Condom Use With Long-Acting Reversible Contraception vs Non–Long-Acting Reversible Contraception Hormonal Methods Among Postpartum Adolescents

Key Points

Question

Among sexually active postpartum adolescents, are long-acting reversible contraceptive (LARC) users as likely to use condoms than users of non-LARC hormonal methods?

Findings

In this cross-sectional analysis of 5480 teenage mothers using LARC or non-LARC hormonal methods, overall, 29% also used condoms; however, condom use was half as common in LARC vs non-LARC hormonal method users. Condom use was lowest among intrauterine device users and highest among pill users compared with users of other hormonal contraceptive methods examined.

Meaning

Evidence-based interventions emphasizing the importance of postpartum condom use for prevention of sexually transmitted infections, particularly among teenage mothers using LARC, may help reduce risk of sexually transmitted infections.

Abstract

Importance

Increased use of long-acting reversible contraception (LARC; intrauterine devices [IUDs] and implants) has likely contributed to declining US teenage pregnancy and birth rates, yet sexually transmitted infection (STI) rates among teenagers remain high. While LARC methods are highly effective for pregnancy prevention, they, as with all nonbarrier methods, do not protect against STIs, including HIV. Studies of the general adolescent population suggest condom use is lower among LARC vs non-LARC hormonal methods users (birth control pill, contraceptive patch, vaginal ring, or injection). Despite the high use of LARC among postpartum teenagers, no studies have examined whether condom use differs by contraceptive method in this population.

Objective

To compare condom use among sexually active postpartum teenagers using LARC vs those using non-LARC hormonal methods.

Design, Setting, and Participants

Cross-sectional analysis using 2012 to 2015 data from the Pregnancy Risk Assessment Monitoring System (PRAMS), a multisite and population-based surveillance system that collects data on maternal attitudes, behaviors, and experiences before, during, and shortly after pregnancy. We used data from 37 sites. Using multivariable survey-weighted logistic regression, we assessed the association of condom use by contraceptive methods. Participants were teenage mothers (≤19 years) with a recent live birth reporting LARC or non-LARC hormonal method use. Data were analyzed between March 2018 and April 2018.

Main Outcomes and Measures

Condom use with LARC vs condom use with non-LARC hormonal methods.

Results

Among the 5480 (weighted N = 245 847) postpartum teenage mothers in our sample, most were aged 18 to 19 years, unmarried, had current Medicaid coverage, were first-time mothers, had reported their pregnancy was unintended, and almost half were non-Hispanic white. Overall, condom use was reported by 28.8% of these teenagers. Users of LARC compared with non-LARC hormonal methods were half as likely to use condoms (17.8% vs 35.6%; adjusted prevalence ratio [aPR], 0.50; 95% CI, 0.41-0.60). Users of IUDs (15.1%) were less likely to report condom use than those using an implant (21.5%; aPR, 0.70; 95% CI, 0.51-0.98), patch, ring, or injection users (24.9%; aPR, 0.61; 95% CI, 0.47-0.79), and pill users (47.2%; aPR, 0.32; 95% CI, 0.25-0.40).

Conclusions and Relevance

Self-reported condom use was low overall among postpartum teenage mothers and lower among users of LARC vs non-LARC hormonal methods. Given the high rates of STIs among teenage mothers combined with higher use of LARC among postpartum teenaged mothers, interventions to promote condom use for STI/HIV prevention during the postpartum period are critically important.

