An Analysis of the Estimated Figure of induced
abortions in the Philippines in 2000
as published in a 2006 Guttmacher Institute report by Susheela Singh et al
By Dr. Roberto De Vera
Summary: In the
2006 Guttmacher Institute report "Unintended Pregnancies and Induced
Abortions in the Philippines: Causes and Consequences", Susheela Singh et
al estimated that there were 473,000 induced abortions completed in the
Philippines in 2000 using a method consisting of three steps. First, based on
reports gathered from 2,039 hospitals which contained the top ten leading
causes of admission in the 1999-2001 period, they arrived at an estimate of the
number of women in 2000 who were hospitalized due to complications from both
induced and spontaneous abortions. Second, they calculated the number of women
hospitalized for induced abortions by subtracting the estimated number of women
hospitalized for spontaneous abortions (or miscarriages) from the estimated
number of women hospitalized for induced and spontaneous abortions. Finally,
they arrived at the estimated number of women who had induced abortions by multiplying
the estimated number of women hospitalized for complications due to induced
abortions by 6 to account for the women who had induced abortions who didn't go
to the hospital.
We find that
their method overestimates the figure of induced abortions in the Philippines
in 2000 because of three flaws. These flaws had the effect of 1) overestimating
the figure for women hospitalized for spontaneous and induced abortions due to
an assumption that is weakly supported by statistical data; 2) underestimating the
number of women hospitalized for complications due to spontaneous abortions (or
miscarriages) because it mistakenly covers only those women with spontaneous
abortions occurring in 12th to 22nd week of pregnancy who were hospitalized for
complications; and 3) using a multiplier which most likely is higher than the
ratio of the number of women who have induced abortions to the number of women
who are hospitalized for complications due to induced abortions.
Using modified
version of the Singh et al methodology (corrected to account for the above
flaws), we arrived at an alternative estimate of 25,924 induced abortions in
the Philippines in 2000 (1.3 abortions per 1,000 women in the reproductive
age). Using a second method, we multiplied 0.0117, the share of induced
abortions to live births by the number of live births in 2000, to arrive at
second estimate of 20,831 induced abortions in the Philippines in 2000 (1.1
abortions per 1,000 women of reproductive age). We consider these two estimates
of induced abortion in the Philippines in 2000 to be more reasonable than the
473,000 estimate (24.5 induced abortions per 1,000 women of reproductive age)
published in the 2006 Guttmacher Institute report.
In the 2006
Guttmacher Institute report "Unintended Pregnancies and Induced Abortions
in the Philippines: Causes and Consequences", Susheela Singh et al
estimated that there were 473,000 induced abortions completed in the
Philippines in 2000 using a method [NOTE 1]
consisting of three steps. First, based on reports gathered from 2,039
hospitals which contained the top ten leading causes of admission in the
1999-2001 period [NOTE 2],
they summed up three figures (i.e. one estimate each for 1) 478 hospitals which
had abortion as among the top ten leading causes of admission; 2) 1,180
hospitals which didn't have abortion as among the top ten leading causes of
admission; 3) 381 hospitals without reports) to arrive at an estimate of the
number of women in 2000 who were hospitalized due to complications from both
induced and spontaneous abortions. Second, they calculated the number of women
hospitalized for induced abortions by subtracting the estimated number of women
hospitalized for spontaneous abortions (or miscarriages) from the estimated
number of women hospitalized for induced and spontaneous abortions. Finally,
they arrived at the estimated number of women who had induced abortions by
multiplying the estimated number of women hospitalized for complications due to
induced abortions by 6 to account for the women who had induced abortions who
didn't go to the hospital. [NOTE 3]
We find that
their method overestimates the figure of induced abortions in the Philippines
in 2000 because of three flaws. First, the method overestimates the figure for
women hospitalized for spontaneous and induced abortions due to an assumption
that is weakly supported by statistical data. Its assumption for 1,180 out of
2,039 hospitals which didn't have abortion among the top ten leading causes for
admissions--that the number of abortions was calculated to be 50% of the figure
corresponding to the tenth leading cause of admissions--was based on data
observed from a small sample of 6 hospitals in Northern Mindanao. [NOTE 4]
Second, the method underestimates the number of women hospitalized for
complications due to spontaneous abortions (or miscarriages) because it
mistakenly covers only those women with spontaneous abortions occurring in 12th
to 22nd week of pregnancy who were hospitalized for complications. Third, the
method uses a multiplier which most likely is higher than the ratio of the
number of women who have induced abortions to the number of women who are
hospitalized for complications due to induced abortions. Based on a 1997
Susheela Singh et al that concluded that the national multiplier most likely is
between 4 and 5, it seems more reasonable to use 4.5 as a multiplier instead 6 that
was used. We used a modified version of the Singh et al methodology (corrected
to account for the above flaws) to arrive at an estimated number of 25,924
induced abortions in the Philippines in 2000. [NOTE 5]
This estimate
seems to be reasonable for it is relatively close to an estimate generated by a
direct method consisting of two steps. First, based on the 1993 Philippine
National Safe Motherhood Survey that estimated that 1% of all pregnancies end
in induced abortion, we estimated that 1.17% of all live births end in an
induced abortion. Second, we multiplied 0.0117 by the number of live births in
2000, 1,766,440 to arrive a second estimate of 20,831. [NOTE 6]
Summing up, due
to several flaws in their methodology, Singh et al's figure of 473,000 induced
abortions in the Philippines in 2000 (24.5 induced abortions per 1,000 women in
reproductive age) grossly overestimates the actual figure. Using two methods,
we think we have arrived at estimates of induced abortions in the Philippines
in 2000: 25,924 and 20,831 (1.3 and 1.1 abortions per 1,000 women in the
reproductive age, respectively) that are more reasonable that those published
in the 2006 Singh et al report.
