Video Abstract

Video Abstract

Close modal
BACKGROUND:

Although lesbian, gay, bisexual and other sexual minority (LGB+) girls are more likely than heterosexual girls to be pregnant during adolescence, relevant pregnancy prevention programming is lacking.

METHODS:

A national randomized controlled trial was conducted with 948 14- to 18-year-old cisgender LGB+ girls assigned to either Girl2Girl or an attention-matched control group. Participants were recruited on social media between January 2017 and January 2018 and enrolled over the telephone. Between 5 and 10 text messages were sent daily for 7 weeks. Both experimental arms ended with a 1-week booster delivered 12 weeks subsequently.

RESULTS:

A total of 799 (84%) participants completed the intervention end survey. Participants were, on average, 16.1 years of age (SD: 1.2 years). Forty-three percent were minority race; 24% were Hispanic ethnicity. Fifteen percent lived in a rural area and 29% came from a low-income household. Girl2Girl was associated with significantly higher rates of condom-protected sex (adjusted odds ratio [aOR] = 1.48, P < .001), current use of birth control other than condoms (aOR = 1.60, P = .02), and intentions to use birth control among those not currently on birth control (aOR = 1.93, P = .001). Differences in pregnancy were clinically but not statistically significant (aOR = 0.43, P = .23). Abstinence (aOR = 0.82, P = .34), intentions to be abstinent (aOR = 0.95, P = .77), and intentions to use condoms (aOR = 1.09, P = .59) were similar by study arm.

CONCLUSIONS:

Girl2Girl appears to be associated with increases in pregnancy preventive behaviors for LGB+ girls, at least in the short-term. Comprehensive text messaging–based interventions could be used more widely to promote adolescent sexual health behaviors across the United States.

What’s Known on This Subject:

Although sexual minority girls are more likely than heterosexual girls to be pregnant as a teenagers, pregnancy prevention programs tailored to the lived experiences of these girls are lacking.

What This Study Adds:

Girl2Girl positively impacts condom use and uptake of other types of contraception for sexual minority girls, at least in the short-term. Findings further suggest that intensive programming delivered over a long period is acceptable to sexual minority girls.

Although adolescent pregnancy rates in the United States have decreased over the past 40 years, just under 200 000 girls, 15 to 19 years old, became pregnant in 2017.1  Cisgender sexual minority girls, who comprise an estimated 10% to 15% of female youth,2  face differential risk: It may seem counterintuitive given assumptions that cisgender lesbian, gay, bisexual, and other sexual minority (LGB+) girls are only having sex with other girls, but researchers consistently find that they are significantly more likely to get pregnant than heterosexual girls.35  Cisgender LGB+ girls also are more likely than their heterosexual female peers to have sex at a younger age,2,4,6,7  report more lifetime and recent numbers of female or male sexual partners,2,4,6,8  and are less likely to use barrier methods to prevent sexually transmitted infections or pregnancy.2,7,9 

Access to school-based sex education varies widely across the United States10 : 19 states require an abstinence-only perspective and 39 require that abstinence is presented as an option; only 20 states require contraception to be included in the sexual education curriculum. Despite the heightened risks of pregnancy and sexually transmitted infections that cisgender LGB+ girls face, sex education that addresses these disparities is often scarce or inadequate.11,12  Indeed, although 11 states and the District of Columbia require LGBT+ inclusive sex education, an additional 7 require that homosexuality be presented negatively and/or that heterosexuality be presented as the only acceptable sexual identity.

To address the lack of sexual health programming for many sexual minority adolescents, we developed and tested Girl2Girl, a text messaging-based pregnancy prevention program tailored to the unique needs of sexually experienced and inexperienced cisgender LGB+ adolescent girls. In this study, we report the preliminary outcomes of the randomized controlled trial (RCT) at intervention end, 5 months after program enrollment. Information about the iterative, user-centered intervention development, which included 269 cisgender LGB+ girls from across the country, is available elsewhere.13  Findings from the current investigation will contribute to the growing literature examining behavioral outcomes associated with text messaging-delivered behavior change content, particularly that which targets complex behaviors, such as pregnancy prevention.

Girls were recruited and enrolled between January 23, 2017, and January 12, 2018. After completing the baseline survey, they were randomized and began the 5-month program. Most participants completed the intervention or control program and their survey by May 31, 2018.

Advarra Institutional Review Board, an Office of Human Research Protections–approved institutional review board, reviewed and approved the research protocol. We were granted a waiver of parental permission so that girls would not have to put themselves in a potentially unsafe situation by disclosing their sexual identity to their parents to obtain permission to participate in the study. The waiver also prevented sampling bias that would have occurred by including only girls who were out to their parents.

Girls were recruited across the United States. To be eligible for the study, participants were (1) aged 14 to 18 years, (2) cisgender female (ie, assigned female sex at birth and endorsed a female gender identity), (3) sexual minority (eg, lesbian, bisexual), (4) in high school or the equivalent (including those who did not finish or dropped out), (5) English speaking, (6) exclusive users of a cell phone (ie, they do not share the phone with someone else) with an unlimited text messaging plan, (7) users of text messaging for at least 6 months, (8) intending to have the same cell number for the next year, and (9) able to provide informed assent for those <18 years of age and consent for those 18 years of age, including an acceptable score on the “capacity to consent” and self-safety assessment.14  Exclusion criteria included knowing another girl enrolled in the RCT and having participated in a previous intervention development activity.

Girl2Girl is a 20-week, text messaging–based teenage pregnancy prevention program. For the first 7 weeks, participants are sent between 4 and 12 messages per day. They then enter a 12 week “latency” period when they receive ∼1 to 2 messages per week. Finally, participants receive 4 to 12 messages daily for a review week. This “booster” delivers messages that reiterate the main concepts discussed in the first 7 weeks of the program.

Intervention content is guided by the information, motivation, behavioral skills model15,16  and focuses on pregnancy prevention information (eg, how one gets pregnant), motivations (eg, reasons to initiate birth control), and behavioral skills (eg, how to use barriers). Additional content describes topics and scenarios that are relevant to sexual decision-making for sexual minority girls (eg, aspects of a healthy relationships).17,18  Content is tailored on the basis of participants’ self-reported sexual experience (ever versus never had sex) and sexual identity (lesbian or gay versus all others, except for girls who identified as queer; in this case, their sexual attractions determined which content they received). Content was not further tailored to the individual.

As described elsewhere,13  most program messages are unidirectional (see Table 1). Bidirectional, interactive components included (1) links to brief online videos in the messages that aimed to reinforce behavioral skills; (2) G2Genie, an on-demand advice text line that provided information about sex, relationships, and the LGBT+ community19,20 ; (3) “leveling up” by answering a text message question correctly about the week’s content; and (4) being awarded “badges,” which was a gif sent over text, for achieving behavioral skills they were presented in the program (eg, getting barriers; talking to a health provider about contraception).

