Vital Statistics, OverviewVital statistics, as a scientific discipline, is asubdomain of demography, the study of thecharacteristics of human populations. Vital statisticscomprises a number of important events in humanlife including birth, death, fetal death, marriage,divorce, annulment, judicial separation, adoption,legitimation, and recognition. The term “vitalstatistics” is also applied to individual measures ofthese vital events. Thus, a birth rate is an example ofa vital statistic and an analysis of trends in birth ratesis an example of an application in the field of vitalstatistics. A vital statistics system is the total processof collecting by civil registration, enumeration, orindirect estimation, information on the frequency ofoccurrence of vital events, selected characteristicsof the events and the persons concerned, and thecompilation, analysis, evaluation, and disseminationof these data in summarized statistical form. Otherlife events of demographic importance such aschange of place of residence (migration), change ofcitizenship (naturalization), and change of name arenot included, mainly because information on these isusually derived from other statistical systems such aspopulation registers [1].

Systems for Collecting Vital Statistics

It is generally accepted that the preferred methodfor individual countries to collect vital statistics isthrough a civil registration system. This is recognizedby the United Nations (UN) and other internationalorganizations, as well as by the many countries thathave had civil registration laws and regulations inplace and in operation for many years [1, 2]. Never-theless, a number of newly emergent and developingnations, facing the difficulties and length of time ittakes to create a satisfactory civil registration sys-tem, have instituted alternative procedures to acquirestatistical data to describe the levels and trends forkey vital events, particularly for fertility and mortalitymeasurements. The UN recognizes the importance ofa civil registration system for each country as the pre-ferred source of vital statistics data for the long run.However, use of an alternative data collection systemis recommended as an interim measure for meetingneeds for essential information where a civil registra-tion system of acceptable quality does not yet exist.

Other systems include, for example, probability areasamples (see Probability Sampling), purposeful areasamples, records-based surveys, and record linkage.Furthermore, the UN recommends a priority order forthe types of vital statistics data to be collected. Thehighest need is given to data on births and deaths, fol-lowed in order by marriages, divorces, fetal deaths,annulments, judicial separations, adoptions, legitima-tions, and recognitions [1].

Uses of Vital Records and Vital Statistics

Vital records created through a civil registration sys-tem have two classes of use. They have value indi-vidually as legal documents for the persons namedthereon; they also constitute the input, when aggre-gated, for the various vital statistics measures thatare used to study the demographics and health ofpopulations and population subgroups.

For the individual, a birth record is a legal docu-ment establishing name, parentage, birth data, orderof birth for multiple births, legitimacy, and citizen-ship, nationality, or geographic place of birth. A widevariety of individual rights and civil entitlementsdepends on these facts, including proof of age forschool entrance, motor vehicle drivers’ licenses, mili-tary service and other age-related activities, establish-ment of eligibility for family allowances, insurancebenefits, tax benefits, inheritance rights, issuance ofpassports, etc. The death record provides documen-tary proof of the facts of death needed for social secu-rity and insurance purposes such as time and place ofdeath and the medical cause of death. Proof of deathand the associated facts are also used for propertyinheritance rights, for remarriage rights of surviv-ing spouses, etc. Marriage and divorce records serveto document rights to special social and economicprograms and benefits for the married, including taxprivileges for couples, alimony, change of national-ity based on marriage, and the right to remarry. Manyrights of children, their parents, and their guardiansare dependent on records of adoption, legitimation,and recognition.

Individual vital records may also be used admin-istratively as the basis for initiating maternal andchild health services, including child immunizationprograms, or for epidemiologic investigations intodisease outbreaks or assessments of causes of acci-dents and injuries. Another important administrative

2 Vital Statistics, Overview

use of individual vital records especially of deathrecords (see Death Certification), is for the updatingor clearing of files such as electoral rolls, social secu-rity files, disease registers, cohort follow-up studies,tax registers, etc.