This study compares condom use among sexually active postpartum teenagers using long-acting reversible contraception LARC vs those using non–long-acting reversible contraception (non-LARC) hormonal methods

Introduction

Teenage birth rates in the United States are at a historic low, having declined 51% since 2007.¹ Among teenagers, uptake of long-acting reversible contraceptive (LARC) methods, which include intrauterine devices (IUDs) and contraceptive implants, has increased.^2,3,4,5 Declines in teenage pregnancy are partially attributable to increasing use of LARC methods.^6,7 Overall, among teenagers at risk for unintended pregnancy, LARC use has increased (1.1% in 2006-2008 to 3.2% in 2011-2013)⁸; however, among postpartum teenagers, this increase has been more marked from 5.3% in 2004/2005 to 25.3% in 2012/2013.⁹ Nevertheless, US teenage pregnancy rates remain considerably higher than in other developed countries,¹⁰ and approximately 210 000 teenagers aged 15 to 19 years gave birth in 2016.¹ Teenaged birth is associated with subsequent adverse health and socioeconomic outcomes for mothers and their children¹¹ and high rates of repeated teenage birth, with 1 in 6 teenage mothers experiencing an additional teenage birth in 2015.⁹

There is variability in the effectiveness of different contraceptive methods for pregnancy prevention. In part because no further user action is required after insertion, LARC methods have a 1% or less failure rate during the first year of both perfect and typical use.¹² In contrast, shorter-acting hormonal contraceptive methods (ie, birth control pill, injection, patch, or ring), which require users to consistently and correctly use their method of choice, have typical use failure rates ranging from 4% to 9%.^12,13 Among sexually active teenagers at risk for unintended pregnancy, LARC methods are safe and highly effective,^14,15,16,17 and increasing their availability has been suggested as an important means for preventing rapid repeated pregnancies¹⁸; however, their promotion needs to be balanced with information that, as with all nonbarrier methods, they do not protect users against sexually transmitted infections (STIs).^14,15,16,17 When used consistently and correctly, alone or in combination with another more effective method for pregnancy prevention, condoms can provide protection against STIs, including HIV.^19,20

Despite declining teenage birth rates, reported rates of chlamydia, gonorrhea, and primary and secondary syphilis increased in both male and female teenagers from 2012 to 2016.²¹ Teenagers and young adults, aged 15 to 24 years, account for more than half of new cases of STIs.²² Sexually transmitted infections can result in long-term adverse health outcomes such as infertility, lead to social stigmatization, and have high economic costs associated with treatment.^21,23,24 In studies of postpartum women, risk factors for STIs include younger age, in particular teenagers,²⁵ non-Hispanic black race, being unmarried, having less than a college education, having a low income, and lack of prenatal insurance coverage.²⁶ In 2016, among women who had recently delivered a live birth, rates of gonorrhea and chlamydia were higher among postpartum teenagers compared with mothers 20 years and older, according to birth certificate data.²⁵ Of particular concern, untreated STIs in pregnant women can increase the risk of maternal (eg, premature rupture of the membranes or endometriosis) and infant complications (eg, stillbirth, preterm delivery, and/or low birth weight).²¹ To prevent both unintended pregnancy and STIs, national recommendations from the US Centers for Disease Control and Prevention (CDC) and the Office of Populations Affairs, the American College of Obstetricians and Gynecologists, and the American Academy of Pediatrics recommend all women at risk for STIs, including teenagers, combine use of a more effective contraceptive method, such as LARC, with condoms.^{14,15,16,17,20}

While several studies^{27,28,29,30,31} have examined condom use with a more effective contraceptive method among women of reproductive age, including teenagers, to our knowledge, no study has examined this association among teenage mothers during the postpartum period. Using a multisite, population-based sample from the Pregnancy Risk Assessment Monitoring System (PRAMS) of sexually active US teenage mothers with a recent live birth, this analysis compared the prevalence of condom use between postpartum users of LARC methods vs users of non-LARC hormonal methods. Growing evidence suggests condom use may be lower among LARC users compared with users of other contraceptive methods.^{27,28,29,30,31} Given the increasing prevalence of STIs^1,25 and LARC use⁹ among postpartum teenagers, examining the prevalence of condom use among postpartum teenagers may highlight the importance of and help inform contraceptive counseling messages for this high-risk population.

Methods

Data Source

We used PRAMS 2012 to 2015 data from 37 sites for this analysis. The PRAMS is a mixed-mode (mail and telephone), site-specific and population-based surveillance system and collects data on attitudes, behaviors, and experiences before, during, and shortly after pregnancy from mothers with a live birth 2 to 6 months after delivery.