TABLE 1

Example Girl2Girl Program Messages That are Unidirectional and Bidirectional

Example Intervention Test Message
Unidirectional messages by domain
 Teenage pregnancy prevention information I get that you’re into girls, not guys. Here’s the thing: LGB+ girls are 2–4 times MORE likely to get pregnant than hetero girls. Totally unexpected, huh?a 
 About 1 in 20 teen girls are on an intrauterine device (IUD) or an implant. They can be great: They last 3–10 years so you get it put in and forget about it. 
 Being on your period doesn’t protect you because sperm can live in your vagina for up to 5 days. So your period can be over, but sperm are still swimming around. 
 Teenage pregnancy prevention motivation There are a lot of reasons why LGB+ girls are more likely to get pregnant. Some assume they’re never going to have sex that involved a penis so they don't have condoms. 
 Maybe surprisingly, women in committed relationships with women can have high STD rates. In fact, 13% to 30% of women having sex with women have HPV (source: bit.ly/1SMlxH6) 
 A girl told me: “Most people won’t look bad at you for buying condoms or going to a gyno to get birth control. No shame. It’s only natural.” Too right! 
 Some women say the female condom has more sensation than a male condom. And it rubs against your clitoris during sex - bonus! 
 Dental dams/condoms prevent STDs and pregnancy but a backup never hurts - like having a car charger in case your phone goes dead. And that's birth control! 
 Teenage pregnancy prevention behavioral skills It's fast and simple: Unroll the male condom. Cut off the tip of the condom. Cut down the length of the tube. Unfold to a square. Done! 
 Unsure how to ask about birth control? Tell the provider you have questions about it and they’ll take it from there. Don’t worry - they talk to teens every day. 
 No need to wait for a crisis. Go online to see where the nearest store is. Too far to get to? Buy the morning after pill online so you have it if you need it. 
 Socio-cultural factors Dating violence can also happen. 2 in 5 LGB+ girls have been victims. 1 in 3 were violent themselves. It's not ok to use hurtful words or physical force. 
 If you or someone you know is assaulted, *please* get help. RAINN is amazing - they have an online chat (ohl.rainn.org/online) and a hotline: 800-656-4673. 
 It’s common for us to question our sexual attractions and identity. Or for these to change over time. It’s OK to take time to explore and discover who you are. 
Bidirectional messages by domain  
 Level-up Initial message: I'm not sure where the last 2 weeks went, but here we are! This is for Level 2. True or False: You need your parents' permission for the morning after pill. 
 Correct response from participant: Yessss! You are absolutely right. You can get emergency contraception at many drug stores or online and you don't need your parent’s ok. Hellllooo Level 2! 
 Incorrect response form participant: Good news! You can get emergency contraception at many drug stores or online and you don’t need your parent’s ok. 
 G2Genie Texting the keyword “condom” to G2Genie: If condoms (aka cut up to be dental dams) are too expensive, go to a clinic and get them for free. The icondom app can help you find free stuff too. The Internet is magical! 
 Badges Initial message: Great! So, what’s your dental dam/condom status at the moment? Do you have one in your possession (like in your bedroom, school bag, etc)? Let me know (yes/no). 
 Participant says yes: Nice! You earned your Go Getter Girl Badge! 
 Participant says no: It might seem really hard to get dental dams/condoms: they’re expensive, they can be embarrassing to buy, and maybe you don’t know where to get them. What stops you: Parents, Cost, IDK where to go, Embarrassed, Not having sex, Not having sex w/a guy, Waiting to have sex, In a relationship, No transport, Other. 
 “No” participant responds she is not having sex with a guy: I totally get that you're not having sex with a guy but practicing now to buy condoms now just makes it easier to them if you need them later on. 
Example Intervention Test Message
Unidirectional messages by domain
 Teenage pregnancy prevention information I get that you’re into girls, not guys. Here’s the thing: LGB+ girls are 2–4 times MORE likely to get pregnant than hetero girls. Totally unexpected, huh?a 
 About 1 in 20 teen girls are on an intrauterine device (IUD) or an implant. They can be great: They last 3–10 years so you get it put in and forget about it. 
 Being on your period doesn’t protect you because sperm can live in your vagina for up to 5 days. So your period can be over, but sperm are still swimming around. 
 Teenage pregnancy prevention motivation There are a lot of reasons why LGB+ girls are more likely to get pregnant. Some assume they’re never going to have sex that involved a penis so they don't have condoms. 
 Maybe surprisingly, women in committed relationships with women can have high STD rates. In fact, 13% to 30% of women having sex with women have HPV (source: bit.ly/1SMlxH6) 
 A girl told me: “Most people won’t look bad at you for buying condoms or going to a gyno to get birth control. No shame. It’s only natural.” Too right! 
 Some women say the female condom has more sensation than a male condom. And it rubs against your clitoris during sex - bonus! 
 Dental dams/condoms prevent STDs and pregnancy but a backup never hurts - like having a car charger in case your phone goes dead. And that's birth control! 
 Teenage pregnancy prevention behavioral skills It's fast and simple: Unroll the male condom. Cut off the tip of the condom. Cut down the length of the tube. Unfold to a square. Done! 
 Unsure how to ask about birth control? Tell the provider you have questions about it and they’ll take it from there. Don’t worry - they talk to teens every day. 
 No need to wait for a crisis. Go online to see where the nearest store is. Too far to get to? Buy the morning after pill online so you have it if you need it. 
 Socio-cultural factors Dating violence can also happen. 2 in 5 LGB+ girls have been victims. 1 in 3 were violent themselves. It's not ok to use hurtful words or physical force. 
 If you or someone you know is assaulted, *please* get help. RAINN is amazing - they have an online chat (ohl.rainn.org/online) and a hotline: 800-656-4673. 
 It’s common for us to question our sexual attractions and identity. Or for these to change over time. It’s OK to take time to explore and discover who you are. 
Bidirectional messages by domain  
 Level-up Initial message: I'm not sure where the last 2 weeks went, but here we are! This is for Level 2. True or False: You need your parents' permission for the morning after pill. 
 Correct response from participant: Yessss! You are absolutely right. You can get emergency contraception at many drug stores or online and you don't need your parent’s ok. Hellllooo Level 2! 
 Incorrect response form participant: Good news! You can get emergency contraception at many drug stores or online and you don’t need your parent’s ok. 
 G2Genie Texting the keyword “condom” to G2Genie: If condoms (aka cut up to be dental dams) are too expensive, go to a clinic and get them for free. The icondom app can help you find free stuff too. The Internet is magical! 
 Badges Initial message: Great! So, what’s your dental dam/condom status at the moment? Do you have one in your possession (like in your bedroom, school bag, etc)? Let me know (yes/no). 
 Participant says yes: Nice! You earned your Go Getter Girl Badge! 
 Participant says no: It might seem really hard to get dental dams/condoms: they’re expensive, they can be embarrassing to buy, and maybe you don’t know where to get them. What stops you: Parents, Cost, IDK where to go, Embarrassed, Not having sex, Not having sex w/a guy, Waiting to have sex, In a relationship, No transport, Other. 
 “No” participant responds she is not having sex with a guy: I totally get that you're not having sex with a guy but practicing now to buy condoms now just makes it easier to them if you need them later on. 
a

This message is written specifically for lesbian or same-sex attracted girls.

Girl2Girl participants also were paired with another participant, their “text buddy.”19,20  As part of the program content, buddies received messages that encouraged them to provide social support to one other and practice skills taught in the program. Intervention participants accepted a Buddy Code of Conduct, which outlined acceptable and unacceptable behavior. Buddy messages were routed through the study server so that participants’ phone numbers were not disclosed to buddies. Buddy conversations were monitored by research staff from 6 am to 10 pm PST to ensure harmful messages were not being sent.

The control arm received a similar intensity and duration of messaging as the intervention. Content addressed topics relevant to adolescents, including diet, exercise, and how to deal with bullying. To help blind this arm, 2 days of pregnancy prevention content readily available online were included. Interactive components (eg, Text Buddy) were not available to this group.

Participants were recruited through online advertisements on Facebook and Instagram. The ads were targeted to users who indicated on their profile that they were female, between 14 and 18 years of age, and “interested in females” or “interested in males and females.” Youth who clicked on the advertisement were directed to the online screener form.

Potentially eligible youth were contacted in order of receipt of screeners while also considering preset demographic targets to ensure sample diversity. For example, once the targeted number of non-Hispanic white girls was enrolled, additional demographically similar candidates were not contacted. Enrollment occurred over the phone with research staff. Once eligibility was confirmed and informed assent and consent were obtained, participants were emailed a unique link to the baseline survey. Girls were randomized to a study arm after they completed the survey.

Randomization

Participants were randomly assigned to the Girl2Girl intervention (n = 473) or the attention-matched control group (n = 475) at a 1-to-1 randomization allocation ratio. The random allocation was stratified by (1) sexual experience and (2) sexual identity.21  Participants, but not researchers, were blind to arm allocation.

