After a live birth, the World Health Organization recommends a minimum of 24 months before attempting a subsequent pregnancy.1 Initiation of long-acting reversible contraception at the time of delivery decreases rapid repeat births,2,3 and immediate postplacental placement of an intrauterine device (IUD) is safe4,5 and cost-effective.6
Intrauterine devices have failure rates of 0.2–0.8 pregnancies per 100 woman-years with perfect use,7 but these rates do not account for IUD discontinuation or expulsion. Although postplacental IUD placement ensures that women leave the hospital with contraception, it also results in a higher proportion of IUD expulsions, with rates ranging from 5.8% to 24%8–15 as compared with 2.9–3.5% (Ogburn T, Espey E, Leeman L, Singh R, Pereda B, Carr S. A randomized trial of immediate postpartum versus interval insertion of an intrauterine device [abstract P57]. Contraception 2013;88:455)10,13,14,16–24 after placement at an outpatient postpartum visit (delayed postpartum IUD placement).4,5,14,25,26 Higher expulsion rates may in turn result in decreased contraceptive efficacy. The proportion of women who attend a postpartum visit ranges from 65% to 90% (Ogburn T et al. Contraception 2013;88:455),13,24,25,27–29 and the proportion of women intending IUD placement at this visit who actually receive an IUD varies from 22.6% to 100%.13,27,28 Actual placement of an intended IUD is, of course, critical to its effectiveness. The term “effectiveness” refers to how well a method works in actual practice, as opposed to “efficacy,” which refers to how well a method works in clinical trials.30
Our objective was to use computational decision analysis to model the effectiveness of immediate and delayed postpartum IUD placement. We compared the 1-year probabilities of pregnancy after intended postplacental IUD insertion after vaginal or cesarean birth with intended delayed postpartum insertion, incorporating factors that influence successful device placement and retention.
MATERIALS AND METHODS
We used a Monte Carlo computational decision model to evaluate 1-year probability of pregnancy in two groups: women intending postplacental IUD placement within 10 minutes of placental delivery and those intending delayed postpartum IUD placement. The Monte Carlo simulation method accounts for the probability distribution of outcomes with inherent uncertainty using decision trees (Figs. 1 and 2). We chose to compare intended IUD placement rather than actual IUD placement to model the common clinical and life circumstances that limit access to IUDs in both populations. Probability of pregnancy in our model was considered to be the inverse of IUD effectiveness, and we chose a 1-year end point so that our effectiveness estimates could be compared with established 1-year estimates of contraceptive efficacy. We built and analyzed our model using TreeAge Pro modeling software.
Before creating our model, we made an a priori selection of the following factors that we hypothesized would affect IUD effectiveness after postplacental or delayed placement: mode of delivery, successful IUD placement, IUD type, postpartum visit attendance, IUD expulsion, detection of expulsion, partial expulsion, IUD discontinuation, and contraceptive use, choice, and efficacy after IUD discontinuation. We included estimates of contraceptive discontinuation and switching into our effectiveness model to simulate the common experiences of women who have IUD expulsion or choose to discontinue a method to provide a realistic estimate of pregnancy probability. Although we did not include estimates of intrapartum events that would affect eligibility for postplacental IUD insertion, we included an estimate of the proportion of intended postplacental IUDs that were actually placed; nonplacement was the result of factors such as withdrawal of consent, infection, precipitous birth, and hemorrhage.31
To identify probability estimates for each event in the model, we conducted a literature review using PubMed and a combination of the following search terms: “IUD,” “intrauterine device,” “contraception failure,” “copper,” “postpartum IUD,” “postplacental,” “levonorgestrel,” “cesarean delivery,” “vaginal delivery,” “continuation,” “discontinuation,” and “expulsion.” In addition, we reviewed the references of relevant articles for additional sources. For each event, two authors (T.H., S.S.) reviewed each article and selected the most likely probability, called the “base case value,” from the trial that was methodologically superior based on robustness of study design and sample size. Randomized controlled trials and prospective cohort studies were selected when possible. For studies that followed participants for less than 1 year, values for each of the estimates were carried forward and assumed to be 1-year estimates.
