Changes in Expectation of Relationship Permanence, Pregnancy Acceptability and Desire, and Contraceptive Use Over Time Among Young Latino/a Women and Men: An Exploratory Analysis

Abstract

Objective:

We longitudinally assess associations between changes in expectation of relationship permanence—a measure that incorporates both relationship duration and commitment—pregnancy desire and acceptability, and highly effective contraception use among young U.S. Latino/as.

Study Design:

We used multivariable logistic regression to analyze associations between changes in relationship factors, pregnancy acceptability and desire, and contraceptive method use over 6 months among 299 U.S. Latino/a women and men (ages 18-34) in relationships in a longitudinal study (retention rate: 32.4%).

Results:

Respondents who found a pregnancy less acceptable at endline than at baseline were more likely to use highly effective contraception at endline (OR 2.97, 95% CI 1.09,8.08). Respondents estimating an increase in relationship permanence more than one standard deviation of the mean were 4.90 (95% CI 1.17,20.55) times more likely to use highly effective contraception at endline, compared to those without a change in estimation of relationship permanence.

Conclusion:

In this longitudinal study of young adult Latino/as in relationships, we found associations between changes in respondents’ perceptions of the acceptability of a pregnancy and expectation of the permanence of their relationship and highly effective contraceptive use. Our findings reiterate the complicated, intertwined links between relationship dynamics, pregnancy orientation, contraception, and time.

Implications:

The significance of decreased pregnancy acceptability (but not desire) for using highly effective contraception suggests the importance of acknowledging pregnancy acceptability in contraceptive counseling, particularly because providers are more likely to direct Latino/a young adults towards methods that may not fulfill their contraceptive desires. Including discussion of young people’s expectations of relationship permanence may also be meaningful in counseling.

1. Introduction

Recent declines in unintended pregnancy in the US have been attributed to increases in overall contraceptive use [1], and in particular, increases in highly effective method use, including intrauterine devices (IUDs) and implants [2]. Research suggests that highly effective method use differs by pregnancy intentions, relationship characteristics, and race/ethnicity [2-5]. However, few studies use longitudinal approaches to understand how relationship contexts, pregnancy intentions, and their connections to contraceptive use change over time or include men’s perspectives [6]. Detailed information on relationship contexts is often not available, and conventional pregnancy intention measurement may be flawed, failing to adequately and accurately capture individuals’ perspectives. Pregnancy acceptability—the degree to which an unexpected pregnancy would be welcomed, manageable, or okay—is an emerging construct that complicates conventional notions of pregnancy intentions and decision-making [7-9]. A qualitative study with young couples found that various factors influenced the acceptability of an unexpected pregnancy, including relational factors, such as happiness with, stability in, and longevity of the relationship [9].

Relationship context, including satisfaction and trust, power, and relationship length, is influential for both pregnancy desires and contraceptive decision-making [3, 5, 10-12]. In general, young women’s conscious desire for pregnancy [11] and openness to pregnancy [13] increases over time as relationships become more serious and committed. Indeed, one study found that women in longer and less casual relationships are more likely to use an effective method of contraception than those in newer or more casual relationships [3]. In general, as relationships lengthen, romantic stress decreases, with couples exhibiting increased love and satisfaction [14]. Furthermore, increased commitment levels, or long-term orientation toward the relationship, can support the development of trust between partners [15]. As couples transition towards more serious relationships, they may forego condom use as a symbol of increased intimacy, as condoms are often equated with casual relationships [5, 6, 10, 16]. Similarly, women’s expectations of their control over sex and contraception change over time with increases in relationship intimacy and commitment; one longitudinal study showed that while expectations of control increase as a relationship moves from casual to a more committed dating relationship, expectations of control over sex and contraception then decrease when women are engaged or married [17].

