Trauma-Informed Personalized Scripts to Address Partner Violence and Reproductive Coercion: Follow-Up Findings from an Implementation Randomized Controlled Trial Study

Abstract

Background: Intimate partner violence (IPV) and reproductive coercion impact women seeking care at family planning (FP) clinics. Interventions to facilitate patient–provider conversations about healthy relationships are needed. We sought to determine the added effect of providing psychoeducational messages to patients compared with tailored provider scripts alone on sexual and reproductive health outcomes at 4–6 months.

Materials and Methods: We randomized participants to Trauma-Informed Personalized Scripts (TIPS)-Plus (provider scripts +patient messages) or TIPS-Basic (provider scripts only) at four FP clinics. Eligible patients included English-speaking females aged 16–29 years. Data were collected at initial visits (T1) and 4–6 months (T2) on IPV, reproductive coercion, fear, condom and other contraceptive use, self-efficacy, harm reduction behaviors, and knowledge/use of IPV-related services. We compared frequencies and summary scores between baseline and follow-up with McNemar's test of paired proportions and Signed Rank-Sum, respectively. We compared the difference in differences over time by treatment arm using two-sample t-tests, and used linear, logistic, and ordinal logistic regression to compare intervention effects at follow-up.

Results: Two hundred forty patients participated (114 TIPS-Plus, 126 TIPS-Basic), 216 completed follow-up. We detected no differences in outcomes between treatment arms. Between T1 and T2, we observed overall reductions in mean summary scores for reproductive coercion (T1 = 0.08 ± 0.02, T2 = 0.02 ± 0.01, p = 0.028) and increases in contraceptive use (69.6%–87.9%, p < 0.001), long-acting reversible contraceptives (8.3%–20.8%, p < 0.001), and hidden methods (20%–38.5%, p < 0.001).

Conclusions: We show no added benefit of patient-activation messages compared with provider scripts alone. Findings suggest potential utility of provider scripts in addressing reproductive coercion and contraceptive uptake (Trial Registration No. NCT02782728).

Introduction

Studies demonstrate that intimate partner violence (IPV) is strongly associated with unintended pregnancy (mistimed, unwanted, or unplanned pregnancies) through multiple mechanisms, including reproductive coercion (birth control sabotage, pregnancy coercion).^1–7 Reproductive coercion is associated with unintended pregnancy even in the absence of physical or sexual violence, and is prevalent among women seeking care in family planning (FP) settings.³ All clinicians can and should address violence and its associated health effects with their patients. FP clinicians, in particular, have a unique opportunity to reach women who have experienced reproductive coercion and offer harm reduction strategies, including contraception that partners cannot interfere with.

Over the last several decades, the response of the health care sector to IPV has predominantly focused on screening. Providers often ask patients “yes” or “no” questions about past or current experiences with different forms of violent and coercive behaviors. If patients answer “yes,” they are screened positive and are eligible to be referred to services. If patients answer “no,” the conversation often ends there. This becomes problematic given the consistent evidence that survivors under-report experiences with IPV and reproductive coercion due to myriad barriers, including fear of an abusive partner, societal stigma, mistrust of the health care system, and concerns about losing custody of their children.^8–11 While well intentioned, this traditional screening paradigm offers patients a difficult choice: disclose, potentially receive resources, and potentially experience consequences related to disclosure or do not disclose and leave empty handed.⁸

To help resolve these unintended consequences of screening, researchers have developed interventions to more comprehensively address IPV and reproductive coercion. One such intervention, ARCHES (Addressing Reproductive Coercion in Healthcare Settings), is an evidence-based approach that encompasses universal education, harm reduction counseling, and warm referrals.^12–14 Compared with a usual care control, ARCHES demonstrated effectiveness in improving knowledge of resources and increasing harm reduction self-efficacy; further, ARCHES reduced reproductive coercion among those with higher levels of reproductive coercion at baseline. However, consistent implementation of ARCHES remains a challenge, particularly given time constraints and productivity pressures that providers face.^14,15 Research has also demonstrated other provider-level barriers, including discomfort, lack of training, inadequate resources, and uncertainty about what to do when a patient discloses.^16,17

In response to both provider- and patient-level barriers that prevent discussions on IPV and reproductive coercion from occurring, we created a new intervention: Trauma-Informed Personalized Scripts (TIPS), founded on the principles of ARCHES. For each clinical encounter, TIPS provides automated scripts to clinicians based on patient responses to a brief tablet-based survey about IPV, reproductive coercion, and contraception needs, completed in private while waiting for a clinic visit. Designed to improve patient–provider communication, the survey can also offer tailored messages directly to patients, intended to encourage patients to talk to their providers about relationships and desired contraceptive methods during the clinical encounter. Patient activation through such messaging is effective in other primary care settings and is based on neuroscience research involving positive reward-related responses.^18,19