Base case values were selected for the 1-year probabilities and ranges of complete and partial IUD expulsion, detection of expulsion, and elective discontinuation of IUDs after postplacental or delayed placement. When data were available, we incorporated likely differences in outcomes of copper compared with levonorgestrel IUDs and vaginal compared with cesarean deliveries. The model also incorporated evidence-based probabilities of choosing a copper compared with a levonorgestrel IUD. In addition, for the delayed postpartum model, we selected base case values for the proportion of women who return for a postpartum visit and for the proportion of women initially desiring a delayed postpartum IUD who actually receive it.
We constructed three decision trees as maps of the possible outcomes after 1) intended postplacental IUD placement after vaginal birth, 2) intended postplacental IUD placement after cesarean birth, and 3) intended delayed postpartum IUD placement. Each branch point is called a “node,” and the outcomes of each node incorporate the base case probabilities determined from the literature (Figs. 1 and 2). At every node in the model where a patient lacked a contraceptive method, we assumed she would restart the contraception selection process, except in the case of undetected IUD expulsion. We modeled this process by first estimating the probability that the patient would use contraception12,32 and then estimating the proportion of women who would select long-acting reversible contraception compared with any other reversible method. Annual effectiveness rates were selected for different forms of contraception.7,33,34
Our simulation models computed the mean effectiveness of postplacental and delayed postpartum IUD insertion in the base case scenario (scenario 1). Our models used a theoretical cohort of 2,500,000 women for the delayed postpartum and vaginal postplacental models and 1,250,000 women for the cesarean model.35
To determine variation around the mean effectiveness, we performed one-way deterministic sensitivity analysis for all probabilities with uncertainty. Deterministic sensitivity assesses how changes in the model inputs affect model outputs, varying a single parameter at a time across its range. Probabilities of contraceptive failure for individual methods have precise estimates from the literature, so the highest reasonable and lowest reasonable values for method failure were estimated only for use of no contraceptive method and the partially expelled IUD. Probabilities were varied individually from the lowest to the highest reasonable value identified from our literature review to identify which variables had the largest effect on the difference in contraceptive outcome between intended delayed postpartum and postplacental insertion. To further explore the influence of the probability of a woman returning for her routine interval postpartum visit, this probability was varied from 0% to 100%. This allowed us to identify the threshold proportion of women returning for a postpartum visit at which a postplacental IUD and a delayed postpartum IUD would have the same expected effectiveness.
We then performed probabilistic sensitivity analysis using a hypothetical sample of 10,000 cycles to account for uncertainty in the true value of our events and to simulate the effect of adjusting for a range of parameters simultaneously. We created β distributions for all probabilities with uncertainty by setting the highest reasonable value and lowest reasonable value as the upper and lower bounds of the 95% CI and the base case value as the mean. These distributions were used to randomly sample all probabilities in the model simultaneously to determine the proportion of the 10,000 scenarios in which intended postplacental IUD placement had higher effectiveness than delayed postpartum insertion. We varied all probabilities in the model in this fashion with the exception of the effectiveness of each contraceptive method for which estimates are precise.
Finally, the probabilities representing the return to clinic rate, postplacental IUD placement rate, and delayed postpartum IUD placement rate were manipulated and the model was run again to evaluate two additional clinical scenarios. In scenario 2, we assessed the effectiveness of delayed postpartum compared with postplacental IUDs if all patients were to return for a clinic follow-up visit, and thus set the probability of a patient returning to the clinic at 100% in both the vaginal and cesarean models. This scenario provided an estimate of IUD effectiveness in a practice with very high rates of postpartum follow-up. Scenario 3 offered an individual-level estimate of contraceptive effectiveness for women who have an IUD inserted immediately postpartum or at a postpartum visit, incorporating the probabilities of discontinuation and expulsion. In this scenario, the probability of a patient returning to the clinic, the probability of delayed insertion, and the probability of immediate insertion were all set at 100%. All other values remained at their baseline values. This scenario provided an estimate of the maximum contraceptive effectiveness an individual IUD user could expect assuming she encountered no external barriers to IUD placement.
Because this study used data solely from previously published work with no identifiable patient information, the University of Pennsylvania institutional review board deemed it to be nonhuman subjects research and approval was not required.
