Background
A decline in the national maternal mortality ratio in Nepal has been observed from surveys conducted between 1996 and 2008. This paper aims to assess the plausibility of the decline and to identify drivers of change.

Methods
National and sub-national trends in mortality data were investigated using existing demographic and health surveys and maternal mortality and morbidity surveys. Potential drivers of the variation in maternal mortality between districts were identified by regressing district-level indicators from the Nepal demographic health surveys against maternal mortality estimates.

Results
A statistically significant decline of the maternal mortality ratio from 539 maternal deaths to 281 per 100,000 (95% CI 91,507) live births between 1993 and 2003 was demonstrated. The sub-national changes are of similar magnitude and direction to those observed nationally, and in the terai region (plains) the differences are statistically significant with a reduction of 361 per 100,000 live births (95% CI 36,686) during the same time period.

The reduction in fertility, changes in education and wealth, improvements in components of the human development index, gender empowerment and anaemia each explained more than 10% of the district variation in maternal mortality. A number of limitations in each of the data sources used were identified. Of these, the most important relate to the underestimation of numbers of deaths.

Conclusion
It is likely that there has been a decline in Nepal's maternal mortality since 1993. This is good news for the country's sustained commitments in this area. Conclusions on the magnitude, pattern of the change and drivers of the decline are constrained by lack of data. We recommend close tracking of maternal mortality and its determinants in Nepal, attention to the communication of future estimates, and various options for bridging data gaps.