Materials and Methods
In this prospective observational study, all adolescents who were attending the Colorado Adolescent Maternity Program (CAMP) and who delivered at the University of Colorado Hospital over the 18 month period of June 1, 2008, to Nov. 30, 2009, were eligible for inclusion in the study cohort. All of the patients expressed the desire to prevent pregnancy for at least 1 year after delivery.
Exclusion criteria included contraindications to etonogestrel use, relinquishing the child for adoption, stillbirth, being delivered at a different hospital, and having no postpartum visits at the CAMP. We chose to exclude women who did not take home infants because they are likely to experience different motivators to prevent or obtain another pregnancy than those who do.
CAMP is a comprehensive, multidisciplinary, adolescent-oriented program that provides prenatal, delivery, postnatal, and infant care to young women aged 13-22 years. The program emphasizes consistent contraceptive use and goal-oriented future planning regarding school completion, job training, and parenting. The importance of consistent contraceptive use is emphasized at prenatal and postpartum visits. Contraceptive counseling begins at the first prenatal visit and is discussed at most subsequent visits by the midwife and the case manager. Counseling occurs in both individual and group settings and is supported by printed materials. During the period of this study, providers emphasized that the subdermal implant could be received immediately after delivery as part of a grant-funded initiative.
Each patient in CAMP is encouraged to have a firm contraceptive plan by 32 weeks’ gestation; however, group assignment was determined by whether the patient had an implant placed before hospital discharge. After delivery, patients continue care in the CAMP clinic at which the mothers and children receive follow-up care together. Adolescent mothers thus are seen more frequently than in traditional care models and have many opportunities to initiate contraception. Participants in CAMP provide informed consent for noninvasive procedures that include medical record review, surveys, and phone contact under an institutional review board–approved protocol that allows minors to consent for themselves.
During the period of this study, young women who were enrolled in CAMP who elected to have an immediate subdermal implant and who received this implant before hospital discharge were included in the IPI group. A few participants elected immediate placement but were unable to receive it because of a lack of provider availability; those participants had placement within 4 weeks after delivery and were included in the IPI group. Those who initiated any other contraceptive method according to standard clinical protocols were included in the comparison group.
The comparison group consisted of participants who chose no contraception, condoms, depot medroxyprogesterone acetate, and progestin-only pills initiated at any time after delivery, combined hormonal contraception (pills, patch, ring) started at any time 4 or more weeks after delivery, implant insertion at 4 or more weeks after delivery, and levonorgestrel-intrauterine system or copper-T 380A (IUD) insertion any time 6 or more weeks after delivery. Depot medroxyprogesterone acetate and IUDs were not available prior to hospital discharge during the study period. All participants remained in their initially assigned study groups, regardless of future contraceptive method discontinuation or change.
The manufacturer currently recommends insertion of the etonogestrel implant in lactating women after the fourth postpartum week. All participants who elected to follow the immediate etonogestrel implant protocol were informed that this practice differs from the package insert recommendations.
All participant demographic and encounter-related variables are maintained in the Electronic Report on Adolescent Pregnancy. Participants’ electronic medical records were reviewed to determine contraceptive use, discontinuation of contraceptive methods, and pregnancy at various time points of interest. Participants with incomplete data were contacted by telephone. Up to 3 attempts were made. Participants who were reached by phone were administered a standard questionnaire that assesses birth control method continuation, pregnancy occurrences, and outcomes. All pregnancies that occurred within the 36-month follow-up period were included.
