Background
Little is known about unintended pregnancy risk and current contraceptive use among women ≥45 years old in the United States.
Objectives
The purpose of this study was to describe the prevalence of women ages 45-50 years old at risk for unintended pregnancy and their current contraceptive use, and to compare these findings to those of women in younger age groups.
Study Design
We analyzed 2006, 2008, and 2010 Massachusetts Behavioral Risk Factor Surveillance System data, the only state in the United States to collect contraceptive data routinely from women >44 years old. Women 18-50 years old (n = 4930) were considered to be at risk for unintended pregnancy unless they reported current pregnancy, hysterectomy, not being sexually active in the past year, having a same-sex partner, or wanting to become pregnant. Among women who were considered to be at risk (n = 3605), we estimated the prevalence of current contraceptive use by age group. Among women who were considered to be at risk and who were 45-50 years old (n = 940), we examined characteristics that were associated with current method use. Analyses were conducted on weighted data using SAS-callable SUDAAN (RTI International, Research Triangle Park, NC).
Results
Among women who were 45-50 years old, 77.6% were at risk for unintended pregnancy, which was similar to other age groups. As age increased, hormonal contraceptive use (shots, pills, patch, or ring) decreased, and permanent contraception (tubal ligation or vasectomy) increased as did non-use of contraception. Of women who were 45-50 years old and at risk for unintended pregnancy, 66.9% reported using some contraceptive method; permanent contraception was the leading method reported by 44.0% and contraceptive non-use was reported by 16.8%.
Conclusion
A substantial proportion of women who were 45-50 years old were considered to be at risk for unintended pregnancy. Permanent contraception was most commonly used by women in this age group. Compared with other age groups, more women who were 45-50 years old were not using any contraception. Population-based surveillance efforts are needed to follow trends among this age group and better meet their family planning needs. Although expanding surveillance systems to include women through 50 years old requires additional resources, fertility trends that show increasingly delayed childbearing, uncertain end of fecundity, and potential adverse consequences of unplanned pregnancy in older age may justify these expenditures.
Little is known about unintended pregnancy (UIP) risk and current contraceptive use among women who are ≥45 years old in the United States. The primary US national surveillance tool that gathers information on family life, pregnancy, and use of contraception—the National Survey of Family Growth (NSFG)—has collected family planning and contraceptive use data only from women who were 15-44 years old since 1973 ; however, starting September 2015, NSFG expanded their age range to 15-49 years old (Anjani Chandra, PhD, personal communication, July 2015). Nonetheless, our ability to understand the fertility desires and contraceptive needs of older women in the United States is limited. Several countries do collect this information from older women, and European data suggest that approximately 30% of women 45-49 years old are not using any contraception.
Fecundity in women significantly declines after 44 years old; the median age at which women in the United States reach natural menopause is 51.4 years old. Nonetheless, conceptions in the later reproductive years do occur. In fact, live births among women in the United States who are 45-49 years old are increasing. In 2013, the US birth rate for women who were ≥45 years old was 0.8 births per 1000 women, which is a small increase from 0.7 births per 1000 women in 2012 and an even larger increase since the early 1990s when the birth rate for women who were ≥45 years old was 0.3 births per 1000 women. Presumably, much of this increase is due to planned births and the increasing use of assisted reproductive technologies; however, to our knowledge, no estimates of the UIP rate among women in the United States who are ≥45 years old have been reported. Among women who are 15-44 years old, proportions of UIP are highest among teenagers and women who are 20-24 years old (82% and 64%, respectively), although the third highest proportion is among women who are 40-44 years old (48%).
According to current US contraception guidelines, contraceptive protection is recommended for women who are ≥45 years old who are at risk for UIP. The American College of Obstetricians and Gynecologists further specifies that women who want to avoid pregnancy should continue contraception until 50-55 years old. All methods of contraception are considered safe or generally safe for women who are ≥45 years old without other risk factors and should not be dismissed from consideration or discontinued based on age alone. Although certain medical conditions that are more common as women age (such as hypertension or diabetes mellitus) may preclude the use of some reversible contraceptive methods (eg, those containing estrogen), many methods such as progestin-only implants or intrauterine devices (IUDs) remain safe, even for women with underlying medical disorders and are among the most effective methods available.
