Background: Maternal mortality (MM) surveillance is conducted at a statewide level and reported nationally. Case ascertainment methods vary. Some states require active surveillance (review of suspect case reports) for maternal deaths; others rely on data from death certificates only. All methods result in some under-reporting.  New ICD 10 coding extends the definition of MM to include deaths up to one year following delivery/termination.  This delineates “pregnancy-related” deaths from “pregnancy-associated” (incidental deaths not due to pregnancy status). Methods: In 1998, the New York City (NYC) Department of Health and Mental Hygiene introduced enhanced surveillance (ES) for maternal deaths. All potential maternal deaths between 1998 and 2000 were reviewed. ES supplemented Bureau of Vital Statistics (BVS) data with medical examiner (ME) and hospital discharge data (SPARCS). ICD 9 and 10 coding was applied. Results: ES of MM cases between 1998-2000 increased overall case ascertainment by 70%. Cases were doubled in 1998, increased by 28% in 1999, and increased by 59% in 2000. <># Cases Maternal Deaths by Referral Source and Classification (1998-2000) 

Source	Pregnancy Related	Pregnancy Associated	Total
BVS	91	17	108
ME	10	22	32
SPARCS	18	8	26
Total	119	47	166 *

* Missing-3

Pregnancy-related cases were increased by 36%. Pregnancy-associated cases were increased by 176%. The use of ES allowed the identification of race, hemorrhage, and obesity as important correlates of MM.

Conclusions: BVS and death certificate data undercount MM. ES increases ascertainment of pregnancy related and associated deaths and improves the assessment of causal associations.