The base case and range values for each node in our decision tree were determined from our literature search (Table 1). Based on these values, our decision model predicted that among a theoretical cohort of 2,500,000 women intending to receive a postplacental IUD after vaginal birth, 1,250,000 women intending a postplacental IUD after cesarean birth, and 2,500,000 women intending a delayed postpartum IUD, 17.3%, 11.2%, and 24.6%, respectively, would become pregnant by 1 year after birth (Table 2). In our model, most women intending postplacental vaginal, postplacental cesarean, and delayed postpartum IUD placement were using some form of contraception at 1 year postpartum. Women who intended to use a delayed postpartum IUD were less likely than those intending to use a postplacental IUD to be using any form of contraception at 1 year. Of women intending a postplacental IUD, 12.4% of women delivering vaginally and 10.3% of women delivering by cesarean were modeled to be using no contraception at 1 year. Of women intending to receive a delayed postpartum IUD, 27.6% were modeled to be using no contraception at 1 year (Table 2).
One-way deterministic sensitivity analysis indicated that the difference in effectiveness of intended postplacental and delayed postpartum IUD placement after vaginal and cesarean birth was driven primarily by the probability of a patient returning to the clinic for the delayed IUD placement visit (Fig. 3). Thus, we varied the probability of a patient returning to the clinic through the expected range, 65–90% (Ogburn T et al. Contraception 2013;88:455).13,24,25,27,28 If the proportion of women returning for a postpartum visit was 65.0%, the 1-year effectiveness of a postplacental IUD after vaginal birth was 19.6% higher than a delayed postpartum IUD; if the proportion of women returning for a postpartum visit was 90%, the effectiveness of a postplacental IUD after vaginal birth was 1.0% higher than a delayed postpartum IUD. Effectiveness of intended IUDs placed at cesarean birth was 25.7% and 7.2% higher than intended delayed IUDs, respectively, if the proportion of women returning for a postpartum visit was varied across the same range. Postplacental IUD insertion was favored until the probability of a patient returning to the clinic was above 91.4% after vaginal birth and 99.7% after cesarean birth (Fig. 3).
The results of the probabilistic sensitivity analysis were consistent with the one-way sensitivity analyses (Fig. 4), demonstrating that intended postplacental IUD insertion increased contraceptive effectiveness in 99.6% of scenarios after vaginal birth and 100% of scenarios after cesarean birth. In the probabilistic sensitivity analysis, the mean improvement in effectiveness resulting from postplacental insertion compared with delayed postpartum insertion was 7.3% and 13.4% for vaginal and cesarean deliveries, respectively (Fig. 4).
In the scenario in which all women were assumed to return to the clinic for their 6-week follow-up visit (scenario 2) and in the scenario in which all women received their intended IUD (scenario 3), 1-year probability of pregnancy rates for delayed postpartum IUDs were 11.0% and 3.9%, respectively (Table 3) and were lower than the 1-year probability of pregnancy after postplacental IUD placement. If all women received their intended IUD, 1-year probabilities of pregnancy after IUD placement at a vaginal birth, cesarean birth, and an outpatient postpartum visit were 15.4%, 6.6%, and 3.9%, respectively.
In this decision analysis, the 1-year probabilities of pregnancy among women intending to receive a postplacental IUD after vaginal or cesarean birth were 17.3% and 11.2%, respectively, and 24.6% for women intending to receive a delayed postpartum IUD. These were much higher than the less than 1% failure rate typically quoted for interval IUDs; that rate does not account for the multitude of factors that affect peripartum IUD initiation and continuation. Given that effectiveness is only one of many attributes of a contraceptive method, these results should be interpreted in the larger clinical context in which postpartum contraception counseling occurs.
Differences in IUD effectiveness by placement timing in our model were driven by higher pregnancy rates in women who did not receive their intended IUD, primarily as a result of follow-up visit nonattendance. Return to clinic rates must exceed 90% to favor the effectiveness of delayed postpartum insertion over immediate postplacental placement from a programmatic perspective.
We manipulated the probabilities of our model to evaluate two additional scenarios: scenario 2, 100% likelihood to return to the clinic for postpartum follow-up to allow for a practice-level determination of benefit of a postplacental IUD program, and scenario 3, 100% adherence to the intended insertion plan to allow for an individual-level determination of the benefit of a postplacental IUD. In both of these scenarios, delayed postpartum placement had higher effectiveness than postplacental placement. These data allow health care providers to make determinations regarding the value of a postplacental IUD program in their practice and to provide individualized counseling to women about postpartum IUD placement in practices that offer postplacental IUD placement.
