Among civilian and military women alike, rates of unintended pregnancy are high.1–3 Such pregnancies, whether undesired or mistimed,4 can have significant emotional, economic, and health consequences for women and families.1,5 For servicewomen, unintended pregnancy can have unique effects, affecting military readiness as well as the career trajectories of female soldiers.6–11 For example, pregnancy during deployment necessitates medical evacuation, and serious complications of pregnancy such as ectopic pregnancy may be more difficult to diagnose and manage while deployed.12
It is estimated that 35–53% of unintended pregnancies in the military are attributable to contraceptive failure.7 Promotion of highly effective long-acting reversible contraceptive (LARC) methods, including the copper-containing intrauterine device (IUD), the levonorgestrel intrauterine system, and the etonogestrel-releasing subdermal implant (implant), is considered key to reducing unintended pregnancies.13 Although 11.6% of contracepting civilian women use a LARC method,14 little is known about LARC use among active-duty servicewomen. Existing studies have examined only subpopulations of deployed or trainee soldiers7,15–17 or broader populations, including retired military and family members.18
Given LARC's potential benefits, understanding patterns of LARC use among servicewomen is critical to their health and well-being.7 This study presents evidence regarding contemporary trends in the use of LARC methods among U.S. Army servicewomen to better understand the present state of LARC use among this underresearched group, identify servicewomen who are less likely to adopt LARC methods, and explore the potential reasons for that low adoption.
MATERIALS AND METHODS
This retrospective cohort study was conducted using the Stanford Military Data Repository, a database containing health and demographic information for all active-duty U.S. Army soldiers between January 2011 and December 2014. Our cohort therefore consisted of all active-duty women aged 18–44 years serving in the U.S. Army at any point during this time period. The Stanford Military Data Repository includes health encounter data comprised of medical information from inpatient and outpatient encounters contained within the Medical Data Repository as well as demographic and health-related data derived from several other sources collectively known as the Defense Manpower Data Center (Table 1). Diagnostic and procedure data arose from health care provider-entered codes that underwent review by professional coders with access to the full documentation of clinical events. Administrative data leveraged in this study were originally entered by military human resources personnel based on official documents such as birth certificates and military service data. These data are used by the military for official purposes including payroll and military unit readiness tracking and constitute the known “gold standard” information available for the participants.
The data we used were obtained from definitive, official military sources and effectively contained no discernible missing values. For some variables such as race, the “other” category covered situations not fitting standard categories, potentially including missing values. This research was approved by the institutional review board at Stanford University and by the Defense Health Agency's Human Research Protection Office. Eligible participants were female U.S. Army soldiers of reproductive age, defined as 18–44 years, who were on active duty at any point between January 2011 and December 2014. The primary outcome of interest was insertion of a LARC method—the copper IUD, levonorgestrel intrauterine system, or implant—at any time during the 4-year observation period.
We identified LARC insertions using Current Procedural Terminology codes, International Classification of Diseases, 9th Revision codes, and Healthcare Common Procedure Coding System codes (Table 1). Our definition of IUD insertion included the presence of an Healthcare Common Procedure Coding System code for the Progestasert IUD, which was discontinued in 2011, as well as codes for levonorgestrel capsules, which have not been available in the United States since 2002. Because these codes were used greater than 90% of the time in conjunction with a current IUD or implant code, we considered them valid indications of LARC insertion. A small number of soldiers (n=781 [4.7% of LARC users]) had codes for both IUD and implant insertion at some point during their active-duty service. We counted these soldiers as IUD users when comparing IUD with non-IUD users and as implant users when comparing implant with nonimplant users, but excluded these “dual users” from direct comparisons between IUD users and implant users given that they would have belonged to both comparison groups.
To assess trends in IUD and implant insertion over time, we calculated cumulative monthly insertion rates stratified into four age categories (18–22, 23–27, 28–35, and 36–44 years of age) and assessed the significance of any changes in insertion rate over time using linear regression. We calculated annual insertion rates by averaging the 12 monthly rates for a given year. For soldiers with multiple recorded insertions, we used only the first IUD or implant insertion.