Power

Effect size estimates for samples that include both sexually experienced and inexperienced LGB+ youth are ill-defined because, as noted above, pregnancy programs for LGB+ adolescent women are lacking. To inform our power estimates about sexual abstinence, we used data from the authors’ Girl2Girl “recruitment pilot” study. Among the 257 14- to 18-year-old sexual minority girls surveyed, 22% had never had penile-vaginal sex (M.Y., unpublished data). A power estimate for condom use was based on data from the authors’ Teen Health and Technology survey.22  Eleven percent of LGB+ girls 14 to 18 years old reported having penile-vaginal or penile-anal sex without a condom in the past 3 months (M.Y., unpublished data). On the basis of these prevalence estimates and assuming 80% power and statistical significance set at P = .05, the minimum detectable odds ratio we could detect with a sample size of 420 girls in each experimental group (n = 840) were 0.49 for condomless sex and 1.66 for abstinence.2327  Recruitment was more successful than anticipated, with 948 participants recruited and randomized.

Incentive

Participants received between $5 and $35 for completing the survey at intervention end; incentives varied on the basis of the mode and length of the survey they completed (see Data Collection below). Youth were not incentivized to complete the baseline survey.

Data Collection

Baseline surveys were collected online. We initially intended to collect intervention end data via text messaging. When funding uncertainties arose, we shifted to an online data collection format. As such, 299 participants completed the intervention end survey via text message and 500 completed it online. Survey mode was balanced by experimental arm: 62% of control and 63% of intervention participants completed the survey online (P = .71).

Program participation length varied for intervention participants on the basis of their responsiveness to programmatic activities. For example, those who did not respond to a level-up question or badge message were sent multiple reminders before the program content resumed. Thus, some had an intervention experience that was longer than 20 weeks.

Main outcomes included (1) condom-protected penile-vaginal sex in the past 3 months, (2) current use of contraception other than condoms, (3) abstinence from penile-vaginal sex in the past 3 months, and (4) pregnancy since program enrollment. Secondary outcomes included behavioral intentions in the next year to (5) use condoms and (6) other forms of contraception and (7) be abstinent. Outcome measures were adapted from those recommended by the Office of Adolescent Health.28  Because of the 160-character limitation in text messages, questions in the text messaging–based intervention end survey varied from the online version. Wording of outcome measures is shown in Table 2.

TABLE 2

Girl2Girl2 RCT Outcome Measures

MeasureOnline Survey Text (Baseline and Intervention End)Text Messaging Survey Text (Intervention End)Outcomes Computation
Abstinence in the past 3 mo In the past 3 months, how many times have you had sex when a human penis went into your vagina? Since the beginning of the G2G program on [date], how many times have you had sex when a human penis (not a toy) went in your vagina? Those who answered 0 were coded as being abstinent. Those who reported 1 or more times at baseline were coded as being recently sexually active. 
No. condom-protected sex acts in the past 3 mo You said you had sex when a human penis (not a toy) went into your vagina [insert number] in the past 3 months. How many times did you use a female or male condom? And of the [×] times you had sex when a human penis went in your vagina since G2G started, how many times did you use female or male condoms? Responses were treated as a count. 
Currently using birth control Youth were asked 6 separate questions about whether they were currently using any types of the following methods of birth control: (1) birth control pills, (2) the shot (like Depo-Provera), (3) the patch (like Xulane), (4) the ring (like NuvaRing), (5) IUD (like Skyla, Mirena, or Paragard), and implants (like Implanon or Nexplanon). Response options were no, yes, I don’t know what this is, and do not want to answer. Those who initially said they did not know about the type of birth control were directed to a pictorial representation and a more detailed description. Are you on birth control? I mean the pill, shot (Depo-Provera), ring (NuvaRing), patch (Xulane), an IUD (like Paragard), or implant (like Implanon). Current use of birth control was indicated if online survey respondents answered yes to any of the 6 types of birth control queried, or text message survey respondents answered yes to the single item. 
Pregnancy since RCT enrollment As far as you know, have you been pregnant, even if no child was born, ever/since Girl2Girl started on [insert start date]? As far as you know, since Girl2Girl started on [date] have you been pregnant even if no child was born? Those who said yes were coded as having been pregnant. 
Intentions to use condoms in the futurea “If I have sex where a human penis (not a toy) goes into my vagina in the next year, I plan to use or have my partner use a female or male condom.” Response options were on a 5-point Likert scale ranging from very untrue to very true. In the next 3 months, if you have sex when a human penis goes into your vagina, do you plan to use female or male condoms? Def not, Prob not, Prob, Def, or IDK. Those who said “somewhat true” or very true” online or “Prob” or “Def’ via text were coded as having intentions to use condoms. 
Intentions to use birth control other than condoms in the futureb “In the next year, do you plan to use these methods of birth control?” The 6 types of birth control described above. Response options were on a 4-point Liker scale ranging from no, definitely not to yes, definitely. “In the next 3 months, do you plan to use birth control (the pill, shot, ring, patch, IUD, implant)? Def not, Prob not, Prob, Def, IDK, or not sure what this is.” Those who said “Yes, probably” or “Yes, definitely” to the online survey or “Prob” or “Def” via text messaging, and said they were not currently on birth control, were coded as having intentions to use birth control. 
Intentions to be abstinent in the future “In the next year, do you think you might have the following types of sex with a guy, or someone with a penis regardless of their gender presentation?” The specific item was: “Sex with a human penis (not a toy) that goes into your vagina.” Response options were on a 4-point Likert scale ranging from “no, definitely not” to “yes, definitely.” “And in the next 3 months, do you think you might have sex when a human penis (not a toy) goes into your vagina? Def not, Prob not, Prob, Def, IDK.” Those who responded “Yes, probably” or Yes, definitely” online, of “Prob” or “Def” via text were coded as having an intention to be abstinent. 
MeasureOnline Survey Text (Baseline and Intervention End)Text Messaging Survey Text (Intervention End)Outcomes Computation
Abstinence in the past 3 mo In the past 3 months, how many times have you had sex when a human penis went into your vagina? Since the beginning of the G2G program on [date], how many times have you had sex when a human penis (not a toy) went in your vagina? Those who answered 0 were coded as being abstinent. Those who reported 1 or more times at baseline were coded as being recently sexually active. 
No. condom-protected sex acts in the past 3 mo You said you had sex when a human penis (not a toy) went into your vagina [insert number] in the past 3 months. How many times did you use a female or male condom? And of the [×] times you had sex when a human penis went in your vagina since G2G started, how many times did you use female or male condoms? Responses were treated as a count. 
Currently using birth control Youth were asked 6 separate questions about whether they were currently using any types of the following methods of birth control: (1) birth control pills, (2) the shot (like Depo-Provera), (3) the patch (like Xulane), (4) the ring (like NuvaRing), (5) IUD (like Skyla, Mirena, or Paragard), and implants (like Implanon or Nexplanon). Response options were no, yes, I don’t know what this is, and do not want to answer. Those who initially said they did not know about the type of birth control were directed to a pictorial representation and a more detailed description. Are you on birth control? I mean the pill, shot (Depo-Provera), ring (NuvaRing), patch (Xulane), an IUD (like Paragard), or implant (like Implanon). Current use of birth control was indicated if online survey respondents answered yes to any of the 6 types of birth control queried, or text message survey respondents answered yes to the single item. 
Pregnancy since RCT enrollment As far as you know, have you been pregnant, even if no child was born, ever/since Girl2Girl started on [insert start date]? As far as you know, since Girl2Girl started on [date] have you been pregnant even if no child was born? Those who said yes were coded as having been pregnant. 
Intentions to use condoms in the futurea “If I have sex where a human penis (not a toy) goes into my vagina in the next year, I plan to use or have my partner use a female or male condom.” Response options were on a 5-point Likert scale ranging from very untrue to very true. In the next 3 months, if you have sex when a human penis goes into your vagina, do you plan to use female or male condoms? Def not, Prob not, Prob, Def, or IDK. Those who said “somewhat true” or very true” online or “Prob” or “Def’ via text were coded as having intentions to use condoms. 
Intentions to use birth control other than condoms in the futureb “In the next year, do you plan to use these methods of birth control?” The 6 types of birth control described above. Response options were on a 4-point Liker scale ranging from no, definitely not to yes, definitely. “In the next 3 months, do you plan to use birth control (the pill, shot, ring, patch, IUD, implant)? Def not, Prob not, Prob, Def, IDK, or not sure what this is.” Those who said “Yes, probably” or “Yes, definitely” to the online survey or “Prob” or “Def” via text messaging, and said they were not currently on birth control, were coded as having intentions to use birth control. 
Intentions to be abstinent in the future “In the next year, do you think you might have the following types of sex with a guy, or someone with a penis regardless of their gender presentation?” The specific item was: “Sex with a human penis (not a toy) that goes into your vagina.” Response options were on a 4-point Likert scale ranging from “no, definitely not” to “yes, definitely.” “And in the next 3 months, do you think you might have sex when a human penis (not a toy) goes into your vagina? Def not, Prob not, Prob, Def, IDK.” Those who responded “Yes, probably” or Yes, definitely” online, of “Prob” or “Def” via text were coded as having an intention to be abstinent. 