Disentangling dimensions of relationship context may provide a more nuanced understanding of the ways in which relationships influence pregnancy desire and contraceptive use. Manlove et al. [18] differentiate between aspects of relationship structure (duration, type) and relationship quality (intimacy, commitment, and conflict), noting that each may influence contraceptive use differently in young adults and that longer relationship duration does not necessarily equate to increased commitment and use of contraception when high levels of intimacy and commitment are not also present. Similarly, other research investigating specific aspects of relationship context has shown that increases in relationship quality, particularly commitment and perceived exclusivity, were related to use of highly effective contraception, while relationship structure was not associated with overall contraceptive use [6] or only associated with use of hormonal methods [10]. Qualitative research with men has described a wide range of relationship structures, which are fluid, not only defined by duration, and which have varying influences on pregnancy intentions and contraceptive use [12].

Expectations about relationship permanence can encompass both relationship structure and quality, as permanence may reflect both duration and commitment, which is in turn related to satisfaction and feelings of dependence and persistence [19]. Therefore, we might expect that greater expectations of relationship permanence could equate to increased use of more effective contraception, as expectations for relationship permanence includes not only expectations for duration (which has a limited influence on contraceptive use), but also the expected continuation of commitment and relationship quality (which has more often been linked to use of highly effective contraception). However, to our knowledge, no studies have investigated the concept of expectation of relationship permanence, with its inherent incorporation of both duration and commitment, and contraceptive use.

Some literature has suggested unique traits among Latino/a populations vis-à-vis reproductive behavior. In nationally representative data, Latina women have higher rates of pregnancies classified as unintended compared to white women [1]. Studies have found that Latina women report being happier about an unintended pregnancy than Black or white women [20] and that Latino men were more likely than white and Black men to report being happy about a pregnancy, regardless of intention [21]. Previous research has also shown that Latino/a women and men are especially likely to report shared decision-making for pregnancy prevention, particularly regarding condom use [e.g., 22, 23-25] and that shared decision-making may increase likelihood of using highly effective contraception [25]. Cultural and familial contexts, gender norms, nativity and generational status, religiosity, acculturation, and experiences with healthcare providers have also influenced positive and negative attitudes towards pregnancy and contraceptive use among Latino/as [e.g., 20, 26-28].

Few studies have examined the connections between relationship factors, pregnancy intentions, and contraceptive behaviors specifically among Latino/as. This study adds to the literature base by examining pregnancy acceptability and desire, relationship permanence, and contraceptive use among a diverse sample of Latino/as at various relationship stages. Specifically, we conduct an exploratory analysis utilizing longitudinal data from young Latino/a adults to assess the associations between changes in expectation of relationship permanence and pregnancy desires and acceptability, and use of highly effective contraceptive methods.

2. Material and methods

2.1. Study overview and data collection

We used data from a longitudinal study of young Latino/as in relationships that aimed to understand how relationship context influences pregnancy preferences and contraceptive use. We conducted two surveys through YouGov’s online panel of 1.8 million+ U.S. residents [29]. A sample of 1,442 women and men aged 18-34 who self-identified as Latino/a, Hispanic, or of Spanish origin was recruited from YouGov’s panel. Eligibility requirements for the baseline survey included: U.S. residence; Latino/a, Hispanic, or Spanish origin self-identification; cisgender and heterosexual self-identification; ever having had sex; not being pregnant or trying to become pregnant; and in a relationship. We classified participants as being in a relationship if they were married, had a boyfriend/girlfriend, had a fiancé/fiancée, were in a self-identified “serious relationship,” or had someone they considered a main partner.

Baseline data collection occurred in June-July 2017. In December 2017, all baseline survey participants received an endline survey invitation; 467 participants completed the endline survey, for a retention rate of 32.4%. Participants received YouGov incentive points redeemable for gift cards upon each survey completion [29]. The Committee for the Protection of Human Subjects at the University of California, Berkeley approved the study protocol.