To assess method continuation, we identified women who had both undergone LARC insertion during the study period and who additionally were still present in the military (as determined by the woman's presence in the Defense Manpower Data Center database) for at least 12 months after their LARC insertion. Within that subpopulation, we calculated the percentage of women whose record also contained a code for LARC removal during those 12 months. We repeated the same percentage calculation for both 24- and 36-month timeframes among those women with sufficient follow-up. Finally, we assessed age category as a predictor of removal for each of the three timeframes using χ2 and univariable regression testing.
To assess predictors of LARC uptake, we calculated relative risks by Poisson regression with robust error variance. We examined associations between the primary outcome of LARC insertion and the demographic and military characteristics listed in Table 1. Additionally, to evaluate characteristics that independently predict uptake of IUDs, and characteristics that independently predict uptake of implants, we evaluated each outcome separately by Poisson regression. In a secondary analysis, we evaluated the subcohort of LARC users only to compare predictors of choosing an IUD compared with an implant. A soldier's number of child dependents was available and was used as a proxy for parity as a result of the absence of more specific data. We controlled for the number of months that each soldier appeared in the data set by including this as a continuous predictor in our regression model. When comparing LARC and non-LARC users, values for predictor variables were drawn from each soldier's first active-duty (baseline) record. To account for changes in some predictor variables over time (such as age, marriage, and child dependents), we additionally conducted a sensitivity analysis drawing the same predictor variables' values from each soldier's last record; no associations or trends changed appreciably in this second analysis (not shown). When comparing IUD and implant users, predictors were drawn from the month of first IUD or implant insertion. All analyses were performed using Stata 14.0.
We analyzed data from all female soldiers aged 18–44 years on active duty anytime during the 48-month observation period (N=114,661). The median observation time for each soldier was 29 months (interquartile range 13–47). Just more than 30% of soldiers were present in the sample for the full 48 months, and more than one third (37.6%) of soldiers were new entrants to active duty during the observation time.
Baseline characteristics for this cohort are shown in Table 2. The mean age among study participants was 26.0 years (standard deviation 6.5). More than half were unmarried, and more than two thirds had no children. The majority (74.8%) reported their faith as Judeo-Christian, and nearly 70% had no more than a high school education. Two thirds of the soldiers in the study population had never been deployed to combat. Among those with a history of deployment, more than half (58.5%) had been deployed only once, 28.3% had been deployed twice, and 13.2% had been deployed three or more times.
A total of 16,623 women (14.5%) received a LARC method during the study period (Table 2). Of these, more than half (55.3%) received an IUD only, 40.0% received an implant only, and 4.7% received both methods during the study period. Among IUD recipients, 21.0% had a code for the levonorgestrel intrauterine system and 5% had a code for the copper IUD. We were unable to categorize the remainder of IUD placements by type.
The overall rate of LARC insertion increased slightly over the 4 years analyzed (from 59.0 to 71.3 insertions per 1,000 women per year, P<.001). The annual rate of first IUD insertion decreased over the study period (from 38.7 to 35.9 insertions per 1,000 women per year, P<.05), whereas implant insertions increased markedly (from 20.3 to 35.4 insertions per 1,000 women per year, P<.001) (Fig. 1). Intrauterine device insertions decreased significantly over time in the two youngest age groups and remained stable in the two oldest age groups. Implant insertions increased significantly in each of the age groups, most dramatically among the youngest women (Fig. 1). Among those with sufficient follow-up to evaluate, 70.8% of IUD users retained their method 1 year after placement, 58.9% at 2 years, and 48.0% at 3 years. Continuation at 1 year was highest for young women with 75.6% of women aged 18–22 years continuing their method at 1 year compared with 56.5% of women aged 36–44 years (P<.001). For implant users, 70.8% continued their method at 1 year, 53.1% at 2 years, and 34.8% at 3 years without significant variation by age group.