Two hundred and ninety-nine participants completed the intervention end survey via text message and 500, online. Online survey questions were based on those recommended by the Office of Adolescent Health. Messages were adapted by the authors for text messaging. IDK, I don’t know.

a

In the text messaging–based survey, this question was asked of everyone except those who said “definitely not” to the question about intentions to have penile-vaginal sex. It was asked of everyone in the online survey.

b

In the text messaging survey, this question was asked only of those who were not currently on birth control. In the online survey, it was asked of everyone.

Participants who did not complete the intervention end survey were excluded from the analyses. Missing data also occurred within the intervention end survey when participants declined to answer an outcome measure. Intentions to use condoms (5%), intentions to have penile-vaginal sex (4%), and intentions to use birth control (4%) had the highest rates of decline to answer. To analyze a consistent sample across outcome measures, decline to answer was treated as outcome failure (eg, no condom-protected sex; no intention to use condoms). As a sensitivity analyses, models also were estimated with these youth eliminated from the analyses.

Analyses were intention to treat; gender identity was assessed during the enrollment call. A small number of youth (n = 10) self-reported a cisgender identity over the telephone and a noncisgender identity in the baseline survey. Because these answers were not again assessed for eligibility, these youth were included in the study sample. One of these participants’ baseline surveys was deleted. Eight completed both the baseline and follow-up surveys and are included in analyses.

Logistic regression was used to quantify the relative odds of dichotomous measures: contraception, abstinence, and behavioral intentions. Poisson regression was used to quantify the relative count of condom-protected sex acts.

Differences in behaviors at baseline were likely equally distributed between treatment and control groups because arm assignment was random. Because statistically significant differences in these characteristics might occur by chance, any baseline characteristics on which the experimental arms were imbalanced were included in multivariate models. Multivariate models also adjusted for the baseline indicator of the outcome in question (eg, condom-protected sex) and the survey mode through which the intervention end survey was completed (ie, online versus text messaging).

Analyses were performed twice: once with all youth and once for youth who reported penile-vaginal sex in the 90 days before baseline, hereafter referred to as “sexually active girls.” Girls who had sex for the first time during the observation period were not included in the latter group. The former provided an estimate of the intervention effect in the target population as a whole, and the latter provided an estimate among those at greater risk for pregnancy. We also examined behavioral intentions among youth who had not had penile-vaginal sex in the past 3 months at baseline (ie, were “not sexually active”).

As shown in Fig 1, 59 of 948 participants either actively terminated their involvement in the RCT or were lost to follow-up. The remaining 94% received all program messages, thereby completing the program. Seven hundred and ninety-nine (84%) completed the intervention end survey and are included in the current analyses. No unintended harms were reported.

FIGURE 1

CONSORT Diagram for Girl2Girl RCT.

FIGURE 1

CONSORT Diagram for Girl2Girl RCT.

Close modal

Intervention and control participants were equivalent on their baseline characteristics except for age, which was borderline statistically significantly different (P = .051; see Table 3).

TABLE 3

Comparison of Baseline Characteristics of Girl2Girl RCT Participants Among Those Who Completed the Intervention End Survey (n = 799)

