index, which quantifies the proportion of patients for whom the death certificate provides the only notification to the registry, is a widely used measure of incompleteness of population-based cancer registration. This paper provides an algebraic assessment and a quantitative illustration of the relationship between the DCO index and incompleteness of cancer registration. It is shown that the relationship between the DCO index and incompleteness of registration is strongly dependent on the case fatality rate and the misclassification rates of cancer deaths among unregistered patients. Therefore, the DCO index is a very poor indicator of incompleteness. Similar limitations apply to the DCN index (proportion of cases first notified by death certificate), which has been proposed as an alternative measure of incompleteness.Keywords: death certificate; cancer registry; disease surveillance; epidemiological method Population-based cancer registries are most valuable tools of both descriptive and analytical cancer epidemiology. Exhaustive ascertainment of all cases can rarely be achieved, however, on the population level. Therefore, attempts to quantify the degree of underascertainment are essential for proper interpretation of cancer registry data.Various approaches have been proposed to quantify incompleteness of population-based cancer registration (Freedman, 1978;Goldberg et al., 1980;Benn et al., 1982;Devesa et al., 1984;Shanmugaratnam, 1989;Jensen et al., 1991;Parkin et al., 1992). A particularly widely used indirect measure of incompleteness is the death certificate only index (DCO index) which quantifies the proportion of cases for whom the death certificate is the only source of notification. For example, the DCO index is routinely reported as an indicator of incompleteness in international cancer incidence statistics published by the International Agency for Research on Cancer (Parkin et al., 1992).One of the major reasons for the popularity of the DCO index is its ease of calculation in situations in which death certificates are routinely linked with registry data. Nevertheless, the DCO index has a variety of limitations, the best known being its dependence on case fatality rates: poor completeness may go along with a low DCO index for cancers which have low case fatality rates. Another major limitation of the DCO index is its sensitivity to the welldocumented inaccuracies of the certified causes of death (Barclay and Phillips, 1962;De Faire et al., 1976;Percy et al., 1981). As an example, endometrial cancer is often misclassified as cervical cancer, and vice versa, on death certificates, and similar misclassification problems have also been documented for other cancers, such as colon and rectum cancer (Percy et al., 1981). Furthermore, non-specific codes, such as cancer of unknown site, are frequently used on death certificates, and cancer often remains undetected as the underlying cause of death among elderly patients. While these limitations have been widely recogn...

Determinants and interpretation of death certificate only proportions in the initial years of newly established cancer registries

“…These include, for example, misreporting of underlying cancer causes of death on death certificates, [20][21][22] problems in record linkage of notifications during lifetime and death certificates, 23,24 or DCO cases arising from immigration to the registration area after cancer diagnosis. Misreporting of cause of death appears to be less common for cancer deaths compared to other causes of deaths, [25][26][27][28][29] and migration across cancer registry borders likewise seems to be rather limited after cancer diagnoses.…”

Comparability, diagnostic validity and completeness of Nigerian cancer registries

“…On the other hand, this could be interpreted favorably in the sense that it does not appear that registries are heavily relying on cause of death from death certificates to populate their registries. Evaluation of this particular method of validity is contentious because of wide variation in the use of death certificates, vital statistics, quality of cause of death statements, and other reasons [46]. The most favored approach seems to be comparison to other registries in the region [24,41].…”

Completeness of cancer registration: a new method for routine use

“…However, this practice can lead to bias. Berrino et al (1995) have shown that patients dying with cancer mentioned on the death certificate have poorer survival than patients dying of other causes, and that the percentage reduction in estimated survival resulting from the inclusion of DCO cases is generally of the same order as the proportion of such cases in the series under study. If possible, DCOs should be included as outlined below.…”