Abstract
Objective
To examine patient characteristics among those who selected the long-acting reversible contraception (LARC) and surgical sterilization methods at vaginal delivery.
Design
Retrospective cohort study.
Setting
The National Inpatient Sample.
Patient(s)
A total of 8,013,785 vaginal deliveries from October 2016 to December 2019.
Intervention(s)
Exposure assignment per LARC (subdermal contraceptive implant [implants] or intrauterine device [IUD]) or surgical sterilization (bilateral salpingectomy [BS] or bilateral tubal ligation [BTL]) type.
Main Outcome Measure(s)
Utilization trends of LARC or surgical sterilization, assessed with linear segmented regression with log-transformation, and differences in patient characteristics per the exposure strata (implants vs. IUD in the LARC group and BS or BTL in the surgical sterilization group), assessed using the multivariate binary logistic regression model.
Result(s)
In a comparison between LARC and surgical sterilization, surgical sterilization use decreased from 1.90% to 1.55% (18.4% relative decrease), whereas LARC use increased from 0.35% to 1.02% (191% relative increase). In the LARC group, implant use (from 0.12% to 0.50%) increased more compared with IUD use (from 0.22% to 0.52%): relative increase, 317% vs. 136%. In the surgical sterilization group, BTL use decreased from 0.66% to 0.18% (72.7% relative decrease), whereas BS use was statistically unchanged (from 1.24% to 1.37%). In a multivariate analysis, recent year remained an independent characteristic for implant use in the LARC group and BS use in the surgical sterilization group. Moreover, in both LARC and surgical sterilization strata, procedure choices significantly differed on the basis of patient, pregnancy, hospital, and delivery factors.
Conclusion(s)
Immediate postpartum contraception choice has evolved in recent years in the United States with an increasing demand for the LARC methods with implants at the time of vaginal delivery.
Key Words: Vaginal delivery, long-active reversible contraception, surgical sterilization, trends, characteristics
Placement of long-acting reversible contraception (LARC) during hospital delivery is increasing in the United States (1, 2). Immediate postpartum LARC placement is associated with high rates of contraceptive satisfaction, continuation of the contraceptive method, and a lower risk of short interpregnancy interval (3, 4). A recent US study found a national-level increase in the use of LARC with either intrauterine device (IUD) or subdermal contraceptive implant (implants) between 2016 and 2018 (1). Furthermore, a decrease in the use of surgical sterilization with bilateral tubal ligation (BTL) at the time of delivery was noted during the same time period (1).
Given the national-level paradigm shift from surgical sterilization to LARC, clarifying the differences in patient characteristics among LARC (IUD or implants) and surgical sterilization (BTL or bilateral salpingectomy [BS]) types at vaginal delivery is of interest. This study aimed to examine patient characteristics associated with the choice of the LARC and surgical sterilization methods in the immediate postpartum period after vaginal delivery.
Materials and methods
Data
The Healthcare Cost and Utilization Project’s National Inpatient Sample was retrospectively queried (5). This program is a population-based all-payer database for hospital admission that randomly selects 20% of inpatient records in each participating center and the weighted data for national estimates represents >90% of the US population. The data set is both publicly available and deidentified, and this study was deemed exempt by the University of Southern California Institutional Review Board. The Strengthening the Reporting of Observational Studies in Epidemiology reporting guidelines were consulted for the cohort study (6).
Study Population
The study population was 8,013,785 vaginal deliveries from October 2016 to December 2019 for national estimates. Patients who underwent hysterectomy after delivery or those who had both LARC and surgical sterilization were excluded from analysis. The starting point was chosen owing to the introduction of the World Health Organization’s International Classification of Disease 10th Revision codes for implants in the program for the exposure factors.
Exposure
Exposure allocation was per LARC (IUD or implants) or surgical sterilization (BTL or BS) type chosen. The IUD and implant cases were identified on the basis of the International Classification of Disease 10th Revision codes of Z30.430 and Z30.017, respectively. This study followed the same coding schema for identifying BTL (0U57, 0UL7, and 0UF7) and BS (0UB7 and 0UT7) (7).