Our study is similar to a 2015 cost-effectiveness study6 in which postplacental and delayed IUD placement resulted in 27 per 1,000 and 78 per 1,000 unintended pregnancies at 1 year, respectively. However, we incorporated additional nodes in our decision tree, including the proportion of women who did not receive an intended postplacental IUD. We also provided data regarding differential effectiveness of a postplacental IUD placed after cesarean compared with vaginal birth.
Strengths of our research include the integration of a large evidence base that incorporates most clinical outcomes that influence IUD effectiveness. In addition, we provide a novel framework for considering contraceptive effectiveness. Previously published estimates of contraceptive effectiveness assume contraceptive continuation at 1 year.7 Our findings illuminate the implications of contraceptive switching and IUD expulsion on pregnancy risk.
Limitations of this study are inherent to its design, which used probability estimates from previously published data. For some probability estimates, including partial expulsion,12 complete expulsion,13 and discontinuation11 of the levonorgestrel IUD after intended postplacental IUD placement, we carried forward 3-11 or 6-month12,13 estimates and assumed them to be 12-month estimates. Although this could have resulted in overestimating effectiveness, our deterministic sensitivity analysis showed that these estimates did not have a significant effect on the outcome throughout their ranges. We did not include in our analyses differential published estimates for the proportions of women who attend postpartum follow-up after vaginal compared with cesarean birth.29 However, our sensitivity analyses demonstrated that postplacental placement resulted in a higher effectiveness than delayed IUD placement across the range of estimates for the proportion of women who return for follow-up (65–90%). Finally, our research does not capture the patient-centered outcomes of women's satisfaction with peripartum contraceptive method choice, the desiredness of subsequent pregnancies, or the clinical implications for women with partially expelled IUDs12 or nonvisible IUD strings9,12 that frequently occur after postplacental IUD placement.
In conclusion, despite higher rates of expulsion, decision analysis showed that intended postplacental IUD insertion was more effective for pregnancy prevention at 1 year postpartum than intended delayed postpartum IUD insertion. However, once placed, the effectiveness of a delayed postpartum IUD was superior to that of a postplacental IUD. Thus, the public health and individual benefits of a postplacental IUD program may conflict, and a high level of attention to the informed consent process is required. Health systems should work to improve access to postplacental IUDs, which is often limited by lack of infrastructure and insurance reimbursement for immediate postpartum long-acting reversible contraception.36 Women should be counseled regarding postpartum contraception during antenatal care rather than during admission for delivery. Counseling should review all potential complications and state that a postplacental IUD is most beneficial in those who are unlikely to obtain an IUD in the delayed postpartum period. The findings from our analysis should be confirmed in a large multicenter, prospective, comparative study of postplacental compared with delayed postpartum IUD insertion.
1. Report of a WHO technical consultation on birth spacing, Geneva, Switzerland, June 13–15, 2005. Geneva (Switzerland): World Health Organization; 2006.
2. Brunson MR, Klein DA, Olsen CH, Weir LF, Roberts TA. Postpartum contraception: initiation and effectiveness in a large universal healthcare system. Am J Obstet Gynecol 2017;217:55.e1–9.
3. Damle LF, Gohari AC, McEvoy AK, Desale SY, Gomez-Lobo V. Early initiation of postpartum contraception: does it decrease rapid repeat pregnancy in adolescents? J Pediatr Adolesc Gynecol 2015;28:57–62.
4. Lopez LM, Bernholc A, Hubacher D, Stuart G, Van Vliet HA. Immediate postpartum insertion of intrauterine device for contraception. The Cochrane Database of Systematic Reviews 2015, Issue 6. Art. No.: CD003036. DOI: 10.1002/14651858.CD003036.pub3.
5. Sonalkar S, Kapp N. Intrauterine device insertion in the postpartum period: a systematic review. Eur J Contracept Reprod Health Care 2015;20:4–18.
6. Washington CI, Jamshidi R, Thung SF, Nayeri UA, Caughey AB, Werner EF. Timing of postpartum intrauterine device placement: a cost-effectiveness analysis. Fertil Steril 2015;103:131–7.
7. Trussell J. Contraceptive failure in the United States. Contraception 2011;83:397–404.
8. Sucak A, Ozcan S, Çelen Ş, Çağlar T, Göksu G, Danişman N. Immediate postplacental insertion of a copper intrauterine device: a pilot study to evaluate expulsion rate by mode of delivery. BMC Pregnancy Childbirth 2015;15:202.