Youth CharacteristicsControl (n = 410)Intervention (n = 389)P
Age, mean (SD) 15.97 (1.2) 16.14 (1.2) .051 
Hispanic ethnicity, n (%)   .25 
 No 300 (73.2) 302 (77.6)  
 Yes 109 (26.6) 85 (21.9)  
 Decline to answer 1 (0.2) 2 (0.5)  
Race, n (%)   .47 
 White 226 (55.1) 232 (59.6)  
 Black or African American 61 (14.9) 57 (14.7)  
Asian American 16 (3.9) 14 (3.6)  
 Native Hawaiian or Other Pacific Islander 3 (0.7) 2 (0.5)  
 American Indian or Alaska native 9 (2.2) 3 (0.8)  
 Multiracial 58 (14.2) 54 (13.9)  
 Some other race 26 (6.3) 23 (5.9)  
 Do not want to answer 11 (2.7) 4 (1.0)  
Rural,an (%) 70 (17.1) 51 (13.1) .12 
Income,bn (%)   .86 
 Lower than the average 122 (29.8) 110 (28.3)  
 Similar to the average 213 (52.0) 200 (51.4)  
 Higher than the average 61 (14.9) 62 (15.9)  
 Do not want to answer 14 (3.4) 17 (4.4)  
Sexual identity,cn (%)    
 Gay 83 (20.2) 78 (20.1) .95 
 Lesbian 178 (43.4) 169 (43.4) .99 
 Bisexual 172 (42.0) 170 (43.7) .62 
 Pansexual 106 (25.9) 121 (31.1) .10 
 Heterosexual 3 (0.7) 2 (0.5) .70 
 Queer 89 (21.7) 77 (19.8) .51 
 Asexual 9 (2.2) 8 (2.1) .89 
 Questioning 50 (12.2) 43 (11.1) .62 
 Unsure 7 (1.7) 8 (2.1) .72 
 Do not want to answer 0 (0) (0) 0 — 
Gender identity, n (%)   .55 
 Cisgender female 406 (99.0) 385 (99.0)  
 Male to female transgenderd 1 (0.2) 0 (0.0)  
 Male to female transgenderd 0 (0.0) 1 (0.3)  
 Genderqueer or pangenderd 0 (0.0) 1 (0.3)  
 I am unsured 2 (0.5) 2 (0.5)  
 Other 1 (0.2) 0 (0.0)  
 Do not want to answer 0 (0) 0 (0)  
Ever penile-vaginal sex, n (%)   .99 
 No 279 (68.1) 263 (67.6)  
 Yes 130 (31.7) 125 (32.1)  
 Do not want to answer 1 (0.2) 1 (0.3)  
Penile-vaginal sex in the past 3 mo, n (%)   .75 
 No 342 (83.4) 319 (82.0)  
 Yes 67 (16.3) 68 (17.5)  
 Do not want to answer 1 (0.2) 2 (0.5)  
No. condom-protected vaginal sex acts in the past 3 mo, mean (SD) 0.8 (5.7) 1.1 (4.9) .42 
 Do not want to answer 0 (0) 0 (0)  
Currently on birth control, n (%)   .45 
 Yes, to any 101 (24.6) 87 (22.4)  
 Birth control pills   .82 
  No 339 (82.7) 324 (83.3)  
  Yes 71 (17.3) 65 (16.7)  
  I don’t know what this is 0 (0) 0 (0)  
  Do not want to answer 0 (0) 0 (0)  
 The shot   .69 
  No 401 (97.8) 382 (98.2)  
  Yes 9 (2.2) 7 (1.8)  
  I don’t know what this is 0 (0) 0 (0)  
  Do not want to answer 0 (0) 0 (0)  
 The patch   .62 
  No 405 (98.8) 385 (99.0)  
  Yes 4 (1.0) 2 (0.5)  
  I don’t know what this is 1 (0.2) 2 (0.5)  
  Do not want to answer 0 (0) 0 (0)  
 The ring   .82 
  No 406 (99.0) 384 (98.7)  
  Yes 3 (0.7) 3 (0.8)  
  I don’t know what this is 1 (0.2) 2 (0.5)  
  Do not want to answer 0 (0) 0 (0)  
 IUD   .15 
  No 400 (97.6) 384 (98.7)  
  Yes 7 (1.7) 1 (0.3)  
  I don’t know what this is 3 (0.7) 3 (0.8)  
  Do not want to answer 0 (0.0) 1 (0.3)  
 Implants   .44 
  No 393 (95.9) 379 (97.4)  
  Yes 16 (3.9) 9 (2.3)  
  I don’t know what this is 1 (0.2) 1 (0.3)  
  Do not want to answer 0 (0) 0 (0)  
Ever pregnant, n (%)   .50 
 No 388 (94.6) 369 (94.9)  
 Yes 15 (3.7) 13 (3.3)  
 I don’t know 5 (1.2) 7 (1.8)  
 Do not want to answer 0.5% (2) 0.0% (0)  
Youth CharacteristicsControl (n = 410)Intervention (n = 389)P
Age, mean (SD) 15.97 (1.2) 16.14 (1.2) .051 
Hispanic ethnicity, n (%)   .25 
 No 300 (73.2) 302 (77.6)  
 Yes 109 (26.6) 85 (21.9)  
 Decline to answer 1 (0.2) 2 (0.5)  
Race, n (%)   .47 
 White 226 (55.1) 232 (59.6)  
 Black or African American 61 (14.9) 57 (14.7)  
Asian American 16 (3.9) 14 (3.6)  
 Native Hawaiian or Other Pacific Islander 3 (0.7) 2 (0.5)  
 American Indian or Alaska native 9 (2.2) 3 (0.8)  
 Multiracial 58 (14.2) 54 (13.9)  
 Some other race 26 (6.3) 23 (5.9)  
 Do not want to answer 11 (2.7) 4 (1.0)  
Rural,an (%) 70 (17.1) 51 (13.1) .12 
Income,bn (%)   .86 
 Lower than the average 122 (29.8) 110 (28.3)  
 Similar to the average 213 (52.0) 200 (51.4)  
 Higher than the average 61 (14.9) 62 (15.9)  
 Do not want to answer 14 (3.4) 17 (4.4)  
Sexual identity,cn (%)    
 Gay 83 (20.2) 78 (20.1) .95 
 Lesbian 178 (43.4) 169 (43.4) .99 
 Bisexual 172 (42.0) 170 (43.7) .62 
 Pansexual 106 (25.9) 121 (31.1) .10 
 Heterosexual 3 (0.7) 2 (0.5) .70 
 Queer 89 (21.7) 77 (19.8) .51 
 Asexual 9 (2.2) 8 (2.1) .89 
 Questioning 50 (12.2) 43 (11.1) .62 
 Unsure 7 (1.7) 8 (2.1) .72 
 Do not want to answer 0 (0) (0) 0 — 
Gender identity, n (%)   .55 
 Cisgender female 406 (99.0) 385 (99.0)  
 Male to female transgenderd 1 (0.2) 0 (0.0)  
 Male to female transgenderd 0 (0.0) 1 (0.3)  
 Genderqueer or pangenderd 0 (0.0) 1 (0.3)  
 I am unsured 2 (0.5) 2 (0.5)  
 Other 1 (0.2) 0 (0.0)  
 Do not want to answer 0 (0) 0 (0)  
Ever penile-vaginal sex, n (%)   .99 
 No 279 (68.1) 263 (67.6)  
 Yes 130 (31.7) 125 (32.1)  
 Do not want to answer 1 (0.2) 1 (0.3)  
Penile-vaginal sex in the past 3 mo, n (%)   .75 
 No 342 (83.4) 319 (82.0)  
 Yes 67 (16.3) 68 (17.5)  
 Do not want to answer 1 (0.2) 2 (0.5)  
No. condom-protected vaginal sex acts in the past 3 mo, mean (SD) 0.8 (5.7) 1.1 (4.9) .42 
 Do not want to answer 0 (0) 0 (0)  
Currently on birth control, n (%)   .45 
 Yes, to any 101 (24.6) 87 (22.4)  
 Birth control pills   .82 
  No 339 (82.7) 324 (83.3)  
  Yes 71 (17.3) 65 (16.7)  
  I don’t know what this is 0 (0) 0 (0)  
  Do not want to answer 0 (0) 0 (0)  
 The shot   .69 
  No 401 (97.8) 382 (98.2)  
  Yes 9 (2.2) 7 (1.8)  
  I don’t know what this is 0 (0) 0 (0)  
  Do not want to answer 0 (0) 0 (0)  
 The patch   .62 
  No 405 (98.8) 385 (99.0)  
  Yes 4 (1.0) 2 (0.5)  
  I don’t know what this is 1 (0.2) 2 (0.5)  
  Do not want to answer 0 (0) 0 (0)  
 The ring   .82 
  No 406 (99.0) 384 (98.7)  
  Yes 3 (0.7) 3 (0.8)  
  I don’t know what this is 1 (0.2) 2 (0.5)  
  Do not want to answer 0 (0) 0 (0)  
 IUD   .15 
  No 400 (97.6) 384 (98.7)  
  Yes 7 (1.7) 1 (0.3)  
  I don’t know what this is 3 (0.7) 3 (0.8)  
  Do not want to answer 0 (0.0) 1 (0.3)  
 Implants   .44 
  No 393 (95.9) 379 (97.4)  
  Yes 16 (3.9) 9 (2.3)  
  I don’t know what this is 1 (0.2) 1 (0.3)  
  Do not want to answer 0 (0) 0 (0)  
Ever pregnant, n (%)   .50 
 No 388 (94.6) 369 (94.9)  
 Yes 15 (3.7) 13 (3.3)  
 I don’t know 5 (1.2) 7 (1.8)  
 Do not want to answer 0.5% (2) 0.0% (0)  

—, not applicable.

a

Based on the 2013 National Center for Health Statistics Urban-Rural Classification Scheme for Counties.29 

b

Self-appraised.

c

Multiple response.

d

Gender identity was assessed during the enrollment call. A small number of youth self-reported a noncisgender identity in the baseline survey. Because these answers were not again assessed for eligibility, they were included in the sample.

At RCT end, 5 months postenrollment, 22% of intervention and 19% of control participants reported penile-vaginal sex in the past 3 months; 17% and 13% reported at least 1 condom-protected sex act; and 10% and 12% reported at least 1 condomless protected sex act, respectively. The number of condom-protected sex acts ranged between 0 and 70 for intervention and 0 to 60 for control participants.

Among those who reported penile-vaginal sex in the past 3 months at baseline (ie, sexually active girls), 46% of intervention and 43% of control participants reported at least 1 condom-protected sex act in the past 3 months at intervention end; 32% and 49%, respectively, reported at least 1 condomless protected sex act over the same time period.

As shown in Table 4, the rate of condom-protected sex acts in the past 3 months was significantly higher at intervention end for those in the intervention versus control group (adjusted odds ratio [aOR] = 1.48, P < .001) after adjusting for survey mode, age, and number of condom-protected sex acts at baseline. The magnitude of association was similar when examined among sexually active girls (aOR = 1.64, P < .001). The intervention also was associated with significantly lower rates of condomless sex acts generally (aOR = 0.79, P = .007) and among sexually active girls specifically (aOR = 0.65, P < .001).

