Accepted Manuscript

Postpartum Contraception: Initiation and Effectiveness in a Large Universal

Healthcare System

Michael R. Brunson, MD, David A. Klein, MD, MPH, Cara H. Olsen, DrPH, Larissa
Weir, MD, Timothy A. Roberts, MD, MPH

PII: S0002-9378(17)30362-9
DOI: 10.1016/j.ajog.2017.02.036
Reference: YMOB 11556

To appear in: American Journal of Obstetrics and Gynecology

Received Date: 19 November 2016

Revised Date: 7 February 2017
Accepted Date: 21 February 2017

Please cite this article as: Brunson MR, Klein DA, Olsen CH, Weir L, Roberts TA, Postpartum
Contraception: Initiation and Effectiveness in a Large Universal Healthcare System, American Journal of
Obstetrics and Gynecology (2017), doi: 10.1016/j.ajog.2017.02.036.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Postpartum Contraception: Initiation and Effectiveness in a Large Universal Healthcare

System

Michael R. BRUNSON MD, David A. KLEIN MD, MPH, Cara H. OLSEN DrPH, Larissa
WEIR, MD, Timothy A. ROBERTS MD, MPH

PT
Affiliations:
From the Department of Pediatrics (Drs. Brunson and Roberts) and Obstetrics and Gynecology
(Dr. Weir), San Antonio Military Medical Center, TX, USA; and Department of Pediatrics (Drs.

RI
Brunson, Klein, and Roberts), Family Medicine (Dr. Klein), and Preventive Medicine and
Biometrics (Dr. Olsen), Uniformed Services University of the Health Sciences, MD, USA

SC
Addresses:
San Antonio Military Medical Center -- 3551 Roger Brooke Dr., Fort Sam Houston, TX, USA,
78234
Uniformed Services University of the Health Sciences -- 4301 Jones Bridge Road, Bethesda,

U
MD, USA, 20814
AN
Disclosure Statement: The authors report no conflict of interest

Financial Support: The authors report no financial support for this study
M

Paper Presentation Information: Abstract accepted for poster presentation at the annual
meeting of the Society for Adolescent Health and Medicine, New Orleans, LA, Mar. 9-12, 2017
D

Disclaimer: The view(s) expressed herein are those of the author(s) and do not reflect the
official policy or position of Brooke Army Medical Center, the U.S. Army Medical Department,
TE

the U.S. Army Office of the Surgeon General, the Department of the Air Force, the Department
of the Army or the Department of Defense or the U.S. Government
EP

Corresponding Author: Michael R. Brunson, MD – Adolescent Medicine Clinic, 3100

Schofield Rd, Bldg 1179, Fort Sam Houston, TX, 78234; Phone: 210-808-2370; Fax: n/a; Email:
michael.r.brunson.mil@mail.mil.
C

All authors have signed the statement of authorship.

AC

Word Count:
Abstract: 473
Main text: 2963

IRB Approval: Brooke Army Medical Center, April 11, 2016, IRB project number C.2016.095n

Preferred Table or Figure: Figure 3

 ACCEPTED MANUSCRIPT

Condensation: Among 373,840 postpartum women in the U.S. military healthcare system, long-
acting reversible contraception initiation is increasing, and is associated with lower hazards of
short interdelivery intervals.

Short version of title: Effectiveness of postpartum contraception

PT
RI
U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Abstract

BACKGROUND: Repeat pregnancies following a short interpregnancy interval are common

and are associated with negative maternal and infant health outcomes. Few studies have
examined the relative effectiveness of postpartum contraceptive choices.

OBJECTIVE: We aim to determine the initiation trends and relative effectiveness of postpartum

PT
contraceptive methods, with typical use, on prevention of short delivery intervals (≤27 months)
among women with access to universal healthcare, including coverage which entails no co-
payments and allows unlimited contraceptive method switching.

RI
STUDY DESIGN: This retrospective cohort study included women enrolled in the United States
military healthcare system admitted for childbirth between October 2010 - March 2015, with ≥6

SC
months postpartum enrollment. Using insurance records, women were analyzed based on the
most effective contraceptive method initiated during the first six months postpartum, even if
subsequently discontinued. Rates of interdelivery intervals ≤27 months, as proxies for
interpregnancy intervals ≤18 months, were determined using the Kaplan-Meier estimator.

U
Women who disenrolled, reached 27 months postpartum without another delivery, or reached the
end of the study period were censored. The influence of sociodemographic variables and
AN
contraceptive choices on time to subsequent delivery was evaluated using Cox regression
analysis, accounting for possible correlation among multiple deliveries by an individual woman.