9. Colwill AC, Schreiber CA, Sammel MD, Sonalkar S. Six-week retention after postplacental copper intrauterine device placement. Contraception 2018;97:215–8.
10. Eroğlu K, Akkuzu G, Vural G, Dilbaz B, Akin A, Taşkin L, et al. Comparison of efficacy and complications of IUD insertion in immediate postplacental/early postpartum period with interval period: 1 year follow-up. Contraception 2006;74:376–81.
11. Goldthwaite LM, Sheeder J, Hyer J, Tocce K, Teal SB. Postplacental intrauterine device expulsion by 12 weeks: a prospective cohort study. Am J Obstet Gynecol 2017;217:674.e1–8.
12. Gurney E, Sonalkar S, McAllister A, Sammel MD, Schreiber CA. Six-month expulsion of postplacental copper intrauterine devices placed after vaginal delivery. Am J Obstet Gynecol 2018;219:183.e1–9.
13. Chen BA, Reeves MF, Hayes JL, Hohmann HL, Perriera LK, Creinin MD. Postplacental or delayed insertion of the levonorgestrel intrauterine device after vaginal delivery: a randomized controlled trial. Obstet Gynecol 2010;116:1079–87.
14. Dahlke JD, Terpstra ER, Ramseyer AM, Busch JM, Rieg T, Magann EF. Postpartum insertion of levonorgestrel–intrauterine system at three time periods: a prospective randomized pilot study. Contraception 2011;84:244–8.
15. Eggebroten JL, Sanders JN, Turok DK. Immediate postpartum intrauterine device and implant program outcomes: a prospective analysis. Am J Obstet Gynecol 2017;217:51.e1–7.
16. Rowe P, Farley T, Peregoudov A, Piaggio G, Boccard S, Landoulsi S, et al. Safety and efficacy in parous women of a 52-mg levonorgestrel-medicated intrauterine device: a 7-year randomized comparative study with the TCu380A. Contraception 2016;93:498–506.
17. Lester F, Kakaire O, Byamugisha J, Averbach S, Fortin J, Maurer R, et al. Intracesarean insertion of the Copper T380A versus 6 weeks postcesarean: a randomized clinical trial. Contraception 2015;91:198–203.
18. O'Neil-Callahan M, Peipert JF, Zhao Q, Madden T, Secura G. Twenty-four-month continuation of reversible contraception. Obstet Gynecol 2013;122:1083–91.
19. Sanders JN, Turok DK, Royer PA, Thompson IS, Gawron LM, Storck KE. One-year continuation of copper or levonorgestrel intrauterine devices initiated at the time of emergency contraception. Contraception 2017;96:99–105.
20. Sivin I, Stern J, Diaz J, Diaz MM, Faundes A, el Mahgoub S, et al. Two years of intrauterine contraception with levonorgestrel and with copper: a randomized comparison of the TCu 380Ag and levonorgestrel 20 mcg/day devices. Contraception 1987;35:245–55.
21. Berenson AB, Tan A, Hirth JM. Complications and continuation rates associated with 2 types of long-acting contraception. Am J Obstet Gynecol 2015;212:761.e1–8.
22. Sivin I, el Mahgoub S, McCarthy T, Mishell DR Jr, Shoupe D, Alvarez F, et al. Long-term contraception with the levonorgestrel 20 mcg/day (LNg 20) and the copper T 380Ag intrauterine devices: a five-year randomized study. Contraception 1990;42:361–78.
23. Sznajder KK, Tomaszewski KS, Burke AE, Trent M. Incidence of discontinuation of long-acting reversible contraception among adolescent and young adult women served by an urban primary care clinic. J Pediatr Adolesc Gynecol 2017;30:53–7.
24. Whitaker AK, Endres LK, Mistretta SQ, Gilliam ML. Postplacental insertion of the levonorgestrel intrauterine device after cesarean delivery vs. delayed insertion: a randomized controlled trial. Contraception 2014;89:534–9.
25. Chen MJ, Hou MY, Hsia JK, Cansino CD, Melo J, Creinin MD. Long-acting reversible contraception initiation with a 2- to 3-week compared with a 6-week postpartum visit. Obstet Gynecol 2017;130:788–94.
26. Levi EE, Stuart GS, Zerden ML, Garrett JM, Bryant AG. Intrauterine device placement during cesarean delivery and continued use 6 months postpartum: a randomized controlled trial. Obstet Gynecol 2015;126:5–11.