2.2. Sample

We included 467 participants who completed both surveys. As we were interested in contraceptive use, we excluded people who were pregnant or had a pregnant partner at endline (n=11), were sterilized or infertile at baseline (n=15), or had not had sex in the last 30 days at baseline (n=34) or endline (n=18). We excluded people who reported a different gender (n=6), history of pregnancy (n=15), or parity (n=15)¹ at endline than baseline. We also excluded people who reported a different partner at endline (n=7), since a change in partner could influence feelings about the relationship’s permanence, desire for pregnancy, and contraceptive use. We excluded respondents missing covariate data (n=47) for a final unweighted sample size of 299.

2.3. Measures

The binary outcome variable captured respondents’ use of a highly effective method of contraception (oral contraceptive pill, patch, ring, implant, IUD, injectable contraception, sterilization, or vasectomy) at endline compared to a less effective method (natural/calendar/rhythm, withdrawal, condoms) or no method.

Table 1 shows the independent variables of interest and their transformations into variables reflecting change from baseline to endline. We included measures of pregnancy desire and acceptability, with response options ordered from positive to negative and negative answers scoring the highest; the development of these measures is described elsewhere [9, 30]. We refer to these together as “pregnancy orientation” to differentiate the concept of intertwined pregnancy desire and acceptability as distinct from the traditional definition and use of the term “pregnancy intentions,” which implies a planning paradigm that may not always be relevant [8]. To create change variables, we assessed whether the endline value was lower than, the same as, or higher than the baseline variable. We also included change in relationship type between baseline and endline, as relationship status is commonly linked to pregnancy attitudes and contraceptive use [3, 4, 31]; categories included no change, in a nonserious relationship (boyfriend/girlfriend, casually dating, hooking-up, or “talking to” a main partner); no change, in a serious relationship (married, engaged, or in a self-described serious relationship); change from serious to nonserious.²

Table 1.

Construction of variables used in regression analyses: pregnancy desire and acceptability measured at baseline and endline and variables measuring change in relationship

Variable	Survey question and response categories	Variable measuring change	Categories
Pregnancy desire	Which of the following best describes your desires about becoming pregnant/getting a partner pregnant right now?  1. I want to get pregnant/get a partner pregnant right now.  2. I might want to get pregnant/get a partner pregnant right now.  3. I am okay either way, whether I get pregnant/get a partner pregnant or not right now.  4. I go back and forth about wanting to get pregnant/get a partner pregnant or not right now.  5. I don’t know if I want to get pregnant/get a partner pregnant right now.  6. I do not want to get pregnant/get a partner pregnant right now.	Change in pregnancy desire	Decrease No change Increase
Pregnancy acceptability	How okay would it be if you got pregnant/got a partner pregnant right now?  1. Very okay  2. Somewhat okay  3. A little okay  4. Not at all okay	Change in pregnancy acceptability	Decrease No change Increase
Relationship type	Which of the following best describes your current relationship status? Nonserious relationship:  I'm "talking to" someone, but we're not officially dating.a  I'm "talking to" more than one person.a  I'm hooking-up with someone.a  I'm hooking-up with more than one person.a  I'm casually dating.a  I'm casually dating more than one person.a  I have a boyfriend or girlfriend. Serious relationship:  I have a fiancé(e).  I'm in a serious relationship with someone.  I'm married.	Change in relationship type	No change, in a serious relationship No change, in a nonserious relationship Change from serious to nonserious relationship
Chances of relationship permanence	Please give us a number from 0 to 100, where 0 means that you think there is absolutely no chance that the relationship will be permanent, and 100 means that you think it is absolutely sure to happen that the relationship will be permanent.  1-100	Change in chances of relationship permanence	More than 1 SD decrease Within 1 SD decrease No change Within 1 SD increase More than 1 SD increase

^aThese respondents were also asked: Do you have someone who you consider to be a main partner? A main partner is defined here as: a person whom you are committed to, above anybody else. Respondents without a main partner were ineligible for the survey.