We have previously published on the primary aim of TIPS: to determine the added effect of patient-activation messages compared with provider scripts alone on intervention implementation.²⁰ We found no significant differences in implementation comparing provider scripts alone to added patient-activation messages, yet there were significant increases in implementation in both arms compared with historical controls.²⁰ However, we are also interested in knowing how these changes in implementation of universal education during clinic visits ultimately impacted participants' recognition and use of relevant resources and harm reduction strategies (including contraceptive use), as well as their exposure to IPV and reproductive coercion. For the purposes of this article, we focus on this secondary aim of TIPS: to determine the added effect of patient activation (TIPS-Plus) compared with provider scripts alone (TIPS-Basic) on IPV, reproductive coercion, fear of partner, contraceptive use, condom use, self-efficacy, harm reduction behaviors, and knowledge and use of IPV-related services.

Materials and Methods

Study design

This was an exploratory, two-arm randomized controlled trial (allocation ratio = 1:1). We used a computer-generated randomization schema programmed through the tablet-based app to assign individuals to TIPS-Plus or TIPS-Basic, concealing patients, providers, and researchers from allocation. Due to the study design (i.e., patients either received or did not receive the patient-activation messages of TIPS-Plus), we were able to blind providers, but not patients, to treatment arm. The study included patients and providers at four FP clinics in western Pennsylvania. These clinics provide services related to contraceptive counseling, sexually transmitted infection (STI) testing and treatment, and general preventive care to reproductive-aged women from diverse racial/ethnic backgrounds and socioeconomic statuses. We included all health care providers, such as physicians, nurse practitioners, registered nurses, and medical assistants. Eligibility criteria for patients included the following: (1) female gender, (2) aged 16–29, (3) ability to read English, and (4) seeking services at a participating FP clinic. We included individuals aged 16–29 given the high prevalence of IPV among this age range, particularly among those who seek care at FP clinics, as well as high prevalence of other adverse sexual and reproductive health outcomes, such as STIs.^21,22 We recruited patients from May 2016 to February 2017 and followed participants from September 2016 to August 2017. The trial ended upon reaching our target sample size. Further details on recruitment and enrollment are described elsewhere.²⁰ All study procedures were approved by the Institutional Review Board at the University of Pittsburgh.

Intervention

In both treatment arms (TIPS-Plus and TIPS-Basic), patients completed a tablet-based survey in private before their clinic appointment, with questions about relationship history, experiences with IPV/reproductive coercion, and sexual and reproductive health needs. Based on patients' responses in both TIPS-Plus and TIPS-Basic, the app notified the research staff to print relevant scripts for the patients' providers to use during their clinic visits. There was a total of four scripts, which have been published in full elsewhere.²⁰ For example, if someone disclosed experiences of IPV, their provider was notified to discuss safety; if someone disclosed experiences of reproductive coercion, their provider was notified to discuss harm reduction strategies; and if someone disclosed being afraid of their partner, their provider was notified to discuss fear. Depending on patients' experiences, multiple provider scripts could be printed. If a patient chose not to disclose IPV or reproductive coercion on the tablet-based survey or they had never experienced IPV or reproductive coercion, a universal education script was printed out for their provider.

Each of the four scripts was no more than seven short sentences, which helped the provider introduce and normalize the topic, ask necessary questions, provide targeted counseling, and discuss resources. Research staff printed these provider scripts and attached them to patients' clinic charts. The scripts reminded providers to address harmful partner behaviors but did not include details of the patients' survey responses, allowing patients to disclose experiences of violence if and when they desired. For example, the universal education script was as follows: “We talk to all of our patients about how you deserve to be treated by people you go out with. We give everyone this card: ‘Are you in a healthy relationship?’ [title of the card given to the participant] This may not be something you need but perhaps you have a friend or family member who could use this information. Please feel free to take these along, if you feel safe taking them with you.” All patients were encouraged to take the card with them if it was safe to do so.

In addition to provider scripts, TIPS-Plus participants received patient-activation messages, viewed as pop-up messages while they were completing the survey on the app. These messages provided psychoeducational feedback and guidance on what the patient might discuss during their clinical encounter tailored to her survey responses. For example, if a patient disclosed lifetime or recent IPV on the survey and she was randomized to TIPS-Plus, she would receive the following pop-up message: “Everyone deserves to be in a healthy and positive relationship and to feel respected. This clinic is a safe place with helpful staff. Please feel free to talk to your providers today about your relationships.” In the TIPS-Plus arm, participants could also receive messages related to reproductive coercion, fear of a partner, or universal education depending on their responses to the tablet-based survey.

Measures

Our primary outcome was intervention implementation, measured immediately postvisit as a summary score through patient-reported discussions on IPV, reproductive coercion, STIs, and a wallet-sized safety card, detailed elsewhere.²⁰ In this article, we present secondary outcomes (patient-reported) measured at baseline and 4–6 months follow-up via computerized surveys: experiences with IPV, reproductive coercion, and fear, use of contraception and condoms, self-efficacy, harm reduction behaviors, knowledge and utilization of IPV-related services, and decisional conflict. We reported IPV, reproductive coercion, fear, harm reduction behaviors, and knowledge and utilization of IPV-related services as summary scores. Self-efficacy and decisional conflict were continuous outcomes; contraceptive use and condom use were dichotomous outcomes. All measures and associated Cronbach's alpha are included in Table 1.

Table 1.

Outcome Measures (Baseline and 4–6 Months Follow-Up)

Outcomesa	Definition	Cronbach's alpha
IPV victimization	Physical violence: hit, pushed, slapped, choked, or otherwise physically hurt by someone you were dating or going out with	0.93b
	Sexual violence: made you have sex (vaginal, oral, anal) with/without force or threats by someone you were dating or going out with
Reproductive coercion victimization	Someone you were dating, going out with, or having sex with has:	0.88b
	Told you not to use birth control
	Taken off a condom while having sex, so you would get pregnant
	Put holes in the condom or broken the condom on purpose so you would get pregnant
	Taken your birth control away from you or kept you from going to the clinic to get birth control
	Made you have sex without a condom so you would get pregnant
Fear of partner	Have you been afraid to:	0.67b
	Ask your sex partner to use a condom
	Discuss birth control with your sex partner
	Refuse sex with a sex partner
	Tell your sex partner if you had an STI
Use of contraceptives	Methods used (if any) to prevent pregnancy in the past 30 days
Condom use	Number of times partner uses condoms when participant has vaginal sex
	Number of times patient has sex without a condom when she wanted to use one
Self-efficacy to use IPV/reproductive coercion harm reduction strategies	Composite score of 17 items with Likert scale (strongly agree → strongly disagree) asking about confidence in recognizing emotionally, physically, and sexually abusive relationships, where to get help if needed, discussing IPV/reproductive coercion with health care providers, discussing health care plans or “hidden” birth control (IUD, implant, emergency contraception)	0.96c
Received information on and utilization of specific IPV/reproductive coercion-related services	Have you received information on or utilized: local domestic violence hotlines or services, local rape crisis hotlines or services, National Domestic Violence Hotline, National Rape Crisis Hotline	0.82 (Knowledge)c; 0.49 (utilization)c
Use of IPV/reproductive coercion harm reduction strategies	Tried to protect yourself from your partner by asking friends or family to call or text you when you were with that partner	0.97c
	Changed your birth control method so that your partner could not mess with your birth control
	Hidden your birth control from your partner because you were afraid he would get upset with you for using it
	Asked your health care provider to let a sexual partner know that they needed to get treated for a STI without using your name
	Taken the morning after pill or emergency contraception (to prevent pregnancy) without the person you were having sex with knowing that you did this
	Taken the morning after pill or emergency contraception (to prevent pregnancy) because your partner was trying to get you pregnant when you didn't want to be

^aAll outcomes measured over past 3 months.

^bCalculated using baseline data.

^cCalculated using follow-up data.

IPV, intimate partner violence; IUD, intrauterine device; STI, sexually transmitted infection.

Victimization outcomes

We assessed recent (past 3 months) IPV (physical and sexual) victimization using one item from the revised Conflict Tactics Scale²³ and two items from the Sexual Experiences Survey²⁴ at baseline and follow-up; researchers have used these three items in several studies.^2,3,11–14 At baseline, we also assessed for lifetime IPV using the same three items. For recent reproductive coercion at baseline and follow-up, we used a 5-item validated scale that included questions on contraceptive sabotage, pregnancy coercion, and condom manipulation.²⁵ We measured recent fear of partner at baseline and follow-up using four items about whether a participant feared negotiating condom or contraceptive use with their partner, refusing sex, or notifying a partner about a STI.

Contraceptive outcomes

We assessed any use of contraceptives, defined as any method to prevent pregnancy over the past 30 days. Participants were allowed to select more than one type of contraceptive option. We categorized implants and intrauterine devices (IUDs) as long-acting reversible contraception (LARC). We also categorized Depo-Provera, IUDs, implants, and emergency contraception as hidden methods.²⁶ Patients also reported the frequency with which their partner used a condom during vaginal sex, dichotomized as consistent condom use (always) or not (usually, sometimes, rarely, or never), and the number of times the patient had sex without a condom when she wanted to use one. All contraceptive outcomes were measured at baseline and follow-up.

Knowledge and utilization outcomes

First, we measured self-efficacy to use harm reduction strategies through a 17-item scale that asked participants their confidence level (Likert scale: strongly agree to strongly disagree) in recognizing abusive partner behaviors, where to get help if needed, and discussing IPV/reproductive coercion with their providers.^13,14 Second, we determined use of IPV or reproductive coercion harm reduction strategies through a 6-item scale. We also asked participants if they received information or used specific IPV/reproductive coercion-related services, including local domestic violence hotlines or services, local rape crisis hotlines or services, the National Domestic Violence Hotline, and/or the National Rape Crisis Hotline. Finally, we measured decisional conflict using a 12-item scale, which assessed an individuals' knowledge of options to and uncertainty about making decisions, as well as values and support.²⁷ Participants reported on a 5-point Likert scale (strongly agree to strongly disagree) on items such as “I know my options for keeping myself safe in an unhealthy or abusive relationship.” The higher the decisional conflict score, the more likely individuals were to be uncertain of their options. Patients reported on decisional conflict and self-efficacy at baseline and follow-up, whereas all other knowledge and utilization outcomes were assessed at follow-up only.

Analysis

Sample size justification is outlined in the baseline study; the study was powered based on secondary outcomes presented in these analyses.²⁰ We compared baseline participant characteristics between treatment arms using chi-squared, Fisher's exact, and t-tests (alpha <0.05). We compared overall mean summary scores and frequencies between baseline and follow-up using Signed Rank-sum and McNemar's test of two proportions, respectively. We used two-sample t-tests to look at the difference in differences over time by treatment arm to increase interpretability of our findings. To adjust for covariates, we used linear (continuous), logistic (dichotomous), and ordinal logistic (summary scores) regression models to measure intervention effects. We decided a priori to adjust for demographic variables (age, race/ethnicity, education, relationship status) and baseline outcome values. We examined the effects of discussing IPV in post hoc analyses, adjusting for demographic variables, intervention arm, baseline IPV, and baseline values. Due to small amounts of missing data, sample size varied minimally between analyses. We did not adjust for multiple testing, as the hypothesis involving each outcome is of independent interest.¹³ We conducted all analyses using StataSE Software (version 15).

Results

Of approached patients, 251 consented to participate (58.0%) and 240 completed the baseline survey (114 TIPS-Plus, 126 TIPS-Basic); 99.2% completed the exit survey immediately after their first clinic visit. Retention at 4–6 months was 91.2% and 88.9% for TIPS-Plus and TIPS-Basic, respectively (Fig. 1). Among the overall sample, individuals who did not complete the follow-up survey tended to be younger (20 vs. 22, p = 0.0028) and more likely to have only completed high school or less (p = 0.001).

FIG. 1.

Flow Diagram of the Trauma-Informed Personalized Scripts (TIPS) Trial. FP, family planning.

There were no clinically significant differences between arms on baseline characteristics (Table 2). Mean age was 22.4 years (standard error = 0.24). Our sample population was predominantly White (70.8%), but also included participants who identified as Black (18.3%), multiracial (5.8%), Asian (2.1%), Hispanic/Latina (1.3%), and American Indian or Alaska Native (1.3%). The majority of participants were college educated (70.8%) with only 9.6% of our sample having completed less than a 12th grade education. Participants most frequently reported dating one person or being in a serious relationship (62.1%). A total of 31.3% reported being single; only 2.1% of our sample indicated they were married. Reporting all methods they currently use, the most commonly used contraceptive method was condoms (38.3%), followed by oral contraceptive pills (32.5%). A total of 29.6% of participants reported using the pullout method, while only 8.3% used a LARC. Some participants (23%) reported using more than one contraceptive method.

Table 2.

Patient Baseline Characteristics

Characteristic	Total (n = 240), % (n)	TIPS-Plusa(n = 114), % (n)	TIPS-Basicb(n = 126), % (n)	p
Race
American Indian or Alaska Native	1.3 (3)	1.8 (2)	0.8 (1)
Asian	2.1 (5)	2.6 (3)	1.6 (2)
Black/African-American	18.3 (44)	17.5 (20)	19.1 (24)
Hispanic or Latina	1.3 (3)	0.9 (1)	1.6 (2)
White	70.8 (170)	71.9 (82)	69.8 (88)
Multiracial	5.8 (14)	5.3 (6)	6.3 (8)
Other	0.4 (1)	0 (0)	0.8 (1)
Fisher's exact p-value				0.966
Age, mean (SE)	22.4 (0.24)	22.2 (0.32)	22.6 (0.34)
Two-sample t-test p-value				0.427
Education
<12th Grade	9.6 (23)	9.7 (11)	9.5 (12)
Finished high school	19.6 (47)	20.2 (23)	19.1 (24)
Some college	40.4 (97)	43.9 (50)	37.3 (47)
College degree or higher	30.4 (73)	26.3 (30)	34.1 (43)
Chi-squared p-value				0.598
Relationship status
Single	31.3 (75)	29.0 (33)	33.3 (42)
Dating more than one person	4.2 (10)	3.5 (4)	4.8 (6)
Dating one person/in a serious relationship	62.1 (149)	64.9 (74)	59.5 (75)
Married	1.3 (3)	0.9 (1)	1.6 (2)
Married with other sex partners than husband	0.8 (2)	1.8 (2)	0 (0)
Other	0.4 (1)	0 (0)	0.8 (1)
Fisher's exact p-value				0.616
Lifetime IPV	47.5 (114)	47.4 (54)	47.6 (60)
Chi-squared p-value				0.969
Type of contraception
Oral contraceptive pills	32.5 (78)	31.6 (36)	33.3 (42)	0.772
Condoms	38.3 (92)	44.7 (51)	32.5 (41)	0.052
Depo-Provera	9.2 (22)	10.5 (12)	7.9 (10)	0.487
Pullout	29.6 (71)	26.3 (30)	32.5 (41)	0.291
Vaginal ring	2.1 (5)	2.6 (3)	1.6 (2)	0.671
Patch	0.42 (1)	0.9 (1)	0 (0)	0.475
IUD	6.3 (15)	7.0 (8)	5.6 (7)	0.640
Emergency contraception	2.6 (6)	0.9 (1)	4.0 (5)	0.216
Implant	2.1 (5)	4.4 (5)	0 (0)	0.023
LARC	8.3 (20)	11.4 (13)	5.6 (7)	0.102
Hidden	20 (48)	22.8 (26)	17.5 (22)	0.301

Hidden method: LARC, Depo-Provera, or emergency contraception.

^aTIPS-Plus = providers received tailored scripts to discuss harmful partner behaviors with their patients + patients received psychoeducational messages.

^bTIPS-Basic = providers received tailored scripts to discuss harmful partner behaviors with their patients.

LARC, long-acting reversible contraception; SE, standard error; TIPS, Trauma-Informed Personalized Scripts.

Victimization outcomes

Intimate partner violence

At baseline, lifetime IPV prevalence was 47.5% and recent (past 3 months) IPV prevalence was 12.5%.²⁰ At follow-up, recent IPV prevalence was 6.9%. Over the course of the study, 25.9% of participants reported ending a relationship in the past 3 months, of whom 30.9% (7.9% of the total sample) did so because she felt it was abusive. Overall, mean IPV summary scores did not significantly decrease from baseline to follow-up (T1 = 0.16, T2 = 0.09, p = 0.061). We observed no difference in differences in recent IPV over time by treatment arm (p = 0.123) (Table 3). In adjusted ordinal logistic regression models, we observed no significant differences in intervention effects for IPV by treatment arm at follow-up (adjusted odds ratio [aOR] = 2.00, 95% confidence interval [CI] = 0.63 to 6.37) or effects of IPV discussion (i.e., participants reported having a discussion with provider about IPV during visit) in post hoc analyses (aOR = 0.81, 95% CI = 0.20 to 3.27) (Table 4).

Table 3.

Change in Outcomes Between Baseline and Follow-Up Between the Trauma-Informed Personalized Scripts-Plus and Trauma-Informed Personalized Scripts-Basic

	Overall sample			TIPS-Plusa		TIPS-Basicb		Intervention comparison
	T1	T2	pc	T1	T2	T1	T2	T2–T1	pd
Summary scores (mean, 95% CI)
IPV summary scoree	0.16 (0.10 to 0.22)	0.09 (0.04 to 0.14)	0.061	0.14 (0.07 to 0.21)	0.13 (0.04 to 0.22)	0.17 (0.08 to 0.27)	0.05 (0.00 to 0.10)	−0.05 (−0.13 to 0.24)	0.123
Reproductive coercion summary scoree	0.08 (0.04 to 0.13)	0.02 (0.00 to 0.04)	0.028	0.05 (0.01 to 0.09)	0.02 (−0.01 to 0.05)	0.11 (0.04 to 0.18)	0.02 (−0.01 to 0.04)	−0.06 (−0.10 to −0.01)	0.150
Fear summary scoree	0.29 (0.21 to 0.38)	0.19 (0.12 to 0.26)	0.110	0.24 (0.13 to 0.35)	0.18 (0.08 to 0.29)	0.34 (0.21 to 0.47)	0.20 (0.11 to 0.29)	−0.06 (−0.14 to 0.02)	0.136
Percentages (n)
Contraceptive usef	69.6 (167)	87.9 (182)	<0.0001	74.6 (85)	90.1 (91)	65.1 (82)	86.0 (91)	0.14 (0.08 to 0.21)	0.605
Use of LARC methodsg	8.3 (20)	20.9 (39)	0.0001	11.4 (13)	19.6 (19)	5.6 (7)	22.2 (20)	0.11 (0.06 to 0.16)	0.169
Use of hidden contraceptivesh	20.0 (48)	38.5 (75)	<0.0001	22.8 (26)	32.7 (32)	17.5 (22)	44.3 (43)	0.16 (0.09 to 0.23)	0.053
Consistent condom usei	19.2 (41)	17.8 (32)	0.572	19.6 (21)	17.4 (16)	18.7 (20)	18.2 (16)	−0.02 (−0.09 to 0.04)	0.986
At least once, partner did not use condom when she wanted to	25.7 (55)	20.0 (36)	0.164	22.4 (24)	18.5 (17)	29.0 (31)	21.6 (19)	−0.06 (−0.14 to 0.02)	0.786

Outcomes evaluated over past 3 months.

^aTIPS-Plus = providers received tailored scripts to discuss harmful partner behaviors with their patients + patients received psychoeducational messages.

^bTIPS-Basic = providers received tailored scripts to discuss harmful partner behaviors with their patients.

^cp-Values are for McNemar's test for paired proportions within treatment arm, between T1 and T2. Sample size varies slightly by outcome due to small amount of missing data.

^dp-Values are for two-sample t-test comparing differences of means over time by treatment arm.

^eSummary scores are compared using Signed Rank-Sum tests, mean summary score presented with 95% CIs.

^fContraceptive use: oral contraceptive pills, condoms, Depo-Provera, patch, vaginal ring, IUD, implant, emergency contraception.

^gLARC methods: IUD, implant.

^hHidden methods: Depo-Provera, IUD, implant, and emergency contraception.

ⁱOnly among participants who reported they had vaginal sex in the past 3 months (n = 214).

CI, confidence interval; T1, baseline; T2, follow-up at 4–6 months.

Table 4.

Unadjusted and Adjusted Odds Ratios to Measure Intervention Effects for Trauma-Informed Personalized Scripts-Plus Compared with Trauma-Informed Personalized Scripts-Basic at 4–6 Months and Post Hoc Analyses of the Effects of Discussing Intimate Partner Violence During the Baseline Visit

Outcomes	Intervention comparison		Post hoc analysis
			Effects of discussing IPV during baseline visit
	OR (95% CI)	aOR (95% CI)a	OR (95% CI)	aOR (95% CI)b
Ordinal logistic regression
Recent IPV	2.42 (0.79 to 7.47)	2.00 (0.63 to 6.37)	0.65 (0.19 to 2.15)	0.81 (0.20 to 3.27)
Recent reproductive coercion	1.03 (0.14 to 7.46)	1.00 (0.13 to 7.66)	N/A	N/A
Recent fear of partner	1.06 (0.46 to 2.41)	0.87 (0.35 to 2.20)	0.69 (0.26 to 1.85)	0.66 (0.20 to 2.16)
Received information on sexual and partner violence related servicesc	0.85 (0.48 to 1.51)	0.90 (0.48 to 1.69)	1.29 (0.61 to 2.72)	1.18 (0.51 to 2.71)
Any use of sexual and partner violence-related servicesc	0.93 (0.18 to 4.74)	0.74 (0.13 to 4.22)	1.31 (0.15 to 11.62)	0.92 (0.09 to 9.56)
Harm reduction behaviorsc	0.82 (0.18 to 3.75)	0.83 (0.17 to 3.98)	1.43 (0.17 to 12.25)	1.89 (0.19 to 18.99)
Logistic regression
Any use of contraceptives	1.28 (0.51 to 3.20)	1.40 (0.46 to 4.27)	4.15 (1.43 to 12.09)	5.59 (1.45 to 21.53)
Use of LARCs	0.61 (0.26 to 1.42)	0.54 (0.22 to 1.34)	5.29 (1.12 to 25.08)	5.81 (0.97 to 34.99)
Use of hidden contraceptives	0.51 (0.27 to 0.99)	0.52 (0.26 to 1.05)	1.91 (0.81 to 4.55)	1.37 (0.53 to 3.53)
Consistent condom usee	0.96 (0.39 to 2.37)	1.01 (0.36 to 2.88)	0.60 (0.20 to 1.75)	0.36 (0.10 to 1.34)
Linear regressiond
Harm reduction self-efficacy	−1.02 (−3.85 to 1.80)	0.01 (−3.00 to 3.01)	−0.55 (−4.14 to 3.05)	−0.02 (−4.06 to 4.03)

OR and aOR are the values associated with the intervention arm; unadjusted models control for baseline values. Recent: past 3 months. TIPS-Plus = providers received tailored scripts to discuss harmful partner behaviors with their patients + patients received psychoeducational messages; TIPS-Basic = providers received tailored scripts to discuss harmful partner behaviors with their patients.

^aAdjusted for outcome at baseline, age, race, education, relationship status.

^bAdjusted for outcome at baseline, age, race, education, relationship status, IPV at baseline, and intervention group.

^cBaseline values not measured.

^dLinear regression, mean score, adjusted models for general self-efficacy at baseline, age, race, education, and relationship status.

^eOnly among participants who reported they had vaginal sex in the past 3 months (n = 214).

aOR, adjusted odds ratio; OR, odds ratio.

Reproductive coercion and fear of partner

Mean reproductive coercion summary scores decreased (T1 = 0.08, T2 = 0.02, p = 0.028) with no difference in differences over time by treatment arm (p = 0.150) (Table 3). We observed no statistically significant difference in mean fear summary scores over time (T1 = 0.29, T2 = 0.19, p = 0.110), nor did we find any differences in differences over time by treatment arm (p = 0.136) (Table 3). Furthermore, we did not find significant intervention effects at follow-up (reproductive coercion aOR = 1.00, 95% CI = 0.13 to 7.66; fear aOR = 0.87, 95% CI = 0.35 to 2.20) (Table 4).

Contraceptive outcomes

Overall use of any contraception increased from 69.6% to 87.9% (p < 0.001). Use of hidden contraceptive methods improved from 20.0% to 38.5% (p < 0.001), as did LARC usage from 8.3% to 20.8% (p < 0.001) (Table 3). To explore whether this finding was due to women seeking care explicitly for contraception, we stratified the models by whether participants self-reported contraceptive use as their reason for visit. This stratified model showed similar intervention effects (results not shown). Consistent condom use did not change, nor did the percentage of participants who reported not using a condom at least once when they wanted to. There was no difference in differences for any condom or contraceptive use outcomes over time by treatment arm. Furthermore, we detected no significant differences by treatment arm in adjusted analyses (Table 4). In post hoc analyses, individuals who reported their providers discussed IPV during their clinic visit at baseline had higher odds of using any contraceptive method at follow-up (aOR = 5.59, 95% CI = 1.45 to 21.53) (Table 4).

Knowledge and utilization outcomes

Harm reduction self-efficacy

We found no intervention effects for harm reduction self-efficacy (adjusted β = 0.01, 95% CI = −3.00 to 3.01) (Table 4). Similarly, we detected no significant effects of IPV discussions on harm reduction self-efficacy in post hoc analyses (adjusted β = −0.02, 95% CI = −4.06 to 4.03).

Knowledge of services

At follow-up, 35.7% of all participants and 30.8% of those who reported recent IPV at baseline reported receiving information on sexual and partner violence-related services since baseline. We did not detect intervention effects at follow-up (aOR = 0.90, 95% CI = 0.48 to 1.69) or any effects of discussing IPV at baseline visit on this outcome at follow-up (aOR = 1.18, 95% CI = 0.51 to 2.71) (Table 4). Overall, mean decisional conflict score decreased (T1 = 15.15, T2 = 12.30, p = 0.011). Linear regression models showed no significant intervention effects (adjusted β = 3.16, 95% CI = −0.13 to 6.44, p = 0.060), when adjusting for baseline IPV and demographic factors.

Use of services and harm reduction behaviors

At follow-up, 2.5% of all individuals (TIPS-Plus = 1.9%, TIPS-Basic = 2.7%) and 3.3% (n = 1) of those who reported IPV at baseline reported using any sexual and partner violence-related service since baseline. We observed no intervention effects (aOR = 0.74 (95% CI = 0.13 to 4.22) or effects of IPV discussion in post hoc analyses (aOR = 0.92 (95% CI = 0.09 to 9.35) (Table 4). In addition, 3.3% of all patients (intervention = 2.9%, comparison = 3.6%) and 6.7% of those who endorsed IPV at baseline reported using a harm reduction strategy in the follow-up period. We detected no intervention effects (aOR = 0.83, 95% CI = 0.17 to 3.98) (Table 4).

Discussion

The TIPS study was designed to assess the impact of adding patient-activation messages to provider scripts, with the hypothesis that providing empowering messages to patients would increase the likelihood of discussions about IPV and reproductive coercion and greater uptake of harm reduction behaviors. We observed no significant intervention effects between participants who received provider scripts and those who received both patient-activation messages and provider scripts for any measured outcome.

