If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
To determine if immediate postpartum (PP) intrauterine device (IUD) placement prevents pregnancy and is cost-effective compared with routine placement.
We developed a decision-analysis model to determine the number of pregnancies prevented and the cost-effectiveness of immediate PP IUD placement defined as within the first 10 minutes of placental expulsion compared with routine placement at the PP visit. Associated costs and probability estimates for adherence to PP follow-up, IUD placement, expulsion, and pregnancy were determined from the literature.
Hospital and outpatient facility.
Women desiring PP IUDs.
Main Outcome Measure(s)
The main outcome measure was the number of pregnancies prevented per 1,000 women. The secondary outcome was an incremental cost-effectiveness ratio (ICER) defined as the marginal cost per quality-adjusted life-year (QALY) gained. An ICER of <$50,000/QALY gained was considered to be cost-effective.
Immediate PP IUD placement prevented 88 unintended pregnancies per 1,000 women over a 2-year time horizon. Immediate PP IUD placement was the dominant strategy. For every 1,000 women who desired a PP IUD, attempted immediate PP placement resulted in a cost savings of $282,540 and a gain of 10 QALYs. The model is most sensitive to the cost of an undesired pregnancy. When the cost of a live birth is <$6,000, immediate placement is no longer cost-saving but remains cost-effective. Monte Carlo simulation demonstrates that immediate PP IUD placement is cost-effective in 99% of simulations.
Immediate PP IUD placement is a dominant strategy that prevents unintended pregnancy.
). During pregnancy, given the frequent nature of prenatal visits, physicians have a unique opportunity to discuss, counsel, and readdress contraceptive planning. Despite this fact, within the first postpartum year, women have a particularly high pregnancy rate, ranging from 6% to 40% depending on the population (
). Intrauterine devices minimize these patient-related factors while providing a long-term effective and reversible method of contraception. The vulnerability of postpartum women to unintended pregnancy makes immediate post-placental placement of intrauterine devices (IUDs) an appealing solution (
). Although the American Congress of Obstetricians and Gynecologists state that the immediate postpartum period is a “favorable time” for insertion, this practice has not been widely adopted in the United States (
). Concerns over the possible risk of infection, bleeding, higher expulsion rates, and even reimbursements have limited the use of immediate placement. Multiple studies have been conducted comparing immediate and routine placement of postpartum IUDs. All of the studies show a higher expulsion rate in the immediate post-placental groups (18%) compared with routine placement (4%); however, infection rates are negligible and use at follow-up is similar (
Given these findings, immediate post-placental IUD placement may be advantageous for women, particularly those who may resume intercourse before their postpartum clinic visit or those who do not attend a postpartum visit. However, in this era of health care reform, an important consideration is the cost-effectiveness of such a medical practice. Earlier research has shown that IUDs are highly cost-effective methods of contraception (
). However, the cost-effectiveness of immediate post-placental IUD insertion is unknown. Therefore, we constructed a decision-analysis model to evaluate both the number of unintended pregnancies prevented and the associated incremental cost-effectiveness of immediate post-placental insertion compared with routine insertion at the postpartum visit.
We designed a decision-tree model with the use of TreeAge Pro 2014 (TreeAge Software) that compared the expected costs and health outcomes of immediate and routine placement of postpartum IUDs. Figure 1 shows a simplified schematic of the decision tree. Immediate placement (strategy 1) was defined as placement within 10 minutes of placental expulsion. Routine placement (strategy 2) was defined as placement at 6–8 weeks postpartum. This study used data from both the levonorgestrel intrauterine system and the copper TCu380 IUD. The decision model was used to simulate health system outcomes and costs from the health care perspective over a 2-year time horizon.
Strategy 1 included women who underwent immediate postpartum IUD placement. It was assumed that 18% of these women experienced IUD expulsion before their postpartum visit (
). The original or second IUD either remained in place or was expelled by the 1-year time point. The women whose IUD remained in place at 1 year and were not pregnant continued with that method until year 2. Outcomes included successful contraception, intrauterine pregnancy, ectopic pregnancy, miscarriage, or pregnancy resulting in termination. The costs associated with each outcome were considered.