TABLE 4

Pregnancy Preventive Behavior and Intentions Outcomes at Girl2Girl RCT Intervention End (n = 799)

Pregnancy Preventive Behaviors and IntentionsControl (n = 410)Intervention (n = 389)IRR/OR95% CIPIRR/aOR95% CIP
No. condom-protected sex acts in the past 3 mo
 All youth (n = 799) mean (SD) 1.2 (6.0) 1.4 (5.9) 1.21 1.07–1.37 .002 1.48 1.30–1.68 <.001 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 135) mean (SD) 4.7 (11.9) 5.3 (11.6) 1.13 0.97–1.31 .13 1.64 1.40–1.93 <.001 
No. condomless sex acts in the past 3 moa         
 All youth (n = 799) mean (SD) 0.7 (2.3) 0.6 (2.1) 0.83 0.70–0.99 .04 0.79 0.66–0.94 .007 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 135) mean (SD) 3.5 (4.4) 2.3 (3.8) 0.64 0.52–0.79 <.001 0.65 0.53–0.80 <.001 
Currently using birth control, n (%)         
 All youth (n = 799) 110 (26.8) 121 (31.1) 1.23 0.91–1.67 .18 1.60 1.08–2.37 .02 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 135) 37 (55.2) 33 (48.5) 0.76 0.39–1.50 .44 0.98 0.41–2.34 .97 
Abstaining from penile-vaginal sex in the past 3 mo, n (%)         
 All youth (n = 799) 330 (80.5) 301 (77.4) 0.83 0.59–1.17 .28 0.82 0.55–1.23 .34 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 135) 22 (32.8) 26 (38.2) 1.27 0.62–2.57 .51 1.20 0.57–2.55 .63 
Pregnancy since program enrollment, n (%)         
 All youth (n = 799) 8 (2.0) 3 (0.8) 0.39 0.10–1.48 .17 0.43 0.11–1.70 .23 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 135) 7 (10.5) 3 (4.4) 0.40 0.10–1.60 .19 0.39 0.09–1.73 .21 
Intentions to use condoms in the next year, n (%)         
 All youth (n = 799) 265 (64.6) 261 (67.1) 1.12 0.83–1.50 .46 1.09 0.79–1.52 .59 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 135) 41 (61.2) 48 (70.6) 1.52 0.74–3.12 .25 1.54 0.72–3.30 .26 
 Youth who have not had sex (n = 661) 223 (65.2) 213 (66.8) 1.07 0.78–1.48 .67 1.04 0.72–1.51 .84 
Intentions to use birth control in the next year,bn (%)         
 All youth (n = 568) 99 (33.0) 128 (47.8) 1.86 1.32–2.61 <.001 1.93 1.31–2.84 .001 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 65) 15 (50.0) 27 (77.1) 3.37 1.16–9.79 .03 3.25 0.99–10.70 .052 
 Youth who have not had sex (n = 500) 83 (30.9) 101 (43.7) 1.74 1.21–2.51 .003 1.87 1.24–2.84 .003 
Intentions to be abstinent in the next year, n (%)         
 All youth (n = 799) 178 (43.4) 157 (40.4) 0.88 0.67–1.17 .38 0.95 0.67–1.35 .77 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 135) 9 (13.4) 7 (10.3) 0.74 0.26–2.12 .57 0.75 0.23–2.42 .63 
 Youth who have not had sex (n = 661) 169 (49.4) 150 (47.0) 0.91 0.67–1.23 .54 0.95 0.65–1.38 .79 
Pregnancy Preventive Behaviors and IntentionsControl (n = 410)Intervention (n = 389)IRR/OR95% CIPIRR/aOR95% CIP
No. condom-protected sex acts in the past 3 mo
 All youth (n = 799) mean (SD) 1.2 (6.0) 1.4 (5.9) 1.21 1.07–1.37 .002 1.48 1.30–1.68 <.001 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 135) mean (SD) 4.7 (11.9) 5.3 (11.6) 1.13 0.97–1.31 .13 1.64 1.40–1.93 <.001 
No. condomless sex acts in the past 3 moa         
 All youth (n = 799) mean (SD) 0.7 (2.3) 0.6 (2.1) 0.83 0.70–0.99 .04 0.79 0.66–0.94 .007 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 135) mean (SD) 3.5 (4.4) 2.3 (3.8) 0.64 0.52–0.79 <.001 0.65 0.53–0.80 <.001 
Currently using birth control, n (%)         
 All youth (n = 799) 110 (26.8) 121 (31.1) 1.23 0.91–1.67 .18 1.60 1.08–2.37 .02 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 135) 37 (55.2) 33 (48.5) 0.76 0.39–1.50 .44 0.98 0.41–2.34 .97 
Abstaining from penile-vaginal sex in the past 3 mo, n (%)         
 All youth (n = 799) 330 (80.5) 301 (77.4) 0.83 0.59–1.17 .28 0.82 0.55–1.23 .34 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 135) 22 (32.8) 26 (38.2) 1.27 0.62–2.57 .51 1.20 0.57–2.55 .63 
Pregnancy since program enrollment, n (%)         
 All youth (n = 799) 8 (2.0) 3 (0.8) 0.39 0.10–1.48 .17 0.43 0.11–1.70 .23 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 135) 7 (10.5) 3 (4.4) 0.40 0.10–1.60 .19 0.39 0.09–1.73 .21 
Intentions to use condoms in the next year, n (%)         
 All youth (n = 799) 265 (64.6) 261 (67.1) 1.12 0.83–1.50 .46 1.09 0.79–1.52 .59 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 135) 41 (61.2) 48 (70.6) 1.52 0.74–3.12 .25 1.54 0.72–3.30 .26 
 Youth who have not had sex (n = 661) 223 (65.2) 213 (66.8) 1.07 0.78–1.48 .67 1.04 0.72–1.51 .84 
Intentions to use birth control in the next year,bn (%)         
 All youth (n = 568) 99 (33.0) 128 (47.8) 1.86 1.32–2.61 <.001 1.93 1.31–2.84 .001 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 65) 15 (50.0) 27 (77.1) 3.37 1.16–9.79 .03 3.25 0.99–10.70 .052 
 Youth who have not had sex (n = 500) 83 (30.9) 101 (43.7) 1.74 1.21–2.51 .003 1.87 1.24–2.84 .003 
Intentions to be abstinent in the next year, n (%)         
 All youth (n = 799) 178 (43.4) 157 (40.4) 0.88 0.67–1.17 .38 0.95 0.67–1.35 .77 
 Youth who have had vaginal sex in the past 3 mo at baseline (n = 135) 9 (13.4) 7 (10.3) 0.74 0.26–2.12 .57 0.75 0.23–2.42 .63 
 Youth who have not had sex (n = 661) 169 (49.4) 150 (47.0) 0.91 0.67–1.23 .54 0.95 0.65–1.38 .79 

Sample size for intentions does not equal 799 because 3 people who declined to answer about recent vaginal sex are not included in either category. Models adjusted for survey mode, age, and baseline indicator of outcome of interest (eg, condom use). aIRR, adjusted incident rate ratio; aOR, adjusted odds ratio; IRR, incident rate ratio; OR, odds ratio.

a

Because of outliers (range: 0–663, SD = 24.2), censored at 10+ acts.

b

Among those not currently on birth control.

The relative odds of current use of birth control other than condoms was 60% higher in the intervention versus control group (aOR = 1.60, P = .02). Differences were not significant among sexually active girls, however (aOR = 0.98, P = .97).

The relative odds of abstaining from penile-vaginal sex in the past 3 months at intervention end were statistically similar for intervention and control participants generally (aOR = 0.82, P = .34) and sexually active girls specifically (aOR = 1.20, P = .63).

The magnitude of the relative odds of pregnancy between intervention and control participants was clinically meaningful but not statistically significant in general (aOR = 0.43, P = .23) or among sexually active girls (aOR = 0.39, P = .21).

As shown in Table 4, the relative odds of intending to use birth control among girls who were not currently on birth control were significantly higher for those in the intervention versus control group at intervention end (aOR = 1.93, P = .001). This also was true among sexually active girls (aOR = 3.25, P = .052) and abstinent girls (aOR = 1.87, P = .003). When girls who intended to use birth control were combined with girls currently on birth control, the relative odds of being in the intervention group were significantly higher among all girls (aOR = 2.08, P < .001), as well as among sexually active (aOR = 3.02, 0.04) and abstinent (aOR = 2.10, P < .001) girls specifically (data not shown).

Intentions to use condoms (aOR = 1.09, P = .59) and intentions to be abstinent did not differ by study arm (aOR = 0.95, P = .77).

Results were similar when youth who declined to answer an outcome question were dropped rather than coded as failure. Data are available on request.