The PRAMS data are weighted for sampling design, noncoverage, and nonresponse to represent each site’s live birth population. We used data that met the established PRAMS response rate threshold of at least 60% for 2012 to 2014 and at least 55% for 2015.³² The overall mean response rate across sites included in this analysis was 64% for 2012 to 2015 (2012: 66%; 2013: 65%; 2014: 65%; and 2015: 62%). The PRAMS methods are described in detail elsewhere.^33,34 The CDC and each site’s institutional review boards reviewed and approved the PRAMS study protocol. A written informed consent was mailed to participants with the PRAMS survey. For women who did not respond via the mail survey, informed consent was completed orally during the telephone call follow-up.

Measures

Postpartum contraceptive use and condom use were assessed using the following PRAMS question: “What kind of birth control are you or your husband or partner using now to keep from getting pregnant?” Response options included: “Tubes tied or blocked (female sterilization, Essure, Adiana)”; “Vasectomy (male sterilization)”; “Birth control pill”; “Condoms”; “Injection (Depo-Provera)”; “Contraceptive implant (Implanon)”; “Contraceptive patch (OrthoEvra) or vaginal ring (NuvaRing)”; “IUD (including Mirena or ParaGard)”; “Natural family planning (including rhythm method)”; “Withdrawal (pulling out)”; “Not having sex (abstinence)”; “Other → Please tell us.” The PRAMS respondents were instructed to select all contraceptive methods that they were currently using. “Other” responses were reviewed and recategorized if they fit into 1 of the listed contraceptive methods.

Our analysis focused on those 19 years or younger who were using either a LARC (ie, contraceptive implants or IUDs) or non-LARC hormonal method (ie, birth control pills [pill], contraceptive patch [patch], vaginal ring [ring], or injection). After excluding those 20 years or older; those who were abstinent (ie, not having sex); those who reported using no contraception or only a method other than a LARC or non-LARC hormonal method; those who reported using both a LARC and a non-LARC hormonal method; or those who were missing data on contraception type or covariates, our final sample included 5480 (weighted N = 245 847) sexually active postpartum teenage mothers who were using either LARC or non-LARC hormonal methods.

Statistical Analyses

Among sexually active teenage mothers, we performed descriptive statistics (population percentage estimate and their 95% confidence intervals) to assess the distribution of selected maternal characteristics overall, separately for LARC and non-LARC hormonal method users. We also examined the distribution of maternal characteristics for condom users and nonusers. Maternal characteristics were selected a priori that have been found to be associated with using condoms with a more effective contraceptive method in previous studies^27,29,30,31 and were available in the PRAMS data set. These included maternal age (≤17 years vs aged 18-19 years), race/ethnicity (non-Hispanic white vs non-Hispanic black, Hispanic, and non-Hispanic other), insurance coverage (private vs Medicaid and none) at time of survey completion, marital status (married vs unmarried), parity (first birth vs second or later birth), and pregnancy intention (intended vs unintended and unsure).

We constructed separate multivariable survey-weighted logistic regression models to examine the association between condom use by each maternal characteristic between LARC and non-LARC hormonal method users. We computed model-based prevalence for all levels of selected maternal characteristics with predicted marginal means and then estimated both the unadjusted prevalence ratios (PRs) and adjusted PRs (aPRs) and their associated 95% confidence intervals for each maternal characteristic.³⁵ Each model was adjusted for all other maternal characteristics because these were specified a priori based on the existing literature.^27,29,30,31 We included all characteristics in the adjusted models to better understand which were associated with condom use, given that to our knowledge this is the first analysis examining condom use with more effective contraceptive methods among teenage mothers. In adjusted models, all variables were thus entered into each model simultaneously. We also compared condom use between LARC and non-LARC hormonal method users and by each method type (IUD; implant; patch, ring, or injection; and pill), adjusting for all previously mentioned maternal characteristics. The model contrasting condom use by each method type was run 3 times to compare IUD users vs 3 different reference groups: (1) implant; (2) patch, ring, or injection; and (3) pill. In regression analyses, a P value less than .05 was considered statistically significant, and the P value was 2-sided. All analyses were conducted with SAS, version 9.3, and SAS Callable SUDAAN, version 11.0 (SAS Institute Inc), using weighted data to account for the complex sampling design of PRAMS.