Strategy 2 included women who received routine IUD placement. Based on earlier studies on postpartum follow-up rates, it was assumed that 81% of these women attended their postpartum visit at 6–8 weeks (
). Although this percentage is based on published research, we acknowledge that there is a wide range based on the population being evaluated. To account for these possibilities, we allowed the pregnancy rate to fluctuate from 0 to 6% in the sensitivity analysis (
). Only women that followed up at the 6–8 week visit and were not pregnant were eligible to receive an IUD, if they still desired it. We assumed that only 64% of women still desired or were able to obtain an IUD (
). This percentage is based on previously published data that took into account the fact that women often change their mind and choose either an alternate method of contraception or no contraception at their postpartum visit. However, again given variation in populations, we varied IUD placement at the postpartum visit from 50% to 100% in our sensitivity analysis. If the IUD was placed, we assumed an expulsion rate of 4.4% (
). Based on earlier studies evaluating IUD use, we assumed that 85% of women in strategy 2 who had an IUD placed continued to use that method at 1 year, compared with 81% in strategy 1. Similarly to strategy 1, the women whose IUD remained in place at 1 year and were not pregnant continued with that method until year 2 (
The model included the following assumptions: 1) All women were sexually active and not attempting to become pregnant during the time horizon of this study; 2) if a woman desired an IUD on the date of placement (either immediate or at the 6–8 week visit), the IUD was placed successfully; 3) in strategy 2, if a woman did not attend her postpartum visit, an IUD was not placed and the only method of contraception used was condoms; 4) all women who followed up at 6–8 weeks but no longer desired an IUD or replacement IUD (in the event of early expulsion) were assumed to use one of three alternate methods of contraception: Depo-Provera, oral contraceptive pills, or condoms; 5) in either strategy, if an IUD was expelled after the 6–8-week postpartum period, the IUD was not replaced and the following assumptions were made: a) In women who did not return to the postpartum visit, we assumed that the only subsequent method of contraception used was condoms; b) in women who initially followed up for the postpartum visit, but whose IUD expelled after 6–8 weeks, we assumed the use one of three alternate methods of contraception: Depo-Provera, oral contraceptive pills, or condoms.
The base-case estimates were chosen from data in the published literature. To obtain these estimations, a bibliographic survey in Pubmed was performed with the use of the following search terms: intrauterine device, postpartum, immediate, delayed, cost-effectiveness, pregnancy rates, follow-up, contraception, and combinations of these terms. Point estimates were determined from randomized controlled trials and prospective cohorts when possible. Retrospective cohorts and expert opinions were used sparingly when prospective data was not available. The probability estimates and references used in support of our model are reported in Table 1. Our search was limited to studies that used either the levonorgestrel intrauterine releasing system or the TCu380 IUD after both vaginal and cesarean-section births.
Table 1Probability estimates and references used in the support of the model.
). For women who followed up for their postpartum visit but no longer desired an IUD or replacement IUD (in the event of early expulsion), we assumed that they selected one of the three most commonly used reversible contraceptive agents: Depo-Provera, oral contraceptive pills, or condoms (
). NVSS did not report on ectopic pregnancy rates. Given that the event rate for ectopic pregnancy is an important outcome when evaluating IUDs, this event rate also needed consideration. The most recent data from the CDC reported an annual ectopic pregnancy rate of 19.7/1,000 pregnancies (
). Using this data and assuming the same distribution as published from the NVSS, we estimated the pregnancy outcomes taking into consideration ectopic pregnancy. A similar method was used when determining rates of pregnancy outcomes among IUD users. Failure rates and ectopic pregnancy rates at 1 year were determined based on manufacturer data and previously reported outcomes in the literature (
). Probabilities of pregnancy outcomes with an IUD in place were again based on annual failure rates and were estimated assuming a similar distribution of pregnancy outcomes as previously reported by the NVSS.
The costs used in the model are presented in Table 2. The cost of the IUDs, Depo-Provera, and oral contraceptive pills were previously published by Trussell et al. and converted to 2014 U.S. dollars by accounting for inflation rates based on the consumer price index from the U.S. Department of Labor (
). The cost of an office visit was incurred for women who presented at the 6–8-week follow-up visit. Except in the circumstance of IUD expulsion, the cost of IUD removal was incurred for any method failures.