Study Covariates
A total of 28 covariates were preselected and examined, including baseline demographics (age, year, admission type, race and ethnicity, primary expected payer, and census-level household income), comorbidity (obesity, tobacco use, grand multiparity, prior uterine scar, uterine myoma, diabetes mellitus, and hypertensive disease), pregnancy factors (multifetal gestation, fetal growth restriction, fetal demise, premature rupture of membrane, and chorioamnionitis), delivery data (gestational age, labor induction, operative delivery with vacuum-assisted or forceps delivery, manual placental removal, postpartum hemorrhage, and severe maternal morbidity per the Centers for Disease Control and Prevention definition), and hospital parameters (bed capacity, location/teaching setting, and region).
Outcome Measures
The primary outcome was the difference in patient characteristics in each exposure strata, comparing IUD vs. implants in the LARC group and BTL vs. BS in the surgical sterilization group, assessed by a multivariate logistic regression model. The effect size was expressed with the adjusted odds ratio and corresponding 95% confidence interval. The co-outcome measures included temporal trends of surgical sterilization and LARC over time, assessed with linear segmented regression with log-transformation (year-quarter increments).
Results
Trends
In a comparison between LARC and surgical sterilization, the number of patients receiving any surgical sterilization decreased from 1.90% to 1.55% (18.4% relative decrease, P trend<.001), whereas that of patient with LARC placement increased from 0.35% to 1.02% (191% relative increase, P trend<.001) during the study period (Fig. 1).
FIGURE 1.
Temporal trends of LARC and surgical sterilization (SS) at vaginal delivery. A total of 8,013,785 vaginal deliveries from October 2016 to December 2019 for national estimates were examined. The temporal trends of any SS (blue) and any LARC (red) are shown (year-quarter increments). In the LARC types, the trends of IUD (brown, n = 32,210) and subdermal contraceptive implant (implants) (light green, n = 23,035) are shown. In the SS types, the trends of BTL (light blue, n = 27,585) and BS (yellow, n = 112,510) are shown. Modeled values based on linear segmented regression with log-transformation (year-quarter increments).
BS = bilateral salpingectomy; BTL = bilateral tubal ligation; IUD = intrauterine device; LARC = long-acting reversible contraception.
Among the LARC types, the number of patients receiving implants (from 0.12% to 0.50%, P trend<.001) as well as IUD (from 0.22% to 0.52%, P trend<.001) both increased during the study period; however, the interval increase was higher for implant placement than for IUD (relative increase, 317% vs. 136%; Fig. 1). In the fourth quarter of 2019, the number of patients receiving IUD only marginally exceeded that of patients with implants (0.52% and 0.50%, respectively).
Among the surgical sterilization types, the number of patients undergoing BTL decreased from 0.66% to 0.18% (72.7% relative decrease, P trend<.001), whereas that of patients undergoing BS were unchanged from 1.24% to 1.37% (P trend=.298; Fig. 1).
Across the 4 exposure groups, nearly 1 in 30 patients who had a vaginal delivery received any 1 of these 4 procedures in the last year quarter of 2019 (3.34%), and the most frequent procedure was BS (1.37%), followed by IUD placement (0.52%), implant placement (0.50%), and BTL (0.18%).
Patient Characteristics
In the LARC group, 32,120 patients who received IUD and 23,035 patients who received implants were compared (Table 1). When compared with patients who received IUD, patients who received implants were more likely to be younger; have a recent year delivery; be a tobacco user; have chorioamnionitis, fetal growth restriction/demise, preterm delivery, operative delivery, manual placental removal, and postpartum hemorrhage; and have delivered at small rural hospitals; however, they were less likely to be obese and Hispanic; have higher household income, private insurance, pregestational hypertension, prior uterine scar and uterine myoma, multifetal gestation, elective admission for delivery, and labor induction; and have delivered at centers at the Northeast and Midwest regions (all, P<.05).
TABLE 1.
Results of the multivariate analysis.
