Objective
Long-acting reversible contraceptive (LARC) methods, such as the intrauterine device and implant, are highly effective but used by less than 10% of US women. The objective of our study was to improve LARC interest by enhancing clinic counseling.
Study Design
A quality improvement methodology was used to evaluate intrauterine device service delivery in 3 Chicago Title X clinics. To address identified barriers, we developed a theory-based iOS application (app) for patients to use in the clinic waiting room using human-centered design. The final prototype was tested in a randomized controlled trial in a Title X clinic with sexually active females (79% African American) under age 30 years. Our sample of 60 was chosen to detect an increase from 10% (baseline) to 45% (app intervention) in the proportion of patients expressing interest in discussing a LARC method during their clinic visit with 80% power and two-sided α = 0.05. After completing demographic and baseline knowledge questionnaires, participants were randomized 1:1 to intervention (app) or standard care arms. App users also completed a posttest. Our primary outcome was expressed interest in discussing a LARC method during the clinic visit. Secondary outcomes were contraceptive knowledge and LARC selection.
Results
App testers (n = 17) preferred interactive, visually appealing design and video testimonials. In the pilot randomized controlled trial (n = 52), app users had a significantly higher knowledge of contraceptive effectiveness ( P = .0001) and increased interest in the implant (7.1-32.1%, P = .02) after the intervention. Users were highly satisfied with the app. Staff reported no problems using the app in the clinic.
Conclusion
Use of a theory-based counseling app offers a novel method to optimize wait time while minimizing clinic flow disruption. Preliminary data demonstrate that app use was associated with improvements in patients’ contraceptive knowledge and interest in the implant.
Long-acting reversible contraceptive (LARC) methods (ie, the intrauterine device [IUD] and contraceptive implant) require few use behaviors and are therefore highly effective. Rates of LARC use in the United States are low overall (<10%) and especially so among young women, who exhibit the highest rates of unintended pregnancy. The lack of awareness of and misinformation about LARC may contribute to the low rates of use.
In a 2009 survey of a nationally representative, racially diverse sample of unmarried adults aged 18-29 years (n = 1800), 87% of the women had heard of the IUD and 52% the implant. However, knowledge was low: 56-80% of the respondents who had heard of the IUD knew “little or nothing” about it. Misconceptions about LARC were also highly prevalent. Thus, more effort is needed to provide accurate information and knowledge of LARC.
Recent reports have called for focused application of behavioral theory to interventions for contraceptive use. The transtheoretical model and its stages of change paradigm posit that in changing behaviors, individuals progress through sequential stages of precontemplation, contemplation, preparation, action, and maintenance and that transition between steps requires a core set of change processes. Patients presenting for contraception will most likely be in the precontemplation or contemplation stages for adopting LARC.
For precontemplation, the key change process is enhancing awareness. The transtheoretical model is often coupled with the theory of planned behavior, which postulates that behavioral intention is informed by the following: (1) attitude toward the behavior and beliefs about outcomes associated with performing the behavior (behavioral beliefs); (2) subjective norms and beliefs about those norms; and (3) perceived behavioral control, or the perception that one has the requisite skills and resources to perform the behavior. Thus, addressing common behavioral, normative, and control beliefs can help women advance from precontemplation or contemplation stages to action in adopting LARC.
Interactive, multimedia, computer-based programs provide a novel platform for delivering information. Tailored, theory-based, computer-based interventions have been effective in reducing repeat births among low-income, African American adolescent mothers, promoting adherence to antiretroviral medications, and supporting preventive health behaviors in vulnerable populations. In a recent randomized controlled trial (RCT), low-income family planning patients using a touch screen computer module were significantly more likely to choose highly effective contraception.
This paper describes the development and evaluation of an iOS application (app) for use in the clinic waiting room to increase LARC awareness and interest prior to the clinic visit.
Materials and Methods
Assessment of IUD service barriers
From January to May 2011, barriers to IUD services for young (15-29 years old) women in 3 Chicago Title X family-planning clinics were assessed, using failure modes effects and criticality analysis. This failure modes effects and criticality analysis is 1 well-developed risk assessment method to systematically assess failure modes in systems of care. A failure mode is a point in a system in which progress toward the desired endpoint is hindered. The failure analysis informs the redesign processes to mitigate the failures.