Given limited information on UIP risk and contraceptive use behaviors among women in the United States who are >44 years old, we sought to describe the prevalence of women who are 45-50 years old and who are at risk for UIP and their current contraceptive use habit and to compare these findings with those of women in younger age groups. Understanding UIP risk and contraceptive use among women who are 45-50 years old, compared with younger age groups, provides insight into UIP and contraceptive trends over the reproductive life span.
Materials and Methods
Overview
The Behavioral Risk Factor Surveillance System (BRFSS) is an ongoing, state-based, random-digit–dialed telephone survey of noninstitutionalized US civilian adults who are ≥18 years old that is conducted annually by state health departments in collaboration with the Centers for Disease Control and Prevention. The BRFSS sampling method samples households rather than individuals; 1 adult in each household is selected randomly to participate, so the likelihood of the same individual participating in multiple BRFSS surveys is low. The BRFSS collects information on health-related risk behaviors, chronic health conditions, and the use of preventive services. The BRFSS questionnaire consists of core questions that are used by all states, optional modules that are supported by Centers for Disease Control and Prevention programs and are available for states to use, and state-added questions. The BRFSS data are weighted to produce estimates that are representative of the state population. More detail on the BRFSS, including methods, is available from the BRFSS website.
We analyzed 2006, 2008, and 2010 Massachusetts BRFSS data; response rates during these years, based on the Council of American Survey and Research Organization guidelines, were 38.6%, 48.2%, and 47.5%, respectively. Although family planning questions were previously part of the 2002 and 2004 core BRFSS questionnaires and were offered as optional modules during other years, the questions were asked only among women who were 18-44 years old. We analyzed Massachusetts data because Massachusetts was the only state to include BRFSS family planning questions and to ask these questions among women who were 18-50 years old, beginning in 2006 and implemented during even-numbered years. We did not include data subsequent to 2010 because the BRFSS methods and weighting methods changed in 2011; therefore, subsequent survey data are not directly comparable to previous years. Analysis of 2012 and 2014 data, independent from previous years, was undesirable because of the low numbers of older women. Institutional review board approval was not needed because the analysis used publicly available data with de-identified participants.
Measures
Our outcome of interest was current contraceptive use and was measured with the use of several questions. Women were first asked, “Are you or your husband/partner doing anything now to keep you from getting pregnant?” Those who indicated “yes” were asked, “What are you or your husband/partner doing now to keep you from getting pregnant?” Response options for numerous contraceptive methods were included, and respondents who were using a method not listed could indicate “other” and specify the method being used. Women who reported the use of multiple methods were asked to report their primary contraceptive method; women who reported multiple partners were asked to consider their usual partner when answering the question. Women who indicated “no” to the first question were classified as using no method and were asked, “What is the main reason for not doing anything now to keep yourself from getting pregnant?” Those who reported “tubes tied” or “partner vasectomy” were recoded as contraceptive users.
Use of any contraceptive method was coded as “yes” or “no.” We also coded current contraceptive use by categories of methods. Use of permanent methods included tubal ligation or vasectomy. Use of long-acting, reversible contraception (LARC) included IUDs or implants. Use of hormonal methods included shots, pills, patch, or ring. Use of barrier methods included condoms, diaphragm, cervical cap, sponge, or shield. Use of some other method included withdrawal, rhythm, emergency contraception, or “other.” Please note that, in 2006 and 2008, ring use was included in the response option for barrier methods rather than its own response category. Although this may have led to underreporting of hormonal methods and overreporting of barrier methods, we do not expect noteworthy misclassification error because of the low rate (1.3%) of ring use reported among women in the United States who were 15-44 years old and an even lower rate (0.4%) among older women who were 40-44 years old.