We initially assessed a number of variables capturing changes in relationship characteristics, including satisfaction [19], commitment [19], quality of alternative partners [19], investment in the relationship [19], financial impact if the relationship ended [32], and trust [33]. However, to permit appropriate analysis with a small sample size, we included only variables associated with highly effective contraceptive method use in bivariate analyses. Therefore, the only relationship variable included was change in respondents’ expectation of the chances of the relationship being permanent. Based on a question in the Relationship Dynamics and Social Life Study [34], respondents were asked, “Please give us a number from 0 to 100, where 0 means that you think there is absolutely no chance that the relationship will be permanent, and 100 means that you think it is absolutely sure to happen that the relationship will be permanent.” We first created a continuous variable measuring the difference between the endline and baseline numbers; then, using the standard deviation (SD) (20.58) of the mean change (0.76) of this variable, we created a categorical change variable showing the relative change in estimation of relationship permanence.

We also included age, gender, household yearly income, education, pregnancy history, and parental status in multivariable analyses.

2.4. Analysis

As the majority of the sample was lost to follow-up, we assessed nonresponse bias between the baseline and endline samples. We found some differences in baseline demographic characteristics between respondents who completed only the baseline survey and those completing both surveys. While more men (62.2%) than women (37.8%) completed the baseline survey, more women (64.7%) than men (35.3%) completed both waves (p<0.001). A slightly larger proportion of respondents who completed both waves were in a serious relationship at baseline (75.6%) compared to those lost to follow-up (64.7%; p<0.001). Respondents who completed both waves were also slightly older (p=0.000) and more educated (p<0.01) at baseline than those lost to follow-up. There were no significant differences in baseline pregnancy desire, acceptability, chances of relationship permanence, or type of contraception used between those retained and lost to follow-up.

We first estimated unadjusted multivariable logistic regression models and then adjusted for sociodemographic and relationship variables. Data were weighted to account for differential panel recruitment and study-specific design. We conducted analyses in Stata statistical software (version 15) using svy commands to account for complex survey design.

3. Results

3.1. Descriptive analyses

At baseline, most respondents were ages 25-29 (32.9%) and 30-34 (38.3%) (Table 2). More than half had children (55.2%) and had ever been pregnant or gotten someone pregnant (58.1%). Between baseline and endline, most (73.5%) respondents did not have a relationship status change (Table 3). 47.2% used a less effective method or no method and 37.0% a highly effective method at both timepoints. 7.0% changed from less effective/no method to highly effective methods; 8.8% switched from using a highly effective method to less effective/no methods. Pregnancy desire was stable for 60.8% of respondents. At endline, 21.9% had more pregnancy desire, and 17.3% had less. Similarly, pregnancy acceptability was the same for 57.4% of respondents at baseline and endline. Pregnancy acceptability decreased for 23.1% and increased for 19.5% of respondents. At endline, 39.0% estimated that their relationship had the same chance of permanence as at baseline. Most participants had a slight change (within one SD from the mean change) in how permanently they viewed their relationship; at endline, 23.2% felt their relationship was slightly less permanent, and 19.6% felt it was slightly more permanent.

Table 2.

Sociodemographic characteristics, pregnancy desire and acceptability, and use of contraception in the sample at baseline (N=299)

	Weighted %	Unweighted n
Gender
Woman	43.8	186
Man	56.2	113
Age
15-19	9.6	12
20-24	19.3	51
25-29	32.9	129
30-34	38.3	107
Educational attainment
High school or less	48.5	62
Some college	39.0	123
Bachelor's degree or higher	12.6	114
Household yearly income
under $20k	13.8	46
$20,000-39,999	34.3	84
$40,000-59,999	25.6	71
$60,000-99,999	17.5	62
$100,000 or more	8.9	36
In a serious relationship	70.8	237
Has children	55.2	154
Ever been pregnant/gotten someone pregnant	58.1	160
Using a highly effective contraceptive method	45.7	148
Pregnancy desire
I want to get pregnant right now.	2.2	6
I might want to get pregnant right now.	6.2	9
I am okay either way, whether I get pregnant or not right now.	18.4	52
I go back and forth about whether I want to get pregnant or not right now.	6.0	16
I don't know if I want to get pregnant right now.	6.3	16
I do not want to get pregnant right now.	60.9	200
Pregnancy acceptability
Very okay	17.5	36
Somewhat okay	13.4	45
A little okay	16.1	60
Not at all okay	53.0	158

Note: Highly effective methods = oral contraceptive pill, patch, ring, implant, IUD, injectable, permanent. Less effective methods = natural/calendar/rhythm, withdrawal, condoms.

