Preventing Unintended Pregnancy

An actuarial approach to long-acting reversible contraceptives Kari D. Berglund and Sara C. Teppema

The field of public health consists of professionals, infrastructure and policy that support solutions and interventions that enhance the lives and well-being of people living in the communities they serve. These professionals study and track these programs, typically focusing on long-term outcomes. As actuaries, we can adapt their findings to the populations we serve and can add further perspective through modeling that quantifies both costs and benefits across a full population.

This article illustrates such an adaptation in the context of long-acting reversible contraception methods (LARCs). LARCs have the potential to reduce unintended pregnancies, an important public health goal, at higher rates than other methods.

We open with a brief review of LARC-related research and conclude by presenting a simple model that enables the practicing actuary to translate LARC research into a quantifiable cost-benefit estimate. The model offers a framework within which the practicing actuary can apply population-specific assumptions and quantify the net cost savings of this public health intervention.

LARC Research and Usage

LARC methods have been available since the late 1960s. Their use in the United States rose quickly and fell precipitously due to health and safety concerns that emerged after their introduction.1 Now that those concerns have been addressed, LARC use is again on the rise, and because they are highly effective, LARCs have great potential to prevent unintended pregnancies.

The American College of Obstetricians and Gynecologists defines LARC methods as intrauterine devices (IUDs) and contraceptive implants. The contraceptive implant is currently approved for use for up to three years.2 A copper IUD can be used for up to 10 years, and a levonorgestrel-releasing IUD can be used for three to five years (depending on the specific model selected).3 LARCs are highly effective and reversible upon removal,4 and they have been shown to be safe in most patient populations.5,6

Research shows that LARCs are the most effective form of birth control, with pregnancy rates of less than 1 percent.7,8 Oral contraceptives and male condoms are more commonly used reversible contraceptive methods in the United States, but both methods have higher failure rates due to user error.9 Despite this, LARC use is less common in the United States compared to many other countries. In a 2015 United Nations study of the prevalence of contraceptive methods among married or in-union women aged 15–49, 13.7 percent of women surveyed around the world utilized an IUD, and 0.7 percent utilized an implant. In the United States, by contrast, only 5.1 percent of women utilized an IUD, and 1.0 percent utilized an implant, despite a high level of overall contraceptive use (75.1 percent).10 After LARC usage decreased between 1982 and 1988 in the United States, LARC usage remained stable from 1988 to 2002 and has increased steadily since 2002. Across women aged 15–44, LARC usage has increased from 1.5 percent in 2002 to 7.2 percent in 2011–2013.11

LARC use, in fact, varies widely by country.12 In Mexico, 72.6 percent of women surveyed used a contraceptive, with 11.8 percent utilizing an IUD and 0.0 percent using an implant.13 In Kenya, despite a lower overall contraceptive use rate (57.4 percent), more women used LARCs (10.8 percent implant, 3.5 percent IUD).14 Perhaps, surprisingly, and despite high contraceptive use overall (73.3 percent), only 1.0 percent of Canadian women used an IUD, and 0.0 percent used an implant.15

The Affordable Care Act (ACA) requires that insurance plans offer FDA-approved contraceptive methods at no cost to members.16 Non-grandfathered commercial plans under the ACA must cover at least one form of contraception within each of the 18 contraception method categories the FDA defines.17 Medicaid pays for contraceptive and family planning services, and all family planning services and contraceptive methods available under a Medicaid plan must be available to members without cost sharing.

A 2015 Kaiser Family Foundation study shows that, despite the ACA mandate, some insurance plans still may not cover specific IUDs or implants without cost sharing.18 One insurance plan did not cover Paraguard, the only non-hormonal IUD available, and some plans required prior authorization for implant or IUD insertion.19 However, for most women, the ACA eliminated out-of-pocket costs for IUDs.20 An early study of the impact of the ACA on LARC initiation rates suggests the ACA did not significantly increase LARC initiation rates during its first year.21 Additional studies may be needed to evaluate the effect of the ACA on LARC initiation rates.

Even if women have access to no- or low-cost LARCs, barriers to LARC initiation remain. Only half of obstetrician-gynecologists offer patients the hormonal implant due to lack of training and patient interest in the implant.22 Women may also have limited knowledge about LARCs. In a recent 2016 study of U.S. women of reproductive age, only 31 percent of women had heard “a lot” about IUDs and implants. Less than half of women knew IUDs were very effective at preventing pregnancy, and only 37 percent of women knew implants were very effective at pregnancy prevention.23

Early programs to increase LARC adoption have shown that LARC adoption increases once LARCs are easier to obtain. Through a program initially funded by a private foundation and now funded by the state budget, Colorado subsidized LARCs for low-income women and girls, increased usage of LARCs, and decreased teen birth and abortion rates faster than national averages.24

LARC use has increased in programs that address some of the residual barriers to LARC initiation:

  • Beginning in 2009, all providers in Gaston County, North Carolina, were trained to insert IUDs or implants, and funding covered the cost for the uninsured. LARC usage increased to nearly 30 percent among teens, and the rate of teen pregnancy dropped 45 percent between 2010 and 2015.25,26
  • In 2007, the Iowa Initiative to Reduce Unintended Pregnancies provided funding to 17 Title X family planning agencies in Iowa. From 2007 to 2012, the percentage of reproductive-age clients at Title X family planning agencies in Iowa using LARC methods increased from less than 1 percent to 15 percent as LARC methods became more available.
  • From August 2007 through September 2011, the Contraceptive CHOICE Project in St. Louis enrolled women aged 14–45 who wanted to avoid pregnancy for at least a year and were interested in new forms of contraception. Seventy-five percent of participants selected a LARC method, and participants in the study had a much lower teen pregnancy rate (34.0 pregnancies per 1,000 teens versus 158.5 per 1,000 nationwide). Rates of teen births were also lower (19.4 per 1,000 teens versus 94.0 per 1,000 nationwide).27