To our knowledge, Girl2Girl is the first pregnancy prevention program developed for and tested among sexual minority girls across the United States and the first comprehensive, technology-based teenage pregnancy prevention program for any group of youth. In this large-scale RCT, outcomes at intervention end suggest that Girl2Girl is associated with increased rates of condom use and increased odds of using other types of contraception. The intervention also appears to be associated with increased intentions to use birth control among girls not on birth control. Although few pregnancies were reported over the 5 months, the intervention group had half the odds of pregnancy versus the control group; given this rare event, the effect size was not statistically significant. Taken together, these findings suggest that Girl2Girl may be associated with multiple pregnancy preventive behaviors, at least in the short-term.

Girl2Girl does not appear to be effective in promoting sexual abstinence or a return to abstinence among recently sexually active girls. This may be in part because the health promotion messages were intentionally sex positive and focused on healthy relationships and use of barriers and testing, rather than solely emphasizing abstinence. Also, other researchers have documented that abstinence-focused interventions appear to have no effect on sexually active girls.30,31  Future research about sex-positive programs that includes sexually experienced and inexperienced youth might consider adding an outcome that reflects self-efficacy to consent to sex when it is wanted and demur to sex when it is not.

Given the nature of the intervention, findings may not be generalizable to girls who do not use social media or have cell phones with a limited messaging plan. Moreover, social media advertising was targeted to girls whose profiles indicated they were “interested in” other girls. Therefore, girls who saw the ads were “out” at least on their social media profile. Findings may not be generalizable to those who are not out to others. Additionally, it is impossible to determine if a particular program message was read. This is not unlike school-based programs in which it is unknown whether students are listening to the intervention facilitator. Finally, because of funding uncertainty, surveys at intervention end were collected either via text messaging or online. Aside from potential mode differences, questions were worded slightly differently and, in some cases, referred to different time frames (eg, next 12 months versus next 3 months). To mitigate the potential impact of this, we adjusted for survey mode in multivariate models. Balancing these limitations, it should be noted that the national sample was diverse in terms of race and ethnicity, sexual identity, rural and urban setting, and age. The intervention also was novel in its intervention target (ie, sexual minority girls), delivery mechanism (ie, comprehensive text messages), and scope (ie, across the United States).

Reviews suggest that text messaging–based interventions can affect and sustain complex health behavior changes across a variety of behaviors, including HIV testing, medication adherence, physical activity, and smoking cessation.3242  Results here provide reason for optimism that this approach can be effective with adolescent pregnancy prevention as well.

The high program completion rate suggests that sexual minority adolescent girls are willing to receive voluminous amounts of sexual health-related text messages over a relatively long period of time (ie, 5 months). Given the relative cost efficiency and wide reach of text messaging as a delivery mechanism compared with more traditional models such as facilitator-based education,43  researchers may consider using this modality to address other adolescent behavior change efforts as well.

Opportunities for future research are noted. First, it is unknown whether some or all components of the intervention (eg, Text Buddy, G2Genie, level-up questions) meaningfully contributed to behavior change. Understanding the relative contributions of these features could inform the future development of technology-based interventions that seek to include game-like program components. Second, girls of various sexual minority identities were analyzed together. Subsequent research may wish to examine outcomes for lesbian and bisexual girls separately given they differ in their sexual behaviors with girls and boys.17,44  Additionally, it would be useful to explore the optimal length of interventions such as this, as well as the timing of the booster delivery. Other important questions include whether the behavioral changes are sustained over time and if behavioral changes are noted in an effectiveness trial.

Sexual minority girls are significantly more likely to be pregnant during their teenage years compared to heterosexual girls and yet, limited teenage pregnancy prevention programming is available that is tailored to their needs. Findings suggest that Girl2Girl is associated with changes in teenage pregnancy preventive behaviors and behavioral intentions with both sexual minority girls who are having penile-vaginal sex and those who are not, at least in the short-term. The latter are key to a comprehensive public health approach that gives all girls the tools they need to make healthy decisions if, and when, they choose to have sex that could lead to pregnancy.

We thank the study participants and research team, particularly Katrina Nardo, Dr Myeshia Price-Feeney, and Desiree Fehmie for their contributions to the study.

This trial has been registered at www.clinicaltrials.gov (identifier NCT03029962).

Deidentified individual participant data will not be made available.

Dr Ybarra made substantial contribution to conception and design, acquisition of data, and analysis and interpretation of data and drafted the article; Ms Prescott made substantial contribution to the acquisition of data and revised the manuscript critically for important intellectual content; Drs Saewyc, Rosario, and Goodenow made substantial contribution to analysis and interpretation of data, provided consultation on the study design, and revised the manuscript critically for important intellectual content; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

FUNDING: Research reported in this publication was supported by the Office of Population Affairs (TP2AH000035) and the National Institute of Child Health and Human Development (R01HD095648). The content is solely the responsibility of the authors and does not necessarily represent the official views of the Office of Population Affairs and the National Institutes of Health. Funded by the National Institutes of Health (NIH).