27. Salcedo J, Moniaga N, Harken T. Limited uptake of planned intrauterine devices during the postpartum period. South Med J 2015;108:463–8.
28. Ogburn JA, Espey E, Stonehocker J. Barriers to intrauterine device insertion in postpartum women. Contraception 2005;72:426–9.
29. Wilcox A, Levi EE, Garrett JM. Predictors of non-attendance to the postpartum follow-up visit. Matern Child Health J 2016;20(suppl 1):22–7.
30. Revicki DA, Frank L. Pharmacoeconomic evaluation in the real world. Effectiveness versus efficacy studies. Pharmacoeconomics 1999;15:423–34.
31. Jatlaoui TC, Marcus M, Jamieson DJ, Goedken P, Cwiak C. Postplacental intrauterine device insertion at a teaching hospital. Contraception 2014;89:528–33.
32. Kavanaugh ML, Jerman J. Contraceptive method use in the United States: trends and characteristics between 2008, 2012 and 2014. Contraception 2018;97:14–21.
33. Anteby E, Revel A, Ben-Chetrit A, Rosen B, Tadmor O, Yagel S. Intrauterine device failure: relation to its location within the uterine cavity. Obstet Gynecol 1993;81:112–4.
34. Winner B, Peipert JF, Zhao Q, Buckel C, Madden T, Allsworth JE, et al. Effectiveness of long-acting reversible contraception. N Engl J Med 2012;366:1998–2007.
35. Martin JA, Hamilton BE, Osterman MJK, Driscoll AK, Mathews TJ. Births: final data for 2015. Atlanta (GA): Centers for Disease Control; 2017.
36. Hofler LG, Cordes S, Cwiak CA, Goedken P, Jamieson DJ, Kottke M. Implementing immediate postpartum long-acting reversible contraception programs. Obstet Gynecol 2017;129:3–9.
37. El-Shafei M, Mashali A, Hassan E, El-Boghdadi, El-Lakkany N. Postpartum and postabortion intrauterine device insertion unmet needs of safe reproductive health: three years experience of a Mansoura University Hospital. Egypt Soc Obstet Gynecol 2000;26:253–62.
38. Letti Müller AL, Lopes Ramos JG, Martins-Costa SH, Palma Dias RS, Valério EG, Hammes LS, et al. Transvaginal ultrasonographic assessment of the expulsion rate of intrauterine devices inserted in the immediate postpartum period: a pilot study. Contraception 2005;72:192–5.
39. Levi E, Cantillo E, Ades V, Banks E, Murthy A. Immediate postplacental IUD insertion at cesarean delivery: a prospective cohort study. Contraception 2012;86:102–5.
40. Ragab A, Hamed HO, Alsammani MA, Shalaby H, Nabeil H, Barakat R, et al. Expulsion of Nova-T380, Multiload 375, and Copper-T380A contraceptive devices inserted during cesarean delivery. Int J Gynaecol Obstet 2015;130:174–8.
41. Grady WR, Hayward MD, Yagi J. Contraceptive failure in the United States: estimates from the 1982 National Survey of Family Growth. Fam Plann Perspect 1986;18:200–9.
42. Braaten KP, Benson CB, Maurer R, Goldberg AB. Malpositioned intrauterine contraceptive devices: risk factors, outcomes, and future pregnancies. Obstet Gynecol 2011;118:1014–20.
43. Secura GM, Allsworth JE, Madden T, Mullersman JL, Peipert JF. The Contraceptive CHOICE Project: reducing barriers to long-acting reversible contraception. Am J Obstet Gynecol 2010;203:115.e1–7.
44. Cohen R, Sheeder J, Arango N, Teal SB, Tocce K. Twelve-month contraceptive continuation and repeat pregnancy among young mothers choosing postdelivery contraceptive implants or postplacental intrauterine devices. Contraception 2016;93:178–83.
45. Chamberlain G. Self-checking the intrauterine device. Fertil Steril 1977;28:1121–2.
46. Loewenberg Weisband Y, Keder LM, Keim SA, Gallo MF. Postpartum intentions on contraception use and method choice among breastfeeding women attending a university hospital in Ohio: a cross-sectional study. Reprod Health 2017;14:45.
47. Singh RH, Rogers RG, Leeman L, Borders N, Highfill J, Espey E. Postpartum contraceptive choices among ethnically diverse women in New Mexico. Contraception 2014;89:512–5.