RESULTS: During the study timeframe, 373,840 women experienced a total of 450,875
M

postpartum intervals. Women averaged 27 (standard deviation=5.3) years of age at the time of
delivery, 33.9% were under the age of 25, 15.5% were active duty service members, and 31.6%
had insurance sponsors of junior enlisted rank (suggesting lower income). Postpartum
D

contraceptive methods initiated included self or partner sterilization (7%), intrauterine device
(13.5%), etonogestrel implant (3.4%), depot medroxyprogesterone acetate (2.5%), and pill,
TE

patch, ring (36.8%). Furthermore, 36.7% did not initiate a prescription method. Etonogestrel
implant initiation increased from 1.7% of postpartum women in the first year of our study to
5.3% in the final year. The estimated short interdelivery interval rate was 17.4%, but rates varied
EP

with contraceptive method: 1% with sterilization, 6% with long-acting reversible contraception,

12% with depot medroxyprogesterone, 21% with pill, patch, ring, and 23% with no prescription
method. In a multivariable analysis, the adjusted hazards of short interdelivery interval was
highest among women who were younger, on active duty, or with officer insurance sponsors.
C

Compared with non-use of any prescription contraceptive, use of an intrauterine device reduced
the hazard of a subsequent delivery (adjusted hazard ratio=0.19, 95% confidence interval=0.18-
AC

0.20), as did etonogestrel-implant (adjusted hazard ratio=0.21, 95% confidence interval=0.19-

0.23), while pill, patch, ring had less effect (adjusted hazard ratio=0.80, 95% confidence
interval=0.78-0.81).

CONCLUSION: Postpartum initiation of long-acting reversible contraception is highly

effective at preventing short interdelivery intervals, whereas pill, patch, ring methods are
associated with rates of short interdelivery intervals similar to users of no prescription
contraception. This study supports long-acting reversible contraception as first-line
 ACCEPTED MANUSCRIPT

recommendations for postpartum women who wish to retain fertility but avoid early repeat
pregnancy.

PT
RI
U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Key Words: effectiveness, interdelivery interval, intrauterine device, LARC, long-acting

reversible contraception, postpartum contraception, subdermal implant

PT
RI
U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Introduction

Repeat pregnancies following a short interpregnancy interval are common in the United States,

with 30% of second or higher-order pregnancies following an interpregnancy interval of less

than 18 months.1-3 Pregnancies after short intervals (<18 months) are associated with negative

PT
health outcomes such as premature rupture of membranes, preterm birth, small for gestation age,

RI
uteroplacental bleeding, neonatal and infant mortality, and negative neurodevelopmental

outcomes.4-7 Most short-interval pregnancies (69%) are either unintended or mistimed.2, 3

SC
Unintended pregnancies are also associated with negative health behaviors and outcomes such as

lower breastfeeding rates, maternal depression, preterm birth, and low birthweight.4, 8 Many of

U
these associations persist after accounting for demographic factors such as young age, lower
AN
income, and minority race.8
M

Prevention of short interval and unintended pregnancies requires initiation of effective

D

contraception during the postpartum period.4, 9, 10 Many studies have demonstrated the greater
TE

contraceptive effectiveness of long-acting reversible contraceptive (LARC) methods, with both

typical and perfect use, than short-acting user-dependent hormonal methods (contraceptive pill,
EP

patch, and ring) and user-dependent non-hormonal methods (e.g. barriers) when initiated outside

the postpartum period.11-13 However, few studies have examined the relative effectiveness of
C

these contraceptive methods initiated within the postpartum period, and are limited by small
AC

sample size, selective population demographics, recall bias, short follow up periods, or lack of

differentiation between LARC methods and surgical sterilization.14-17

This study examined postpartum contraception initiation and relative theoretical effectiveness

under real-world conditions using a large group of postpartum mothers with access to
 ACCEPTED MANUSCRIPT

contraception at no personal financial cost. We specifically aimed to: (1) determine postpartum

contraception initiation trends within the military healthcare system, (2) evaluate the effect of

postpartum contraceptive method choice on short interdelivery intervals (≤27 months) in a large

actual-use cohort including method discontinuation or imperfect use, and (3) measure the

PT
relationship between demographic variables and short interdelivery intervals (IDI).

RI
Materials and Methods

SC
Design and Setting

We used data from the Military Health System Management Analysis and Reporting Tool (M2)

U
to conduct a retrospective cohort study of postpartum contraceptive use and repeat deliveries
AN
among women enrolled in the military healthcare insurance program, TRICARE Prime, between

October 1, 2010 and September 30, 2015. TRICARE Prime includes all active duty service
M

members, military retirees under the age of 65, and members of the national guard and reserve
D

who have been called to active duty. Sponsor’s spouses and unmarried dependent children under
TE

the age of 26 are also eligible. Beneficiaries receive medical care at no personal financial cost

through military treatment facilities or contracted civilian providers world-wide. The M2

EP

database contains medical and pharmacy billing records of all individuals who used TRICARE

Prime, and includes enrollment status for each month. Information is collected directly from the
C

military’s electronic medical record or from civilian network clinic, hospital, and pharmacy
AC

billing statements. The institutional review board of Brooke Army Medical Center, Fort Sam

Houston, Texas, approved this study.

 ACCEPTED MANUSCRIPT

We identified TRICARE Prime enrollees admitted to the hospital for childbirth, abortion, ectopic

pregnancy, or a miscarriage between October 1, 2010 and March 30, 2015 using the following

Medicare Severity-Diagnosis Related Group (MS-DRG) codes: 765, 766, 767, 768, 770, 774,

775, 777, and 779. Live births accounted for 99.5% of all index admissions. Some women gave

PT
birth multiple times during the study period. The shortest interdelivery interval was 6 months,

RI
therefore, only postpartum periods of ≥6 months were included in our analysis. We examined

450,875 postpartum periods occurring among 373,840 women in this study.

