National Mortality Followback Survey

The National Mortality Followback Survey (NMFS) has been conducted sporadically since 1961 by the National Center for Health Statistics (NCHS) to collect information on various specific topics related to mortality in the United States, including the events and circumstances that preceded death, and relevant characteristics of the decedent. The NMFS has been conducted six times: four times in the 1960s, in 1986, and most recently, in 1993.

The main national database for the United States mortality statistics (the National Death Index) is maintained by the NCHS and is compiled from the official death certificate registries maintained by individual states. The NMFS was created to enrich the national mortality database by collecting information not available from death certificates. Each of the six surveys has been unique; each has focused on different topics and has used a variety of survey instruments and sources. The objective has been to focus on specific topics of current interest to public health researchers and policymakers, using instruments designed to fit the purposes of the individual survey rather than using a uniform instrument to gather data on a consistent group of topics.

All the surveys have obtained information from persons identified on the death certificates as informants (next of kin or a person familiar with the decedent). The surveys have collected information by means of mail questionnaires, personal interviews, and/or telephone interviews. The sample for each survey has been [Page 718]drawn from the Current Mortality Sample, a systematic 10% of the states’ death certificates.

The topics covered have varied from survey to survey. The 1961 survey sought information on the use of hospital and institutional care during the last year of life. The 1962 to 1963 survey focused on socioeconomic factors. The 1964 to 1965 survey obtained data on health care expenditures during the last year of life, sources of payment, and health insurance coverage. The 1966 to 1968 survey focused on the link between smoking and cancer mortality.

The two most recent surveys have been more comprehensive (covering a larger sample and more information items and sources) than the surveys conducted in the 1960s, and they have attempted to provide comparability with the data obtained by previous surveys.

The 1986 survey covered three topics: socioeconomic factors, risk factors related to premature death, and health care provided in the last year of life. Brief questionnaires were mailed to all hospitals, nursing homes, and other health care facilities reportedly used by decedents in their last year of life. The sample (18,733) was an approximate 1% sample of all deaths of adults over age 25 in 1986.

The 1993 survey included the 1986 topics and added a fourth topic, disability in the last year of life. The 1993 survey sought information from medical examiner/coroner offices if death was due to homicide, suicide, or accidental injury. The 1993 sample included 22,957 deaths in 1993 of persons aged 15 and older.

Both the 1986 and 1993 surveys examined the reliability of information reported on the death certificate by comparing it with the information reported by the survey respondent. The comparable items included age, race, gender, veteran status, education, occupation, and industry.

...