Cost-effectiveness was determined by calculating expenditures that would be encountered by a Colorado Medicaid-supported IPI program in the following way. For the IPI group, we included the cost of implant insertion, removal, and gynecological/obstetrical outcomes. For the comparison group, we included the cost of gynecological/obstetrical outcomes. We did not include the cost of the contraceptive methods used in the comparison group, although Colorado Medicaid covers all hormonal contraceptive methods and LARC when initiated in the traditional outpatient setting. Adolescent users of short-acting contraceptives frequently discontinue and change their methods. Our study methodology was not designed to capture these high-frequency changes. Furthermore, we made the assumption that if the IPI program did not exist, the contraceptive costs of all participants would be similar. By not including these costs in our analysis, we present the additional costs over a presumed baseline cost of contraception incurred by Medicaid when IPI is not available ( Table 1 ). These additional expenses are compared between the IPI and comparison groups. Outpatient implant insertion and removal cost estimates were utilized because inpatient estimates did not exist. A schematic of incorporated costs and outcomes are shown in Figure 1 .
| Variable | Payment in dollars |
|---|---|
| Etonogestrel implant | $614 |
| Implant insertion | $83 |
| Implant removal | $97 |
| Prenatal care and NSVD | $6420 |
| Prenatal care and cesarean delivery | $10,523 |
| Prenatal care and unknown mode of delivery | $9689 |
| Infant’s medical care for first year of life | $2243 |
| Management of spontaneous abortion | $1001 |
| Treatment of ectopic pregnancy | $29,310 |
The potential costs of a repeat pregnancy included prenatal care and delivery (vaginal [$6420] or caesarean section [$10,523]), management of miscarriage ($1001), and treatment of an ectopic pregnancy ($29,310). There was no standard reimbursement available for management of miscarriage and treatment of ectopic, so estimates were created using the mean Medicaid payment for those patients in the cohort with these diagnoses.
The mode of delivery was unable to be determined in a few cases. For these deliveries, a composite cost ($7446) was created by weighting the cost of vaginal and cesarean delivery with our cesarean section rate for the entire cohort (25%). For delivered pregnancies, Medicaid reimbursement for the care of a well-child during the first year of life was also included ($2200).
Because Colorado Medicaid does not cover the termination of pregnancy, the costs of abortion were not included in this analysis. In some cases the outcome of the pregnancy (ie, abortions or deliveries outside our institutions) was known, but the date the pregnancy ended was not. Given this circumstance, we chose to attribute the cost of all pregnancies to the incident date of the pregnancy rather than the end date. The frequency for these outcomes for both groups at each time period were then normalized to a 1000 woman cohort and multiplied by their respective costs to obtain a per-1000 women total cost.
A sensitivity analysis was conducted by varying the repeat pregnancy rates of the comparison group. Using the actual pregnancy rates in the IPI group, we incrementally adjusted pregnancy rates in our comparison group until cost equivalency was reached between the 2 groups. The calculated rates of pregnancy at which costs are equal were then compared with the actual rates of pregnancy in our cohort as well as the pregnancy rates in the literature to evaluate the robustness of our results.
Summary statistics were used to describe the population as well as the outcomes. The Student t tests were used to compare means; the χ 2 test was used to compare proportions, and the Fisher exact tests were used when cell sizes were less than 5. SPSS statistical software (version 19.0.0; SPSS Inc, Chicago, IL) and Microsoft Excel (version12.3.6; Microsoft, Redmond, WA) were used for the analyses. The study was approved by the Colorado Multiple Institutional Review Board.
Results
During the study period, 396 CAMP patients were eligible for the study cohort. No one declined study participation. Participant demographics are illustrated in Table 2 . Most were mid- to late adolescents (mean age, 18.6 years; range, 13–23 years) and primiparous (77%). The population was ethnically diverse. There were no significant differences in baseline characteristics between the 2 study groups aside from parity ( Table 2 ); the IPI group members were more likely to be primiparous (82.5% vs 72%; P = .01).