In aggregated form, vital records become a col-lection of vital statistics, most often in the form ofmeans, medians, and various ratios such as propor-tions and rates. Whether collected by civil registra-tion or by other means, vital statistics serve as keydemographic variables in the analysis of populationsize, growth and geographic distribution, especiallywhen used in conjunction with periodic populationcensuses. When census data are used as a base, cur-rent intercensal estimates of population size can bemade, and projections into the future can be pre-pared using estimates of future trends in fertility,natality, and mortality linked with estimates of netmigration. In addition to the importance of vitalstatistics to the study of population size and growthtrends, other national and subnational economic andsocial concerns such as health, welfare, education,occupation, housing, urbanization, family structure,and income are also affected by these measures. Inthe fields of public health and medicine, for example,levels and trends of infant and perinatal mortal-ity are often used as surrogate measures of levelsand trends in the overall health and well-being ofnations. Life expectancy at birth is also frequentlyused to compare the overall effects of mortality andits determinants. Cause of death information pro-vides a foundation upon which much research intodiseases and disease prevention is based.

Differentials in mortality by sex, age, racial groups,and other variables are often the basis for the planningof health and medical intervention programs. In addi-tion, the planning and provision of public and privatehousing, educational facilities, social security and pri-vate insurance plans, medical facilities, and consumergoods of all kinds are examples of activities depen-dent on vital statistics data. At the international level,vital statistics provide a basis for comparing impor-tant demographic, social, and economic differencesand trends over time among countries or regions ofthe world.

Definitions of Selected Vital Events

Standard statistical definitions of vital events havebeen promulgated by international agencies [1, 5]. In

some cases, legal definitions may differ from theinternational standards in varying degrees, but, inmany cases, national vital statistics reports are eitherbased on the standard statistical definitions or donot differ in principle. In cases where comparabilityamong countries is compromised because of the useof nonstandard definitions, international agencies andothers presenting national comparisons of tabular,graphical or descriptive vital statistics usually provideappropriate cautions to users. Nevertheless, users ofvital statistics data need to ascertain the comparabilityof the data before drawing reliable conclusions aboutnational differences. The World Health Organiza-tion (WHO) promulgates a number of vital statisticsdefinitions as part of the International Classificationof Diseases (ICD). These definitions are incorporatedin regulations adopted by the World Health Assemblyand which each WHO member country has agreed tofollow [4]. Nevertheless, it is still necessary to ensurethat the standard definitions have been followed for agiven data set. The international standard definitionsfor selected vital events are given below.

Live Birth. This is the complete expulsion or extrac-tion from its mother of a product of conception,irrespective of the duration of the pregnancy, which,after such separation, breathes or shows any otherevidence of life, such as beating of the heart, pulsa-tion of the umbilical cord, or definite movement ofvoluntary muscles, whether or not the umbilical cordhas been cut or the placenta is attached; each productof such a birth is considered liveborn [5].

Fetal Death. This is death prior to the completeexpulsion or extraction from its mother of a productof conception, irrespective of the duration of preg-nancy; the death is indicated by the fact that after suchseparation the fetus does not breathe or show anyother evidence of life, such as beating of the heart,pulsation of the umbilical cord, or definite movementof voluntary muscles [5].

Maternal Death. This is the death of a woman whilepregnant or within 42 days of termination of preg-nancy, irrespective of the duration and the site ofthe pregnancy, from any cause related to or aggra-vated by the pregnancy or its management, but notfrom accidental or incidental causes. Maternal deathsmay be subdivided into two groups: direct obstetric

Vital Statistics, Overview 3

deaths which are the result of obstetric complica-tions of the pregnant state (pregnancy, labor, and thepuerperium), from interventions, omissions, incorrecttreatment, or from a chain of events resulting fromany of these; and indirect obstetric deaths which arethe result of previously existing disease or diseasethat developed during pregnancy and which was notdue to direct obstetric causes, but which was aggra-vated by physiologic effects of pregnancy [5].

Infant Death. This is the death of a liveborn infantwho dies before completing its first year of life.

Neonatal Death. This is the death of a liveborninfant who dies during the first 28 completed daysof life. These may be subdivided into early neonataldeaths, occurring during the first seven days of life,and late neonatal deaths, occurring after the comple-tion of the seventh day but before the completion of28 days [5].

Perinatal Death. This is the death of a fetus ornewborn infant occurring after 22 completed weeks(154 days) of gestation (the time when fetal weightis normally about 500 g), but prior to the completionof seven days after birth [5].