| Characteristic | Implants vs. IUD |
BS vs. BTL |
||
|---|---|---|---|---|
| aOR (95% CI) | P value | aOR (95% CI) | P value | |
| Age (y) | 0.97 (0.96–0.97) | <.001 | 1.01 (1.01–1.01) | <.001 |
| Year | ||||
| Q4/2016 | 0.97 (0.88–1.08) | .583 | 0.57 (0.54–0.59) | <.001 |
| 2017 | 1 | 1 | ||
| 2018 | 1.33 (1.26–1.40) | <.001 | 1.44 (1.39–1.49) | <.001 |
| 2019 | 1.85 (1.76–1.94) | <.001 | 1.95 (1.88–2.02) | <.001 |
| Admission type | ||||
| Nonelective | 1 | 1 | ||
| Elective | 0.89 (0.86–0.93) | <.001 | 1.02 (0.99–1.05) | .215 |
| Unknown | 0.90 (0.60–1.37) | .626 | 2.34 (1.46–3.74) | <.001 |
| Race/ethnicity | ||||
| Black | 1 | 1 | ||
| White | 0.80 (0.76–0.85) | <.001 | 0.95 (0.91–0.99) | .017 |
| Hispanic | 1.32 (1.26–1.39) | <.001 | 1.03 (0.98–1.08) | .201 |
| Asian | 1.26 (1.13–1.39) | <.001 | 1.37 (1.23–1.52) | <.001 |
| Native American | 1.37 (1.16–1.62) | <.001 | 1.24 (1.07–1.43) | .003 |
| Others | 0.80 (0.73–0.87) | <.001 | 0.97 (0.89–1.05) | .406 |
| Unknown | 1.34 (1.20–1.50) | <.001 | 0.81 (0.75–0.88) | <.001 |
| Primary expected payer | ||||
| Medicaid | 1 | 1 | ||
| Private including HMO | 0.55 (0.52–0.58) | <.001 | 1.04 (1.01–1.08) | .007 |
| Medicare | 0.84 (0.72–0.98) | .028 | 1.39 (1.21–1.59) | <.001 |
| Self-pay | 1.06 (0.95–1.18) | .333 | 1.03 (0.93–1.14) | .535 |
| No charge | 0.73 (0.44–1.19) | .203 | 2.86 (1.49–5.48) | .002 |
| Others | 1.20 (1.01–1.43) | .036 | 0.83 (0.75–0.91) | <.001 |
| Unknown | 3.09 (1.89–5.06) | <.001 | 3.51 (2.06–5.99) | <.001 |
| Median household income | ||||
| QT1 (lowest) | 1 | 1 | ||
| QT2 | 1.02 (0.98–1.07) | .331 | 1.13 (1.09–1.17) | <.001 |
| QT3 | 0.81 (0.77–0.85) | <.001 | 1.20 (1.15–1.25) | <.001 |
| QT4 (highest) | 0.67 (0.63–0.72) | <.001 | 1.28 (1.21–1.35) | <.001 |
| Unknown | 1.03 (0.84–1.27) | .750 | 1.18 (1.03–1.35) | .015 |
| Hp bed capacity | ||||
| Small | 3.00 (2.81–3.20) | <.001 | 1 | |
| Mid | 1.56 (1.49–1.63) | <.001 | 1.00 (0.96–1.04) | .976 |
| Large | 1 | 1.16 (1.12–1.20) | <.001 | |
| Hp location/teaching | ||||
| Rural | 1.58 (1.33–1.87) | <.001 | 0.60 (0.58–0.63) | <.001 |
| Urban nonteaching | 1 | 1 | ||
| Urban teaching | 1.08 (0.99–1.18) | .067 | 1.14 (1.10–1.18) | <.001 |
| Hp region | ||||
| Northeast | 0.50 (0.47–0.53) | <.001 | 1.01 (0.95–1.06) | .777 |
| Midwest | 0.56 (0.53–0.60) | <.001 | 1.29 (1.23–1.35) | <.001 |
| South | 1.02 (0.96–1.07) | .603 | 1.10 (1.05–1.14) | .201 |
| West | 1 | 1 | ||
| Obesity | ||||
| No | 1 | 1 | ||
| Yes | 0.79 (0.75–0.83) | <.001 | 1.10 (1.05–1.14) | <.001 |
| Grand multiparity | ||||
| No | 1 | 1 | ||
| Yes | 1.18 (0.88–1.58) | .259 | 0.64 (0.59–0.70) | <.001 |
| Tobacco use | ||||
| No | 1 | 1 | ||
| 1.31 (1.23–1.40) | <.001 | 1.04 (0.99–1.08) | .118 | |
| Prior uterine scar | ||||
| No | 1 | 1 | ||
| Yes | 0.83 (0.77–0.90) | <.001 | 1.04 (0.98–1.11) | .163 |
| Uterine myoma | ||||
| No | 1 | 1 | ||
| Yes | 0.73 (0.59–0.91) | .005 | 1.28 (1.06–1.54) | .010 |
| Diabetes mellitus | ||||
| No | 1 | 1 | ||
| Pregestational | 1.00 (0.87–1.17) | .957 | 1.01 (0.89–1.15) | .844 |
| Gestational | 0.94 (0.87–1.02) | .124 | 1.01 (0.97–1.06) | .571 |
| Hypertensive disease | ||||
| No | 1 | 1 | ||
| Pregestational | 0.81 (0.74–0.88) | <.001 | 0.95 (0.88–1.01) | .115 |
| Gestational | 0.95 (0.89–1.02) | .133 | 0.98 (0.92–1.04) | .476 |
