Temporal trends in maternal mortality in Canada I: estimates based on vital statistics data

Abstract

Objectives: Vital Statistics and World Health Organization reports show a recent increase in maternal mortality in Canada. We carried out a study of temporal trends, regional variations, and causes of death in Canadian maternal mortality using Vital Statistics data.

Methods: We used Vital Statistics death registrations to ascertain maternal deaths between 1981 and 2007. Maternal mortality rates, risk ratios, and 95% confidence intervals were estimated, and the Cochran-Armitage test was used to evaluate temporal trends. We used hospitalization data from the Canadian Institute for Health Information from 1996 to 2007 to confirm maternal mortality trends observed in the Vital Statistics data.

Results: Maternal mortality rates increased significantly from 4.5 (95% CI 3.3 to 5.8) in 1981 to 1983 to 4.7 (95% CI 3.5 to 6.2) in 1996 to 1998 and to 7.2 (95% CI 5.7 to 9.0) per 100 000 live births in 2005 to 2007 (P value for trend < 0.001). The most common causes of maternal death were diseases of the circulatory system, obstetric embolism (venous thromboembolism and amniotic fluid embolism), and hypertension. Deaths due to diseases of the circulatory system and puerperal infection increased significantly from 1981 to 2007. Maternal mortality rates in the hospitalization data were higher and did not show an increase over time. Provincial and territorial maternal mortality rates from Vital Statistics data showed varying degrees of under-ascertainment (12% to 70%) compared with hospitalization data.

Conclusion: Temporal increases in maternal mortality in Canada observed in Vital Statistics data do not correspond with stable temporal trends observed in hospitalization data, and appear to be an artefact of changes in the coding and ascertainment of maternal deaths.