Quantifying the fall in mortality associated with interventions related to hypertensive diseases of pregnancy

BACKGROUND: In this paper we review the evidence of the effect of health interventions on mortality reduction from hypertensive diseases in pregnancy (HDP). We chose HDP because they represent a major cause of death in low income countries and evidence of effect on maternal mortality from randomised studies is available for some interventions. METHODS: We used four approaches to review the evidence of the effect of interventions to prevent or treat HDP on mortality reduction from HDP. We first reviewed the Cochrane Library to identify systematic reviews and individual trials of the efficacy of single interventions for the prevention or treatment of HDP. We then searched the literature for articles quantifying the impact of maternal health interventions on the reduction of maternal mortality at the population level and describe the approaches used by various authors for interventions related to HDP. Third, we examined levels of HDP-specific mortality over time or between regions in an attempt to quantify the actual or potential reduction in mortality from HDP in these regions or over time. Lastly, we compared case fatality rates in women with HDP-related severe acute maternal morbidity with those reported historically in high income countries before any effective treatment was available. RESULTS: The Cochrane review identified 5 effective interventions: routine calcium supplementation in pregnancy, antiplatelet agents during pregnancy in women at risk of pre-eclampsia, Magnesium sulphate (MgS04) for the treatment of eclampsia, MgS04 for the treatment of pre-eclampsia, and hypertensive drugs for the treatment of mild to moderate hypertension in pregnancy. We found 10 studies quantifying the effect of maternal health interventions on reducing maternal mortality from HDP, but the heterogeneity in the methods make it difficult to draw uniform conclusions for effectiveness of interventions at various levels of the health system. Most authors include a health systems dimension aimed at separating interventions that can be delivered at the primary or health centre level from those that require hospital treatment, but definitions are rarely provided and there is no consistency in the types of interventions that are deemed effective at the various levels. The low levels of HDP related mortality in rural China and Sri Lanka suggest that reductions of 85% or more are within reach, provided that most women give birth with a health professional who can refer them to higher levels of care when necessary. Results from studies of severe acute maternal morbidity in Indonesia and Bolivia also suggest that mortality in women with severe pre-eclampsia or eclampsia in hospital can be reduced by more than 84%, even when the women arrive late. CONCLUSIONS: The increasing emphasis on the rating of the quality of evidence has led to greater reliance on evidence from randomised controlled trials to estimate the effect of interventions. Yet evidence from randomised studies is often not available, the effects observed on morbidity may not translate in to mortality, and the distinction between efficacy and effectiveness may be difficult to make. We suggest that more use should be made of observational evidence, particularly since such data represent the actual effectiveness of packages of interventions in various settings.