Results

Among 5480 (weighted N = 245 847) sexually active postpartum teenage mothers using a LARC method or a non-LARC hormonal method, most were aged 18 or 19 years, unmarried, had current Medicaid coverage, were first-time mothers, had reported that their pregnancy was unintended, and almost half were non-Hispanic white (Table 1). Overall, 61.7% reported using a non-LARC hormonal method (pill, 29.4%; patch, ring or injection, 32.3%), and 38.3% used LARC (IUD, 21.1%; implant, 17.2%). Among teenage mothers, 28.8% reported using condoms in conjunction with LARC or non-LARC hormonal methods during the postpartum period. The distribution of maternal characteristics was similar between teenage mothers reporting condom use and those not reporting condom use (Table 2).

Table 1. Maternal Characteristics of Teenage Mothers Reporting Use of LARC Compared With Non-LARC Hormonal Methods, 37 Sites, Pregnancy Risk Assessment Monitoring System, 2012-2015.

Maternal Characteristics	Total No. (N = 5480)a		LARC (n = 2111)a		Non-LARC Hormonal Methods (n = 3369)a
	No.a	% (95% CI)b	No.a	% (95% CI)b	No.a	% (95% CI)b
Age, y
≤17	1628	30.6 (28.3-32.9)	608	29.1 (25.6-32.9)	1020	31.5 (28.6-34.5)
18-19	3852	69.4 (67.1-71.7)	1503	70.9 (67.1-74.4)	2349	68.5 (65.5-71.4)
Race/ethnicity
Non-Hispanic white	2146	45.6 (43.2-48.0)	812	47.6 (43.8-51.5)	1334	44.4 (41.3-47.4)
Non-Hispanic black	1289	19.9 (18.2-21.7)	349	14.4 (12.1-17.2)	940	23.3 (21.0-25.8)
Hispanic	1337	26.9 (24.6-29.2)	636	29.9 (26.5-33.7)	701	24.9 (22.1-28.0)
Non-Hispanic other	708	7.6 (6.5-8.9)	314	8.0 (6.3-10.1)	394	7.4 (6.0-9.0)
Marital status
Married	554	10.2 (8.8-11.8)	233	12.1 (9.5-15.3)	321	9.0 (7.5-10.9)
Unmarried	4926	89.8 (88.2-91.2)	1878	87.9 (84.7-90.5)	3048	91.0 (89.1-92.5)
Insurance
Private	1189	24.1 (22.1-26.3)	492	25.6 (22.3-29.3)	697	23.2 (20.6-26.0)
Medicaid	3589	61.8 (59.4-64.2)	1330	58.5 (54.5-62.3)	2259	63.9 (60.8-66.9)
None	702	14.1 (12.4-15.9)	289	15.9 (13.2-19.1)	413	12.9 (10.8-15.3)
Parity
First birth	4643	85.7 (84.0-87.3)	1747	83.4 (80.3-86.1)	2896	87.1 (85.0-89.0)
Second or later birth	837	14.3 (12.7-16.0)	364	16.6 (13.9-19.7)	473	12.9 (11.0-15.0)
Pregnancy intention for current birthc
Intended	1072	19.8 (17.9-21.8)	394	19.8 (16.8-23.2)	678	19.8 (17.5-22.4)
Unintended	3195	60.7 (58.3-63.0)	1259	60.9 (57.1-64.6)	1936	60.6 (57.5-63.5)
Unsure	1213	19.5 (17.8-21.4)	458	19.3 (16.6-22.4)	755	19.6 (17.4-22.0)

Abbreviation: LARC, long-acting reversible contraception.

^aUnweighted sample size.

^bWeighted prevalence (expressed as a percentage).