Table 3.

Use of contraception at endline and changes in pregnancy desire and acceptability, contraceptive use, and relationship characteristics from baseline to endline (N=299)

	Weighted %	Unweighted n
Using a highly effective contraceptive method at endline	44.0	145
Change in contraceptive method
No change, using a less effective method or no method at both timepoints	47.2	129
Change from less effective or no method to highly effective method	7.0	22
Change from highly effective method to less effective or no method	8.8	25
No change, using a highly effective method at both timepoints	37.0	123
Change in pregnancy desire
Decrease	17.3	39
No change	60.8	196
Increase	21.9	64
Change in pregnancy acceptability
Decrease	23.1	55
No change	57.4	181
Increase	19.5	63
Change in relationship type
No change, in a nonserious relationship	29.2	62
No change, in a serious relationship	44.4	149
Change from serious to nonserious relationship	26.5	88
Change in chances of relationship permanence
No change	39.0	129
More than 1 SD decrease	6.5	15
Within 1 SD decrease	23.2	69
Within 1 SD increase	19.6	61
More than 1 SD increase	11.8	25

Note: Highly effective methods = oral contraceptive pill, patch, ring, implant, IUD, injectable, permanent. Less effective methods = natural/calendar/rhythm, withdrawal, condoms.

3.2. Multivariable regression analyses

In the multivariable model, odds of highly effective contraceptive method use at endline varied significantly by changes in pregnancy acceptability and perceived relationship permanence (Table 4). Respondents for whom pregnancy acceptability decreased were more likely to be using a highly effective method (OR 2.97; 95% CI 1.09,8.08) compared to those with no change in acceptability. Respondents reporting increased estimation of relationship permanence were more likely to be using highly effective methods, compared to those without a change: respondents with an increase of more than one SD were 4.90 (95% CI 1.17,20.55) times more likely to be using highly effective contraception. Education was also associated with likelihood of using highly effective contraception: respondents with some college education (OR 2.72; 95% CI 1.25,5.91) and a Bachelor’s degree or higher (OR 2.47; 95% CI 1.03,5.95) were more likely to be using highly effective contraception.

Table 4.

Multivariable logistic regression models for change in pregnancy desire, acceptability, and relationship and endline use of highly effective contraceptive methods (N=299)

	Model 1 (unadjusted)		Model 2 (adjusted)
	OR	95% CI	OR	95% CI
Change in pregnancy desire (ref.: no change)
Decrease	0.38	0.09,1.66	0.35	0.11,1.09
Increase	0.85	0.34,2.13	1.03	0.44,2.41
Change in pregnancy acceptability (ref.: no change)
Decrease	1.99	0.61,6.46	2.97	1.09,8.08
Increase	0.86	0.33,2.23	0.67	0.27,1.64
Change in relationship type (ref.: no change, in a nonserious relationship)
No change, in a serious relationship			2.21	0.77,6.33
Change from serious to nonserious relationship			1.17	0.40,3.37
Change in chances of relationship permanence (ref.: no change)
More than 1 SD decrease			1.73	0.48,6.24
Within 1 SD decrease			1.28	0.57,2.84
Within 1 SD increase			2.3	0.90,5.88
More than 1 SD increase			4.9	1.17,20.55
Educational attainment (ref.: high school or less)
Some college			2.72	1.25,5.91
Bachelor's degree or higher			2.47	1.03,5.95
Household yearly income (ref.: under $20k)
$20,000-39,999			0.61	0.24,1.59
$40,000-59,999			0.67	0.23,1.95
$60,000-99,999			0.61	0.18,2.03
$100,000 or more			0.47	0.11,2.10
Has children (ref.: no)			2.47	0.65,9.41
Age (ref.: 15-19)
20-24			1.25	0.27,5.74
25-29			0.93	0.22,3.92
30-34			0.73	0.16,3.30
Ever been pregnant/gotten someone pregnant (ref.: no)			0.31	0.07,1.30
Gender (ref.: woman)
Man			1.02	0.51,2.03

