Contraceptive use projections: 1990 to 2010 James Trussell, PhD, and Barbara Vaughan Princeton, New Jersey Factors that will affect both contraceptive use and choice of method during the next 20 years are reviewed. Two factors are predictable: the changing age distribution of women and the revised upper-age limits for oral contraceptive use, with the effect of the latter as yet unknown. Less predictable factors include the number of women in each age group at risk for pregnancy, the effects of delayed childbearing and sterilization, and the impact of new contraceptive methods. Unpredictable factors include adverse publicity about oral contraceptives and breast cancer, concern about sexually transmitted diseases and acquired immunodeficiency syndrome, and changes in the availability of legal abortion. Numbers of women using oral contraception, other reversible methods, sterilization, and no method are projected from 1990 to 2010 under the assumption that use patterns in each age group resemble those observed in 1988. We conclude that discrepancies between projections of contraceptive use and fact are likely to occur because of the unpredictable nature of these few important variables. (AM J OBSTET GVNECOL 1992;167:1160-4.)
Key words: Contraceptive use projections, fertility, National Survey of Family Growth, abortion, sterilization, public perceptions Forecasts of future events or conditions are seldom correct. Demographic projections are no exceptionprojections of population size and distribution by age are notoriously prone to error because of the inability to forecast accurately the future course of mortality and, more importantly, fertility. I. 2 Projecting numbers of contraceptive users and patterns of contraceptive use during the next two decades involves even more uncertainty because the age distribution of women is only one of the many factors affecting use. Nevertheless, it is instructive to identify the factors likely to be relevant and to speculate about likely trends.
Factors affecting contraceptive use Factors affecting the numbers of women who will use various methods of contraception include sexual and reproductive behavior, the age distribution of women in their reproductive years, perceptions about contraception that are shaped by the media and clinicians, and legislation or regulations concerning abortion and contraception. Some factors are predictable or already known, others are less predictable, and still others are entirely unpredictable (Table I). Changing age structure. Changes in the distribution of ages among women in the reproductive age-span during the next 20 years can be predicted with great accuracy because these women are mostly already born and there is little death during the reproductive years. Recent U.S. Bureau of the Census projections, which take into account both death rates and predicted net
From the Office of Population Research, Princeton University. Reprint requests: James Trussell, PhD, Office ofPopulation Research, 21 Prospect Ave., Princeton, NJ 08544. 6/0/28063
immigration rates, show that in 1990 there were 65.9 million women, aged 15 to 49, and that this number will increase to 68.6 million in the year 2000 and then fall gradually thereafter. 3 Changes in age distribution will be more dramatic (Table II). Currently the largest cohort is the group of women aged 30 to 34. This cohort, representing the vanguard of the baby-boom generation, will move through and pass out of the reproductive age-span during the next 20 years. The cohort immediately younger than age 30 is also fairly large and in 20 years time will be near the end of the reproductive age-span. Younger cohorts are smaller in number, yet still large enough to ensure continuity in the total number of women of reproductive age during the next two decades. Thus a general aging of the female population will occur, and 20 years from now more women will be concentrated in the older cohorts, who are most likely to use sterilization rather than reversible contraceptive methods. Change in upper age limits for oral contraceptive use. One factor that may have an impact on patterns of contraceptive use is the upward revision of the upper age limits for oral contraceptive (OC) use. In October 1989 the Fertility and Maternal Health Drugs Advisory Committee of the Food and Drug Administration recommended the removal of all references to age limits for OC use by healthy, nonsmoking women. The potential impact of this change from the previous age limit of 35 years for smokers and 40 years for nonsmokers remains unknown. Conservative medical practice may possibly minimize the impact, as may the opinions of older women. It is possible that older women may choose OCs rather than permanent sterilization if they are aware of that alternative and of the health benefits ofOC use.
Contraceptive use projections
Volume 167 l>iumber 4. Part 2
Table I. Factors affecting contraceptive use: 1990-2010 Predictable Changing numbers of women aged 15-49 Revised upper age limits for OC use Less predictable Numbers of women at risk of unintended pregnancy Delayed childbearing, therefore delayed sterilization? Effect of revised upper age limits for OC use Impact of new methods CuT 380A Norplant Female condoms Unpredictable Publicity about contraceptive methods (especially OCs and breast cancer) Concern about AIDS and STDs Availability of legal abortion
1161
Table II. Number of women aged 15 to 49: 1990-2010 No. (millions)
Age (yr)
1990
15-19 20-24 25-29 30-34 35-39 40-44 45-49 15-49
8.5 9.3 10.7 11.2 10.1 9.0 7.1 65.9
2010
8.6 8.7 9.4 10.9 11.1 10.1 8.9 67.8
9.3 8.7 8.8 9.7 10.9 1l.l 10.1 68.6
9.5 9.5 8.9 9.7 9.6 10.9 11.0 68.4
9.2 9.6 9.6 9.1 9.0 9.6 10.8 66.9
U.S. Bureau of the Census 1989; series P-26, no. 1018, Table 4. 3
STDs, Sexually transmitted diseases.