SC
We recorded the following demographic variables for each woman at the time of admission: age

U
(12-19, 20-24, 25-29, 30-34, 35-39, and 40+ years), eligibility status (active duty service member
AN
versus family members or military retirees), and the sponsor’s rank at the time of delivery (from

most junior to most senior: Junior Enlisted (E-1 to E-4), Senior Enlisted (E5 to E-9), Warrant
M

Officer (W-1 to W-5), Junior Officer (O-1 to O-3), and Senior Officer (O-4 to O-10)). Military
D

servicewomen are their own insurance sponsor, and dependents are sponsored by their military
TE

spouse or parent. Sponsor’s rank for retirees and their family members is the rank at the time of

the sponsor’s retirement. More senior rank is associated with higher level of sponsor education,
EP

more years of service, higher military pay, and was used as a proxy for socioeconomic status.
C

To examine the relative theoretical effectiveness of postpartum contraceptive methods on short

AC

IDI regardless of continuation or method switching, we identified the most effective method of

contraception initiated in the first 6 months after each index delivery. Postpartum contraceptive

options were grouped in the following order: tubal ligation, partner vasectomy, intrauterine

device (IUD), etonogestrel implant (ENG-implant), depot medroxyprogesterone acetate

 ACCEPTED MANUSCRIPT

(DMPA), short-acting user-dependent hormonal methods (contraceptive pills, patch, and ring),

and no prescription contraceptive method (e.g. barrier, withdrawal, and no method). We

identified episodes of tubal ligation of the mother using the following Current Procedural

Terminology (CPT) codes: 58565, 58600, 58605, 58611, 58615, and 58670. To identify partner

PT
vasectomy, we identified all postpartum women with a male spouse who had evidence of a

RI
vasectomy occurring between eight months before and six months after the delivery using the

CPT code 55250. We were not able to identify vasectomies occurring among non-spousal sexual

SC
partners.

U
We identified women who obtained an intrauterine device during the postpartum period by
AN
reviewing records for the 58300 CPT code and Healthcare Common Procedure Coding System

(HCPCS) codes J7300, J7302, and S4981. To identify women who obtained an ENG-implant,
M

we reviewed patient records for the 11975 and 11981 CPT codes and the J7307 HCPCS code.
D

We identified prescriptions for contraceptive pills, patches, and rings by reviewing pharmacy
TE

records for prescriptions using the American Hospital Formulary Service (AHFS) therapeutic

classification code of 681200 (contraceptives).

EP

To assess for postpartum periods that contained a short interpregnancy interval (≤18 months) but
C

were not captured in our analysis of short IDI, we conducted a post-hoc analysis examining
AC

repeat pregnancies among women who delivered in 2012 identified by outpatient pregnancy-

related billing codes (International Classification of Diseases (ICD-9) codes V22.x, V23.x,

V28.x, 632, 633-638, 640-649, 651-659, 671, 673-674, and 678-679) at least 2 months after

index delivery.
 ACCEPTED MANUSCRIPT

Statistical Analysis

Interdelivery intervals were estimated using Kaplan-Meier analysis, measuring the time in

months from an index admission for childbirth, abortion, ectopic pregnancy, or miscarriage, to a

PT
following admission for childbirth. Each postpartum period was followed forward in time until

RI
reaching one of four endpoints after the index admission: (1) another admission for childbirth,

(2) disenrollment from TRICARE Prime, (3) reaching 27 months postpartum without childbirth,

SC
or (4) reaching the end of the study period in September 2015. Abortion, ectopic pregnancy, and

miscarriage were included only in identifying index cases to attain more generalizable results.18

U
We determined an endpoint of <27 months to best approximate interpregnancy intervals of <18
AN
months used in prior postpartum contraception studies.1, 14, 16 Rates of short IDI were calculated

for each age group, eligibility status, sponsor’s rank, and contraceptive type. Postpartum
M

intervals that reached an endpoint other than childbirth were censored.

D
TE

Descriptive statistics and Kaplan-Meier analyses were conducted using SPSS version 22 (IBM

Corp., Armonk, NY). Cox proportional hazard models were used to assess the independent
EP

association of age, sponsor’s rank, eligibility status, and most effective contraception type with

short IDI. Demographic variables were included in the multivariate analysis if they demonstrated
C

a significant effect (p<0.05) on short IDI in a bivariate analysis. To account for the possible
AC

correlation among multiple observations per individual in our multivariable analyses, Cox

proportional hazards regression was used to estimate the conditional model of Prentice et al., in

which follow-up time starts at zero following each successive pregnancy (the "gap-time"

model).19 Robust variance estimates were calculated to account for possible correlation among
 ACCEPTED MANUSCRIPT

multiple observations per subject. The robust variance estimates use the sandwich estimator

described by Lin and Wei and implemented in the PHREG procedure in SAS version 9.3.20 An

alpha level of 0.05 was set for all comparisons.