Close to half of all pregnancies in the United States are unintended, so increased LARC initiations can decrease unintended pregnancies in the United States and also decrease cost and utilization related to maternity services.28,29 A 2012 study found that if 10 percent of women aged 20–29 switched from oral contraceptives to LARCs, total medical costs in the United States would decrease by $288 million per year by preventing unintended pregnancies.30 Such results prompt the need for an actuarial cost-benefit analysis.

A Simple Actuarial Cost-Benefit Model

Figure 1 outlines a simple actuarial model, which can be constructed for a population of women of childbearing age who are avoiding pregnancy. This model shows the significant opportunity for cost savings if members of the population switch contraception methods from either oral contraceptives or condoms to IUDs.

Figure 1: Actuarial Cost-benefit Analysis of Switching Contraception Methods
In a given year for each woman of childbearing age:
IUD Pill Condoms
Contraception Costs
Three-year cost $1,0091 $7202 $3603
Annual cost per member per month (PMPM) for women using contraception $28.03 $20.00 $10.00
Cost of Contraception Failure
Annual contraception failure rate4   1%   9% 18%
Percentage of unintended pregnancies carried to term5 58% 58% 58%
Maternity and delivery costs6 $13,430 $13,430 $13,430
Maternity and delivery cost PMPM per woman using contraception     $6.49   $58.42  $116.84
Net Cost of Contraception
Contraception cost PMPM $28.03   $20.00     $10.00
Maternity and delivery costs PMPM +$6.49 +$58.42 +$116.84
Total cost PMPM per woman using contraception $34.52   $78.42   $126.84
Savings from IUD use over using oral contraception ($43.90)
Savings from IUD use over using condoms ($92.32)
Annual Savings for Example Population
Assumed population using contraception 10,000
Assumed portion of the population using oral contraceptives7 26%
Assumed portion of the population using condoms8 15%
Potential upper bound savings from IUD use over other methods ($3,031,391)

Notes
1 “Birth Control—Intrauterine Device.” Guroo. National average in 2017 including device and office visit, assuming three years.

2 Assume $20.00 per monthly prescription, 12 per year, for three years.

3 Assume $1.00 each, 10 per month, for three years.

4 Aisch, Gregor, and Bill Marsh. 2014. “How Likely Is It That Birth Control Could Let You Down?” The New York Times. September 13.

5 “Unintended Pregnancy in the United States.” 2016. Guttmacher Institute. September 20.

6Childbirth—Vaginal Delivery and Newborn Care.” Guroo. The national average vaginal birth and pregnancy cost is $12,560 (accounts for 75 percent of births). The national average cost for cesarean section birth and pregnancy is $16,038 (accounts for 25 percent of births). Therefore the average cost of pregnancy and delivery is $13,430. Although the source provides Chicago-specific costs (the Guroo tool requires geographic specific information), only national costs were incorporated into the averages.

7 “Contraceptive Use in the United States.” 2016. Guttmacher Institute. September 21.

8 Supra note 7.

The model incorporates publicly available national statistics for contraception costs, contraception utilization, contraception failure rates, delivery rates and maternity/delivery costs representative of the national average for a commercially insured population. These statistics may not be appropriate for a given specific population, and the practicing actuary should replace these with more appropriate measures for the specific population to be studied. Much of the public data used in this model is available for specific states or regions.

Further enhancements to this model might include the incorporation of statistics for a specific age group (for example, adolescents), or for a differing geographic area that can vary in cost and utilization patterns. The model might also be enhanced to include participant cost sharing, to the extent applicable.

How Might Practicing Actuaries Use This Model?

Despite assumption caveats, the model is likely directionally correct: IUDs are a safe and cost-effective alternative to oral contraceptives and condoms. After selecting appropriate assumptions, how might the practicing actuary use findings such as these? One application might be to justify an investment in promotion of IUD use among providers for Medicaid patients. Another application might be to evaluate the removal of member cost-sharing barriers in a commercial population.

Beyond these straightforward applications, actuarial cost-benefit analyses of LARC replacement could be extended to include a broader social impact. A model might apply the outcomes of family planning programs (such as those described earlier) to long-term societal savings. For example, a reduction in adolescent pregnancy rates would result in higher high school and college graduation rates among young women, resulting in higher lifetime incomes, lower burden on public assistance programs and even better outcomes for their later-born children.

Other uses of this model may also be appropriate, further demonstrating how a simple actuarial model can be leveraged beyond its traditional application to bring a new dimension to health services research, optimizing public health resources and improving the lives of people and society as a whole. The authors welcome future discussion of this topic and ideas to extend the model beyond the traditional view.

Kari D. Berglund, MSc, is an advanced analytics consultant on the Provider Payment Innovation and Analytics team at Blue Cross Blue Shield of Illinois, Montana, New Mexico, Oklahoma and Texas in Chicago.
Sara C. Teppema, FSA, MAAA, is DVP and actuary on the Provider Payment Innovation and Analytics team at Blue Cross Blue Shield of Illinois, Montana, New Mexico, Oklahoma and Texas in Chicago.