Marriage. This is the act, ceremony or process bywhich the legal relationship of husband and wife isconstituted. The legality of the union may be estab-lished by civil, religious, or other means recognizedby the laws of each country [1].

Divorce. This is a final legal dissolution of a mar-riage which confers on the parties the right toremarriage under civil, religious, or other provisions,according to the laws of each country [1].

Definitions of Selected Vital StatisticsMeasures

Raw vital statistics most often are comprised ofcounts of how often a specified vital event hasoccurred, rather than on measurements of continuousvariables such as height, weight, or blood pressure.The analysis of vital data depends mainly on the con-version of observed frequencies into indices, ratios,and probabilities. Counts of vital events often do have

utility, but, for the majority of uses, absolute frequen-cies are not sufficient and it becomes necessary tocompute relative numbers, including crude rates, var-ious types of specific rates, percentages, probabilities,and other ratios.

Some of the more commonly encountered vitalstatistics relative numbers are defined and calculatedas follows.

Crude Death Rate

The most common form of mortality measurementis the crude death rate. It is computed from thefollowing formula [3]:

mcd =!

D

P

"k,

where mcd is the crude death rate, D is the total num-ber of deaths for a given area and time period, usuallya calendar year, P is the size of population at riskof dying, usually taken as the estimated population atthe midpoint of the calendar year, and k is a constant,usually taken as 1000.

The crude rate is so named to differentiate it fromvarious specific and adjusted rates and representsthe total or overall death rate without regard tothe various component elements which combine toproduce the total figure. The crude death rate isusually expressed as “the number of deaths per 1000persons” for a specified place (country, city, state,etc.) for a given year.

Specific Death Rate

Detailed analyses of vital statistics frequently gobeyond the overall risk of death in the populationas a whole. Many studies deal with subsets of thepopulation or with particular classes of deaths. Epi-demiologists often focus on deaths from a particulardisease or class of diseases. Actuaries and demogra-phers are concerned with differences in mortality bysex and in different age groups within the population.Environmental and occupational health specialistsare interested in the differential risks of dying inselected occupations, and in different geographic sub-divisions such as urban and rural areas. To meet thesekinds of needs, various specific death rates are calcu-lated. Specific rates for different age groups are calledage-specific death rates; rates for males and females

4 Vital Statistics, Overview

are called sex-specific death rates, rates for particularcauses of death are called cause-specific death rates.Rates may be specific for combinations of charac-teristics. For example, age–sex–race-specific deathrates are computed separately for each age group byrace and sex. Specific death rates are approximationsof true probabilities. That is, the denominator of theratio is an estimate of the total number of events of aparticular type that could happen, while the numeratoris a count of those that did happen.

Specific death rates are computed as follows [3]:

msd =!

di

pi

"k,

where msd is the specific rate for any defined ith class,di is the number of deaths occurring in the ith classfor a given area and time, pi is the number of personsin the ith class of the population for the same areaand time, and k is a constant, usually 100 000.

For cause-specific death rates, the denominator,pi , in the above formula is replaced by P , the totalpopulation exposed to the risk of death. Therefore, acause-specific death rate measures the risk in the totalpopulation of dying from a specified cause of death.

Infant Mortality Rate

The infant mortality rate is considered by manyas one of the important indicators of the overalllevel of health and social well-being of a countryor other geopolitical area. This is, in part, becausea large proportion of deaths in the first year of lifeare considered to be preventable through adequateprenatal care, good nutrition for women and infants,and improved control of the environment, includinginjury prevention.

The infant mortality rate is computed as fol-lows [3]:

mi =!

d<1

B

"k,

where mi is the infant mortality rate, d<1 is thenumber of deaths to liveborn infants under one yearof age during a specified time period, usually oneyear, B is the total number of live births during thesame time period, and k is a constant, usually 1000.