PT
Results

RI
Our sample contained 373,840 women who accounted for 450,875 postpartum periods. Women

in our sample were an average of 27.3 (SD=5.3) years of age at the time of delivery, 33.9% were

SC
between 12-24 years, 15.5% were service members, 31.6% were junior enlisted personnel or had

junior enlisted insurance sponsors, and 11.4% were junior officers or had junior officer insurance

U
sponsors. The rates of short IDI varied among age groups, with the highest rates occurring
AN
among women ages 20-24 years (21.2%), women with junior enlisted sponsors (20%), and

women with junior officer sponsors (24.3%). When compared to active duty women, dependents
M

and retirees had similar rates of short interdelivery intervals (17.9% vs 17.3%) (Table 1).
D
TE

Of all postpartum women in our sample, 7.0% relied on tubal ligation or their partner’s

vasectomy as a contraceptive method, 16.9% initiated a LARC method, 2.5% were prescribed
EP

DMPA, 36.8% were prescribed a pill, patch or contraceptive ring, and 36.7% did not initiate any

prescription contraceptive method in the first 6 months postpartum. (Table 2 and Figure 1) Use
C

of ENG-implants increased from 1.7% of postpartum women in Fiscal Year 2010 to 5.3% in
AC

Fiscal Year 2015; otherwise, the initiation rates for other contraceptive methods remained

relatively stable over our study period. (Figure 2)

In univariable analysis, women ages 20-24 were more likely than all other age groups to initiate

any form of prescription contraception (p<0.0001). Compared to women over 24 years old,
 ACCEPTED MANUSCRIPT

younger women were more likely to initiate ENG-implant (12-19 years, 7.3%; 20-24 years,

5.9%; p <0.0001) and IUD (12-19 years, 15.2%; 20-24 years, 16.5%; p<0.0001). Postpartum

LARC method initiation declined with increasing age beyond 24 years, and use of either tubal

ligation, partner vasectomy, or no prescription contraception was highest among women ages

PT
>30 years. Contraception methods initiated by active duty servicewomen were similar to those of

RI
retirees and dependents (Table 2).

SC
For all postpartum women in our dataset, the estimated probability of experiencing a short IDI

was 17.4%. Rates of short IDI varied significantly among contraceptive options: 1.2% of women

U
who received tubal ligation or partner vasectomy, 5.8% of women who had received IUD, 7.1%
AN
of women who had received ENG-implant, 11.6% of women who had received DMPA, 20.7% of

women who had been prescribed a pill, patch, or ring, and 22.9% of women using no
M

prescription contraception method (Table 3 and Figure 3). In the post-hoc analysis of outpatient
D

encounters, an additional 6.0% of deliveries were followed by a short interpregnancy interval not
TE

identified in our IDI analysis. Of these pregnancies, 28% were spontaneous abortions and 0.5%

were ectopic pregnancies. The remainder did not have an outcome recorded. TRICARE does not
EP

cover elective terminations except in cases of rape, incest, or when the life of the mother is at

risk.
C
AC

In multivariable analysis controlling for demographic factors and multiple observations, hazards

of short IDI varied among contraception methods. Compared to no prescription contraception

method, women who initiated a LARC method in the first six months postpartum experienced

approximately an 80% lower hazard of having a short IDI, while women prescribed user
 ACCEPTED MANUSCRIPT

dependent hormonal methods experienced only a 20% lower hazard (ENG-Implant: adjusted HR

0.21, 95% confidence interval 0.19-0.23; IUD: HR adj 0.19, 95% CI 0.18-0.20; Pill, patch, ring:

HR adj 0.80, 95% CI 0.78-0.81). Women who initiated DMPA in the postpartum period reduced

their hazard of short IDI by 60% (HR adj 0.39 95% CI 0.36-0.42). Finally, women who used

PT
tubal ligation or partner vasectomy had the lowest hazards of experiencing short IDI (HR adj

RI
0.008, 95% CI 0.005-0.012 and HR adj 0.05, 95% CI 0.04-0.06 respectively) (Table 4 and Figure

3). Compared to women more than 40 years of age, women ages 20-24 (HR adj 5.03, 95% CI

SC
4.50-5.63) and ages 12-19 (HR adj 4.79, 95% CI 4.26-5.40) had higher hazards of short IDI.

Among all rank groups, women of higher socioeconomic status (junior and senior officers, and

U
their beneficiaries) had higher hazards of short IDI compared to women with the lowest
AN
socioeconomic status (junior enlisted and their beneficiaries) (Table 4).
M

Comment
D

This analysis shows that the majority of postpartum women in the U.S. military healthcare
TE

system use less effective contraceptive methods such as short-acting user-dependent hormonal

methods or no prescription contraceptive method. With typical use, LARC methods initiated in
EP

the postpartum period are the most effective reversible contraception methods for preventing

short IDI. Our study population included women with universal access to healthcare, which
C

includes access to contraception and postpartum care at no personal financial cost, free and
AC

unlimited contraceptive method switching, and a comprehensive healthcare records system.

Therefore, this is an ideal population to evaluate postpartum contraceptive choices and

effectiveness with typical use.