| Characteristic | Comparison group (n = 225) | IPI (n = 171) | P value |
|---|---|---|---|
| Age at conception, y | 18.7 ± 1.8 | 18.5 ± 1.6 | .22 |
| Race/ethnicity | |||
| Black | 37.3% | 33.3% | .41 |
| Hispanic | 37.8% | 43.3% | .27 |
| White | 17.8% | 19.9% | .60 |
| Gravidity | 1.5 ± 0.8 | 1.4 ± 0.7 | .16 |
| Primigravid | 61.3% | 61.9% | .11 |
| Parity | 0.3 ± 0.6 | 0.2 ± 0.5 | .05 |
| Primiparous | 72.0% | 82.5% | .01 |
| Prior TAbs | 0.0 ± 0.2 | 0.0 ± 0.2 | .53 |
| Prior SAbs | 0.2 ± 0.4 | 0.2 ± 0.5 | .81 |
| BMI, kg/m 2 | 24.5 ± 5.8 | 24.1 ± 5.1 | .50 |
| Underweight (BMI <18.5) | 8.1% | 12.3% | .22 |
| Average weight (BMI 18.5-24.9) | 56.5% | 49.0% | .18 |
| Overweight/obese (≥BMI 25.0) | 35.4% | 38.7% | .54 |
| In school, graduated, or has GED | 73.4% | 71.6% | .71 |
| Has a relationship with the father of the child | 85.3% | 87.9% | .47 |
Loss to follow-up rates were similar between the IPI and comparisons at 6 months (0.94 vs 0.95), 1 year (0.90 vs 0.92), and 2 years (0.79 vs 0.71). At 3 years, the follow-up rates were 72.5% in the IPI group and 49.3% in the comparison group with a χ 2 statistic of 21.6 ( P < .001). Women who were lost to follow-up did not differ from the rest of the cohort on any demographic or reproductive factors.
A total of 171 adolescents (43.2%) received immediate postpartum etonogestrel implants. In the IPI group, implant continuation was 97%, 86%, 65%, and 48% at 6, 12, 24, and 36 months, respectively. The 225 participants who delivered within the study period but did not choose IPI insertion (56.8%) were classified as the comparison group. Of the comparison group, 14.7% chose LARC at their outpatient postpartum visit, 4-8 weeks after delivery. By 6 months, 43.7% of the comparison group (93 of 213) were using either implants or IUDs, 34% (72 of 213) were using short-acting methods, and 16% of the comparison group (34 of 213) had discontinued or elected to not use birth control.
At each time period of interest, pregnancy rates were significantly higher in the comparison group compared with the IPI group. Repeat pregnancy rates at 6, 12, 24, and 36 months of the IPI group were 0%, 2.6%, 8.1%, and 17.7%, respectively, compared with 9.9%, 20.1%, 46.5%, and 83.7% in the non-IPI group ( P < 0.001 between groups for each time point). Pregnancy outcomes entered into the model are presented in Table 3 . Cost equivalency is achieved if pregnancy rates in the comparison group were limited to 13.9%, 18.6%, and 31% at 12, 24, and 36 months. The IPI program resulted in cost savings between 6 and 12 months.
| Variable | 0-6 months | 0-12 months | 0-24 months | 0-36 months | ||||
|---|---|---|---|---|---|---|---|---|
| Comparison (n = 213) | IPI (n = 161) | Comparison (n = 204) | IPI (n = 153) | Comparison (n = 159) | IPI (n = 135) | Comparison (n = 111) | IPI (n = 124) | |
| NSVD | 9 | 0 | 25 | 2 | 44 | 6 | 53 | 12 |
| C/S | 2 | 0 | 5 | 2 | 11 | 3 | 12 | 4 |
| TAb | 5 | 0 | 7 | 0 | 8 | 0 | 11 | 1 |
| SAb | 1 | 0 | 4 | 0 | 8 | 1 | 12 | 1 |
| Unk delivery model | 3 | 0 | 3 | 0 | 7 | 1 | 11 | 4 |
| Ectopic pregnancy | 0 | 0 | 0 | 0 | 1 | 0 | 1 | 0 |
| Second repeat pregnancies a | 3 | 5 | 7 | |||||
| Total number of pregnancies | 20 | 0 | 44 | 4 | 79 | 11 | 100 | 22 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