During 13 meetings with clinicians, staff, and patients, 33 failures in the processes of care that could undermine IUD service delivery were identified and further substantiated by clinic administrative data. It revealed that contraceptive counseling was limited by the lack of time, the nonuse of waiting room pamphlets, and failure to counsel all women on LARC. Finally, waiting room down time was recognized as a missed opportunity for patient engagement. We therefore opted to create a counseling app to promote interest in and awareness of LARC.
App development
The app development was guided by the principles of human-centered design, the theory of planned behavior, and our failure analysis findings, along with continued input from clinic staff and clinicians. Human-centered design is an innovative approach to designing interventions whereby the designer and end user are brought together early in the design process. Design solutions are created in an iterative process of rapid, low-fidelity prototyping (eg, pen and paper) with key stakeholder participation.
The goal was to create an app that accomplished the following: (1) reflected the tenets of the theory of planned behavior; (2) filled gaps in LARC awareness while providing information on the full range of contraceptive options; (3) appealed to a variety of learning styles with text, video, and imagery; (4) featured unbiased, evidence-based content; and (5) complemented existing clinic flow.
Simple app prototypes were created using Microsoft PowerPoint (Richmond, WA) to test the individual aspects of the app in a controlled setting with targeted users by observing, recording, and taking notes about the user’s experience. From August 2012 to March 2013, iterative rounds of usability testing were conducted at the University of Chicago and at another Chicago-based family-planning clinic.
Testers (n = 17) were African American women demographically similar to patients of the Title X clinics included in the failure analysis. As testers used the app, they narrated their experience and gave their opinions regarding the specific areas being studied. Design adjustments were made following each round of testing. Once tester feedback became consistent, a programming team at the University of Chicago built a native iOS app prototype. Video testimonials were filmed with 2 young women recruited from the study clinics who were using the IUD or implant. Videos were inserted into the tablet prototype and assessed in the final rounds of usability testing.
App evaluation
Procedure
An RCT was conducted to evaluate the app’s impact on contraceptive knowledge, LARC interest, and selection of a LARC method. Eligible participants were sexually experienced, English-speaking women, aged 15-30 years presenting for contraceptive services at a Chicago Title X clinic. Exclusion criteria were currently pregnant, desiring pregnancy within the next 12 months, currently using LARC, scheduled for a LARC-related visit, and reliance on male partner sterilization for contraception. A sample size of 60 was chosen to detect an increase from 10% (baseline) to 45% (app intervention) in the proportion of patients expressing interest in discussing a LARC method during their clinic visit with 80% power and 2-sided α = 0.05 based on a checklist of methods.
Potential study participants were identified by clinic staff, based on age and appointment type, and referred to on-site research staff who then confirmed eligibility, explained the study, and obtained informed consent. All participants completed a baseline survey after which they were randomized to the intervention (app) or standard care. Randomization was performed in advance using a random numbers generator ( www.random.org ), with assignments concealed in sequentially numbered opaque envelopes.
Women randomized to the intervention arm were given a tablet computer programmed with the app and instructions to use the app for up to 15 minutes. After completing a brief postintervention survey, participants returned to the clinic waiting room for standard care. Standard care included contraceptive counseling by a clinic counselor followed by a visit with a nurse practitioner to receive (or be prescribed) their chosen method. Participants randomized to standard care returned to the clinic waiting room following completion of the baseline survey for the same sequence of care. All participants received $10 for participation.
After the visit, clinic staff reviewed the participant’s medical chart and her chosen contraceptive method, reconfirmed the inclusion criteria, and relayed results to the research team. Clinic counselors and clinicians were blinded to participants’ treatment assignment throughout. Study data were managed using REDCap electronic data capture tools hosted at the University of Chicago. On March 28, 2013, the study, number 13-0145, was approved by the University of Chicago Institutional Review Board, which also provided a waiver of parental consent for minors under the age of 18 years.