 ACCEPTED MANUSCRIPT

Overall, postpartum LARC method initiation is increasing over time in our study population and

IUDs account for the largest proportion of LARC initiation, which is consistent with national and

military studies among all women.13, 21, 22 However, there are few large studies examining

postpartum contraception initiation. Our cohort initiated LARC methods more often (17%) than

PT
a national survey study (6%) and a prospective study of Texas women with public and private

RI
health insurance (13%).14, 23 The latter study demonstrated that women with higher income were

more likely to initiate LARC, however, we found that lower income was associated with

SC
increased LARC initiation. This may reflect the elimination of cost and availability barriers in

this universal healthcare system, as previous studies have found that 34% of postpartum women

U
desire LARC, and postpartum LARC initiation is highest among populations for which cost or
AN
availability barriers are minimized.16, 23-27
M

Our study, like previous studies, found that short-acting user-dependent hormonal contraceptive
D

methods are associated with lower hazards of short interval outcomes than non-prescription
TE

methods or no contraception.14, 16 However, user-dependent methods have much lower

continuation rates and are much less effective with typical use than with ideal use.11, 28-30 We
EP

found that postpartum women prescribed short-acting user-dependent hormonal contraception

had significantly higher hazards of short IDI when compared to women initiating longer-acting
C

methods like DMPA, LARC, tubal ligation, and partner vasectomy, which is consistent with the
AC

results of prior studies.14, 16 We also found lower hazards of short IDI with intrauterine

contraceptives than subdermal implants, which is likely due to higher continuation rates of IUDs

than of ENG-implants.11, 22
 ACCEPTED MANUSCRIPT

Our focus on measuring interdelivery intervals rather than interpregnancy intervals makes

comparison to previous studies challenging. Our rate of short IDI (17.4%) is lower than the

national rate of short interval pregnancies (e.g. 30% during 2006-2010), but our analysis does not

include spontaneous and elective abortions, or deliveries occurring after TRICARE

PT
disenrollment.1-3 We estimated that 6.0% of our deliveries were followed by a short interval

RI
pregnancy not captured in our IDI analysis, however, TRICARE does not cover elective

terminations, and data on self-funded terminations is not known in our population. 17, 32, 33

SC
Like previous studies, this study demonstrated a strong relationship between younger age and

U
increased hazards, and rate, of short interval outcomes.1-3, 14 Previous studies have also
AN
demonstrated an association between low income or education and short interval outcomes, but

our data found the highest hazards of short IDI among women with the highest socioeconomic
M

status.2, 3, 14 Our population has access to contraception at no personal-financial cost, which

D

might minimize the confounding effects of income on pregnancy intervals. Alternatively, these
TE

higher-income women in our population were less likely to use any form of postpartum

contraception, which may indicate different family planning desires.

EP

Active duty military servicewomen make postpartum contraception choices similar to those of
C

military retirees and dependents, but have higher hazards of short interdelivery intervals. Prior
AC

studies have shown that military servicewomen have higher rates of unintended pregnancy than

the general U.S. population, and approximately ten percent of servicewomen report an

unintended pregnancy during a year-long combat deployment.34, 35 Also, military servicewomen

report less use of contraception during deployment than at home, citing difficulty with initiating
 ACCEPTED MANUSCRIPT

or continuing contraception due to logistical constraints in a deployed environment.36 This is

important because increasing initiation of LARC in the postpartum period could reduce

unintended pregnancies in this military population as it has in the general U.S. population,

leading to improved military readiness, cost savings, and population health outcomes.32, 37

PT
RI
Strengths of this study include the large number of postpartum women identified, six years of

continuous enrollment information, and data on services received from both military healthcare

SC
and military-covered civilian healthcare providers. This allowed us to examine postpartum

contraception choices and subsequent childbirth based on typical use among a large sample of

U
women receiving free health care. These results are important information for health care
AN
providers to consider while counseling their patients, as the effectiveness of a chosen

contraception begun in the postpartum period, with typical use, might be lower than intended.
M

Accurate discussions of relative effectiveness are an important part of counseling on postpartum

D

contraceptive options.10
TE

Limitations of this study include its retrospective design and dependence on coding accuracy.38
EP

Also, our ability to differentiate combined oral contraceptive pills from progestin-only pills, and

barrier methods from no method, was limited because of the nature of the dataset. Although our
C

dataset contained limited race and ethnicity information, it contained accurate income and
AC

education data of insurance sponsors via rank. Finally, while many studies used repeat pregnancy

as their defined endpoint, we used repeat delivery because this was a more reliable measurable

outcome, and also because the military financial and operational costs are more strongly affected

by completed pregnancies than those not followed by childbirth.

 ACCEPTED MANUSCRIPT

Overall, this study supports LARC as a first-line recommendation for postpartum women who

wish to retain fertility but avoid early repeat pregnancy. Postpartum LARC remains underutilized

in our universal healthcare system and measures to increase its initiation, such as improved

PT
prenatal counseling and immediate postpartum placement, should be utilized. Our study also

RI
identifies a potential area for intervention to improve cost and military readiness outcomes of

those in the healthcare system studied. Further study is needed to identify methods to increase

SC
LARC initiation in postpartum women nationally, such reducing financial barriers, immediate

postpartum initiation, and improved prenatal contraceptive counseling.