Data analysis
We combined 2006, 2008, and 2010 Massachusetts BRFSS data for women who were 18-50 years old and who participated in the version of the survey that included family planning questions (n = 4930). Women were considered to be at risk for UIP unless they reported current pregnancy, hysterectomy, or not being sexually active in the past 12 months, which was ascertained by 3 separate questions, or reporting a same-sex partner or wanting to become pregnant as reasons for not using contraception. We were unable to identify and subsequently exclude women who were no longer at risk of UIP because of menopause because the survey did not ask about menopausal status.
Missing data are typically a concern in secondary analysis of large surveys. We examined differences between women with and without missing data (or “don’t know” or refused responses) on at least 1 of the 3 main questions that were used to determine whether women were at risk for UIP. We found that older women who were 45-50 years old had significantly ( P < .05) more missing data (28.3%) than women in other age groups (8.7%, 12.5%, 17.0%, and 13.3% for women who were 18-24, 25-34, 35-39, and 40-44 years old, respectively). A large proportion of older women with missing data had missing information on a current pregnancy; this was in part due to women who were ≥45 years old not being asked about current pregnancy in the 2006 survey. Among the entire sample (n = 4930), we assumed at risk for UIP status for 900 women (16.2%) because of missing data. We conducted a sensitivity analysis excluding these 900 women to explore how findings may have differed.
We estimated the prevalence of at risk for UIP, overall, and by age group (18-24, 25-34, 35-39, 40-44, and 45-50 years old). Among women considered to be at risk for UIP, we estimated the prevalence of current contraceptive use overall and by age group. We reported use of any method and use of specific methods by category. Given the large percentage of women (16.2%) with “don’t know” responses or missing data on current contraceptive use, we included a category for “don’t know”/missing data in the analysis. We also compared characteristics of women with and without missing data on current contraceptive use.
Among the subset of women who were 45-50 years old who were considered to be at risk for UIP (n = 940), we examined characteristics that were associated with current contraceptive use using chi-square tests of independence and probability values <.05 to determine statistical significance. Characteristics of interest included marital status, education level, race/ethnicity, smoking status, and lifetime diagnosis of diabetes mellitus or a cardiovascular medical condition (ie, heart attack, angina, coronary heart disease, or stroke). All analyses were performed on weighted data with SAS-callable SUDAAN (RTI International, Research Triangle Park, NC) to account for the complex sampling design of the BRFSS.
Results
Among all women in the sample, 18.3% were 45-50 years old; the majority were married (64.3%) and had >12 years of education (75.0%; Table 1 ). Most women classified themselves as white and non-Hispanic (80.4%), and 16.7% reported being a smoker (ie, smoked at least 100 cigarettes in their lifetime and smoked regularly at the time of interview). Less than 5% of women reported ever being diagnosed with diabetes mellitus or a cardiovascular medical condition.
| Variable | N a | % a |
|---|---|---|
| Age, y | ||
| 18-24 | 374 | 10.2 |
| 24-34 | 1201 | 25.5 |
| 35-39 | 945 | 21.1 |
| 40-44 | 1034 | 24.8 |
| 45-50 | 1376 | 18.3 |
| Married | ||
| Yes | 2595 | 64.3 |
| No | 2308 | 35.7 |
| Education, y | ||
| ≤12 | 1526 | 25.0 |
| >12 | 3386 | 75.0 |
| Race/ethnicity | ||
| White, non-Hispanic | 3553 | 80.4 |
| Other | 1342 | 19.6 |
| Smoker b | ||
| Yes | 1033 | 16.7 |
| No | 3877 | 83.3 |
| Ever diagnosed with diabetes mellitus or cardiovascular medical condition c | ||
| Yes | 291 | 4.6 |
| No | 4621 | 95.4 |
a Numbers based on unweighted sample. Percentages based on weighted sample. Percentages calculated excluding observations with missing values
b Has smoked at least 100 cigarettes in lifetime and is now smoking regularly
c Includes heart attack, angina, coronary health disease, or stroke.