Measures
The baseline survey collected demographic information; mobile phone use; sexual and reproductive health history (sexual experience and pregnancy history); and contraception-related variables (awareness of methods, knowledge of effectiveness, past and current method use, self-efficacy, method interest). Contraceptive knowledge was measured by 3 questions in which the participants were presented with a method pair (eg, oral contraceptives [OCs] and condoms) and asked which method was more effective.
Individual question scores (1, correct; 0, incorrect) were combined into a single summary score ranging from 0 to 3. A 5-item contraceptive self-efficacy scale previously validated by Galavotti et al (α = 0.84) with responses measured on 5-point Likert scales (anchors: not at all confident to extremely confident) was also administered. Participants indicated all of the contraceptive methods that they were interested in discussing during counseling. The postintervention survey reassessed contraceptive knowledge and the indicated methods that they wanted to discuss during counseling and gauged intervention satisfaction. Participants indicated their agreement with 7 statements (eg, “The app was easy to use,” “I would recommend to a friend”) using 5-point Likert scales (strongly disagree to strongly agree). Selection of contraceptive method was ascertained through chart review 1 month after the visit. The method selected by the participant, as well as how the method was provided (dispensed directly, prescribed, or referred for follow-up visit), was obtained.
Statistical analysis
Descriptive statistics were used to assess the satisfaction with the app. Baseline survey responses and the selection of a LARC method were compared using the Wilcoxon-Mann-Whitney test for ordinal or continuous variables and the χ 2 test statistic for categorical variables between the 2 groups. Within the intervention arm, pre-post differences in contraceptive knowledge and LARC interest were evaluated using a Wilcoxon’s signed-rank test (if ordinal) or McNemar’s χ 2 (if binary). For self-efficacy, summary scores ranging from 0 to 25 were computed; the interitem correlation was assessed using a standardized Cronbach’s alpha. We performed a per-protocol analysis, excluding ineligible patients who were mistakenly enrolled. However, for the outcomes of LARC interest and LARC selection, we also present findings from an intent-to-treat (ITT) analysis.
Results
App development
The primary app design challenge faced by the research team and clinic staff was to present content that was LARC focused yet comprehensive for all contraceptive methods. The app needed to provide an enriching experience for clients who were interested in LARC as well as those seeking non-LARC methods. The extent to which the app screen directed the user from 1 screen to the next vs allowing free navigation was also taken into consideration. The final app prototype began with prescribed pathways in which all users see a landing page comparing the effectiveness of all methods and then view LARC peer testimonials but allowed for self-navigation of all other information ( Figure 1 ).
The human-centered design process revealed a number of desired features. Testers preferred contraceptive effectiveness conveyed in absolute numbers (eg, <1 pregnancy per 100 women) rather than percentages (eg, >99% effective). They rated peer testimonials very highly and related well to the stories. Testers were presented with multiple options for the information categories to be included on all method-specific pages. The most popular categories (“How does it work,” “What can I expect,” and “What will he think?”) were incorporated into the final design.
RCT
Study flow
Research staff screened 73 presumptively eligible patients ( Figure 2 ) referred by the clinic staff. Of these, 4 declined to participate and 7 were ineligible because of an age older than 30 years. On subsequent chart review by clinic staff following the visit, an additional 10 participants were excluded from analysis because of missing or ineligible age (n = 4), current implant use (n = 3), currently desiring pregnancy (n = 2), and presenting for IUD insertion (n = 1). Thus, the final per-protocol sample was comprised of 52 participants (intervention, n = 28; standard care, n = 24).
Participant characteristics
Participants had a median age of 22.5 years ( Table 1 ). The majority (78.9%) self-identified as black/African American with half reporting their relationship status as single. Most of the participants had mobile phones and had experience using mobile apps. Almost all participants (98%) were sexually experienced, with a median age of first sex of 16 years. Pregnancy histories were similar between the 2 arms, although intervention participants were significantly younger at their first pregnancy. Method-specific contraceptive awareness was high for non-LARC methods (>85%, data not shown) and 69% and 50% for the IUD and implant, respectively. Less than half (45.1%) reported currently using birth control, with condoms being the most commonly used method.