The infant mortality rate is a proxy for the age-specific death rate for the “under one year of agegroup” and is intended to be a measure of the riskof dying during the first year of life. The numerators

of the infant mortality rate and the “under one yearof age” age-specific death rate are the same. For adenominator, however, a reliable estimate of the sizeof the population under one year of age for a giventime period is hard to obtain, even in a census year.As a proxy measure, the denominator may be consid-ered to be the number of births occurring during theperiod. For either of these choices of denominator,there is some mismatch with the numerator in termsof a true probability number. Not all events in thenumerator arise from the events in the denominator.For example, in the infant mortality rate, some of thedeaths under one year of age in a given year andcounted in the numerator were actually born in theprevious year and are not represented in the denom-inator, while some of the births represented in thedenominator will die before their first birthday butthe deaths will occur in the next year and are notincluded in the numerator. However, when the birthrate is fairly stable from one year to the next, cal-culation of the infant mortality rate results in a ratiothat closely approximates the probability of a live-born infant dying within the first year of life. Whenthe birth rate is not stable from year to year, a moreaccurate mortality rate may be computed by follow-ing each live birth occurring during a one year periodand measuring how many of them die before theirfirst birthday.

Neonatal, Early Neonatal and PostneonatalMortality Rates

The neonatal mortality rate is defined as follows [5]:

mn =!

d<1 mo

B

"k,

where mn is the neonatal mortality rate, d<1 mo isthe number of deaths of infants under 1 month ofage during a specified time period, B is the numberof live births occurring during the same time period,and, k is a constant, usually 1000.

The neonatal mortality rate, like the infant mor-tality rate, is a proxy for an age-specific death rate.It approximates the risk of dying in the first monthof life. The relative importance of an infant mortal-ity rate compared with the corresponding neonatalmortality rate depends on the proportionate age distri-bution of the deaths under one year of age. Generally,when the infant mortality rate is low, a large pro-portion of infant deaths occur during the first month

Vital Statistics, Overview 5

of life. The neonatal mortality rate then reflects animportant measure of the mortality risk for infants.Conversely, when the infant mortality rate is high,larger proportions of deaths fall into the older agegroups under a year. Often it is useful to partitionthe deaths of infants under one year of age into twogroups: those dying before one month of age, andthose dying between one month and their first birth-day. The former comprise the numerator, d<1 mo, ofthe neonatal mortality rate, while the latter can beused to calculate the postneonatal mortality rate:

mpn =!

d1 mo – 1 yr

B

"k,

where mpn is the postneonatal mortality rate,d1 mo – 1 yr is the number of deaths occurring between1 month and 1 year of age during a specified timeperiod, B is the number of live births occurringduring the same time period, and k is the sameconstant used in the neonatal mortality rate, usually1000.

In similar fashion, the neonatal deaths may bepartitioned into those dying within the first week oflife and the remainder that survive the first seven daysbut die before one month of age. The risk of dyingin the first week of life is measured by the earlyneonatal mortality rate, men, as follows [5]:

men =!

d<7 days

B

"k,

where the components of the calculation are the sameas in the neonatal and postneonatal mortality rates,except that the numerator contains only those deathsto infants occurring during the first week of life.

Perinatal Mortality Rate

The perinatal period, as defined earlier, is the periodof time surrounding the event of birth. It includes thetime that a fetus spends in utero after it has reached22 weeks of gestation and continues through the birthprocess until the end of the first week of life afterbirth. The perinatal mortality rate measures mortalityoccurring during this period. The rate, therefore,combines deaths of fetuses of specified gestationalage with deaths of liveborn infants who die in theirfirst week of life. The determination of whether afetus is born dead or whether it shows any sign oflife before expiring is not always clear-cut; social,

economic, and cultural factors, as well as medicaland biological considerations, tend to push the fetaldeath rate in one direction or the other in differentsocieties, thus making comparisons of neonatal orinfant mortality among countries difficult. By usingthe perinatal mortality rate for comparisons, thisdifficulty is minimized since fetuses dying just beforeor during the birth process as well as those born alivebut dying shortly thereafter are all included in thecalculation [5]:

mperi =#

dperi

F + B

$k,

where mperi is the perinatal mortality rate, dperi is thenumber of deaths of fetuses of 22 or more weeksof gestation plus deaths of liveborn infants of lessthan 7 days of age during a specified period, usuallya calendar year, F is the number of fetal deaths of 22or more weeks of gestation during the same period,B is the number of live births during the same period,and, k is a constant, usually 1000.