| Multifetal gestation | ||||
| No | 1 | 1 | ||
| Yes | 0.71 (0.59–0.85) | <.001 | 1.40 (1.20–1.64) | <.001 |
| Fetal growth restriction | ||||
| No | 1 | 1 | ||
| Yes | 1.10 (1.01–1.2) | .026 | 1.14 (1.04–1.24) | .003 |
| Intrauterine fetal demise | ||||
| No | 1 | 1 | ||
| Yes | 1.39 (1.10–1.75) | .006 | 1.14 (0.79–1.65) | .480 |
| Premature rupture of membrane | ||||
| No | 1 | 1 | ||
| Preterm | 1.05 (0.94–1.17) | .393 | 1.18 (1.05–1.32) | .005 |
| Term | 0.99 (0.92–1.06) | .739 | 1.27 (1.17–1.37) | <.001 |
| Chorioamnionitis | ||||
| No | 1 | 1 | ||
| Yes | 2.87 (2.52–3.26) | <.001 | 2.18 (1.67–2.85) | <.001 |
| Gestational age at delivery | ||||
| ≥39 | 1 | 1 | ||
| 37–38 | 1.16 (1.11–1.21) | <.001 | 0.97 (0.94–0.99) | .030 |
| 34–36 | 1.22 (1.13–1.31) | <.001 | 0.97 (0.91–1.03) | .314 |
| <34 | 1.64 (1.46–1.84) | <.001 | 0.94 (0.83–1.06) | .316 |
| Unknown | 0.54 (0.43–0.67) | <.001 | 0.96 (0.85–1.08) | .499 |
| Labor induction | ||||
| No | 1 | 1 | ||
| Yes | 0.86 (0.83–0.90) | <.001 | 1.04 (1.01–1.07) | .035 |
| Operative delivery | ||||
| No | 1 | 1 | ||
| Forceps | 1.75 (1.46–2.10) | <.001 | 1.02 (0.84–1.24) | .840 |
| Vacuum-assisted | 1.18 (1.07–1.30) | .001 | 0.89 (0.83–0.96) | .001 |
| Manual removal | ||||
| No | 1 | 1 | ||
| Yes | 1.30 (1.10–1.54) | .002 | 1.09 (0.91–1.30) | .355 |
| Postpartum hemorrhage | ||||
| No | 1 | 1 | ||
| Yes | 1.52 (1.40–1.65) | <.001 | 1.19 (1.10–1.30) | <.001 |
| Severe maternal morbidity | ||||
| No | 1 | 1 | ||
| Yes | 0.97 (0.83–1.13) | .667 | 1.40 (1.18–1.65) | <.001 |
Note: A binary logistic regression model was used for the multivariate analysis. The initial covariate selection was P<.05 in the univariate analysis. The parsimonious conditional backward method was used for the final modeling. aOR = adjusted odds ratio; BS = bilateral salpingectomy; BTL = bilateral tubal ligation; CI = confidence interval; HMO = health maintenance organization; Hp = hospital; implant = subdermal contraceptive implant; IUD = intrauterine device.
In the surgical sterilization group, 112,510 patients who underwent BS and 27,585 patients who underwent BTL were compared (Table 1). When compared with patients who underwent BTL, patients who underwent BS were more likely to be older; have a recent year delivery; be obese; have higher household income, uterine myoma, premature rupture of membrane, chorioamnionitis, multifetal gestation, fetal growth restriction, labor induction, postpartum hemorrhage, and severe maternal morbidity; and have delivered at large urban hospitals in the Midwest regions; however, they were less likely to be grand multiparity and have a vacuum-assisted delivery (all, P<.05).
Discussion
Findings
The present study adds more detailed information to the recent US national-level observation of a shift from surgical sterilization to LARC in that this shift is mainly because of the decrease in BTL procedures and increase in implant placements (1). The observed temporal trend projects that the number of patients receiving LARC will likely exceed that of patients receiving surgical sterilization by the first quarter of 2021 and that in the LARC group, the number of patients receiving implants will likely exceed that of patients receiving IUD by the first quarter of 2020.