Model 2 (adjusted model) adds all relationship and sociodemographic variables shown

4. Discussion

In this longitudinal study of young adult Latino/as in relationships, we found associations between changes in respondents’ pregnancy acceptability and estimation of the permanence of their relationship, and highly effective contraceptive use. Our findings could be seen as somewhat contradictory. On one hand, respondents whose pregnancy acceptability decreased at endline were more likely to use highly effective contraception. On the other, participants who perceived themselves as moving towards a more serious relationship also were more likely to use highly effective methods at endline. Yet, both are logical reactions to situations in which young adults may find themselves: when pregnancy is unacceptable, individuals may consequently seek out highly effective contraception to ensure they do not become pregnant, but as their expectations of relationship permanence increase, there may be an increase in the intimacy and communication skills necessary for decision-making about contraceptive methods that can be seen as more invasive or longer-term [35].

Literature documents similar relationships between strong avoidance of pregnancy and consistent contraceptive use [e.g., 13, 31]. Yet studies focusing on pregnancy ambivalence (researchers’ perceptions of inconsistencies in pregnancy intentions, feelings about pregnancy, and contraceptive use) have not always found such associations [7, 36]. However, some studies have found that pregnancy ambivalence may actually capture pregnancies that would be acceptable, even if not actively desired—and contraceptive use would therefore be less important [8, 30, 37]. Our finding that changes in pregnancy acceptability but not pregnancy desire was related to use of highly effective contraception lends weight to the importance of the construct of acceptability in public health and clinical practice [7-9]. For Latino/a young adults, who are more likely to receive contraceptive counseling directing them towards longer-term and permanent methods that may not fulfill their contraceptive desires [38, 39], this is an important distinction. Pregnancy acceptability may be a more nuanced way of understanding patients’ contraceptive desires and needs, particularly for groups whose desires have historically been ignored or devalued, or who may feel societal pressure to accept a pregnancy after the fact [40].

Our finding that Latino/a young adults’ increased expectation of relationship permanence correlated with increased likelihood of using highly effective contraception reflects extant literature noting that couples tend to reduce use of condoms and switch to more effective methods as relationships increase in length and seriousness [3, 16]. Women’s perceptions about highly effective methods (e.g., IUDs are invasive and require long-term commitment) may cause them to equate these methods with longer-term relationships [35]; discussing IUD insertion with a partner requires intimacy and frank discussion about the future that a casual relationship may not support. In the sense that the expectations of relationship permanence reflects both actual duration and perceived longevity of the relationship, as well as commitment, our study is consistent with existing research that acknowledges relationship duration as important but emphasizes more nuanced indications of relationship nature and quality [3, 6, 10, 12, 18]. However, outside of intimate partner violence assessment and basic discussion of partner preferences and barriers in contraceptive use, relationship context is rarely addressed in an indepth way during contraceptive and pregnancy counseling or in contraceptive decision-making apps [17, 41, 42]. Previous studies’ reports of high levels of shared pregnancy prevention decision-making in Latino/a couples note that relationship duration, commitment, and communication were associated with effective contraceptive use [22-25]. Incorporating increased description and discussion of clients’ relationship contexts and expectations in contraceptive counseling conversations may provide an alternate entry point for meaningful discussion about pregnancy and contraceptive choices between providers and clients and perhaps provide another mechanism to increase clients’ satisfaction with both counseling and method choice. Notably, gender was not significantly associated with contraceptive use; though further research with larger samples disaggregated by gender may yield additional insights, this study joins previous research in indicating that inclusion and discussion of contraceptive communication and joint decision-making may be particularly valued for both women and men [23].