Sexual activity. Among all reproductive-age women, those at risk of an unintended pregnancy constitute a select demographic group-women who are fecund, sexually active, not pregnant or postpartum, and not seeking to become pregnant. Many factors affect the size and age composition of this risk pool, making projections of contraceptive use less certain. An important variable is the number of teenagers who are sexually active. The 1988 National Survey of Family Growth revealed that 51 % of girls aged 15 to 19 years had engaged in intercourse, an increase of 8 percentage points from the 43% recorded in the 1982 National Survey of Family Growth.' Apparently, media attention to the risk of acquired immunodeficiency syndrome (AIDS), some focused specifically on the risk to teenagers, has at best had minimal impact on the sexual activity of teenagers. Continuation of this trend of increased sexual activity among teens may be expected to increase OC use because young, sexually active women use this method with greatest frequency. Among unmarried women over 20 years old, the impact of the fear of sexually transmitted diseases or AIDS remains speculative. We simply do not know whether these fears are decreasing sexual activity in this group. There has not been a national survey of sexual activity in the United States since the work of Kinsey in the 1950s.' Given the political pressures, it is doubtful when, if ever, another national survey of sexual practices will be conducted. 67 The 1988 National Survey of Family Growth revealed the percentage of women at risk of pregnancy in all post-teenage groups to be fairly constant through age 44, varying between 67% and 75%. There was considerably greater variability between age groups in the percentage of women at risk who used specific contraceptive methods (Table III). OC use was highest in the 20- to 24-year age group (59.8%), decreased to 39.7%
at ages 25 to 29, and then fell rapidly thereafter. Strikingly, 19.8% of women at risk at ages 15 to 19 used no contraceptive method. As would be expected, virtually no teenagers had undergone contraceptive sterilization, an option first chosen at 20 to 24 years old and increasingly exercised in subsequent age groups as OC use diminished. Delayed childbearing and sterilization. The current trend toward delayed childbearing may also affect the number of women at risk for unintended pregnancy. Fertility rates are increasing among women in older age groups. Between 1980 and 1988, births per 1000 women increased from 19.8 to 27.9 among women aged 35 to 39 and from 3.9 to 4.8 among women aged 40 to 44. 8 This increase represents a significant reversal of a dramatic trend toward lower birth rates among women between 35 to 45 years old that was observed during previous decades. Consequently, fewer women in this age group are exposed to risk because they are either pregnant, trying to become pregnant, or are postpartum. The incidence of breast-feeding, with its negative effect on fecundity, also appears to be increasing in all age groups. The statistics for childless women tell a similar story (Table IV). In 1975 only 8% of women aged 30 to 34 years were childless; by 1988, the figure had jumped to 22%. However, of the women in this age group who were childless in 1975, only one third expected to have a child thereafter, whereas in 1988, 51 % of childless women aged 30 to 34 still expected to have a child. g )2 As a consequence of delayed childbearing, the very large group of women over age 30 may be delaying sterilization. Sterilization traditionally has been regarded as the preferred contraceptive option for women who face 20 or more years of exposure to the risk of unwanted fertility after their last wanted birth. Women who have perhaps only 8 more years of exposure to risk after their final wanted pregnancy may elect to forego sterilization altogether, especially if OC
1162 Trussell and Vaughan
October 1992 Am J Obstet Gynecol
Table III. Contraceptive method use by age group Users (% of women at risk)
Age (yr)*
15-19 20-24 25-29 30-34 35-39 40-44 45-49
Women at riskt (%)
39.4 66.5 72.1 73.6 75.0 71.0 65.0
OCs
Other reversible methods
No method
Sterilization
47.7 59.8 39.7 20.4 4.6 3.0 1.0
32.1 22.5 28.9 29.8 28.8 21.4 13.0
19.8 1l.8 10.4 7.4 6.3 7.6 9.0
0.4 5.9 21.0 42.4 60.3 68.0 77.0
*Ages 15 to 44: 1988 National Survey of Family Growth; ages 45 to 49: authors' estimate. tHaving sex and not pregnant, seeking pregnancy, or postpartum and not infecund.
Table IV. Childless women in the United States Planning to have children (% among those childless)
Childless women (%) Yr
18-24 yr
1975 1980 1985 1988
45 66 66 70
I
I
25-29 yr
20 34 38 39
30-34 yr
Yr
18-24 yr
8 17 23 22
1975 1980 1985 1988
91 83 85 86
I
25-29 yr
75 68 77 76
I
30-34 yr
33 37 45 51
U.S. Bureau of the Census, series P-20, nos. 301, 375, 406, and 436. 9.'2
Table V. Favorable opinions of birth control methods* Method Pill
Condom Vasectomy Female sterilization Diaphragm Sponge Foam Vaginal suppository Intrauterine device Rhythm Cream Jelly Withdrawal Menses extraction
Favorable opinion (%)
76 62 57 54 30 22 18 18 16 16 14 14 10 3
*1989 Ortho Survey of 12,500 women (7861 replies).