In the multivariable Poisson regression models incorporating all predictor variables, younger age persisted as a strong positive predictor of LARC uptake (Table 3). The probability of IUD insertion among the youngest women (18–22 years) was 2.5 times that of the oldest (36–44 years) (relative risk [RR] 2.56, 95% confidence interval [CI] 2.33–3.81) (Table 3). An even stronger effect was observed for implants, in which the probability of insertion for young women was seven times that of women in the oldest group (RR 7.12, 95% CI 5.92–8.55).
Having one child doubled the probability of IUD placement (RR 1.94, 95% CI 1.85–2.04) and modestly increased the probability of implant placement (RR 1.25, 95% CI 1.17–1.34). Intrauterine device insertion among married women was more than 1.5 times that among never-married women (RR 1.52, 95% CI 1.44–1.59), whereas marital status was not associated with implant placement (Tables 4 and 5).
Intrauterine device insertion was significantly more common among white women than among women in other racial groups. Implant insertions were slightly less likely among Asian and Pacific Islander as compared with white women, but otherwise there were no significant associations between race and implant insertion (Tables 4 and 5). Hispanic women were slightly less likely to receive an IUD (RR 0.93, 95% CI 0.88–0.99) and slightly more likely to receive an implant (RR 1.10, 95% CI 1.03–1.17) than non-Hispanic women (Tables 4 and 5). Implant insertions were higher among women with only a high school education compared with other education levels, but no clear association was observed between IUD placement and education. Notably, religious faith was not significantly associated with either IUD or implant placement. Previous combat deployment at any time, including before the study period, was associated with slightly decreased risk of IUD placement and was not associated with implant placement.
In addition to evaluating all participants' baseline characteristics, we conducted a secondary analysis of the subcohort of LARC users only to explore differences in those who opted for IUD compared with implant. Here, we used data from the month of first LARC insertion. At the time of LARC insertion, the mean age of IUD users was 28.6 years (range 18–44 years, standard deviation 5.7), and the mean age of implant users was 24.9 years (18–44 years, standard deviation 4.4). Among women who used either LARC method, implant users were more likely to be young, unmarried, nonwhite, and without dependents compared with IUD users. Multivariable regression among all LARC users (not shown) confirmed that these differences remained significant after adjustment. There was no clear trend differentiating IUD from implant users in terms of religious faith, education level, or number of past deployments.
There are wide-ranging consequences of unintended pregnancy for servicewomen. Pregnancy carries a risk of complications in the field12 and necessitates an end to deployment: of all nonbattle medical evacuations of women in combat in a 15-month period during Operation Iraqi Freedom, 74% were pregnancy-related.9 Long-acting reversible contraceptive methods provide an effective, nonuser-dependent contraceptive option for women in an unpredictable military environment. In addition to preventing unintended pregnancy, LARC's noncontraceptive benefits, including decreased menstrual bleeding,19,20 may be well suited to soldiers' needs. These characteristics suggest that LARC methods are underutilized in this population.
We observed higher rates of LARC insertion than in a study of civilians with employer-sponsored insurance21 but rates similar to those in a study of military TRICARE insurance.18 It is possible that the military's no-cost contraceptive coverage contributes to this difference. In comparison, the CHOICE project, a study of civilian women seeking contraception, found that two thirds chose LARC when financial barriers were removed and LARC was presented as a first-line option.22 Although the CHOICE cohort is not directly comparable with our own, this difference is noteworthy and suggests that, apart from issues of cost, offering LARC as a first-line option may affect its uptake. A similar portrayal of LARC as first line in the military could serve to provide a valuable option to a population for whom unintended pregnancy can be particularly high risk.12
In assessing predictors of LARC use, we observed an increased likelihood of LARC insertion among women with at least one dependent child and an increased likelihood of IUD insertion among married women. In civilian populations, health care provider and patient misconceptions about the safety of IUDs for nulliparous and nonmonogamous women are believed to be a central contributor to LARC underuse.19,23–25 Despite professional guidelines encouraging LARC as a first option for adolescents and nulliparous women,13,20 historical concerns about infection and infertility resulting from IUD placement persist among health care professionals as well as patients.19,20,24 Limited data show related misconceptions regarding implants.26 Our findings could be explained if military providers and soldiers share the same biases as civilians. If so, both health care provider and patient education about the safety of LARC methods, particularly IUDs, may allow more women in this demographic to receive highly effective contraception. Alternative explanations for LARC uptake among these demographics such as patient preference should also be considered.