Overall, 76.6% of women who were 18-50 years old were considered to be at risk for UIP; prevalence estimates were similar across age groups ( Table 2 ). Not being sexually active in the past 12 months or having a same-sex partner was the most common reason women who were 18-24 (21.7%), 35-39 (6.7%), 40-44 (11.6%), and 45-50 (12.6%) years old were considered not to be at risk for UIP. Older women (45-50 years old) more frequently reported hysterectomy (9.2%). Among women who were 25-34 years old, current pregnancy (8.8%), not being sexually active in the past 12 months or having a same-sex partner (8.7%), and wanting to become pregnant (7.6%) were common reasons for being considered not to be at risk for UIP.
| Variable | Unintended pregnancy | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| At risk a | Not at risk | |||||||||
| Currently pregnant | Hysterectomy | Not sexually active or same-sex partner | Want to become pregnant | |||||||
| N b | % c | N b | % c | N b | % c | N b | % c | N b | % c | |
| Total | 3605 | 76.6 | 153 | 3.7 | 276 | 4.3 | 746 | 11.0 | 150 | 4.3 |
| Age, y | ||||||||||
| 18-24 | 260 | 72.0 | 24 | 3.6 | 0 | 0 | 76 | 21.7 | 14 | 2.7 |
| 25-34 | 896 | 74.1 | 88 | 8.8 | 15 | 0.8 | 135 | 8.7 | 67 | 7.6 |
| 35-39 | 733 | 79.9 | 35 | 4.1 | 29 | 3.1 | 107 | 6.7 | 41 | 6.2 |
| 40-44 | 776 | 77.6 | 5 | 0.8 | 73 | 7.2 | 161 | 11.6 | 19 | 2.8 |
| 45-50 | 940 | 77.6 | 1 | 0.0 | 159 | 9.2 | 267 | 12.6 | 9 | 0.6 |
a Women were considered to be at risk for unintended pregnancy, unless they reported current pregnancy, hysterectomy, not being sexually active in the past 12 months, having a same-sex partner, or wanting to become pregnant
b Based on the unweighted sample
Table 3 summarizes current contraceptive use among women who were 18-50 years old and considered to be at risk for UIP. Use of any contraceptive method was highest for women who were 18-24 years old (79.9 %) and decreased as age increased with a low of 66.9% of women who were 45-50 years old using some method. Other trends by age group were observed. As age increased, so did the use of permanent contraception (tubal ligation or vasectomy), from 4.7% among women who were 18-24 years old to 44.0% among those who were 45-50 years old. Non-use of contraception also generally increased as age group increased, from 9.3% among women who were 18-24 years old to 16.8% among those who were 45-50 years old. For hormonal contraceptives (shots, pills, patch, or ring), use decreased as age group increased from 46.9% among women who were 18-24 years old to 5.9% among women who were 45-50 year old. Although the use of LARCs (IUDs or implants) was low overall (6.6%), its use peaked among women who were 25-34 years old (9.0%) and was lowest (3.3%) among women who were 45-50 years old. Reported use of barrier methods essentially plateaued after age 34 years, with a prevalence of 11.7% among women who were 35-39 years old, 10.9% among women who were 40-44 years old, and 11.6% among women who were 45-50 years old. Similar percentages of women among age groups 25-34, 35-39, 40-44, and 45-50 years old reported “don’t know” or gave no response to the question regarding contraceptive use. Significant differences ( P < .05) in characteristics of women with and without missing data on current contraceptive use were race/ethnicity and smoking status. Non-Hispanic, white women had a lower proportion of missing data (15.0%) compared with women of other racial/ethnic backgrounds (21.6%), and smokers had a lower proportion of missing data (10.5%) than nonsmokers (17.3%; data not shown).