Note that, unlike the infant and neonatal mortalityrates, the denominator of the perinatal mortality ratecombines both the number of live births and thenumber of fetal deaths of 22 or more weeks ofgestation. This denominator is called “total births”and better approximates the population from whichthe numerator could arise than would a denominatorrestricted to only live births. On the other hand, it isrecognized that it is easier to collect reliable countsof live births than of fetal deaths, thus introducinganother source of error into the calculation of theperinatal mortality rate.

Maternal Mortality Rate

The maternal mortality rate is calculated as fol-lows [5]:

mm =#

dmd + dmi

B

$k,

where mm is the maternal mortality rate, dmd is thenumber of direct maternal deaths in a specified timeperiod, usually 1 year, dmi is the number of indirectmaternal deaths in the same period, B is the numberof live births in the same period, and k is a constant,usually 10 000 or 100 000.

A related measure, the direct obstetric mortalityratio, may be calculated from the above formula

6 Vital Statistics, Overview

but using in the numerator only the direct maternaldeaths, dmd.

Proportionate Mortality

Proportionate mortality, sometimes known as thedeath ratio (see Proportional Mortality Ratio(PMR)), is defined as [3]:

pd =!

di

D

"k,

where pd is the proportionate mortality, di is thenumber of deaths in a specified class during a statedtime period, D is the total number of deaths in thesame time period, and k is a constant, usually 100 or1000.

Proportionate mortality ratios may be calculatedfor any class of deaths, but their most common usesare for given causes or group of causes of deathexpressed as percentages of deaths from all causes, orfor deaths at a specified age expressed as percentagesof deaths at all ages.

Crude Birth Rate

The crude birth rate is the most frequently usedoverall measure of the reproduction of a population.Like its counterpart, the crude death rate, it isinfluenced by many factors and represents a proxy formore specific fertility measurements. It is calculatedas follows [3]:

mcb =!

B

P

"k,

where mcb is the crude birth rate, B is the totalnumber of live births for a given area and time period,P is the total population at the midpoint of the timeperiod, and, k is a constant, usually 1000.

Comparing Vital Statistics Data

Aggregated vital statistics data, whether in tabular orgraphical form, often appear as time trends for par-ticular variables such as causes or groups of causesof death, or for age and sex groups of the population.They also appear frequently as comparisons betweencountries or other geographical entities for a point intime, usually a particular year. In either case, great

care must be taken to ensure that the quality of thedata in the groups being compared warrants makingthe comparisons. In registration based systems, mea-sures or estimates of completeness of reporting ofvital events should be known. In sample based sys-tems, the representativeness of the sample and thenonresponse rate is important. In the comparison ofdata between two or more geographic places, it isimportant to ascertain if common definitions and pro-cedures were used to collect, process, analyze, andpresent the data; in looking at time trends, it is essen-tial to know if the definitions of the events and theprocedures for classifying the data remained constantover the entire time period being studied. This lat-ter point is particularly important when looking attrends in causes of death since the instrument forgrouping diseases into categories for study, the ICD,is revised approximately every 10 years (see Mor-bidity and Mortality, Changing Patterns in theTwentieth Century; Mortality, International Com-parisons). Vital statistics data are often presented instatistical compendia published by official nationaland international organizations that attempt to includeimportant notes for interpretation of the data in head-notes and footnotes to tables, appendices, etc. (seeData Access, National and International). The useris cautioned to pay careful attention to such explana-tory or cautionary notes.

References

[1] United Nations (1973). Principles and Recommendationsfor a Vital Statistics System. Statistical Papers, Series M,No. 19, Rev. 1. United Nations, New York.

[2] United States Bureau of the Census (1971). The Methodsand Materials of Demography, H. Shryock, J. Siegel et al.,eds. US Government Printing Office, Washington.

[3] United States Department of Health, Education, andWelfare (1965). Techniques of Vital Statistics. Reprintof Chapters I-IV, Vital Statistics Rates in the UnitedStates, 1900–1940. National Center for Health Statistics,Washington.

[4] World Health Organization (1967). WHO NomenclatureRegulations. World Health Organization, Geneva.

[5] World Health Organization (1992). International Statisti-cal Classification of Diseases and Related Health Prob-lems: 10th revision, 3 vols. World Health Organization,Geneva.

ROBERT A. ISRAEL