use, the reversible contraceptive most effective in preventing pregnancy, is perceived as a real alternative. In summary, both delayed childbearing and delayed sterilization may be expected to affect the mix of contraceptive choices among older women, enlarging the scope for reversible methods. Fewer women in older age groups may elect sterilization and more may choose OC use. New contraceptive methods. The availability of new methods of contraception is another factor affecting
patterns of contraceptive use that is difficult to predict. This factor would include both the impact of products that we know will become available'3-the new intrauterine device, CuT 380A; Norplant, an alternative form of hormonal contraception; and female condoms-and the impact of other products, some perhaps not yet imagined, that will become available by the year 2010. Public perceptions. Virtually unpredictable but vitally important is the impact of public perceptions as shaped by the media. Foremost among the perceptions likely to influence patterns of contraceptive use are those engendered by adverse publicity about OCs and cancer and by concern about sexually transmitted diseases and AIDS. Regardless of whether a link between OC use and breast cancer is proved, OC use will almost certainly decline if the public perceives a link. Concerns about sexually transmitted diseases and AIDS may promote greater use of condoms and other barrier methods among young women, who have been the group most likely to use OCs. The interesting question is whether such women will substitute barrier methods for OCs (thereby raising their risk of pregnancy but reducing their risk of sexually transmitted diseases) or will supplement OC use with barrier methods (thereby reducing risks of both pregnancy and sexually transmitted diseases). What perceptions currently exist among women? An-
Contraceptive use projections
Volume 167 Number 4, Part 2
1163
Table VI. Projected number of contraceptive users: Effect of changes in age distribution* Women (millions) Yr
oes
Other reversible methods
No method
Sterilization
1990 1995 2000 2005 2010
10.6 10.1 9.9 10.1 10.2
11.3 11.4 11.3 11.1 10.9
4.2 4.2 4.3 4.3 4.3
18.0 19.5 20.1 19.7 18.8
*Assumes 1988 age-specific proportions at risk and method mix (Table III).
nual contraceptive-user surveys by the Ortho Pharmaceutical Corp. (data on file) provide some insights. The 1989 Ortho survey of 15,000 women, of whom 7861 (52%) replied, showed that OCs (at 76%) have the highest approval rating of all contraceptive methods (Table V). This statistic compares with a favorable rating of only 65% in a 1982 Ortho survey. 14 When analyzed by age group, the only approval rating that showed a substantial age-related trend was that for sterilization, with older women favorably disposed to this method of contraception and younger women regarding it as highly unfavorable. There was a slight age-related trend shown in attitudes toward OC use, with the peak approval at ages 20 to 24. However, in both cases, age-related trends may reflect misunderstanding by the women surveyed. The survey was intended to elicit attitudes about contraceptive methods as used by women in general, but many respondents apparently interpreted the question as asking how they felt about each method for their personal use. Consequently, the peak of OC approval occurred in the age group most likely to use OCs. A negative attitude toward sterilization was strongest in younger women, almost certainly because these women anticipated future pregnanCies. Opinion was related to experience. Among current OC users, 97% approved of this method, although a cause-and-effect relationship in the approval rating is difficult to establish because only women who approve of OCs are likely to continue using them. Similarly, a low approval rating of 44% was observed among women who had never used OCs, which probably reflects the prevailing attitude among those who are unlikely to adopt this method. These responses seemed to reflect personal attitudes; thus a woman who never used an OC because of problems with embolisms would have rated OCs unfavorably because they were not a suitable choice for her. Certainly, women have strong opinions about OC use; few respondents indicated no opinion or failed to reply to this question. Annual OC discontinuation rates among users reported in Ortho surveys conducted from 1987 through 1989 averaged 23%, compared with 29% for all other
methods. Overwhelmingly, the reason cited for discontinuation was side effects, either actually experienced or feared. Interestingly, only 5% of those who discontinued OC use attributed their discontinuation to adverse publicity. However, it is difficult to distinguish between "fear of side effects" and "adverse publicity," because the latter may have caused the former. Moreover, adverse publicity would more likely be cited as a reason for discontinuation in a survey conducted immediately after any unusually negative media attention. Among other findings of the 1989 Ortho survey, 70% of respondents strongly agreed that health risks are associated with taking OCs for long periods of time, 15% strongly disagreed that a woman taking the pill without problems and seeing her physician regularly has no reason not to stay on the pill for a long time, and 22% strongly disagreed that the pill may protect against developing certain types of cancer. Opinions about whether women who use the pill are at increased risk of cancer were essentially split; most respondents indicated no opinion. Also in this survey, 21 % strongly agreed that women who take the pill have difficulty becoming pregnant once they stop. The question involves ambiguous wording. Studies indicate that there is reduced fecundity when women stop using hormonal contraceptives, but the reduction is only temporary, and by 18 months cumulative rates of conception are the same as in women who have not used this method. I5 To which effect, short-term or longer-term, were interviewees responding? It is probable that respondents did not interpret this question as referring to a time-limited effect, but rather that many indeed believed OC use would permanently affect a woman's ability to have a child. Although the extent of misperceptions about OC safety by representative women may be surprising, the misperceptions observed in a survey of sex educators are even more striking. Fifty-two percent of public school teachers who offer sex education agreed that pill use should be stopped periodically "to give the body a rest," whereas a further 25% were unsure; only 34% knew that pill use by teenagers does not have a bad