The strengths of this study include its large sample size and comprehensive examination of inpatient, outpatient, military, and nonmilitary sources of reimbursed care, likely covering the majority of services sought by these soldiers and therefore presenting an accurate picture of LARC insertion rates over the study period. Furthermore, because military insurance provides free contraception, we were able to assess LARC uptake without cost as a barrier to access.
This study was limited by a dependence on health care provider coding accuracy to determine LARC insertions and removals. Additionally, we did not attempt to evaluate the reasons for LARC insertion or removal. We confined our analysis to an examination of first LARC insertions and did not attempt to count LARC users who entered the study period with an existing LARC method or tubal ligation. Finally, some servicewomen may seek LARC, or LARC removal, outside of the military system, and these events were not captured in our data. For this reason, the actual prevalence of women using LARC during the study period may be either higher or lower than the percentage of insertions we have reported here.
There are many aspects of LARC provision—including the timing of insertion in relation to childbirth and deployment, the types and locations of health care providers performing insertions, and the reasons for LARC insertion and removal—which we were unable to address within the scope of this article or as a result of limitations in the data. Further research should answer these important questions. In addition, qualitative work investigating military patient and health care provider experiences using and providing LARC will be key to better understanding the reasons for quantitative trends, including the increased implant uptake we have observed here.
The need to effectively address servicewomen's reproductive health has become critical in light of the growing number of female servicemembers in the military27 as well as increased servicewomen in combat positions after the Department of Defense's 2013 lift of the ban on women in combat.28 With these changes comes greater responsibility to empower servicewomen to prevent unintended pregnancy. Future efforts should focus on identifying and decreasing any existing barriers to the use of LARC for our soldiers and on optimizing access to and knowledge of all contraceptive choices.
1. Finer LB, Zolna MR. Declines in unintended pregnancy in the United States, 2008–2011. N Engl J Med 2016;374:843–52.
2. Grindlay K, Grossman D. Unintended pregnancy among active-duty women in the United States military, 2011. Contraception 2015;92:589–95.
3. Lindberg LD. Unintended pregnancy among women in the US military. Contraception 2011;84:249–51.
4. Sonfield A, Hasstedt K, Kavanaugh ML, Anderson R. The social and economic benefits of women's ability to determine whether and when to have children. New York (NY): Guttmacher Institute; 2013.
5. Gipson JD, Koenig MA, Hindin MJ. The effects of unintended pregnancy on infant, child, and parental health: a review of the literature. Stud Fam Plann 2008;39:18–38.
6. Burrelli DF. Abortion services and military medical facilities. Available at: https://fas.org/sgp/crs/misc/95-387.pdf
. Retrieved May 10, 2016.
7. Holt K, Grindlay K, Taskier M, Grossman D. Unintended pregnancy and contraceptive use among women in the U.S. military: a systematic literature review. Mil Med 2011;176:1056–64.
8. Tepe V, Yarnell A, Nindl BC, Van Arsdale S, Deuster PA. Women in combat: summary of findings and a way ahead. Mil Med 2016;181(suppl):109–18.
9. Belmont PJ Jr, Goodman GP, Waterman B, DeZee K, Burks R, Owens BD. Disease and nonbattle injuries sustained by a U.S. Army Brigade Combat Team during Operation Iraqi Freedom. Mil Med 2010;175:469–76.