^cRespondents were asked “Thinking back to just before you got pregnant with your new baby, how did you feel about becoming pregnant?” Respondents selecting “I wanted to be pregnant sooner” or “I wanted to be pregnant then” were categorized as intended. Those who selected “I wanted to be pregnant later” or “I didn’t want to be pregnant then or at any time in the future” were categorized as unintended. Those who selected “I wasn’t sure what I wanted” were categorized as unsure.

Table 2. Maternal Characteristics Among Teenage Mothers Using LARC or Non-LARC Hormonal Methods by Reported Condom Use, 37 Sites, Pregnancy Risk Assessment Monitoring System, 2012-2015.

Maternal Characteristics	Did Not Use Condoms (n=3882)a		Used Condoms (n=1598)a
	No.a	% (95% CI)b	No.a	% (95% CI)b
Age, y
≤17	1142	31.1 (28.4-33.9)	486	29.3 (25.3-33.6)
18-19	2740	68.9 (66.1-71.6)	1112	70.7 (66.4-74.7)
Race/ethnicity
Non-Hispanic white	1488	43.7 (40.9-46.5)	658	50.4 (45.8-54.9)
Non-Hispanic black	879	20.0 (18.0-22.2)	410	19.7 (16.8-23.1)
Hispanic	1008	28.1 (25.5-30.9)	329	23.7 (19.7-28.3)
Non-Hispanic other	507	8.2 (6.8-9.8)	201	6.2 (4.8-8.0)
Marital status
Married	409	10.6 (9.0-12.6)	145	9.1 (6.6-12.5)
Unmarried	3473	89.4 (87.4-91.0)	1453	90.9 (87.5-93.4)
Insurance
Private	806	22.6 (20.2-25.0)	383	28.1 (23.9-32.6)
Medicaid	2570	62.9 (60.0-65.7)	1019	59.1 (54.4-63.6)
None	506	14.5 (12.5-16.9)	196	12.8 (10.1-16.3)
Parity
First birth	3255	84.5 (82.3-86.5)	1388	88.7 (86.1-90.9)
Second or later birth	627	15.5 (13.5-17.7)	210	11.3 (9.1-13.9)
Pregnancy intention for current birthc
Intended	825	21.5 (19.2-24.0)	247	15.6 (12.7-19.1)
Unintended	2168	57.8 (55.0-60.6)	1027	67.8 (63.5-71.8)
Unsure	889	20.7 (18.6-22.9)	324	16.6 (13.6-20.0)

Abbreviation: LARC, long-acting reversible contraception.

^aUnweighted sample size.

^bWeighted prevalence (expressed as a percentage).

^cRespondents were asked “Thinking back to just before you got pregnant with your new baby, how did you feel about becoming pregnant?” Respondents selecting “I wanted to be pregnant sooner” or “I wanted to be pregnant then” were categorized as intended. Those who selected “I wanted to be pregnant later” or “I didn’t want to be pregnant then or at any time in the future” were categorized as unintended. Those who selected “I wasn’t sure what I wanted” were categorized as unsure.

In adjusted analyses (Table 3), LARC users were half as likely to report use of condoms compared with users of non-LARC hormonal methods (17.8% vs 35.6%; aPR, 0.50; 95% CI, 0.41-0.60). For each maternal characteristic examined, reported condom use was significantly lower among LARC users compared with non-LARC hormonal method users. However, there was no significant interaction between LARC use and any of the maternal characteristics examined. By contraceptive method (Table 4), condom use was lower among those who used an IUD compared with those using an implant (15.1% vs 21.5%, aPR, 0.70; 95% CI, 0.51-0.98) and among those who used a patch, ring, or injection compared with those who used the pill (24.9% vs 47.2%; aPR, 0.53; 95% CI, 0.44-0.63). The IUD users were less likely to report using condoms than both pill (aPR, 0.32; 95% CI, 0.25-0.40) and patch, ring, or injection (aPR, 0.61; 95% CI, 0.47-0.79) users. While implant users were also less likely to report using condoms compared with pill users (aPR, 0.45; 95% CI, 0.35-0.60), condom use did not differ between implant users and patch, ring, or injection users (aPR, 0.86; 95% CI, 0.64-1.15).

