Quarterly Provisional Estimates - Technical Notes - Infant Mortality, 2014–Quarter 3, 2016

Nature and sources of data

Provisional estimates are based on all complete death and birth records received and processed by the National Center for Health Statistics (NCHS) as of a specified cutoff date. National provisional estimates include events occurring to U.S. residents within the 50 states and District of Columbia. NCHS receives the birth and death records and monthly provisional occurrence counts from state vital registration systems through the Vital Statistics Cooperative Program.

Individual records are weighted, when necessary, to independent provisional counts of deaths or births occurring in each state by month. These monthly state-specific provisional counts serve as control totals and are the basis for the record weights used for computing provisional estimates. If the number of complete records is greater than the provisional count received from the state, the state-specific number of complete records is used instead, and the weight is set at 1.

Table I shows the percent completeness of the provisional data for infant deaths by month for the United States and each jurisdiction, based on where the deaths occurred. The percent completeness is obtained by dividing the number of complete records from each state for each month by the corresponding provisional count and multiplying by 100. The record weights described above are calculated by taking the inverse of the percent completeness (divided by 100) by state and month. Although data by place of occurrence are used to compute the weights, all rate estimates are for the residents of the 50 states and District of Columbia. Denominators for infant mortality rates (counts of live births) are based on final birth data for 2014 and 2015 (4,5) and provisional birth data for 2016.

 

… Category not applicable; no infant deaths reported.

¹ Excludes New York City.

NOTE: Percent completeness equals 100 times the number of infant death records received by NCHS divided by an individual provisional count of infant deaths reported by each jurisdiction. NCHS receives the death records and monthly provisional occurrence counts from state vital registration systems through the Vital Statistics Cooperative Program.

 

Computing rates

The “12 months ending with quarter” infant mortality rates are calculated by dividing the number of infant deaths in the 12 months ending with quarter by the number of births over the same time period and are presented as rates per 1,000 or per 100,000 live births. Denominator estimates used for computing rates are based on the number of births for the corresponding time period. These estimates represent rolling period estimates of infant mortality rates. Quarterly (i.e., 3-month) estimates are not provided due to the potential instability of quarterly estimates as well as seasonal fluctuations in the number of infant deaths and live births.  Rates presented here are based on deaths and births to residents of the 50 states and District of Columbia

The neonatal mortality rate refers to the total number of deaths to infants under age 28 days per 1,000 births. The postneonatal mortality rate refers to the number of deaths to infants aged 28 days through 11 months per 1,000 births.

 

Cause-of-death classification

Mortality statistics are compiled in accordance with World Health Organization (WHO) regulations specifying that WHO member nations classify and code causes of death in accordance with the current revision of the International Statistical Classification of Diseases and Related Health Problems (ICD). ICD provides the basic guidance used in virtually all countries to code and classify causes of death. It provides not only disease, injury, and poisoning categories but also the rules used to select the single underlying cause of death for tabulation from the conditions reported on the death certificate, as well as definitions, tabulation lists, the format of the death certificate, and regulations on use of the classification. Causes of death for data presented in this report were coded according to ICD guidelines described in annual issues of the NCHS Instruction Manuals (7).

Cause-specific infant mortality rates are presented per 100,000 live births. The five leading causes of infant death are: 1) Congenital malformations, deformations and chromosomal abnormalities (Q00–Q99); 2) Disorders related to short gestation and low birth weight, not elsewhere classified (P07); 3) Newborn affected by maternal complications of pregnancy (P01); 4) Sudden infant death syndrome (SIDS) (R95); and 5) unintentional injuries (V01–X59).

 

Accuracy of estimates

There are several reasons why estimates presented here may differ from other published reports of infant mortality. Because the provisional data have not undergone the standard processing, evaluation, imputations, and recoding done in the course of producing a final data file for release, estimates may differ slightly from those based on the final general mortality data files.

Provisional estimates are subject to some nonrandom sampling error. The quarterly provisional estimates are based on data that is potentially more incomplete for the most recent months. Data are weighted to independent control counts of deaths, and no imputations have been performed. It is assumed that the data are missing at random (i.e., the degree of missing data is unrelated to estimates of infant mortality), certain states may have more delayed reporting, and it is unknown whether infant mortality rates may be different for these states compared with states having complete reporting. Even without a differential delay, some sampling error would still exist for rate estimates, because they are based on incomplete data. A guideline for the size of this sampling error is given by deriving the variation that would occur if the data were missing at random (1); standard errors for infant mortality rates have been calculated according to these methods, accounting for sampling error.