U
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

References

[1] Copen CE, Thoma ME, Kirmeyer S. Interpregnancy Intervals in the United States: Data
From the Birth Certificate and the National Survey of Family Growth. Natl Vital Stat Rep
2015;64:1-10.

[2] Cheslack-Postava K, Winter AS. Short and long interpregnancy intervals: correlates and

PT
variations by pregnancy timing among U.S. women. Perspect Sex Reprod Health 2015;47:19-26.

[3] Gemmill A, Lindberg LD. Short interpregnancy intervals in the United States. Obstet

RI
Gynecol 2013;122:64-71.

[4] Institute of Medicine. Clinical preventive services for women: Closing the gaps. Washington,

SC
DC: The National Academies Press; 2011:102-110.

[5] Conde-Agudelo A, Rosas-Bermudez A, Castano F, Norton MH. Effects of birth spacing on

maternal, perinatal, infant, and child health: a systematic review of causal mechanisms. Stud

U
Fam Plann 2012;43:93-114.
AN
[6] Kozuki N, Lee AC, Silveira MF, et al. The associations of birth intervals with small-for-
gestational-age, preterm, and neonatal and infant mortality: a meta-analysis. BMC Public Health
2013;13 Suppl 3:S3.
M

[7] Zerbo O, Yoshida C, Gunderson EP, Dorward K, Croen LA. Interpregnancy Interval and
Risk of Autism Spectrum Disorders. Pediatrics 2015;136:651-7.
D

[8] Kost K, Lindberg L. Pregnancy intentions, maternal behaviors, and infant health:
investigating relationships with new measures and propensity score analysis. Demography
TE

2015;52:83-111.

[9] Gynecology TaCOOA. Committee Opinion No. 670: Immediate Postpartum Long-Acting
Reversible Contraception. Obstet Gynecol 2016;128:e32-7.
EP

[10] Gavin L, Moskosky S, Carter M, et al. Providing quality family planning services:
Recommendations of CDC and the U.S. Office of Population Affairs. MMWR Recomm Rep
C

2014;63:1-54.
AC

[11] Diedrich JT, Zhao Q, Madden T, Secura GM, Peipert JF. Three-year continuation of
reversible contraception. Am J Obstet Gynecol 2015;213:662 e1-8.

[12] O'neil-Callahan M, Peipert JF, Zhao Q, Madden T, Secura G. Twenty-four-month

continuation of reversible contraception. Obstet Gynecol 2013;122:1083-91.

[13] Berenson AB, Tan A, Hirth JM. Complications and continuation rates associated with 2
types of long-acting contraception. Am J Obstet Gynecol 2015;212:761 e1-8.
 ACCEPTED MANUSCRIPT

[14] White K, Teal SB, Potter JE. Contraception after delivery and short interpregnancy intervals
among women in the United States. Obstet Gynecol 2015;125:1471-7.

[15] Cohen R, Sheeder J, Arango N, Teal SB, Tocce K. Twelve-month contraceptive

continuation and repeat pregnancy among young mothers choosing postdelivery contraceptive
implants or postplacental intrauterine devices. Contraception 2016;93:178-83.

PT
[16] Thiel De Bocanegra H, Chang R, Howell M, Darney P. Interpregnancy intervals: impact of
postpartum contraceptive effectiveness and coverage. Am J Obstet Gynecol 2014;210:311 e1-8.

RI
[17] Tocce KM, Sheeder JL, Teal SB. Rapid repeat pregnancy in adolescents: do immediate
postpartum contraceptive implants make a difference? Am J Obstet Gynecol 2012;206:481 e1-7.

SC
[18] Shachar BZ, Lyell DJ. Interpregnancy interval and obstetrical complications. Obstet
Gynecol Surv 2012;67:584-96.

[19] Prentice RL, Williams BJ, Peterson AV. On the Regression Analysis of Multivariate Failure

U
Time Data. Biometrika 1980;68:373-79.
AN
[20] Lin DY, Wei LJ. The Robust Inference for the Cox Proportional Hazards Model. Journal of
the American Statistical Association 1989;84:1074-78.

[21] Branum AM, Jones J. Trends in long-acting reversible contraception use among U.S.
M

women aged 15-44. NCHS Data Brief 2015:1-8.

[22] Chiles DP, Roberts TA, Klein DA. Initiation and continuation of long-acting reversible
D

contraception in the United States military healthcare system. Am J Obstet Gynecol 2016.
TE

[23] Potter JE, Hopkins K, Aiken AR, et al. Unmet demand for highly effective postpartum
contraception in Texas. Contraception 2014;90:488-95.
EP

[24] Salcedo J, Moniaga N, Harken T. Limited Uptake of Planned Intrauterine Devices During
the Postpartum Period. South Med J 2015;108:463-8.

[25] Trussell J, Lalla AM, Doan QV, Reyes E, Pinto L, Gricar J. Cost effectiveness of
C

contraceptives in the United States. Contraception 2009;79:5-14.

AC

[26] 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.

[27] Tang JH, Dominik R, Re S, Brody S, Stuart GS. Characteristics associated with interest in
long-acting reversible contraception in a postpartum population. Contraception 2013;88:52-7.