The cost of method failures that resulted in pregnancy termination were reported by Jerman et al., assuming that the management occurred in a nonhospital setting and that 23% of terminations were managed medically (
). The cost of an ectopic pregnancy was calculated based on published data on the average costs for laparoscopic and single-dose methotrexate treatment, assuming that 67% of ectopic pregnancies are managed medically and 33% managed surgically (
). The cost of a pregnancy was estimated from a 2004 March of Dimes study that reported costs of 9 months of prenatal care, delivery-related costs, and 3 months of postnatal care and converted to 2014 dollars (
A utility is a way of assigning value to a particular health state, where “perfect health” is equal to 1 and death is equal to 0. Depending on the severity of a particular disease state, “perfect health” is discounted. A quality-adjusted life-year (QALY) is a summary of utilities over a particular time period. The utility values for the various health states in the present model were derived from data published by Sonnenberg et al. using the time-tradeoff technique (
). Year 2 utilities were discounted at a baseline rate of 3%. The range was varied from 1% to 5% in the sensitivity analysis. Utility data are outlined in Table 2. We included only maternal utilities.
For a cohort of 1,000 women, we calculated the cost of care for each strategy and the number of pregnancies prevented. The secondary outcome of the study was cost-effectiveness measured as the incremental cost-effectiveness ratio (ICER). ICER is the cost incurred to gain QALYs (a sum of the utilities). A strategy is considered to be cost-saving if the ICER is negative (decreased costs for increased QALYs). Additionally, a strategy is cost-effective if the ICER is less than society's willingness to pay for 1 QALY and not cost-effective if the ICER is more than society is willing to pay for 1 QALY. Although some debate exists in the literature, for the present study an ICER of $50,000/QALY gained was considered to be cost-effective.
We performed univariate analyses by varying the values of the variables in the model to their plausible extremes. We also performed a Monte Carlo simulation (a computational algorithm that relies on repeated random sampling) with 10,000 trials to assess the robustness of the model. The primary outcome was cost per unintended pregnancy. Other outcomes calculated included ICER, total cost of each strategy, total QALYs per strategy, incidence of unintended pregnancy, and total IUDs consumed in each strategy.
Institutional Review Board approval was not necessary for this study. All data points are based on previously published literature. The authors have no conflicts of interest.
Our base-case analysis suggests that immediate postpartum IUD placement (strategy 1) is not only cost-effective but also cost-saving (Table 3). Immediate postpartum IUD placement results in cost savings of ∼$282,540 per 1,000 women over a 2-year time horizon. Although unintended pregnancies occurred in both groups, immediate placement compared with routine placement is expected to prevent an additional 88 unintended pregnancies over 2 years per 1,000 women. Fifty-one of these pregnancies are prevented within the 1st year, and 37 additional pregnancies are prevented in year 2. According to our data, this translates into a cost savings of >$3,200 for each unintended pregnancy. In terms of IUD use, our data also suggests that more IUDs are used in the immediate strategy compared with routine. For every 1,000 women who desire a postpartum IUD, 1,090 IUDs are placed in strategy 1 compared to 510 in strategy 2.
Given that many of the variables used in the model change significantly depending on the population, we performed univariate analyses examining each variable at its extremes. The results of these analyses demonstrate that the model is robust. There is no plausible change in a single variable that makes immediate placement not cost-effective compared with routine placement. The model is most sensitive to cost of pregnancy. When the cost of a live birth is <$6,000, immediate placement is no longer cost-saving but remains cost-effective. For immediate placement to not be cost-effective, variables have to shift to implausible values. For example, when the probability of IUD expulsion at 6 weeks exceeds 56%, immediate placement is no longer cost-effective. In addition, when the cost of an IUD exceeds $10,000, immediate placement is not cost-effective.
Given the robustness in one-way sensitivity analysis, a Monte Carlo simulation was performed that allows all of the variables in the model to vary simultaneously to their extremes as listed in Tables 1 and 2. This simulation demonstrates that immediate postpartum IUD placement is cost-saving in 87% of simulations and cost-effective but not cost-saving in an additional 12% of simulations. As stated above, this study assumed a willingness to pay $50,000 to prevent one additional unintended pregnancy.