This study has several limitations, primarily due to sample size. The retention rate from baseline to endline (32.4%) likely compromised the sample. Other longitudinal internet surveys covering contraception have had retention rates ranging from 22% to 84% [11, 31, 43], depending on the length of time between survey waves and extent of follow-up with participants. While we used the standard YouGov panel incentives (including a larger incentive for the second survey) and recontact methods (a series of email survey invitations), we were not able to utilize approaches that could have decreased attrition, including offering larger incentives, prepaid incentives, or differential incentives to individuals lost to follow-up [44] or utilizing heterogenous recontact methods that could allow for more intensive contact to minimize attrition [45]. At approximately 25 minutes, the survey was fairly long for a YouGov survey; it is possible that panel members who took the baseline survey decided that the incentive was not sufficient enough to take the endline survey. Additionally, the December timing of the endline survey may have influenced participants’ ability and willingness to complete the endline survey. Comparison of demographics, pregnancy, and contraceptive use variables between baseline-only and endline survey respondents showed variation in several baseline demographic variables (gender, age, education, relationship seriousness), suggesting some possible confounding. It is possible that baseline survey respondents who were in more serious, stable, and longer-term relationships were more likely to take the endline survey due to particular interest in the relational topic. The older age and more serious relationships of the respondents who completed both survey waves could have had a confounding effect, as relationships typically become more stable and committed after adolescence and throughout early adulthood [46, 47] and women’s use of long-acting and permanent methods of contraception generally increases with age, relationship length, and relationship quality [6, 10, 18, 48].

The sample size restricted the analytic methods we could use (i.e., we could not use repeated measures analysis), and likely contributed to several slightly wide CIs. However, as this was an exploratory analysis, estimates should be interpreted with caution. Small cell sizes also restricted our ability to construct an outcome variable with more than two categories, so we could not capture change in contraceptive use; over one-third of respondents were using highly effective contraception at both timepoints. Further research with larger sample sizes could ameliorate this and allow for more advanced longitudinal analytic techniques. Additionally, we advise caution in the interpretation of the variable measuring change in chances of relationship permanence. Questions requiring numeracy and estimation of percent can be tricky for survey-takers; this may inhibit the reliability of the change in chances of relationship permanence variable. While we and others acknowledge the importance of multi-dimensional measurement of pregnancy intentions [e.g., 49, 50], we use the single-item measures of pregnancy desire and acceptability in this study in recognition that the constructs are distinct (particularly acceptability)[8] and should be uniquely highlighted. One of our intents when originally developing these indicators was to create short, easy-to-use measures that captured the ideas around acceptability that we saw in the young couples that we interviewed but which were not reflected in existing pregnancy intentions measures [9]. However, single-item measures likely miss nuanced aspects of these constructs. Finally, we collected data only in English, restricting generalizability to the Latino/a population. Due to the study’s restrictive eligibility criteria, the Spanish-speaking sample size would have been too small for analysis.

Despite the limitations, we see several strengths in this exploratory study. Much of the previous research on Latino/as relationships and contraceptive use focuses on use of condoms or whether a method is male- or female-controlled [e.g., 22, 23, 24] as opposed to highly effective. Our study focuses on highly effective contraception and takes into account changes in relationships and the changes in pregnancy orientation that may accompany them among Latino/a young adults. Our findings that both increased expectation of relationship permanence and decreased acceptability of pregnancy were associated with use of highly effective contraception suggest the complicated, intertwined relationships between relationship dynamics, pregnancy orientation, contraception, and time. As people move in and out of relationships, the type and quality of those relationships fluctuate and so does both the acceptability of a future pregnancy and contraceptive use. Increased acknowledgment and incorporation of the roles of alternative and more holistic conceptualizations of pregnancy intentions (e.g., acceptability) and relationship context (e.g., expectations of permanence) is critical for the provision of high-quality contraceptive care for young Latino/as and all patients.