[28] Reeves MF, Zhao Q, Secura GM, Peipert JF. Risk of unintended pregnancy based on
intended compared to actual contraceptive use. Am J Obstet Gynecol 2016;215:71 e1-6.
 ACCEPTED MANUSCRIPT

[29] Peipert JF, Zhao Q, Allsworth JE, et al. Continuation and satisfaction of reversible
contraception. Obstet Gynecol 2011;117:1105-13.

[30] Trussell J. Contraceptive failure in the United States. Contraception 2011;83:397-404.

[31] Finer LB, Zolna MR. Declines in Unintended Pregnancy in the United States, 2008-2011. N

PT
Engl J Med 2016;374:843-52.

[32] Peipert JF, Madden T, Allsworth JE, Secura GM. Preventing unintended pregnancies by

RI
providing no-cost contraception. Obstet Gynecol 2012;120:1291-7.

[33] Baldwin MK, Edelman AB. The effect of long-acting reversible contraception on rapid

SC
repeat pregnancy in adolescents: a review. J Adolesc Health 2013;52:S47-53.

[34] Lindberg LD. Unintended pregnancy among women in the US military. Contraception
2011;84:249-51.

U
[35] Grindlay K, Grossman D. Unintended pregnancy among active-duty women in the United
AN
States military, 2011. Contraception 2015;92:589-95.

[36] 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.
M

[37] Harper CC, Rocca CH, Thompson KM, et al. Reductions in pregnancy rates in the USA
with long-acting reversible contraception: a cluster randomised trial. The Lancet 2015;386:562-
D

68.
TE

[38] Yoshihara H, Yoneoka D. Understanding the statistics and limitations of large database
analyses. Spine (Phila Pa 1976) 2014;39:1311-2.
C EP
AC
 ACCEPTED MANUSCRIPT

Table 1: Population Characteristics and Short Interdelivery

Interval (IDI) Rates by Demographic Group
Short IDI Rate*
Demographic Group Total (n=450,875) (95% CI)
Age at delivery (y)

PT
12-19 21003 ( 4.7%) 19.8% (19.1-20.5)
20-24 131780 (29.2%) 21.2% (21.0-21.4)
25-29 151122 (33.5%) 19.0% (18.9-19.1)

RI
30-34 100775 (22.4%) 14.8% (14.8-14.9)
35-39 38157 ( 8.5%) 10.3% ( 9.9-10.6)
40+ 8038 ( 1.8%) 5.1% ( 4.5-5.6)

SC
Eligibility Status
Active Duty 69936 (15.5%) 17.9% (17.6-18.1)
Dependents and Retirees 380939 (84.5%) 17.3% (16.5-18.1)
Sponsor's Rank†
Junior Enlisted
U
142699 (31.6%) 20.0% (19.6-20.5)
AN
Senior Enlisted 227297 (50.4%) 15.1% (14.6-15.7)
Warrant Officers 6182 ( 1.4%) 13.4% (12.4-14.4)
Junior Officers 51179 (11.4%) 24.3% (23.8-24.8)
Senior Officers 23403 ( 5.2%) 15.4% (14.9-15.9)
M

CI, confidence interval; IDI, interdelivery interval

* Kaplan-Meier estimated rates of repeat delivery by 27-months postpartum
D

with 95% confidence interval

†
Records listed as ‘Other’ (n=115) not included.
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Table 2 : Most Effective Contraceptive Method Started Within 6 Months Postpartum

by Demographic Group*

Demographic Total Tubal Vasect- ENG- No Rx

PT
Group (n) Ligation omy IUD Implant DMPA PPR Method
Age at delivery (y)
12-19 21003 0.1% 0.1% 15.2% 7.3% 5.1% 36.8% 35.5%

RI
20-24 131780 1.3% 1.1% 16.5% 5.9% 3.7% 40.4% 31.0%
25-29 151122 3.5% 3.4% 14.0% 2.8% 2.3% 38.6% 35.4%
30-34 100775 5.6% 5.1% 11.1% 1.4% 1.6% 34.3% 40.9%

SC
35-39 38157 9.2% 6.2% 8.1% 0.8% 1.3% 27.5% 47.0%
40+ 8038 12.1% 5.2% 6.1% 0.4% 0.9% 20.4% 54.9%
All Ages 450875 3.8% 3.2% 13.5% 3.4% 2.5% 36.8% 36.7%

U
Duty Status
Active Duty 69936 2.8% 1.2% 15.5% 5.6% 4.1% 39.1% 31.5%
AN
Servicewomen
Retirees and 380939 4.0% 3.6% 13.2% 3.0% 2.3% 36.4% 37.6%
Dependents
M

Sponsor's Rank†
Junior Enlisted 142699 2.7% 1.6% 15.2% 5.4% 3.6% 38.3% 33.1%
Senior Enlisted 227297 4.7% 3.9% 13.5% 3.0% 2.4% 36.3% 36.2%
D

Warrant 6182 5.4% 6.3% 10.9% 2.2% 1.8% 33.3% 40.1%

Officer
TE

Junior Officer 51179 2.3% 3.2% 11.1% 1.2% 0.9% 37.7% 43.5%
Senior Officer 23403 4.4% 5.8% 9.5% 0.7% 0.7% 31.5% 47.4%
DMPA, depot medroxyprogesterone; ENG-Implant, etonogestrel implant; IUD, intrauterine device; PPR, pills,
EP

patch and ring;