1164 Trussell and Vaughan
effect on later fertility (and 25% were sure it does), and only 24% knew that cigarette smoking is not a contraindication to teenage OC use. '6 Availability of legal abortion. Perhaps the least predictable factor that will affect contraceptive use during the next decades is the continued availability of legal abortion. In many parts of the United States, abortion, while still legal, is increasingly unavailable, with fewer physicians willing to perform the procedure. '7 Women who use barrier methods or other less effective methods of contraception often regard abortion as a fallback option in the event of contraceptive failure. Factors making legal abortion more difficult or more expensive to obtain may cause contraceptive efficacy to weigh more heavily in a woman's choice of method. Projections to the year 2010
Using the latest available data base, the 1988 National Survey of Family Growth, we attempt to project future patterns of contraceptive use. The distributions of women by age in the years 1990, 1995,2000,2005, and 2010 (Table II) are taken from Census Bureau projections. If we assume that the proportions of women at risk of unintended pregnancy in each age group in 1988 and the fractions of women at risk in each age group who chose each contraceptive method in 1988 remain constant over time, the following forecasts can be made. Levels of OC use will decline moderately through the end of the century and then will recover slightly in the twenty-first century. Levels of use of other reversible methods will change little in this century and decline slightly thereafter. There will be virtually no change in the number of women who use no method. Use of sterilization will rise until the year 2000 and decline afterward, mirroring the passage of the large birth cohorts through the latter part of the reproductive span (Table VI). However, projections based on the assumption that the mix of contraceptive choices made by women will be the same in 2010 as it was in 1988 will almost certainly be wrong. Indeed, projections of contraceptive use in 1990 (not shown) based on results of the 1982 National Survey of Family Growth, when compared with projections based on the 1988 National Survey of Family Growth, underestimate OC use by 2.3 million and sterilization use by 3.2 million and overestimate the use of other reversible methods by 2.2 million. Furthermore, the estimate of the number of OC users based on the 1991 Ortho Pharmaceutical Corp. survey (16.8 million)* is 58% greater than our projection for
October t 992 Am J Obstet Gynecol
1990 based on the data from the 1988 National Survey of Family Growth (l0.6 million). These discrepancies are not surprising, given the many factors that affect contraceptive choice and the difficulty in predicting the impact of changes in these factors on the choices women make regarding contraception. *Data from The Ortho Annual Birth Control Study, Ortho Pharmaceutical Corp., Raritan, N.J. (1969-1991). Request for Ortho 1991 Annual Birth Control Study reprints: Public Affairs Department, Ortho Pharmaceutical Corp., Route 202, P. O. Box 300, Raritan, N] 08869.
REFERENCES 1. Keyfitz N. The limits of population forecasting. Pop Dev Rev 1981;7:579-94. 2. Stoto MA. The accuracy of population projections. J Am Stat Assoc 1983;78:13-20. 3. Spencer G. Projections of the population of the United States by age, sex and race: 1988-2080. Current population reports 1989; series P-25, no 1018. Washington: Government Printing Office. 4. Trussell], Vaughan B. Unpublished tabulations from the public-use samples of 1982 and 1988 National Survey of Family Growth. 5. Kinsey AC, Pomeroy WB, Martin CE, Gebhard PH. Sexual behavior in the human female. Philadelphia: WB Saunders, 1953. 6. Marshall E. Stymied sex survey. Science 1991 ;252:497. 7. Marshall E. Sullivan overrules NIH on sex survey. Science 1991;253:502. 8. National Center for Health Statistics. Advance report of final natality statistics, 1988. Monthly Vital Stat 1990;39(4). 9. U.S. Bureau of the Census. Fertility of American women: June 1975. Current Population Reports 1976; series P20, no 301. 10. U.S. Bureau of the Census. Fertility of American women: June 1980. Current Population Reports, 1982; series P20, no 375. 11. U.S. Bureau of the Census. Fertility of American women: June 1985. Current Population Reports 1986; series P20, no 406. 12. U.S. Bureau of the Census. Fertility of American women: June 1988. Current Population Reports 1989; series P20, no 436. 13. Hatcher RA, Stewart F, Trussell], et al. Contraceptive technology 1990-1992. New York: Irvington Publishers, 1990. 14. Forrest JD, Fordyce RR. U.S. women's contraceptive attitudes and practice: how have they changed in the 1980s? Fam Plann Perspec 1988;20: 112-8. 15. Trussell J, Kost K. Contraceptive failure in the United States: a critical review of the literature. Stud Fam Plann 1987;18:237-83. 16. Forrest JD, Silverman]. What public school teachers teach about preventing pregnancy, AIDS, and sexually transmitted diseases. Fam Plann Perspec 1989;21:65-72. 17. Henshaw SK, Van Vort]. Abortion services in the United States, 1987 and 1988. Fam Plann Perspec 1990;22: 1028, 142.