Table 3. Adjusted Prevalence Estimates and Prevalence Ratios of Condom Use Among Teenage Mothers by Maternal Characteristics Between LARC and Non-LARC Hormonal Method Users, 37 Sites, Pregnancy Risk Assessment Monitoring System, 2012-2015.

Maternal Characteristic	Condom Use
	LARC (n = 2111)a		Non-LARC Hormonal Methods (n = 3369)a		aPR (95% CI)	P Valuec
	No.a	% (95% CI)b	No.a	% (95% CI)b
Overalld	393	17.8 (15.0-21.1)	1205	35.6 (32.7-38.7)	0.50 (0.41-0.60)	NA
Age, ye
≤17	119	16.6 (11.8-22.9)	367	32.4 (27.4-37.9)	0.51 (0.35-0.74)	.77
18-19	274	18.4 (15.1-22.3)	838	37.1 (33.5-40.7)	0.50 (0.40-0.62)
Race/ethnicitye
Non-Hispanic white	157	19.6 (15.5-24.5)	501	38.6 (34.2-43.1)	0.51 (0.39-0.66)	.92
Non-Hispanic black	77	18.4 (12.5-26.2)	333	32.9 (27.9-38.2)	0.56 (0.37-0.84)
Hispanic	97	15.8 (10.6-23.1)	232	34.2 (27.5-41.5)	0.46 (0.30-0.72)
Non-Hispanic other	62	13.8 (8.7-21.4)	139	30.5 (22.4-39.9)	0.45 (0.27-0.78)
Marital statuse
Married	27	12.7 (5.4-26.8)	118	37.0 (27.9-47.0)	0.34 (0.15-0.80)	.33
Unmarried	366	18.5 (15.5-21.8)	1087	35.5 (32.4-38.7)	0.52 (0.43-0.63)
Insurancee
Private	100	18.7 (13.1-26.0)	283	41.1 (34.6-47.8)	0.46 (0.31-0.66)	.72
Medicaid	251	17.5 (14.4-21.3)	768	33.7 (30.2-37.5)	0.52 (0.42-0.65)
None	42	17.9 (10.0-29.8)	154	34.4 (26.9-42.9)	0.52 (0.29-0.94)
Paritye
First birth	332	18.3 (15.1-22.0)	1056	36.4 (33.3-39.8)	0.50 (0.41-0.62)	.97
Second or later birth	61	14.8 (9.9-21.5)	149	30.6 (23.9-38.2)	0.48 (0.31-0.76)
Pregnancy intention for current birthe,f
Intended	49	8.9 (5.6-13.9)	198	31.3 (25.2-38.1)	0.28 (0.17-0.47)	.10
Unintended	266	21.2 (17.2-25.8)	761	38.9 (35.0-42.9)	0.55 (0.44-0.68)
Unsure	78	16.2 (10.8-23.6)	246	29.9 (24.3-36.2)	0.54 (0.35-0.84)

Abbreviations: aPR, adjusted prevalence ratio; LARC, long-acting reversible contraception; NA, not applicable.

^aUnweighted sample size.

^bAdjusted weighted prevalence (expressed as a percentage).

^cP value based on F test for 2-way interaction between LARC use and the maternal characteristic.

^dAdjusted for all maternal characteristics listed in the table.

^eWe constructed separate multivariable survey–weighted logistic regression models to examine the association between condom use by each maternal characteristic between LARC and non-LARC hormonal method users. Each model was adjusted for all maternal characteristics listed in the table and included an interaction term between the respective maternal characteristic being examined by LARC use to estimate the prevalence ratios of condom use by the maternal characteristic subgroup.

^fRespondents were asked “Thinking back to just before you got pregnant with your new baby, how did you feel about becoming pregnant?” Respondents selecting “I wanted to be pregnant sooner” or “I wanted to be pregnant then” were categorized as intended. Those who selected “I wanted to be pregnant later” or “I didn’t want to be pregnant then or at any time in the future” were categorized as unintended. Those who selected “I wasn’t sure what I wanted” were categorized as unsure.