* All comparisons are statistically significant; P < 0.0001
† Records listed as ‘Other’ (n=115) not included.
C
AC
 ACCEPTED MANUSCRIPT

Table 3: Short IDI Rate by the Most Effective Contraceptive Method

Started Within 6 Months Postpartum
Months Postpartum (95% CI)
Contraceptive Method 9 months 18 months 27 months
Tubal Ligation 0.0% 0.1% (0.0-0.1) 0.2% ( 0.1-0.2)

PT
Partner Vasectomy 0.0% 0.6% (0.5-0.8) 1.0% ( 0.8-1.1)
Intrauterine Device 0.0% 1.0% (0.9-1.1) 5.8% ( 5.6-6.1)
Etonogestrel Implant 0.0% 1.2% (1.0-1.5) 7.1% ( 6.5-7.7)

RI
Depot Medroxyprogesterone 0.0% 2.5% (2.2-2.8) 11.6% (10.9-12.4)
Pill, Patch, or Ring 0.0% ( 0-0.1) 6.7% (6.4-7.1) 20.7% (20.3-21.1)

SC
No Prescription Method 0.2% ( 0-0.3) 9.0% (8.9-9.1) 22.9% (22.6-23.2)
CI, confidence interval; IDI, interdelivery interval
Data are Kaplan-Meier estimated rates of repeat delivery by 27-months postpartum with 95%
confidence interval.

U
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Table 4: Multivariable Analysis of Characteristics

Associated With Having a Short Interdelivery
Interval
Adjusted
Characteristic HR 95% CI

PT
Age Group
12-19 4.79 (4.26-5.40)
20-24 5.03 (4.50-5.63)

RI
25-29 4.20 (3.75-4.70)
30-34 3.14 (2.80-3.51)

SC
35-39 2.14 (1.91-2.41)
40+ Reference
Eligibility Status
Active Duty 1.08 (1.05-1.10)

U
Dependents and Retirees Reference
AN
Sponsor's Rank
Junior Enlisted Reference
Senior Enlisted 0.85 (0.83-0.87)
Warrant Officers 0.86 (0.79-0.93)
M

Junior Officers 1.32 (1.28-1.35)

Senior Officers 1.11 (1.06-1.16)
D

Contraceptive Method
No Prescription Method Reference
TE

Pill, Patch, or Ring 0.80 (0.78-0.81)

Depot Medroxyprogesterone 0.39 (0.36-0.42)
Etonogestrel-Implant 0.21 (0.19-0.23)
EP

Intrauterine Device 0.19 (0.18-0.20)

Partner Vasectomy 0.05 (0.04-0.06)
Tubal Ligation 0.008 (0.005-0.012)
C

CI, confidence interval; HR, hazard ratio

Hazards for having a repeat delivery prior to 27-months
AC

postpartum based on the marginal Cox proportional hazard

regression adjusting for repeated events using robust variance
estimates; P ≤0.0001
 ACCEPTED MANUSCRIPT

Supplemental Table 1: MS-DRG Code Definitions

Code Description
765 Cesarean Section with Complications/Comorbidities
766 Cesarean Section without Complications/Comorbidities
767* Vaginal Delivery with Sterilization and/or Dilation and Curettage
768 Vaginal Delivery with Operating Room Procedure except

PT
Sterilization and/or Dilation and Curettage
770 Abortion with Dilation and Curettage, Aspiration Curettage Or
Hysterotomy

RI
774 Vaginal Delivery with Complicating Diagnoses
775 Vaginal Delivery without Complicating Diagnoses
777† Ectopic Pregnancy

SC
779† Abortion without Dilation and Curettage
* Used only to identify repeat deliveries
†
Used only to identify index deliveries

U
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Figure 1: Most Effective Contraception Method Initiated Within 6 Months

Postpartum by Age at Delivery
Brown = Tubal Ligation
Black = Partner Vasectomy
Orange = Intrauterine Contraception

PT
Purple = Etonogestrel Implant
Green = Depot Medroxyprogesterone

RI
Red = Pill, Patch, or Ring
Blue = No Prescription Method

SC
Figure 2: Most Effective Contraceptive Method Initiated in the First 6 Months
Postpartum in the Military Healthcare System
Blue = No Prescription Method
U
AN
Brown = Pill, Patch, or Ring
Light Green = Intrauterine Contraception
Red = Etonogestrel Implant
M

Dark Green = Tubal Ligation

Purple = Partner Vasectomy
Black = Depot Medroxyprogesterone
D
TE

Figure 3: Kaplan-Meier Analysis of Subsequent Delivery by Postpartum

Contraceptive Method
EP

Time to delivery by most effective contraception method initiated within 6 months of delivery.
Orange = Tubal Ligation
Purple = Partner Vasectomy
C

Green = Intrauterine Contraception

AC

Red = Etonogestrel Implant

Black = Depot Medroxyprogesterone
Brown = Pill, Patch, or Ring
Blue = No Prescription Method
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC