Materials and Methods
This study was a randomized (1:1) parallel-group, single-blind trial comparing the rates of LARC initiation immediately after abortion between women assigned to a brief, theory-based video intervention and those assigned to a control video, both prior to routine contraceptive counseling. All study protocols and procedures were approved by The University of Chicago Biological Sciences Division Institutional Review Board.
Video intervention development
The video intervention was developed using the TTM as a framework, assuming that most abortion patients would be in TTM stages of precontemplation or contemplation for LARC uptake. The video featured messages delivered by a health care provider as well as peers (real women sharing their postabortal LARC experiences). Consistent with the TTM, video content was designed to facilitate LARC initiation by increasing women’s awareness of LARC, helping women weigh the pros and cons of LARC use and gain self-efficacy for using LARC in the postabortal period.
The video intervention was comprised of 3 segments and delivered on an iPAD. The first segment featured a health care provider delivering basic information about the 3 LARC methods (ie, levonorgestrel IUD [LNG-IUD], copper IUD, and contraceptive implant). The second and third segments featured narrative comments from patients who had used a LARC following a surgical abortion.
To create the health care provider segment, a physician was invited to read a scripted text describing the mechanisms of action, side effects, and efficacy of the 3 LARC methods. The physician showed replicas of each device and emphasized the safety, ease of use, and effectiveness of all 3 methods in the immediate postabortal period.
To create the patient narrative comments, women who were demographically similar to the study population and who had initiated LARC immediately after abortion were recruited to participate in the development of the intervention. Three women agreed to participate. Patient narrative comments were elicited with interview questions scripted in accordance with the TTM. Women discussed how and why they decided to use LARC, their impression of the insertion procedure, and their overall experience with the method including how they had managed any negative aspects of LARC use.
The videos were then edited by 2 videographers, resulting in 2 unique versions of each segment, for a total of 8 video segments. The physician segments were each 4 minutes in length, whereas the patient narratives ranged from 2 to 4 minutes in length. These 8 segments were pilot tested for acceptability, likability, and length by 13 women presenting for a medical abortion follow-up appointment at a reproductive health clinic in Chicago, IL.
This population was chosen because it was demographically similar to the RCT study population, yet the women had more time to participate than women presenting for surgical abortion. The participants viewed all 8 video segments and then gave written and verbal feedback about each video. Based on the pilot testing, 3 segments were selected and combined to create the final video intervention: the health care provider and 2 patient narratives, one about the LNG-IUD and one about the implant. The copper IUD was described in the provider segment; however, to keep the intervention brief, a testimonial from a woman using a copper IUD was not incorporated into the final video. The final video was 7 minutes in duration. The individual video segments are available for viewing online. Participants received $50 for creating the video and $10 for pilot testing.
Setting and participants
The video intervention was evaluated in an RCT conducted between June and September 2013 in a free-standing clinic in Chicago, IL. The study clinic provides more than 5000 surgical abortions annually. Approximately one-third of women seen at the study clinic have publicly funded health insurance (Medicaid), approximately one-third have private insurance, and the remainder are uninsured. Medicaid recipients are eligible for all contraceptive methods without copay, and privately insured patients receive contraception as stipulated by their insurance policy, many of whom in 2013 had copays or deductibles. Uninsured women receive 1 pack of pills, patch, or ring at their abortion visit free of charge.
Inclusion criteria for the study were age 18-29 years, presenting for a surgical abortion, and not desiring pregnancy in the next 12 months. Exclusion criteria were nonviable or anomalous pregnancy, pregnancy as a result of sexual assault, or an inability to speak, read, and/or write in English.
Study procedures
Clinic personnel informed eligible women about the study when they presented for standard blood draw before the abortion. Women who expressed an interest in study participation were referred to 1 of 3 research team members. We did not track the proportion of patients who were not referred to the study team or their demographics. Research activities took place in a private counseling room in the clinic. There the research team member verified eligibility, described the study in detail, and obtained written informed consent.
Immediately following informed consent, all participants completed a 17 question baseline survey administered on a tablet computer using REDCap software (version 5.8.2, 2014; Vanderbilt University, Nashville, TN). The baseline survey consisted of demographic questions and obstetric and contraceptive history. RANDOM.org ( www.random.org ) was used to create randomization, determining a 1:1 sequence allocation constrained by blocks of 10. This sequence was entered into RedCap, which automatically performed randomization following completion of the baseline survey. Participants were randomized to the intervention or control video. The control video was the same length as the intervention video and featured a doctor discussing stress management. The research team member was present in the room while the participants viewed the videos.
After viewing their assigned video, all participants underwent usual care, including contraception and abortion counseling by clinic staff (during which time all contraceptive methods were discussed with the participant) and then underwent the abortion procedure. Following the abortion procedure, all participants completed a 5 item questionnaire that addressed their overall satisfaction with the contraceptive counseling and their perceived autonomy in their contraceptive decision making. All questions were answered using a 5 point Likert scale (1, strongly disagree; 5, strongly agree). Contraceptive method selection was determined by chart review. All contraceptive methods were provided at no cost to ensure that all study participants would have equal access to LARC, regardless of insurance status or whether insurance paid for contraception. Other women were provided a 1 month supply of the contraceptive pill, patch, or ring or a depot medroxyprogesterone acetate injection. Each of these was provided the participant free of charge.
Counselors and clinicians were blinded to the study arm allocation. Study participants were instructed not to discuss the video with clinic staff. To assess blinding procedures, 10% of the clinic counselors were randomly assigned to complete a brief questionnaire asking which arm of the study they believed the participant to have been allocated and their level of certainty in their belief.
Statistical analysis
The primary outcome was the initiation of LARC immediately after the abortion. In 2011 (at the time of the study and the most recent year in which complete data were available), 6.3% of surgical abortion patients at the study site had initiated a LARC. Sample size was calculated based on the assumption that rates of LARC initiation immediately after the abortion would be 6% in the control arm and 21% in the intervention arm. To detect a 15% increase in the LARC initiation (from 6% to 21%), using a 2-sided alpha P = .05 and 80% power, 188 subjects were needed (94 in each arm). Because the method initiation was assessed on the same day of the procedure, a loss to follow-up was not anticipated. However, an interim analysis revealed unexpected protocol deviations, and thus, the decision was made to overrecruit 5 additional participants for the potential loss to follow up.
An intent-to-treat analysis was performed, analyzing all participants according to the arm to which they were randomized. Demographic characteristics between groups were compared using Student t tests for continuous variables and χ 2 for categorical variables. Initiation of a LARC immediately after the abortion was compared using a χ 2 test. χ 2 and Student t tests were used to examine bivariate associations between LARC initiation and variables including age, race, parity, prior abortion, income, education, and health insurance status. Satisfaction scores were calculated (means [SD]) for each item and were compared across group assignment using a Student t test and confirmed using Wilcoxon signed-rank test. A kappa statistic was calculated to evaluate for successful blinding. All data were analyzed using Stata/SE version 11.2 (StataCorp, College Station, TX).
Because of unexpectedly high rates of LARC initiation in both study arms, a post-hoc analysis was conducted to evaluate selection bias by comparing rates of LARC initiation among our study participants to rates among nonparticipants of eligible age during the same period as the trial enrollment. Additionally, the study site’s rates of LARC initiation among all potentially eligible women 2 months prior and 2 months following study enrollment were calculated and compared with trial participants.
Results
A total of 201 women were screened for enrollment and 193 were randomized. Of the 8 who were screened but not randomized, 3 were ineligible because of a desire for pregnancy within the next year, 3 declined participation, 1 was lost to follow-up after the screening, and 1 did not have time to complete the study prior to her abortion procedure. Two participants were excluded from analysis after randomization due to ineligible age and did not complete the abortion procedure. This yielded a final sample of 191 participants: 96 in the intervention arm and 95 in the control arm ( Figure 1 ). Demographic characteristics did not differ between the two groups ( Table 1 ). Fifty-three percent of the study participants were African American. Approximately half (51%) had an annual income less than $10,000. The majority (93%) was single. One-third (32%) was uninsured. Half (51%) had had a prior abortion.
| Demographic | Intervention (n = 97) a | Control (n = 95) |
|---|---|---|
| Mean age (SD) | 23.1 (3.1) | 23.6 (3.0) |
| Race or ethnicity | ||
| African American | 53 (54.6) | 49 (51.6) |
| White | 22 (22.7) | 25 (26.3) |
| Hispanic | 17 (17.5) | 13 (13.7) |
| Other b | 5 (5.2) | 6 (6.32) |
| Not reported | 0 (0) | 2 (2.1) |
| Annual income | ||
| <$10,000 | 46 (47.4) | 51 (54.3) |
| $10,001–$30,000 | 30 (30.9) | 26 (27.7) |
| $30,001–$50,000 | 10 (10.3) | 12 (12.6) |
| >$50,000 | 11 (11.3) | 6 (6.3) |
| Insurance | ||
| None | 30 (31.3) | 30 (32.6) |
| Public | 36 (37.5) | 33 (35.9) |
| Private/other | 30 (31.3) | 29 (31.5) |
| Education, y | ||
| <12 | 4 (4.1) | 1 (1.1) |
| 12 | 44 (45.4) | 41 (43.2) |
| >12 | 49 (50.5) | 53 (55.8) |
| Mean gestational age, d | 59.6 | 60.7 |
| Parity ≥1 | 74 (76.3) | 78 (82.1) |
| Prior abortion (yes) | 52 (53.6) | 45 (47.4) |
| Single, never married | 87 (89.7) | 92 (96.8) |
a Includes participant who did not have abortion procedure
b includes Asian, American Indian, and self-described other.
LARC initiation, satisfaction, and blinding
A total of 106 women (55.5%) chose a LARC immediately after the abortion. The rate of LARC initiation did not differ significantly between study arms, with 59.3% in the intervention arm and 51.6% in the control arm choosing one of these methods ( P = .27) ( Table 2 ). The LNG-IUD was the most commonly selected contraceptive method in both groups; 40.6% in the intervention and 33.7% in the control groups chose this method ( Table 2 ). No significant associations between LARC initiation and age, race, parity, prior abortion, income, education, or health insurance status were found.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