In our analysis of patient-reported intervention implementation immediately postvisit, published elsewhere, we demonstrated that patient activation did not significantly increase the likelihood that a discussion on IPV/reproductive coercion occurred.²⁰ However, in post hoc analyses, presented here, we showed significantly improved outcomes among those who discussed IPV compared with those who did not. Therefore, it is likely that the provider scripts component of TIPS contributed more strongly to reductions in reproductive coercion and improvements in contraceptive use compared with the patient-activation component. Provider scripts show promise to reduce barriers that health care professionals face, including limited time, discomfort, and lack of training, through facilitating tailored discussions on harmful partner behaviors.^16,17

The lack of difference between arms may suggest that the patient-activation component may not be as critical as provider-led interventions. However, it is important to reflect that one-time patient-activation messages may not be sufficient, especially for patients who face fear of retaliation from abusive partners, judgment from health care providers, and stigma.^8,9 To overcome these obstacles, we need sustained engagement and multisector interventions that address societal power dynamics and challenge gender norms. While such larger scale interventions are designed and implemented, TIPS may be helpful to assist providers in increasing implementation of IPV/reproductive coercion assessments, potentially leading to reductions in victimization and improvements in sexual and reproductive health outcomes. It is also possible that patient activation may lead to changes over a longer time period, unobserved by our study with such a short follow-up interval.

Given the overall improvement in both study arms from baseline to follow-up, these results highlight the potential of tailored provider scripts—provided for all participants in this study—to reduce reproductive coercion victimization and increase uptake of effective contraception. However, it is possible that the improvement in contraceptive use between time points was due to the fact that individuals sought care at FP clinics. To examine this, we stratified results by those who reported that their reason for visit was contraception, and noted similar intervention effects, indicating that baseline care seeking does not fully explain the increase in contraceptive use. While it cannot be concluded that these differences are attributable to our study given we did not have a true control, TIPS provides preliminary evidence on how discussions about IPV between patients and health care providers may be associated with increased use of contraception, particularly LARCs.

To investigate this relationship further, we conducted exploratory post hoc analyses to determine the impact of IPV discussions at baseline on contraceptive use at follow-up. Our findings show statistically significant improvements in contraceptive use, LARC use specifically, among those who reported discussing IPV with their providers at baseline. By addressing IPV/reproductive coercion in the context of fear, safety, and harm reduction with all patients, participants may have felt empowered by the information that they could choose methods of contraception that their partners did not have to know about.^2,12 It is important to note that no statistically significant differences were observed with condom use; condom use is largely a partner behavior not addressed in the TIPS intervention. Further research is needed to better understand how provider conversations on healthy and unhealthy relationships may impact sexual and reproductive health outcomes more broadly.

TIPS emerges from over a decade of research on strategies to integrate IPV and reproductive coercion discussions into FP clinic visits.^2,3,12–14 Our results help strengthen the evidence from the ARCHES intervention, which provided an alternative to traditional screening approaches that rely heavily on patient disclosure of IPV. TIPS, like ARCHES, uses a more comprehensive approach through universal education, providing information and resources to all patients regardless of disclosure during clinic visits. Closely related is another approach, “CUES” (confidentiality, universal education and empowerment, and support), which is increasingly being promoted as an evidence-based strategy to facilitate patient–provider conversations about healthy relationships.²⁸ However, TIPS provides an added benefit to improve implementation through tailored scripts with exact phrasing, designed to overcome provider barriers of insufficient time and discomfort. Larger studies are needed with control clinics providing usual care (i.e., IPV screening only) and with audio-recorded visits to more rigorously evaluate the effectiveness of provider scripts in addressing IPV, reproductive coercion, and contraceptive uptake.

The study has limitations. The small sample size (n = 240) reduces generalizability beyond these FP clinic settings. Since previous IPV/reproductive coercion interventions were conducted at the clinics, no true control site was available, making it difficult to attribute our results to TIPS. We only followed participants for 4–6 months, and do not know whether results were sustained past this time frame. Finally, it is possible that providers who used tailored scripts to guide discussions on IPV/reproductive coercion during the visit were also more adept at counseling in general, including talking to patients about contraceptive options.

Conclusion

TIPS is a novel study that explores the added benefit of patient-activation messages together with tailored provider scripts to reduce victimization from IPV and reproductive coercion and improve contraceptive and condom use. We did not observe any difference between our treatment arms regarding the use of patient-activation messages, but provide preliminary evidence that patient–provider conversations, facilitated by brief provider scripts provided to all TIPS patients, may have broader implications on sexual and reproductive health.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

The research in this study was supported by the National Institute of Child Health and Human Development (R01HD064407 to E.M., K24HD075862 to E.M.), as well as by the National Center for Advancing Translational Science of the National Institutes of Health (TL1TR001858 to A.L.H.; principal investigator: Kevin Kraemer).