Table 4. Prevalence Estimates and Prevalence Ratios of Condom Use Among Teenage Mothers by Contraceptive Method, 37 Sites, Pregnancy Risk Assessment Monitoring System, 2012-2015.

Contraceptive Methoda	Condom Use (N = 5480)b
	No.b	% (95% CI)c	PR (95% CI)c	aPR (95% CI)d,e
IUD vs implantf
IUD	203	15.1 (12.2-18.5)	0.73 (0.53-1.03)	0.70 (0.51-0.98)
Implant	190	21.5 (16.5-27.4)	1 [Reference]	1 [Reference]
IUD and implant vs patch, ring, or injection
IUD	203	15.1 (12.2-18.5)	0.63 (0.49-0.81)	0.61 (0.47-0.79)
Implant	190	21.5 (16.5-27.4)	0.85 (0.63-1.15)	0.86 (0.64-1.15)
Patch, ring, or injection	488	24.9 (21.4-28.8)	1 [Reference]	1 [Reference]
IUD, implant, and patch, ring, or injection vs birth control pill
IUD	203	15.1 (12.2-18.5)	0.32 (0.26-0.41)	0.32 (0.25-0.40)
Implant	190	21.5 (16.5-27.4)	0.44 (0.33-0.58)	0.45 (0.35-0.60)
Patch, ring, or injection	488	24.9 (21.4-28.8)	0.52 (0.43-0.62)	0.53 (0.44-0.63)
Birth control pill	717	47.2 (42.8-51.7)	1 [Reference]	1 [Reference]

Abbreviations: aPR, adjusted prevalence ratio; IUD, intrauterine device; LARC, long-acting reversible contraception; PR, prevalence ratio.

^aThe IUD and implant are classified as LARC methods and the patch, ring, or injection and birth control pill are classified as non-LARC hormonal methods.

^bUnweighted sample size.

^cAdjusted weighted prevalence (expressed as a percentage).

^dThe model was run 3 times to compare IUD users vs 3 different reference groups: (1) implant; (2) patch, ring, or injection; and (3) birth control pill.

^eAdjusted for maternal age, race/ethnicity, marital status, insurance coverage, parity, and pregnancy intention.

^fP value based on F test across the 4 contraceptive methods (IUD; implant; patch, ring, or injection; and pill) was less than .05.

Discussion

Despite recommendations from the American College of Obstetricians and Gynecologists and American Academy of Pediatrics for sexually active teenagers to use condoms concurrently with a more effective contraceptive method,^14,15,16,17 our analysis of PRAMS data found only 3 in 10 postpartum teenagers reported use of condoms combined with LARC or non-LARC hormonal methods. Additionally, we found that condom use was 50% lower among LARC users compared with users of non-LARC hormonal methods. This lower prevalence of condom use combined with use of the most effective reversible contraceptive methods has been consistently demonstrated in cross-sectional studies with women of reproductive age, including teenagers.^27,28,30 Our findings extend previous studies by demonstrating that condom use is also lower among sexually active teenage mothers using LARC in the postpartum period.

The variation in condom use by contraceptive type among sexually active postpartum teenagers is notable. In a cross-sectional analysis using CDC’s Youth Risk Behavior Survey (YRBS) of high school students, condom use was approximately 60% lower among teenagers using LARC compared with those using birth control pills.²⁸ Similarly, in our study population of teenage mothers, condom use was nearly 70% lower among IUD users and 55% lower among implant users compared with pill users. However, the YRBS study found no significant difference in condom use between LARC users and Depo-Provera, patch, or ring users.²⁸ This aligns with our findings that implant users had a similar prevalence of condom use compared with patch, ring, or injection users. By contrast, we found that IUD users had a 40% lower prevalence of condom use compared with patch, ring, or injection users. The differences in our findings were likely associated with the fact the YRBS study was unable to distinguish between IUD and implant users when comparing them with patch, ring, or injection users. We were able to examine LARC methods separately and also found that IUD compared with implant users were 30% less likely to use condoms.