Key words: Contraceptive use projections, fertility, National Survey of Family Growth, abortion, sterilization, public perceptions Forecasts of future events or conditions are seldom correct. Demographic projections are no exceptionprojections of population size and distribution by age are notoriously prone to error because of the inability to forecast accurately the future course of mortality and, more importantly, fertility. I. 2 Projecting numbers of contraceptive users and patterns of contraceptive use during the next two decades involves even more uncertainty because the age distribution of women is only one of the many factors affecting use. Nevertheless, it is instructive to identify the factors likely to be relevant and to speculate about likely trends.
Factors affecting contraceptive use Factors affecting the numbers of women who will use various methods of contraception include sexual and reproductive behavior, the age distribution of women in their reproductive years, perceptions about contraception that are shaped by the media and clinicians, and legislation or regulations concerning abortion and contraception. Some factors are predictable or already known, others are less predictable, and still others are entirely unpredictable (Table I). Changing age structure. Changes in the distribution of ages among women in the reproductive age-span during the next 20 years can be predicted with great accuracy because these women are mostly already born and there is little death during the reproductive years. Recent U.S. Bureau of the Census projections, which take into account both death rates and predicted net
From the Office of Population Research, Princeton University. Reprint requests: James Trussell, PhD, Office ofPopulation Research, 21 Prospect Ave., Princeton, NJ 08544. 6/0/28063
immigration rates, show that in 1990 there were 65.9 million women, aged 15 to 49, and that this number will increase to 68.6 million in the year 2000 and then fall gradually thereafter. 3 Changes in age distribution will be more dramatic (Table II). Currently the largest cohort is the group of women aged 30 to 34. This cohort, representing the vanguard of the baby-boom generation, will move through and pass out of the reproductive age-span during the next 20 years. The cohort immediately younger than age 30 is also fairly large and in 20 years time will be near the end of the reproductive age-span. Younger cohorts are smaller in number, yet still large enough to ensure continuity in the total number of women of reproductive age during the next two decades. Thus a general aging of the female population will occur, and 20 years from now more women will be concentrated in the older cohorts, who are most likely to use sterilization rather than reversible contraceptive methods. Change in upper age limits for oral contraceptive use. One factor that may have an impact on patterns of contraceptive use is the upward revision of the upper age limits for oral contraceptive (OC) use. In October 1989 the Fertility and Maternal Health Drugs Advisory Committee of the Food and Drug Administration recommended the removal of all references to age limits for OC use by healthy, nonsmoking women. The potential impact of this change from the previous age limit of 35 years for smokers and 40 years for nonsmokers remains unknown. Conservative medical practice may possibly minimize the impact, as may the opinions of older women. It is possible that older women may choose OCs rather than permanent sterilization if they are aware of that alternative and of the health benefits ofOC use.
Contraceptive use projections
Volume 167 l>iumber 4. Part 2
Table I. Factors affecting contraceptive use: 1990-2010 Predictable Changing numbers of women aged 15-49 Revised upper age limits for OC use Less predictable Numbers of women at risk of unintended pregnancy Delayed childbearing, therefore delayed sterilization? Effect of revised upper age limits for OC use Impact of new methods CuT 380A Norplant Female condoms Unpredictable Publicity about contraceptive methods (especially OCs and breast cancer) Concern about AIDS and STDs Availability of legal abortion
1161
Table II. Number of women aged 15 to 49: 1990-2010 No. (millions)
Age (yr)
1990
15-19 20-24 25-29 30-34 35-39 40-44 45-49 15-49
8.5 9.3 10.7 11.2 10.1 9.0 7.1 65.9
2010
8.6 8.7 9.4 10.9 11.1 10.1 8.9 67.8
9.3 8.7 8.8 9.7 10.9 1l.l 10.1 68.6
9.5 9.5 8.9 9.7 9.6 10.9 11.0 68.4
9.2 9.6 9.6 9.1 9.0 9.6 10.8 66.9
U.S. Bureau of the Census 1989; series P-26, no. 1018, Table 4. 3
STDs, Sexually transmitted diseases.