| Characteristic | Total (n = 52) | Intervention (n = 28) | Standard care (n = 24) | P value |
|---|---|---|---|---|
| Sociodemographic | ||||
| Age, y (median, range) | 22.5 (16–30) | 23 (17–30) | 21.5 (16–30) | .36 |
| Race/ethnicity | .62 | |||
| African American, non-Hispanic | 41 (78.9) | 23 (82.1) | 18 (75.0) | |
| White, non-Hispanic | 6 (11.5) | 3 (10.7) | 3 (12.5) | |
| Hispanic | 3 (5.8) | 2 (7.1) | 1 (4.2) | |
| Multiple/other/unknown | 2 (3.9) | 0 (0.0) | 2 (8.3) | |
| Relationship status | .12 | |||
| Living with partner or spouse | 4 (7.7) | 4 (14.3) | 0 (0.0) | |
| Dating/casual partner(s) | 22 (42.3) | 13 (46.4) | 9 (37.5) | |
| Single | 26 (50.0) | 11 (39.3) | 15 (62.5) | |
| Education status | .42 | |||
| High school graduate or less | 11 (21.2) | 4 (14.3) | 7 (29.2) | |
| Some college | 27 (51.9) | 16 (57.1) | 11 (45.8) | |
| College degree or higher | 14 (26.9) | 8 (28.6) | 6 (25.0) | |
| Owns mobile phone | 48 (92.3) | 27 (96.4) | 21 (87.5) | .32 |
| Uses mobile apps | 44 (91.7) | 24 (88.9) | 20 (95.2) | .62 |
| Pregnancy/sexual history | ||||
| Ever had sexual intercourse | 51 (98.1) | 27 (96.4) | 24 (100.0) | 1.00 |
| Age at first intercourse, y (median) | 16 | 16 | 16 | .58 |
| Pregnancy history | .56 | |||
| Never been pregnant | 31 (59.6) | 15 (53.6) | 16 (66.7) | |
| 1 pregnancy | 9 (17.3) | 5 (17.9) | 4 (16.7) | |
| ≥2 pregnancies | 12 (23.1) | 8 (28.6) | 4 (16.6) | |
| Age at first pregnancy, y, median | 19 | 17 | 20 | .034 |
| Parity | .12 | |||
| No births | 31 (73.8) | 15 (71.4) | 16 (76.2) | |
| 1 birth | 5 (11.9) | 1 (4.8) | 4 (19.1) | |
| ≥2 births | 6 (14.3) | 5 (23.8) | 1 (4.8) | |
| Contraception: awareness | ||||
| Ever heard of IUD | 36 (69.2) | 20 (71.4) | 16 (66.7) | .71 |
| Ever heard of implant | 26 (50.0) | 13 (46.4) | 13 (54.2) | .58 |
| Contraception: past use | ||||
| IUD | 2 (3.9) | 2 (7.1) | 0 (0.0) | .49 |
| Implant | 1 (1.9) | 0 (0.0) | 1 (4.2) | .46 |
| DMPA | 16 (30.8) | 4 (14.3) | 12 (50.0) | .005 |
| OCs | 29 (55.8) | 17 (60.7) | 12 (50.0) | .44 |
| Patch | 12 (23.1) | 8 (28.6) | 4 (16.7) | .31 |
| Ring | 8 (15.4) | 4 (14.3) | 4 (16.7) | .81 |
| Condom | 49 (94.2) | 27 (96.4) | 22 (91.7) | .59 |
| Contraception: current use (primary method) | ||||
| DMPA | 4 (7.8) | 1 (3.6) | 3 (13.0) | .50 |
| OCs | 6 (11.8) | 4 (14.3) | 2 (8.7) | |
| Condom | 13 (25.5) | 6 (21.4) | 7 (30.4) | |
| None | 28 (54.9) | 17 (60.7) | 11 (47.8) | |
| Contraception: intention | ||||
| Intends to start in next 30 d | .51 | |||
| Yes | 36 (70.6) | 20 (74.1) | 16 (66.7) | |
| No | 4 (7.8) | 1 (3.7) | 3 (12.5) | |
| Unsure | 11 (21.6) | 6 (22.2) | 5 (20.8) | |
| Contraception: self-efficacy | ||||
| Five-item score: contraceptive use, median | 20 | 19 | 20 | .17 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