| Variable | Any method | Permanent method b | Long-acting reversible contraception method c | Hormonal method d | Barrier method e | Other method f | No method | Do not know/no response | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| N g | % h | N g | % h | N g | % h | N g | % h | N g | % h | N g | % h | N g | % h | N g | % h | |
| Total | 2478 | 70.6 | 1046 | 27.4 | 217 | 6.6 | 612 | 20.6 | 473 | 13.2 | 130 | 2.8 | 489 | 13.1 | 638 | 16.2 |
| Age, y | ||||||||||||||||
| 18-24 | 190 | 79.9 | 8 | 4.7 | 24 | 6.7 | 101 | 46.9 | 51 | 21.0 | 6 | 0.7 | 32 | 9.3 | 38 | 10.8 |
| 25-34 | 648 | 72.2 | 119 | 12.9 | 84 | 9.0 | 273 | 32.4 | 135 | 15.1 | 37 | 2.9 | 92 | 11.7 | 156 | 16.0 |
| 35-39 | 529 | 71.3 | 260 | 32.1 | 46 | 7.5 | 107 | 17.0 | 92 | 11.7 | 24 | 3.1 | 75 | 11.3 | 129 | 17.3 |
| 40-44 | 511 | 67.9 | 281 | 34.0 | 36 | 6.0 | 77 | 13.2 | 85 | 10.9 | 32 | 3.8 | 115 | 14.7 | 150 | 17.3 |
| 45-50 | 600 | 66.9 | 378 | 44.0 | 27 | 3.3 | 54 | 5.9 | 110 | 11.6 | 31 | 2.1 | 175 | 16.8 | 165 | 16.3 |
a Women were considered to be at risk for unintended pregnancy unless they reported current pregnancy, hysterectomy, not being sexually active in the past 12 months, having a same-sex partner, or wanting to become pregnant
c Intrauterine devices or implants
d Shots, pills, patch, or ring; ring was included in this group in 2010 only
e Condoms, diaphragm, cervical cap, sponge, or shield; this group also included vaginal ring in 2006 and 2008
f Withdrawal, rhythm, emergency contraception, or other
Results of our sensitivity analysis, which excluded women in our sample assumed to be at risk for UIP because of missing data, showed fewer women to be at risk for UIP than detected in the main analysis (72.1% vs 76.6%); this difference was more substantial among women who were 45-50 years old (68.7% vs 77.6%). Patterns of reasons women were considered not to be at risk for UIP were similar (eg, not being sexually active or having a same-sex partner was the top reason for not being at risk among older women). Additionally, contraceptive use (any method and use of specific methods) was higher overall and by age group (data not shown). Patterns in contraceptive use by age group remained the same (eg, use of sterilization increased as age group increased; LARC use was highest among women who were 25-34 years old and lowest among women who were 45-50 years old).
In our sample, most women who were 45-50 years old and at risk for UIP (n = 940) were married (80.0%), had >12 years of education (76.8%), were white non-Hispanic (89.3%), were nonsmokers (84.9%), and had never been diagnosed with diabetes mellitus or a cardiovascular medical condition (94.8%; Table 4 ). None of these characteristics were significantly ( P < .05) associated with current use of contraceptive method when categorized as “any method.” However, the use of permanent contraception was significantly ( P < .05) associated with being married (46.7%) vs unmarried (34.3%), and the use of a LARC or hormonal method was associated significantly ( P < .05) with having >12 (10.8%) vs <12 (4.0%) years of education, being a nonsmoker (10.0%) vs smoker (3.2%), and never (9.4%) vs ever (2.6%) being diagnosed with diabetes mellitus or a cardiovascular medical condition. Contraceptive non-use was associated significantly ( P < .05) with being a smoker (26.7%) vs nonsmoker (15.3%) and being non-Hispanic white (18.0%) vs other race/ethnicity (8.2%). “Don’t know” or no response was associated significantly ( P < .05) with being a nonsmoker (18.3%) vs smoker (6.0%).