Sexual activity. Among all reproductive-age women, those at risk of an unintended pregnancy constitute a select demographic group-women who are fecund, sexually active, not pregnant or postpartum, and not seeking to become pregnant. Many factors affect the size and age composition of this risk pool, making projections of contraceptive use less certain. An important variable is the number of teenagers who are sexually active. The 1988 National Survey of Family Growth revealed that 51 % of girls aged 15 to 19 years had engaged in intercourse, an increase of 8 percentage points from the 43% recorded in the 1982 National Survey of Family Growth.' Apparently, media attention to the risk of acquired immunodeficiency syndrome (AIDS), some focused specifically on the risk to teenagers, has at best had minimal impact on the sexual activity of teenagers. Continuation of this trend of increased sexual activity among teens may be expected to increase OC use because young, sexually active women use this method with greatest frequency. Among unmarried women over 20 years old, the impact of the fear of sexually transmitted diseases or AIDS remains speculative. We simply do not know whether these fears are decreasing sexual activity in this group. There has not been a national survey of sexual activity in the United States since the work of Kinsey in the 1950s.' Given the political pressures, it is doubtful when, if ever, another national survey of sexual practices will be conducted. 67 The 1988 National Survey of Family Growth revealed the percentage of women at risk of pregnancy in all post-teenage groups to be fairly constant through age 44, varying between 67% and 75%. There was considerably greater variability between age groups in the percentage of women at risk who used specific contraceptive methods (Table III). OC use was highest in the 20- to 24-year age group (59.8%), decreased to 39.7%
at ages 25 to 29, and then fell rapidly thereafter. Strikingly, 19.8% of women at risk at ages 15 to 19 used no contraceptive method. As would be expected, virtually no teenagers had undergone contraceptive sterilization, an option first chosen at 20 to 24 years old and increasingly exercised in subsequent age groups as OC use diminished. Delayed childbearing and sterilization. The current trend toward delayed childbearing may also affect the number of women at risk for unintended pregnancy. Fertility rates are increasing among women in older age groups. Between 1980 and 1988, births per 1000 women increased from 19.8 to 27.9 among women aged 35 to 39 and from 3.9 to 4.8 among women aged 40 to 44. 8 This increase represents a significant reversal of a dramatic trend toward lower birth rates among women between 35 to 45 years old that was observed during previous decades. Consequently, fewer women in this age group are exposed to risk because they are either pregnant, trying to become pregnant, or are postpartum. The incidence of breast-feeding, with its negative effect on fecundity, also appears to be increasing in all age groups. The statistics for childless women tell a similar story (Table IV). In 1975 only 8% of women aged 30 to 34 years were childless; by 1988, the figure had jumped to 22%. However, of the women in this age group who were childless in 1975, only one third expected to have a child thereafter, whereas in 1988, 51 % of childless women aged 30 to 34 still expected to have a child. g )2 As a consequence of delayed childbearing, the very large group of women over age 30 may be delaying sterilization. Sterilization traditionally has been regarded as the preferred contraceptive option for women who face 20 or more years of exposure to the risk of unwanted fertility after their last wanted birth. Women who have perhaps only 8 more years of exposure to risk after their final wanted pregnancy may elect to forego sterilization altogether, especially if OC
1162 Trussell and Vaughan
October 1992 Am J Obstet Gynecol
Table III. Contraceptive method use by age group Users (% of women at risk)
Age (yr)*
15-19 20-24 25-29 30-34 35-39 40-44 45-49
Women at riskt (%)
39.4 66.5 72.1 73.6 75.0 71.0 65.0
OCs
Other reversible methods
No method
Sterilization
47.7 59.8 39.7 20.4 4.6 3.0 1.0
32.1 22.5 28.9 29.8 28.8 21.4 13.0
19.8 1l.8 10.4 7.4 6.3 7.6 9.0
0.4 5.9 21.0 42.4 60.3 68.0 77.0
*Ages 15 to 44: 1988 National Survey of Family Growth; ages 45 to 49: authors' estimate. tHaving sex and not pregnant, seeking pregnancy, or postpartum and not infecund.
Table IV. Childless women in the United States Planning to have children (% among those childless)
Childless women (%) Yr
18-24 yr
1975 1980 1985 1988
45 66 66 70
I
I
25-29 yr
20 34 38 39
30-34 yr
Yr
18-24 yr
8 17 23 22
1975 1980 1985 1988
91 83 85 86
I
25-29 yr
75 68 77 76
I
30-34 yr
33 37 45 51
U.S. Bureau of the Census, series P-20, nos. 301, 375, 406, and 436. 9.'2
Table V. Favorable opinions of birth control methods* Method Pill
Condom Vasectomy Female sterilization Diaphragm Sponge Foam Vaginal suppository Intrauterine device Rhythm Cream Jelly Withdrawal Menses extraction
Favorable opinion (%)
76 62 57 54 30 22 18 18 16 16 14 14 10 3
*1989 Ortho Survey of 12,500 women (7861 replies).