Table
1: Calculations for arriving at number of induced abortions in the Philippines
in 2000 using the methodology found in 2006 Guttmacher report
Quick
guide to calculations below: Note that abortions in categories a1, a2, a3 and b
include both induced and spontaneous abortions. Thus, to arrive at the number
of women hospitalized for induced abortion complications (d), we need to subtract the number of women
hospitalized for spontaneous abortions (c) from the estimated number of women
hospitalized for both induced and spontaneous abortions (b). The figure in
(d) is multiplied by a multiplier that accounts for the women that may have had
spontaneous and induced abortions that were not hospitalized for complications.
This figure refers to the year 2000.
3 Methods in determining the
number of women hospitalized for complications due to spontaneous and induced
abortions
Estimated figure
a1 Method:For the 478 hospitals in which abortion was one of the ten leading
causes of admission, the number of women hospitalized for abortion
complications was estimated directly from the hospital's information.
71,837
a2 Method:For the 1,180 hospitals in which abortion was not among the top ten
causes of admission, they estimated the
number of women hospitalized for abortion complications as one-half the
figure of the tenth leading cause of admission since they argued that since
abortion complications had different admission codes it was possible that
these separate figures would be too small to reach the top ten.
26,845
a3 Method:For the 381 hospitals "whose basic characteristics were known,
but for which there were no data on cause of discharge or no report was
available", "[t]he likely number of abortion complications treated
annually in these hospitals was estimated using regression analysis."
The analysis was based on the number of
abortions as the dependent variable and "ownership (public vs. private),
hospital level (primary, secondary, or tertiary), hospital size (number of
beds) and region (according to the 16-region classification)" as the
dependent variables.
6,311
b)Estimated number of women with
abortion (spontaneous and induced) with complications. (Add figures in
a1, a2 and a3).
104,993
Less (or
subtract from the above estimates)
c)Number of women with spontaneous
abortions hospitalized for complications.
"An assumption was made that late spontaneous abortions (those
occurring at 13-22 weeks) are likely to be accompanied by complications that
will require hospital care. In the absence of induced abortion, these
miscarriages account for 2.89% of all recognized pregnancies, and the ratio
of spontaneous pregnancy losses to live births (which represent 84.8% of all
observed pregnancies) is 3.41 per 100 (2.89/84.8)." After estimating the
number of women with late pregnancy spontaneous abortions, they applied this
to the percentage of women of had their baby delivered at a health facility
at the regional level.
26,092
d)Estimated number of women with
induced abortion hospitalizedfor complications (Subtract the figure in c from the figure in b.)
78,901
e) Estimated number of women who have had induced abortions ina year. (Multiply figure in d by the
corresponding multiplier.)
This was arrived at by applying a
multiplier to 78,901. This multiplier isarrived at by dividing 1 by the share of women who have had abortion
who were hospitalized for complications.
multiplierestimate
4315,604
5394,505
6473,406
Table
2: Calculations for arriving at number of induced abortions in the Philippines
in 2000 using 2006 Singh et al and modified methodology. (See Table 1 for
details on categories below.)
3
Methods in determining the number of women hospitalized for complications due
to spontaneous and induced abortions
Singh et al
Method
Modified
Method
a1 Method:For the 478 hospitals in which abortion was one of the ten leading
causes of admission
71,837
71,837
a2 Method:For the 1,180 hospitals in which abortion was not among the top ten
causes of admission
26,845
17,897(a)
a3 Method:For the 381 hospitals "whose basic characteristics were known,
but for which there were no data on cause of discharge or no report was
available".
6,311
6,311
b)Estimated number of women with
abortion (spontaneous and induced) with complications. (Add figures in
a1, a2 and a3).
104,993
96,045
Less (or subtract from the above
estimates)
c) Number
of women with spontaneous abortions hospitalized for complications.
26,092
90,284(b)
d) Estimated number of women with induced abortion hospitalized for complications (Subtract
the figure in c from the figure in b.)
