The MDAC is a study developed to permit research on social, economic, demographic, and occupational differentials in mortality. These objectives are achieved through record linkage of the large, nationally representative, 2008 American Community Survey (ACS) with mortality outcomes derived from the National Death Index (NDI) housed at the National Center for Health Statistics (NCHS) and Medicare/Medicaid information obtained from the Centers for Medicare and Medicaid Services (CMS). The database also includes data on local neighborhood and environmental characteristics such as: parks and recreation, crimes, hospitals and medical care facilities.  The MDAC database currently consists of approximately 4.5 million person records with over 308,000 mortality cases identified between the years 2008 and 2015. 

The ACS is designed as a replacement for the Decennial Census long form.  ACS is based on a complex, stratified sample of the 2000 Master Address File, therefore, it represents the population of the United States. All ACS responses are self-reported and cover a number of domains related to socio-economic status. In addition to demographic characteristics (age, race/ethnicity, sex, detailed geography), the ACS includes questions on place of birth, citizenship, years living in the US, primary language, years at residence, occupation, income, health insurance, work commuting habits, military service, and housing characteristics (type of housing, age of residence, number of rooms, heating fuel, cost of utilities, estimated value). The demographic and socio-economic variables from the ACS offer researchers the potential to answer questions on mortality differentials for a variety of important subgroups not covered as extensively in other databases.

Mortality information is obtained from death certificates of decedents through the use of the National Death Index (NDI) housed at the National Center for Health Statistics (NCHS). In order to conduct a match to the NDI, the NCHS algorithm requires that a Social Security Number (SSN) or Name and date of birth be provided. Approximately 91% of the MDAC records can be matched via an SSN, which indicates a high potential for matching to the NDI. Of the remaining records, name and date of birth are used. The 2008 ACS is linked to the NDI for years 2008 through 2015 providing a follow-up period of 7+ years. The total number of deaths identified to date is over 308,000. 

The MDAC is designed to enhance previous research by not only ascertaining the independent contributions of person-level and neighborhood characteristics, but will allow exploration of interactions across a broad spectrum of mortality causes. This research may be used for informing policy makers and by scientists designing clinical trials targeting interventions in population subgroups defined by both neighborhood and person-level traits. The design of the MDAC also allows researchers to address questions concerning public health. In addition to its large size and comprehensive information on individual socio-economic characteristics, detailed geography (county, tract, and block level) is available on the ACS allowing linkage to standard census neighborhood characteristics such as median income, poverty rates, education, and multi-unit housing. Linkage opportunities also include area resources such as local and state parks, neighborhood traffic density, fast food restaurant density, and healthcare facilities. The ACS is one of the few surveys, due to its large sample size, permitting examination of not only person-level and neighborhood characteristics, but their interaction as well. Linkage to CMS Medicare and Medicaid data expand the possibilities of research to include morbidity outcomes and healthcare utilization among the elderly and poor. The 2008 ACS has over 700,000 respondents aged 65+ of which 575,000 indicate Medicaid as a source of health insurance.

1. To provide an analytical research database to study the effect of demographic and socio-economic differentials on U.S. mortality rates.
2. To provide answers to specific public health questions which are not available from other sources.
3. To publish research findings in scholarly, scientific and health related journals.
4. To release a public-use file after extensive review to ensure that the confidentiality of respondents has been protected. The U.S. Census Bureau’s Disclosure Review Board and the organizations contributing data to the file must approve the release of the public-use file.