# Age-specific fertility rates

The number of births occurring during a given year or reference period per 1,000 women of reproductive age classified in single-or five-year age groups

The ASFR is calculated as:

ASFRa = (Ba/Ea) x1000

Where:

Ba = number of births to women in age group a in a given year or reference period; and

Ea = number of person-years of exposure in age group a during the specified reference period.

Illustrative Computation

Estimates of Annual ASFRs for  all women 15-49, Egypt 1997-2000

 Age Group Births (Ba) Person- Years of Exposure (Ea) Rate/Woman Rate/1000 person years 15-19 764 14893.2 .051 51 20-24 2304 11747.2 .196 196 25-29 1994 9602.3 .208 208 30-34 1295 8805.5 .147 147 35-39 564 7549.5 .075 75 40-44 161 6643.2 .024 24 45-49 19 4498.8 .004 4

The number of births in a given year or reference period classified by age of mother and the number of women of reproductive age (i.e., 15-44 or 15-49 years), in 1-or 5-year age groups

Vital statistics (numerator only), population censuses or population-based surveys

The ASFR has two primary uses: (1) as a measure of the age pattern of fertility, that is of the relative frequency of childbearing among women of different ages within the reproductive years, and (2) as an intermediate computation in the derivation of the total fertility rate (TFR).

Evaluators may derive ASFRs from several sources. When evaluators estimate ASFR from vital statistics, they use population projections or estimates of the number of women in each age group 15-49 for the denominator in the rate. When using population censuses or surveys, evaluators obtain both the numerator and denominator of the rate from the census or survey. Estimates from censuses are derived from questions on births during a specified period preceding the census (usually 12 months), while survey estimates may be derived either from questions on births within a specified prior period or from partial or complete birth histories.

A simpler, although less precise, procedure for computing the denominator of the rate is to take the average of the number of women in each age group during the reference period covered by the measure (i.e., the average of the numbers of women in each age group at the beginning and end of the reference period).

Reference periods of more than one year are frequently used to compute ASFRs from survey data, the rationale being to decrease sampling variability associated with relatively small numbers of annual births occurring to women in single or five-year age groups and the distorting effects of reference period reporting errors. Various analyses of DHS fertility data, for example, alternately use the three- or five-year period prior to the survey in calculating ASFRs (Arnold and Blanc, 1989; Lutz, 1990). When multiple years are used for computational purposes, average annual rates are normally presented.

ASFRs are sometimes presented for different groups of women; for example, ASFRs are for women currently married or in union and for all women of reproductive age in DHS country reports. In societies where fertility is largely confined to marriage, ASFRs for women currently married or in union will provide more or less complete coverage of recent fertility. Where a large share of fertility occurs outside of recognized unions, however, the restriction of the ASFR to currently married women will result in an under estimate of the level of current fertility.

The ASFR is of particular interest in countries, cities, or districts with adolescent RH interventions designed to reduce unintended pregnancy. Although the ASFR is rarely used as an outcome measure in evaluating such programs (due to the human, financial, and logistic resources needed to collect the data), it is a variable that program administrators and policy makers track over time as a macro-level indicator of program effectiveness combined with non-program influences.

Unlike the crude birth rate, the ASFR is unaffected by differences or changes in population age composition, and thus is more useful in comparing different populations or sub-groups and in measuring changes over time. The ASFR is, however, affected by differences or changes in the number or percent of women exposed to the risk of pregnancy. Thus, changes in ASFRs may provide misleading information regarding the impact of family planning programs on fertility when other factors affecting risk of pregnancy are changing (for example, for the 15-19 and 20-24 age groups when age at marriage is rising quickly).

To address this problem, evaluators may calculate ASFRs only for women who were continuously married or in union during the reference period of the measure. The resulting measure is known as the marital age-specific fertility rate (MASFR). However, to calculate this measure, evaluators require data on duration of marriage or marriage histories. In actual practice, MASFRs are more often approximated by calculating ASFRs for women married or in union at the time of a survey, although evaluators should recognize that this figure only approximates the MASFR because women who are married or in union at the time of a given survey may not have been continuously married or in union over the entire reference period of the measure (e.g., for the three to five years prior to the survey).