Even in the same treatment category of LARC, there was substantial variability in the patient choice between IUD and implants on the basis of patient, pregnancy, delivery, and hospital factors. This also applied to surgical sterilization cases between BS and BTL. Taken together, these data suggest that increasing accessibility to all LARC and surgical sterilization methods is useful for shared decision-making between providers and patients.
Although the absolute percentage rate of BS was unchanged, because of the decrease in BTL use, there was a relative increase in BS use over time that this association was independent after controlling other factors including hospital parameters (Table 1). This nationwide paradigm shift from BTL to BS was also reported in cesarean deliveries and benign hysterectomies (7, 8). It is most likely that consensus for opportunistic salpingectomy for ovarian cancer risk reduction may change the landscape of surgical sterilization in both obstetric and gynecologic surgeries.
Limitations
The limitations of the present study include the unmeasured bias with the lack of information on the exact reason for procedure selection, patient and physician’s knowledge and experience, routine prenatal care, hospital practice, home delivery, patient satisfaction, quality-of-life measures, regret, and long-term complications. Lower likeliness of implant placement vs. IUD placement in uterine myoma also suggests a possible unmeasured confounder effect, warranting further investigation.
Accuracy of data, particularly for the BTL and BS procedures, was not assessable because actual medical record review was not performed. Ascertainment bias because of the data capturing schema in the program and generalizability to other populations are also recognized as possible limitations of this study. Despite these limitations, the results of this study suggest that the immediate postpartum contraception choice at vaginal delivery has evolved in recent years in the United States.
Footnotes
A.C.Y. has nothing to disclose. A.M.M. has nothing to disclose. H.E.S. has nothing to disclose. R.S.M. has nothing to disclose. J.G.O. has nothing to disclose. K.M. has nothing to disclose.
A.C.Y., A.M.M., and K.M. should be considered similar in author order.
Supported by Ensign Endowment for Gynecologic Cancer Research (to K.M.).
References
- 1.Fang N.Z., Westhoff C.L. Update on incidence of inpatient tubal ligation and long-acting reversible contraception in the United States. Am J Obstet Gynecol. 2022;227 doi: 10.1016/j.ajog.2022.05.021. 477.e1–7. [DOI] [PubMed] [Google Scholar]
- 2.Fang N.Z., Advaney S.P., Castano P.M., Davis A., Westhoff C.L. Female permanent contraception trends and updates. Am J Obstet Gynecol. 2022;226:773–780. doi: 10.1016/j.ajog.2021.12.261. [DOI] [PubMed] [Google Scholar]
- 3.Wu M., Eisenberg R., Negassa A., Levi E. Associations between immediate postpartum long-acting reversible contraception and short interpregnancy intervals. Contraception. 2020;102:409–413. doi: 10.1016/j.contraception.2020.08.016. [DOI] [PubMed] [Google Scholar]
- 4.Wallace Huff C., Potter J.E., Hopkins K. Patients' experiences with an immediate postpartum long-acting reversible contraception program. Womens Health Issues. 2021;31:164–170. doi: 10.1016/j.whi.2020.11.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Agency for Healthcare Research and Quality Overview of the National (Nationwide) Inpatient Sample (NIS) https://hcup-us.ahrq.gov/nisoverview.jsp Available at:
- 6.Ghaferi A.A., Schwartz T.A., Pawlik T.M. STROBE reporting guidelines for observational studies. JAMA Surg. 2021;156:577–578. doi: 10.1001/jamasurg.2021.0528. [DOI] [PubMed] [Google Scholar]
- 7.Mandelbaum R.S., Matsuzaki S., Sangara R.N., Klar M., Matsushima K., Roman L.D., et al. Paradigm shift from tubal ligation to opportunistic salpingectomy at cesarean delivery in the United States. Am J Obstet Gynecol. 2021;225 doi: 10.1016/j.ajog.2021.06.074. 399.e1–32. [DOI] [PubMed] [Google Scholar]
- 8.Mandelbaum R.S., Adams C.L., Yoshihara K., Nusbaum D.J., Matsuzaki S., Matsushima K., et al. The rapid adoption of opportunistic salpingectomy at the time of hysterectomy for benign gynecologic disease in the United States. Am J Obstet Gynecol. 2020;223 doi: 10.1016/j.ajog.2020.04.028. 721.e1–18. [DOI] [PubMed] [Google Scholar]