10. Grindlay K, Yanow S, Jelinska K, Gomperts R, Grossman D. Abortion restrictions in the U.S. military: voices from women deployed overseas. Womens Health Issues 2011;21:259–64.
11. McGraw K, Koehlmoos TP, Ritchie EC. Women in combat: framing the issues of health and health research for America's servicewomen. Mil Med 2016;181(suppl):7–11.
12. Krulewitch CJ. Reproductive health of active duty women in medically austere environments. Mil Med 2016;181(suppl):63–9.
13. Increasing access to contraceptive implants and intrauterine devices to reduce unintended pregnancy. Committee Opinion No. 642. American College of Obstetricians and Gynecologists. Obstet Gynecol 2015;126:e44–8.
14. Daniels K, Daughtery J, Jones J, Mosher W. Current contraceptive use and variation by selected characteristics among women aged 15–44: United States, 2011–2013. National Health Statistics Reports, Centers for Disease Control and Prevention 2015;86:1–14.
15. Battista RM, Creedon JF Jr, Salyer SW. Knowledge and use of birth control methods in active duty army enlisted medical trainees. Mil Med 1999;164:407–9.
16. Nielsen PE, Murphy CS, Schutz J, Deering SH, Truong V, McCartin T, et al. Female soldiers' gynecologic healthcare in Operation Iraqi Freedom: a survey of camps with echelon three facilities. Mil Med 2009;174:1172–6.
17. Thomson BA, Nielsen PE. Women's health care in Operation Iraqi Freedom: a survey of camps with echelon I or II facilities. Mil Med 2006;171:216–9.
18. Chiles DP, Roberts TA, Klein DA. Initiation and continuation of long-acting reversible contraception in the United States military healthcare system. Am J Obstet Gynecol 2016;215:328.e1–9.
19. Stanwood NL, Garrett JM, Konrad TR. Obstetrician-gynecologists and the intrauterine device: a survey of attitudes and practice. Obstet Gynecol 2002;99:275–80.
20. Smith SA. The use of intrauterine devices (IUDs) in adolescents and nulliparous women: a systematic review. J Womens Health Care 2015;4:277.
21. Xu X, Macaluso M, Ouyang L, Kulczycki A, Grosse SD. Revival of the intrauterine device: increased insertions among US women with employer-sponsored insurance, 2002–2008. Contraception 2012;85:155–9.
22. Secura GM, Allsworth JE, Madden T, Mullersman JL, Peipert JF. The Contraceptive CHOICE Project: reducing barriers to long-acting reversible contraception. Am J Obstet Gynecol 2010;203:115.e1–7.
23. Kavanaugh ML, Frohwirth L, Jerman J, Popkin R, Ethier K. Long-acting reversible contraception for adolescents and young adults: patient and provider perspectives. J Pediatr Adolesc Gynecol 2013;26:86–95.
24. Spies EL, Askelson NM, Gelman E, Losch M. Young women's knowledge, attitudes, and behaviors related to long-acting reversible contraceptives. Womens Health Issues 2010;20:394–9.
25. Tyler CP, Whiteman MK, Zapata LB, Curtis KM, Hillis SD, Marchbanks PA. Health care provider attitudes and practices related to intrauterine devices for nulliparous women. Obstet Gynecol 2012;119:762–71.
26. Collier CH, Rosenthal M, Harris K, Lucas G, Stanwood N. Contraceptive implant knowledge and practices of providers serving an urban, low-income community. J Health Care Poor Underserved 2014;25:1308–16.
27. Patten E, Parker K. Women in the US military: growing share, distinctive profile. Washington, DC: Pew Social and Demographic Trends; 2011.
28. Rosenberg M, Phillips D. All combat roles now open to women, Defense Secretary says. Available at: http://www.nytimes.com/2015/12/04/us/politics/combat-military-women-ash-carter.html?_r=0
. Retrieved December 3, 2016.