Regardless of the measure used, infant mortality is the number of infants who died before
they were 365 days old (from 0 through 364 days of age).
The differences in the three measures derive from three factors:
- Whether the number of infant deaths is measured from the infant DEATHS during a particular time period,
or the number of deaths among infants who were BORN during a particular time period
- The denominator used to calculate a rate
- The loss of cases when birth and death files are linked
The three different measures are described below with their pros and cons.
For more information on this issue, please go to the
NCHS web site.
Measure 1: Infant Mortality
The first measure, Infant Mortality, provides the number of infant deaths during a year (regardless of birth
year). The individual infant death records are not linked to the birth file. The number of births in the year
are used as a denominator, however, in calculating the rate. The concise definition for Measure 1 is as
follows: The number of infant deaths during the year per 1,000 live births during the same year.
- Includes all infant deaths, regardless of ability to link records to birth records (because no
linkage is performed).
- The data are available as soon as the death and birth data files are final for a given year.
- No information from the birth record is available for analysis.
- The infant mortality rate that is calculated is adequate at the state level, but becomes less
accurate for smaller areas and subpopulations.
| Measure 1 |
1999 |
2000 |
2001 |
| Death Record (numerator) |
Infant deaths during 2000 |
| Birth Record |
No linkage to birth record |
| Denominator for Rate |
Live births during 2000 |
| Availability of data file |
At end of calendar year |
Measure 2: Birth-Death Linked Infant Mortality (Birth Cohort)
The second measure, Birth-Death Linked Infant Mortality (Birth Cohort), provides the number of infant deaths
that occurred among infants born during a particular year (regardless of the year of death). The individual
infant death records ARE linked to the birth records, so that characteristics of the infant and mother may be
analyzed. The concise definition for Measure 2 is as follows: The number of infant deaths among infants born
in a given calendar year.
- The infant mortality rate that is calculated is most accurate, even for smaller areas and
subpopulations.
- Information from the birth record is available for analysis.
- The data file is not available until 364 days FOLLOWING the last birth in the calendar year, in
other words, this data file is available one year later than the other two.
- The infant mortality rate will not be counted for infants whose death records are not available.
This occurs rarely.
| Measure 2 |
1999 |
2000 |
2001 |
| Death Record (numerator) |
Infant deaths among infants born in 2000 |
Infant deaths among infants born in 2001 |
| Birth Record |
Live births during 2000 |
| Denominator for Rate |
Live births during 2000 |
| Availability of data file |
Linked file complete at end of calendar year 2001 |
Measure 3: Birth-Death Linked Infant Mortality (Death Cohort, or Period File)
The third measure, Birth-Death Linked Infant Mortality (Death Cohort, or Period File), provides the number of
infant deaths that occurred among infants born during a particular year (regardless of the year of death).
The individual infant death records ARE linked to the birth records, so that characteristics of the infant
and mother may be analyzed. The concise definition for Measure 3 is the same as for Measure 1: The number of
infant deaths during the year per 1,000 live births during the same year.
- The data are available as soon as the death and birth data files are final for a given year.
- The infant mortality rate that is calculated is adequate at the state level, but becomes less accurate for smaller areas and subpopulations
- Records will not be included in the linked file in those cases where a birth record can not be found. This occurs rarely.
- The infant mortality rate that is calculated is adequate at the state level, but becomes less accurate for smaller areas and subpopulations.
| Measure 3 |
1999 |
2000 |
2001 |
| Death Record (numerator) |
Infant deaths among infants born in 2000 |
Infant deaths among infants born in 2001 |
| Death Record (numerator) |
Infant deaths during 2000 |
| Birth Record |
Birth records for infants who died in 1999 |
Birth records for infants who died in 2000 |
| Denominator for Rate |
Live births during 2000 |
| Availability of data file |
Linked file complete at end of calendar year 2001 |