use, the reversible contraceptive most effective in preventing pregnancy, is perceived as a real alternative. In summary, both delayed childbearing and delayed sterilization may be expected to affect the mix of contraceptive choices among older women, enlarging the scope for reversible methods. Fewer women in older age groups may elect sterilization and more may choose OC use. New contraceptive methods. The availability of new methods of contraception is another factor affecting
patterns of contraceptive use that is difficult to predict. This factor would include both the impact of products that we know will become available'3-the new intrauterine device, CuT 380A; Norplant, an alternative form of hormonal contraception; and female condoms-and the impact of other products, some perhaps not yet imagined, that will become available by the year 2010. Public perceptions. Virtually unpredictable but vitally important is the impact of public perceptions as shaped by the media. Foremost among the perceptions likely to influence patterns of contraceptive use are those engendered by adverse publicity about OCs and cancer and by concern about sexually transmitted diseases and AIDS. Regardless of whether a link between OC use and breast cancer is proved, OC use will almost certainly decline if the public perceives a link. Concerns about sexually transmitted diseases and AIDS may promote greater use of condoms and other barrier methods among young women, who have been the group most likely to use OCs. The interesting question is whether such women will substitute barrier methods for OCs (thereby raising their risk of pregnancy but reducing their risk of sexually transmitted diseases) or will supplement OC use with barrier methods (thereby reducing risks of both pregnancy and sexually transmitted diseases). What perceptions currently exist among women? An-
Contraceptive use projections
Volume 167 Number 4, Part 2
1163
Table VI. Projected number of contraceptive users: Effect of changes in age distribution* Women (millions) Yr
oes
Other reversible methods
No method
Sterilization
1990 1995 2000 2005 2010
10.6 10.1 9.9 10.1 10.2
11.3 11.4 11.3 11.1 10.9
4.2 4.2 4.3 4.3 4.3
18.0 19.5 20.1 19.7 18.8
*Assumes 1988 age-specific proportions at risk and method mix (Table III).
nual contraceptive-user surveys by the Ortho Pharmaceutical Corp. (data on file) provide some insights. The 1989 Ortho survey of 15,000 women, of whom 7861 (52%) replied, showed that OCs (at 76%) have the highest approval rating of all contraceptive methods (Table V). This statistic compares with a favorable rating of only 65% in a 1982 Ortho survey. 14 When analyzed by age group, the only approval rating that showed a substantial age-related trend was that for sterilization, with older women favorably disposed to this method of contraception and younger women regarding it as highly unfavorable. There was a slight age-related trend shown in attitudes toward OC use, with the peak approval at ages 20 to 24. However, in both cases, age-related trends may reflect misunderstanding by the women surveyed. The survey was intended to elicit attitudes about contraceptive methods as used by women in general, but many respondents apparently interpreted the question as asking how they felt about each method for their personal use. Consequently, the peak of OC approval occurred in the age group most likely to use OCs. A negative attitude toward sterilization was strongest in younger women, almost certainly because these women anticipated future pregnanCies. Opinion was related to experience. Among current OC users, 97% approved of this method, although a cause-and-effect relationship in the approval rating is difficult to establish because only women who approve of OCs are likely to continue using them. Similarly, a low approval rating of 44% was observed among women who had never used OCs, which probably reflects the prevailing attitude among those who are unlikely to adopt this method. These responses seemed to reflect personal attitudes; thus a woman who never used an OC because of problems with embolisms would have rated OCs unfavorably because they were not a suitable choice for her. Certainly, women have strong opinions about OC use; few respondents indicated no opinion or failed to reply to this question. Annual OC discontinuation rates among users reported in Ortho surveys conducted from 1987 through 1989 averaged 23%, compared with 29% for all other
methods. Overwhelmingly, the reason cited for discontinuation was side effects, either actually experienced or feared. Interestingly, only 5% of those who discontinued OC use attributed their discontinuation to adverse publicity. However, it is difficult to distinguish between "fear of side effects" and "adverse publicity," because the latter may have caused the former. Moreover, adverse publicity would more likely be cited as a reason for discontinuation in a survey conducted immediately after any unusually negative media attention. Among other findings of the 1989 Ortho survey, 70% of respondents strongly agreed that health risks are associated with taking OCs for long periods of time, 15% strongly disagreed that a woman taking the pill without problems and seeing her physician regularly has no reason not to stay on the pill for a long time, and 22% strongly disagreed that the pill may protect against developing certain types of cancer. Opinions about whether women who use the pill are at increased risk of cancer were essentially split; most respondents indicated no opinion. Also in this survey, 21 % strongly agreed that women who take the pill have difficulty becoming pregnant once they stop. The question involves ambiguous wording. Studies indicate that there is reduced fecundity when women stop using hormonal contraceptives, but the reduction is only temporary, and by 18 months cumulative rates of conception are the same as in women who have not used this method. I5 To which effect, short-term or longer-term, were interviewees responding? It is probable that respondents did not interpret this question as referring to a time-limited effect, but rather that many indeed believed OC use would permanently affect a woman's ability to have a child. Although the extent of misperceptions about OC safety by representative women may be surprising, the misperceptions observed in a survey of sex educators are even more striking. Fifty-two percent of public school teachers who offer sex education agreed that pill use should be stopped periodically "to give the body a rest," whereas a further 25% were unsure; only 34% knew that pill use by teenagers does not have a bad