The larger association observed in our analysis and the YRBS analysis when comparing LARC users with pill as compared with patch, ring, or injection users is consistent with the higher prevalence of condom use among pill users as compared to other hormonal contraceptive methods users found in previous studies.^5,28,30,31 One possible explanation for a higher prevalence of condom use among pill users is that condoms are being used as a backup method for pregnancy prevention.^31,36 Among all women, particularly teenagers, inconsistent pill use is common,³⁷ and condom use may be encouraged by health care professionals more often when low compliance is anticipated.

Owing to the higher contraceptive failure rate when using a birth control pill, patch, ring, or injection vs a LARC method, it is especially critical to provide counseling on the effectiveness of all contraceptive methods for pregnancy prevention in combination with counseling for condom use for STI prevention.^14,15 Regardless of the contraceptive method selected by sexually active teenage mothers, counseling should be provided on the importance of consistent and correct use of condoms for optimal protection against STIs to all postpartum teenagers. Prenatal and postpartum visits offer a unique opportunity to provide counseling to teenage mothers on methods to prevent both unintended pregnancy and STIs.^38,39 This opportunity is particularly important given the high LARC use among postpartum teenagers⁹ and high incidence of STIs generally observed among teenagers.²¹

Limitations

This study is subject to several limitations. The PRAMS data are based on maternal self-report and may be subject to social desirability or recall bias. Recall bias may have been minimized because the focus of this analysis was on current contraceptive use at the time the PRAMS survey was completed in the postpartum period (typically 2-6 months after delivery). However, it is possible that contraceptive method use may have varied by time of assessment. The PRAMS survey also currently does not routinely collect information on the clinical or self-reported history of STIs, the number of sexual partners during pregnancy or postpartum, relationship characteristics such as partner type, or the frequency or type of sexual encounters, all factors that may influence condom use.⁴⁰ Among postpartum teenagers, engagement in unprotected sexual behaviors (ie, no method use, including no condoms) have been found to be higher among those reporting a regular sexual relationship with their infant’s father.^41,42 The variation in condom use we found among teenage mothers using LARC compared with non-LARC hormonal methods may be confounded by relationship characteristics that were not measured in PRAMS. It is important to note that while we found lower levels of condom use among LARC users, some evidence suggests condom use is already lower at method initiation among those who choose a LARC compared with other methods and remains unchanged after initiation.²⁹ However, PRAMS is cross-sectional and does not collect data on condom use prior to or at the time of initiation of other contraceptive methods or changes in method use over time. In addition, because PRAMS asks specifically about contraceptive use, including condom use, for pregnancy prevention, condom use may not have been reported by those using condoms only for STI prevention. Additionally, it is possible that condom use was inherently lower among LARC users compared with non-LARC hormonal method users because LARC methods are more effective at preventing pregnancy, and the use of condoms as a backup method to prevent pregnancy may be more common with non-LARC hormonal methods. Finally, while we found associations between reported condom use and type of postpartum contraceptive method, it is important to note these findings are not causal.

Conclusions

Our finding that less than 30% of sexually active teenage mothers using LARC or non-LARC hormonal methods also reported using condoms suggests the need for enhanced efforts to increase condom use among teenage mothers. In particular, there is a pronounced need among LARC users, who were 50% less likely to use condoms compared with non-LARC hormonal contraceptive users. These findings can be used to inform clinician counseling that sexually active teenage mothers have low uptake of condom use combined with more effective contraceptive methods and may need additional counseling on the importance of consistent and correct condom use for the prevention of STIs. Regardless of the contraceptive method selected by sexually active teenage mothers, counseling should be provided on the importance of consistent and correct use of condoms for protection against STIs. Improved understanding of what motivates sexually active postpartum teenagers to use condoms with other contraceptive methods could help inform strategies to better address the reproductive health needs of this unique, high-risk population.