     
  • aOR

    adjusted odds ratio

  •  
  • LGB+

    lesbian, gay, bisexual and other sexual minority

  •  
  • RCT

    randomized controlled trial

1
Martin
JA
,
Hamilton
BE
,
Osterman
MJK
,
Driscoll
AK
,
Drake
P
.
Births: Final Data for 2017
.
Hyattsville, MD
:
National Center for Health Statistics
;
2018
2
Kann
L
,
McManus
T
,
Harris
WA
, et al
.
Youth risk behavior surveillance - United States, 2017
.
MMWR Surveill Summ
.
2018
;
67
(
8
):
1
114
3
Tornello
SL
,
Riskind
RG
,
Patterson
CJ
.
Sexual orientation and sexual and reproductive health among adolescent young women in the United States
.
J Adolesc Health
.
2014
;
54
(
2
):
160
168
4
Goodenow
C
,
Szalacha
LA
,
Robin
LE
,
Westheimer
K
.
Dimensions of sexual orientation and HIV-related risk among adolescent females: evidence from a statewide survey
.
Am J Public Health
.
2008
;
98
(
6
):
1051
1058
5
Saewyc
EM
,
Bearinger
LH
,
Blum
RW
,
Resnick
MD
.
Sexual intercourse, abuse and pregnancy among adolescent women: does sexual orientation make a difference?
Fam Plann Perspect
.
1999
;
31
(
3
):
127
131
6
Lindley
LL
,
Walsemann
KM
.
Sexual orientation and risk of pregnancy among New York City high-school students
.
Am J Public Health
.
2015
;
105
(
7
):
1379
1386
7
Saewyc
EM
,
Poon
CS
,
Homma
Y
,
Skay
CL
.
Stigma management? The links between enacted stigma and teen pregnancy trends among gay, lesbian, and bisexual students in British Columbia
.
Can J Hum Sex
.
2008
;
17
(
3
):
123
139
8
Rosario
M
,
Corliss
HL
,
Everett
BG
, et al
.
Sexual orientation disparities in cancer-related risk behaviors of tobacco, alcohol, sexual behaviors, and diet and physical activity: pooled Youth Risk Behavior Surveys
.
Am J Public Health
.
2014
;
104
(
2
):
245
254
9
Charlton
BM
,
Corliss
HL
,
Missmer
SA
, et al
.
Reproductive health screening disparities and sexual orientation in a cohort study of US adolescent and young adult females
.
J Adolesc Health
.
2011
;
49
(
5
):
505
510
10
Guttmacher Institute
.
Sex and HIV education
.
2020
. Available at: https://www.guttmacher.org/state-policy/explore/sex-and-hiv-education#. Accessed November 9, 2020
11
Kosciw
J
,
Greytak
EA
,
Zongrone
AD
,
Clark
CM
,
Truong
N
.
The 2017 National School Climate Survey: The Experiences of Lesbian, Gay, Bisexual, Transgender, and Queer Youth in Our Nation’s Schools
.
New York, NY
:
GLSN
;
2018
12
Hodel
D
,
Levi
J
,
De Biasi
A
.
Furthering the Adaptation and Implementation of LGBTQ-Inclusive Sexuality Education: Issue Report
.
Washington, DC
:
Trust for America’s Health
;
2016
13
Ybarra
ML
,
Price-Feeney
M
,
Prescott
T
,
Goodenow
C
,
Saewyc
E
,
Rosario
M
.
Girl2Girl: how to develop a salient pregnancy prevention program for cisgender sexual minority adolescent girls
.
J Adolesc
.
2020
;
85
:
41
58
14
Ybarra
ML
,
Prescott
TL
,
Philips
GL
 II
,
Bull
SS
,
Parsons
JT
,
Mustanski
B
.
Iteratively developing an mHealth HIV prevention program for sexual minority adolescent men
.
AIDS Behav
.
2016
;
20
(
6
):
1157
1172
15
Fisher
JD
,
Fisher
WA
. Theoretical Approaches to Individual-Level Change in HIV Risk Behavior. In:
Peterson
JL
,
DiClemente
RJ
, eds.
Handbook of HIV Prevention
.
New York, NY
:
Kluwer Academic/Plenum Press
;
2000
:
3
55
16
Fisher
JD
,
Fisher
WA
.
Changing AIDS-risk behavior
.
Psychol Bull
.
1992
;
111
(
3
):
455
474
17
Ybarra
ML
,
Santana Parilla
J
,
Wolowic
J
,
Rosario
M
,
Goodenow
C
,
Saewyc
E
.
A national, qualitative study of sexual decision making by teenage girls who are lesbian, gay, bisexual, or who have another nonheterosexual identity
.
J Pediatr
.
2020
;
217
:
177
183
18
Doull
M
,
Wolowic
J
,
Saewyc
E
,
Rosario
M
,
Prescott
T
,
Ybarra
ML
.
Why girls choose not to use barriers to prevent sexually transmitted infection during female-to-female sex
.
J Adolesc Health
.
2018
;
62
(
4
):
411
416
19
Rodgers
A
,
Corbett
T
,
Bramley
D
, et al
.
Do u smoke after txt? Results of a randomised trial of smoking cessation using mobile phone text messaging
.
Tob Control
.
2005
;
14
(
4
):
255
261
20
Ybarra
ML
,
Holtrop
JS
,
Prescott
TL
,
Rahbar
MH
,
Strong
D
.
Pilot RCT results of stop my smoking USA: a text messaging-based smoking cessation program for young adults
.
Nicotine Tob Res
.
2013
;
15
(
8
):
1388
1399
21
Taves
DR
.
Minimization: a new method of assigning patients to treatment and control groups
.
Clin Pharmacol Ther
.
1974
;
15
(
5
):
443
453
22
Center for Innovative Public Health Research
.
Teen health and technology
.
2015
. Available at: http://innovativepublichealth.org/projects/teen-health-and-technology/. Accessed November 10, 2020
23
Machin
D
,
Campbell
M
,
Fayers
P
,
Pinol
A
.
Sample Size Tables for Clinical Studies
.
Malden, MA
:
Blackwell Science
;
1997
24
Lachin
JM
.
Biostatistical Methods
.
New York, NY
:
John Wiley & Sons
;
2000
25
Fleiss
JL
,
Levin
B
,
Paik
MC
.
Statistical Methods for Rates and Proportions
, 3rd ed.
New York, NY
:
John Wiley & Sons
;
2003
26
D’Agostino
RB
,
Chase
W
,
Belanger
A
.
The appropriateness of some common procedures for testing the equality of two independent binomial populations
.
Am Stat
.
1988
;
42
:
198
202
27
Chow
SC
,
Shao
J
,
Wang
H
.
Sample Size Calculations in Clinical Research
.
New York, NY
:
Marcel Dekker
;
2003
28
Office of Adolescent Health
.
Performance measures for replication of evidence-based programs (tier 1) c/d, research and demonstration programs (tier 2), and PREIS grantees
.
2014
. Available at: www.hhs.gov/ash/oah/oah-initiatives/for-grantees/program-guidance/Assets/cd-performancemeasures2014.pdf. Accessed November 10, 2020
29
Centers for Disease Control and Prevention National Center for Health Statistics
.
NCHS urban-rural classification scheme for counties
.
2017
. Available at: https://www.cdc.gov/nchs/data_access/urban_rural.htm. Accessed November 10, 2020
30
Lieberman
L
,
Su
H
.
Impact of the choosing the best program in communities committed to abstinence education
.
SAGE Open
.
2012
;
2
(
1
):
1
12
31
Pedlow
CT
,
Carey
MP
.
Developmentally appropriate sexual risk reduction interventions for adolescents: rationale, review of interventions, and recommendations for research and practice
.
Ann Behav Med
.
2004
;
27
(
3
):
172
184
32
Whittaker
R
,
McRobbie
H
,
Bullen
C
,
Borland
R
,
Rodgers
A
,
Gu
Y
.
Mobile phone-based interventions for smoking cessation
.
Cochrane Database Syst Rev
.
2012
;
11
:
CD006611
33
Preston
KE
,
Walhart
TA
,
O’Sullivan
AL
.
Prompting healthy behavior via text messaging in adolescents and young adults
.
Am J Lifestyle Med
.
2011
;
5
(
3
):
247
252
34
Wei
J
,
Hollin
I
,
Kachnowski
S
.
A review of the use of mobile phone text messaging in clinical and healthy behaviour interventions
.
J Telemed Telecare
.
2011
;
17
(
1
):
41
48
35
Cole-Lewis
H
,
Kershaw
T
.
Text messaging as a tool for behavior change in disease prevention and management
.
Epidemiol Rev
.
2010
;
32
(
1
):
56
69
36
Ybarra
ML
,
Jiang
Y
,
Free
C
,
Abroms
LC
,
Whittaker
R
.
Participant-level meta-analysis of mobile phone-based interventions for smoking cessation across different countries
.
Prev Med
.
2016
;
89
:
90
97
37
Ybarra
ML
,
Prescott
TL
,
Phillips
GL
 II
,
Bull
SS
,
Parsons
JT
,
Mustanski
B
.
Pilot RCT results of an mHealth HIV prevention program for sexual minority male adolescents
.
Pediatrics
.
2017
;
140
(
1
):
e20162999
38
Garofalo
R
,
Kuhns
LM
,
Reisner
SL
,
Mimiaga
MJ
.
Behavioral interventions to prevent HIV transmission and acquisition for transgender women: a critical review
.
J Acquir Immune Defic Syndr
.
2016
;
72
(
suppl 3
):
S220
S225
39
Blake
SM
,
Ledsky
RA
,
Sawyer
RJ
, et al
.
Local school district adoption of state-recommended policies on HIV prevention education
.
Prev Med
.
2005
;
40
(
2
):
239
248
40
Santelli
JS
,
Jacobson
MS
.
Birth weight outcomes for repeat teenage pregnancy
.
J Adolesc Health Care
.
1990
;
11
(
3
):
240
247
41
Buchholz
ES
,
Korn-Bursztyn
C
.
Children of adolescent mothers: are they at risk for abuse?
Adolescence
.
1993
;
28
(
110
):
361
382
42
Bull
S
,
Devine
S
,
Schmiege
SJ
,
Pickard
L
,
Campbell
J
,
Shlay
JC
.
Text messaging, teen outreach program, and sexual health behavior: a cluster randomized trial
.
Am J Public Health
.
2016
;
106
(
suppl 1
):
S117
S124
43
Anderson
M
,
Jiang
J
.
Teens, social media & technology 2018
. Available at: www.pewinternet.org/2018/05/31/teens-social-media-technology-2018/. Accessed November 10, 2020
44
Rosario
M
,
Goodenow
C
,
Ybarra
M
,
Saewyc
E
,
Prescott
T
.
Reasons lesbian and bisexual adolescent girls have or might have sex with females or males: implications for discordance between sexual identity and behaviors and for prevention of pregnancy and STIs
.
J Sex Res
.
2020
;
57
(
8
):
966
978

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

COMPANION ARTICLE: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/2020-029801.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.