Earlier studies have documented that contraception, compared with no contraception, not only gives rise to significant financial savings, but is also cost-effective and results in increased QALYs (
). Our study echoes these findings, because the cost-effectiveness of immediate IUD placement depends significantly on the cost of an undesired continued pregnancy. The results of this model emphasize that immediate postpartum IUD placement is a cost-saving strategy in most circumstances, leading to more pregnancies prevented. Similarly to previously published data, our study found that immediate postpartum IUDs prevent an additional 88 pregnancies per 1,000 women over routine placement in the first 2 years postpartum while also providing significant cost savings to the health care system.
Much of the reluctance to utilize immediate post-placental IUD placement stems from concerns over the significantly higher expulsion rate in the immediate postpartum period (
). Our model confirms that more than twice as many IUDs are used in strategy 1. Despite the increased number of IUDs consumed in the immediate postpartum group, this additional cost is offset by a reduction in unintended pregnancies and their associated costs. According to our data, IUD expulsion in the immediate postpartum period would have to exceed 38% to eliminate the cost savings associated with immediate placement. Moreover, immediate IUD expulsion would need to exceed 56% to no longer be cost-effective. These expulsion rates are much higher than any reported in the literature (
Although our model shows that immediate postpartum placement is cost-effective in the majority of situations, there are limitations to this study. The first limitation is that cost-effectiveness models are based on multiple assumptions, and the outcomes of the study can vary considerably based on the initial assumptions made. Our model made several assumptions regarding alternate contraceptive use based on previously published data examining patterns of contraceptive use among women of reproductive age (
). We assumed that women who declined IUDs opted for condoms, oral contraceptives, or Depo-Provera. Alternatively, if we had assumed women opted for tubal ligations or other more reliable contraceptive options, the model would show different results. A second important weakness is that cost-effectiveness studies are also based on previously published data and therefore depend on the quality of those data. Unfortunately, there are very few randomized controlled trials comparing the efficacy and risks associated with delayed versus immediate postpartum IUDs. The majority of the data were derived from comparative studies and a Cochrane review (
). That study tested the validity of the utilities with a sensitivity analysis, and the results were consistent. The limited utility data are part of the reason we chose a primary outcome of cost per pregnancy prevented. There are also very little published data on pregnancy rates in the immediate postpartum period. The range used in the present study was based on a study conducted by Ogburn et al., which examined barriers to IUD placement in postpartum women and found a pregnancy rate of ∼6%, the maximum used in our study (
Another significant limitation is that our study was based on a 2-year time horizon. Therefore, we can not specifically state that immediate postpartum placement remains cost-effective over multiple years. However, the price of acquisition comprises the bulk of the cost of IUD placement. Failure rates either remain the same, as in the levonorgestrel intrauterine system, or decrease over the years, as in the TCu380; therefore, it is likely that immediate postpartum placement would only become increasingly cost-effective if a longer time interval were considered. Our model design allowed for women to choose an alternate method of contraception in the event of an IUD expulsion within the first 6 weeks; however, it assumed that women with an IUD in place used that method continuously over the 2-year period. This assumption was made because there are very sparse data on the rates of switching between contraceptive methods. A recent study published by O'Neil-Callahan et al. found that IUDs have the highest continuation rates at 24 months compared with all other methods of contraception (
). Finally, we attempted to account for all direct costs of IUD placement and pregnancy outcomes. Indirect costs were not considered. However, the indirect costs of IUD acquisition and placement are likely minimal compared with the indirect costs of pregnancy outcomes. Therefore, by not including indirect costs we likely underestimated cost-effectiveness.
The immediate postpartum period represents an ideal time for IUD placement, because women often have epidurals, there are no additional appointments required, and the women have no risk of being pregnant. Although immediate IUDs have a higher expulsion rate, this study suggests that immediate placement is a cost-saving strategy from a health care perspective.
Use of contraception in the United States: 1982–2008. National Center for Health Statistics.