1164 Trussell and Vaughan
effect on later fertility (and 25% were sure it does), and only 24% knew that cigarette smoking is not a contraindication to teenage OC use. '6 Availability of legal abortion. Perhaps the least predictable factor that will affect contraceptive use during the next decades is the continued availability of legal abortion. In many parts of the United States, abortion, while still legal, is increasingly unavailable, with fewer physicians willing to perform the procedure. '7 Women who use barrier methods or other less effective methods of contraception often regard abortion as a fallback option in the event of contraceptive failure. Factors making legal abortion more difficult or more expensive to obtain may cause contraceptive efficacy to weigh more heavily in a woman's choice of method. Projections to the year 2010
Using the latest available data base, the 1988 National Survey of Family Growth, we attempt to project future patterns of contraceptive use. The distributions of women by age in the years 1990, 1995,2000,2005, and 2010 (Table II) are taken from Census Bureau projections. If we assume that the proportions of women at risk of unintended pregnancy in each age group in 1988 and the fractions of women at risk in each age group who chose each contraceptive method in 1988 remain constant over time, the following forecasts can be made. Levels of OC use will decline moderately through the end of the century and then will recover slightly in the twenty-first century. Levels of use of other reversible methods will change little in this century and decline slightly thereafter. There will be virtually no change in the number of women who use no method. Use of sterilization will rise until the year 2000 and decline afterward, mirroring the passage of the large birth cohorts through the latter part of the reproductive span (Table VI). However, projections based on the assumption that the mix of contraceptive choices made by women will be the same in 2010 as it was in 1988 will almost certainly be wrong. Indeed, projections of contraceptive use in 1990 (not shown) based on results of the 1982 National Survey of Family Growth, when compared with projections based on the 1988 National Survey of Family Growth, underestimate OC use by 2.3 million and sterilization use by 3.2 million and overestimate the use of other reversible methods by 2.2 million. Furthermore, the estimate of the number of OC users based on the 1991 Ortho Pharmaceutical Corp. survey (16.8 million)* is 58% greater than our projection for
October t 992 Am J Obstet Gynecol
1990 based on the data from the 1988 National Survey of Family Growth (l0.6 million). These discrepancies are not surprising, given the many factors that affect contraceptive choice and the difficulty in predicting the impact of changes in these factors on the choices women make regarding contraception. *Data from The Ortho Annual Birth Control Study, Ortho Pharmaceutical Corp., Raritan, N.J. (1969-1991). Request for Ortho 1991 Annual Birth Control Study reprints: Public Affairs Department, Ortho Pharmaceutical Corp., Route 202, P. O. Box 300, Raritan, N] 08869.
REFERENCES 1. Keyfitz N. The limits of population forecasting. Pop Dev Rev 1981;7:579-94. 2. Stoto MA. The accuracy of population projections. J Am Stat Assoc 1983;78:13-20. 3. Spencer G. Projections of the population of the United States by age, sex and race: 1988-2080. Current population reports 1989; series P-25, no 1018. Washington: Government Printing Office. 4. Trussell], Vaughan B. Unpublished tabulations from the public-use samples of 1982 and 1988 National Survey of Family Growth. 5. Kinsey AC, Pomeroy WB, Martin CE, Gebhard PH. Sexual behavior in the human female. Philadelphia: WB Saunders, 1953. 6. Marshall E. Stymied sex survey. Science 1991 ;252:497. 7. Marshall E. Sullivan overrules NIH on sex survey. Science 1991;253:502. 8. National Center for Health Statistics. Advance report of final natality statistics, 1988. Monthly Vital Stat 1990;39(4). 9. U.S. Bureau of the Census. Fertility of American women: June 1975. Current Population Reports 1976; series P20, no 301. 10. U.S. Bureau of the Census. Fertility of American women: June 1980. Current Population Reports, 1982; series P20, no 375. 11. U.S. Bureau of the Census. Fertility of American women: June 1985. Current Population Reports 1986; series P20, no 406. 12. U.S. Bureau of the Census. Fertility of American women: June 1988. Current Population Reports 1989; series P20, no 436. 13. Hatcher RA, Stewart F, Trussell], et al. Contraceptive technology 1990-1992. New York: Irvington Publishers, 1990. 14. Forrest JD, Fordyce RR. U.S. women's contraceptive attitudes and practice: how have they changed in the 1980s? Fam Plann Perspec 1988;20: 112-8. 15. Trussell J, Kost K. Contraceptive failure in the United States: a critical review of the literature. Stud Fam Plann 1987;18:237-83. 16. Forrest JD, Silverman]. What public school teachers teach about preventing pregnancy, AIDS, and sexually transmitted diseases. Fam Plann Perspec 1989;21:65-72. 17. Henshaw SK, Van Vort]. Abortion services in the United States, 1987 and 1988. Fam Plann Perspec 1990;22: 1028, 142.