Provisional or preliminary mortality data have generally been found to closely approximate final data (1). Due to the weighting of incomplete data, provisional estimates will not necessarily underestimate the true rate and may be higher in some cases than rates estimated using final data. Based on simulations of various levels of data completeness, ranging from 50% through 90%, 12 month-ending estimates of the infant mortality rates included in this release can be expected to be within 2% of the estimates based on complete data. Notably, these simulations assumed data were missing at random. It is likely that the data are not missing at random, and that certain causes of death such as SIDS and unintentional injuries are more likely to be underreported in recent quarters. Comparisons of provisional data with final data for the 2015 data year indicate that infant mortality rates due to SIDS and unintentional injuries may be underestimated by as much as 11%; thus, provisional estimates for these causes of death should be interpreted with caution.

Additionally, there are two sources of infant mortality data—the general mortality file (presented here) and the linked file of live births and infant deaths (8). Provisional data from the linked file is not currently available. Differences may occur between the data presented here and the data in the linked file, although the differences are expected to be negligible (6,8). Differences are mainly due to geographic coverage differences or the application of record weights (8). Specifically, the general mortality file includes infant deaths that occurred in the 50 states and District of Columbia., regardless of where the infant was born. In contrast, the linked file requires both the birth and the death to have occurred in the 50 states or District of Columbia (8).

Because the timeliness of death and birth reporting has been improving, accuracy of the estimates may change over time. Estimates for previously released quarters are revised to include data and updates received since the previous release. As a result, the reliability of estimates for a 12-month period ending with a specific quarter will improve with each quarterly release, and estimates for previous time periods may change with the addition of updated data.

 

Interpretation of changes over time

Unless otherwise specified, a difference in infant mortality rates is reported only if statistically significant at the 0.05 level by the z test given in reference 1. That z test uses a variance estimate that includes both the random fluctuation in the true number of deaths and the random sampling error.

 

Acknowledgement

The interactive dashboard was designed by Anthony Lipphardt.

 

References

1. Hoyert DL, Xu JQ. Deaths: Preliminary data for 2011 [PDF – 890 KB]. National vital statistics reports; vol 61 no 6. Hyattsville, MD: National Center for Health Statistics. 2012.
2. Middleton O, Baxter S, Demo E, Honeywell C, Jentzen J, Miller F, et al. National Association of Medical Examiners Position Paper: Retaining postmortem samples for genetic testing [PDF – 243 KB]. Acad Forensic Pathol 3(2):191–4..
3. Erck Lambert AB, Parks SE, Camperlengo L, Cottengim C, Anderson RL, Covington TM, Shapiro-Mendoza CK. Death scene investigation and autopsy practices in sudden unexpected infant deaths. J Pediatr 174:84–90.e1. 2016.
4. Martin JA, Hamilton BE, Osterman MJK et al. Births: Final data for 2015. National vital statistics reports; vol 66, no 1. Hyattsville, MD: National Center for Health Statistics. 2017.
5. Hamilton BE, Martin JA, Osterman MJK, et al. Births: Final data for 2014 [PDF – 2.8 MB]. National vital statistics reports; vol 64 no 12. Hyattsville, MD: National Center for Health Statistics. 2015.
6. Kochanek KD, Murphy SL, Xu JQ, Tejada-Vera B. Deaths: Final data for 2014. National vital statistics reports; vol 65 no 4. Hyattsville, MD: National Center for Health Statistics. 2016.
7. National Center for Health Statistics. National Vital Statistics System, instruction manuals.
8. Mathews TJ, MacDorman MF, Thoma ME. Infant mortality statistics from the 2013 period linked birth/infant death data set [PDF – 993 KB]. National vital statistics report; vol 64 no 9. Hyattsville, MD: National Center for Health Statistics. 2015.
9. Spencer MR, Ahmad F. Timeliness of death certificate data for mortality surveillance and provisional estimates [PDF – 210 KB]. National Center for Health Statistics. January 2017.