78,901
5,761
e) Multiplier
6
4.5(c)
f) Estimated number of women who have had induced abortions in a year.
(Multiply figure in d by the corresponding multiplier in e.)
473,406
25,924
Notes: a)First, for hospitals which did not include
abortion as among the top ten causes for admission in their report, the method
used to estimate the number of induced and spontaneous abortions (taken to be
50% of the number of admissions of the tenth cause) rests on an assumption (see
explanation in a2 above) that was based on data gathered from a small sample of
6 hospitals in Northern Mindanao.Quote: For
six hospitals where abortion ranked below the 10th cause and for which data
were available on both the 10th cause and the number of abortion patients,
abortion complications accounted for about 60% of the total number of patients
hospitalized for the 10th-ranking diagnosis (taken from a footnote on p. 101
of Singh S et al, Estimated the level of abortions in the Philippines and
Bangladesh, International Family Planning
Perspectives, 23(3), pp. 101-107 & 144, 1997).We find that the correlation used to support their
assumption was based on too small a sample--6 out of 2,039 hospitals.
So it is reasonable to reduce the Guttenmacher Instutitute estimated figure of
26,845 by a third, or 17,897. b)Second, the figure calculated for women
hospitalized for spontaneous abortion complications may have been
underestimated because it was based on spontaneous abortions (or miscarriages)
that occurred in the 13 to 22 weeks of pregnancy, and disregarded the
spontaneous abortions that happened in the first 22 weeks, which is when most
of them take place. Since "[s]tudies reveal that anywhere from 10-25% of
all clinically recognized pregnancies will end in miscarriage" (taken from
the American Pregnancy Association website at
http://www.americanpregnancy.org/pregnancycomplications/miscarriage.html), in
this case it would be reasonable to take 10% (the lower figure of the range) as
the share of pregnancies that end in spontaneous abortions or miscarriages.
This translates into 11.79% share of of spontaneous abortions (or miscarriages)
to live births. By ratio and proportion, this translates into 90,284 (i.e.
26,092 times 11.79 divided 3.41) women hospitalized for complications due to
spontaneous abortions (or miscarriages).
c)
It should be emphasized that the 473,000 induced abortion figure was based on a
multiplier value of 6 which covers the case for rural areas where women have
poorer access to hospitals, when in fact the 1997 Singh et al paper mentioned
that the multiplier applied to the nation would probably be between 4 and 5. So it
would be reasonable to apply the multiplier 4.5 to the adjusted figure in d) of
5,761 in column of modified method, or 25,924 which is the new estimated figure
for spontaneous and induced abortions done in year 2000.
NOTE 1. The discussion of the methodology is drawn
from Appendix of Singh S, "Unintended Pregnancy and Induced Abortion in
the Philippines", Guttmacher Institute, 2006, pp. 31-36. Unless pointed
out otherwise, text in quotes are taken from the appendix of the 2006 report.
This methodology is based mainly from Singh S. et al, Estimating the level of
abortion in the Philippines and Bangladesh, International Family Planning
Perspectives, 23(3), pp. 100-107 & 144, 1997.
NOTE 2.
Consideration 1.The number of abortions were estimated using
data gathered from 2,039 Philippine hospitals in 2003-2004, using the
methodology of a 1994 paper.
Consideration2.Data gathering.They gathered data for six years (1996-2001)
including figures for the top ten leading causes of hospital admissions, but
they only used the data for 1999-2001 "because data for these years are
the most complete and most current.”
Consideration3....All estimates presented
here are based on averaged data for these three years."
NOTE 3. See Table 1 for details of the calculation
using Singh et al methodology.
NOTE 4. Quote: For six hospitals where abortion ranked
below the 10th cause and for which data were available on both the 10th cause
and the number of abortion patients, abortion complications accounted for about
60% of the total number of patients hospitalized for the 10th ranking diagnosis”
(taken from a footnote on p. 101 of Singh S et al, Estimated the level of
abortions in the Philippines and Bangladesh, International Family Planning
Perspectives, 23(3), pp. 101-107 & 144, 1997).We find that the correlation used to support
their assumption was based on too small a sample--6 out of 2,039 hospitals.
NOTE 5. See Table 2 for details of the calculation
using modified technology.
NOTE 6. Based on the 1993 Philippine National Safe
Motherhood Survey, it was estimated that 1% of women who were pregnant had a
successful induced abortion (See chapter seven of the report). As a share of
live births, we estimate that 1.17% (i.e. 1/84.8 times 100%) of all live births
ended in induced abortions. When we apply this share to the 1,766,440 live
births in 2000 (taken from the National Statistics Office website at
http://www.census.gov.ph/data/sectordata/2000/lb0001.htm), this alternative
method estimates that there were 20,831 induced abortions in the year 2000.
FILIPINO FAMILY FUND: STRONG FAMILIES FOR A STRONG NATION