The following is a guest post by Daniel Liebman, a research assistant for Dr. Ashish Jha at the Harvard School of Public Health, and a part-time research assistant for The Incidental Economist. He graduated from Brandeis University in 2012 with degrees in Health Policy and American Studies, and will begin at Harvard Medical School in Fall 2014. He tweets about good policy and bad puns at @D_Liebman.
[Note from Austin and Daniel: Contraceptives’ primary benefit is the betterment of women’s health, not cost savings. Nothing below or in Austin's Upshot post that links to this should imply otherwise!]
Birth control is back in the spotlight again (again). By now you are probably well-aware of the Supreme Court’s ruling in the Hobby Lobby case, which found that closely-held corporations can disregard the federal requirement to provide contraceptive coverage on religious grounds. (Prior attention to the case and issues on TIE here, here, here, here, and here.)
The ensuing public discussion (read: “firestorm”) has focused primarily on the legal and sociopolitical ramifications of this decision. A health economics question raised briefly by Austin lurks in the background: does coverage for contraception pay for itself in better women’s health and avoided pregnancies?
A few studies relevant to this question are referenced in the administration’s final rule providing an accommodation to religious non-profit organizations who do not wish to cover contraception. I suspected there might be more work on the topic out there, however, and took a look at the literature. Here’s what I found in full; below are the main takeaways.
Not everyone is convinced that the mandate is completely cost-neutral.
For Factcheck.org, Ben Finley performed an analysis of the contraceptive mandate back in 2012, and concluded that the evidence regarding cost-neutrality for insurers was “conflicting and inconclusive”. His analysis looked at a number of studies cited by both the Obama Administration and by critics of the mandate, most of which are summarized here as well. I found some other studies and additional points that are also worth considering.
From a population-wide perspective, contraception is cost-effective and cost-saving.
The arithmetic is fairly straightforward: last year, Elisabeth Rosenthal reported average commercial insurance payouts of $18,329 for a vaginal delivery and $27,866 for a C-section. Prices without insurance are even higher: $30,000-$50,000. All told, unintended pregnancies generate direct costs of over $5 billion annually (Trussell, 2007).
Contraception, on the other hand, normally costs roughly $100-$600 per year. Typical contraceptive use drops a woman’s risk of an unintended pregnancy from 85% within one year to 9% with the “Pill”, and to under 1% for hormone implants and IUDs, directly saving $19.3 billion annually (Trussell, 2007 and 2011). Sonnenberg and colleagues (2004) estimated societal savings of approximately $6,000-$10,000 per person every two years from all forms of contraception, while Trussell et al. (1995) calculated savings of approximately $13,000 (circa 1995 dollars) over five years.
A number of studies have also demonstrated that societal investment in contraceptive coverage is cost-saving among low-income women, for whom the up-front cost of most contraceptives might otherwise be a deterrent. Sawhill et al. (2010), for example, estimated that an expansion of Medicaid-subsidized contraceptive services saves $4.26 for every dollar invested; Frost et. al (2008) calculate a savings of $3.74 per dollar, and a 2012 Brookings report pegs the savings at $5.60-to-$1. Similar savings are reported by Foster et. al (2009), Frost, Finer & Tapales (2008), and the CBO. Just days ago, in fact, Colorado reported that implementation of its Family Planning Initiative—which provides long-acting reversible contraceptives to young low-income women—coincided with a 40% reduction in teen births and a 35% drop in teen abortions, at an estimated public savings of $42.5 million in 2010.
From a typical insurer’s perspective, contraceptive coverage might not be cost-saving.
While the cost-savings of contraception on a societal scale are clear, it is less apparent whether an individual insurer will necessarily reap these savings. This is because many women (with, perhaps, the exception of very low-income women) will choose to purchase birth control out of pocket even if it’s not covered, precisely because it is so cost-effective. When an insurer picks up the tab for contraception, it often assumes the cost without much change in savings, since most potential pregnancies would be averted regardless of whether or not the insurer covered the contraception.
One paper cited by ASPE for its prediction of cost-savings—Trussell et al. 1995—specifically notes in its discussion “if a payer simply finances the contraceptives that would otherwise have been purchased by individuals…the payer’s net costs are likely to increase”. The same conclusion was drawn in a 2000 Milliman Study assessing Texas’ contraceptive mandate.
This notion of demand inelasticity for contraception was specifically tested in a preliminary study by Collins & Hershbein (2011), which found that an overnight several-fold increase in the price of the Pill on college campuses after the Deficit Reduction Act of 2005 only led to a 2-4% reduction in usage, with no reported increase in rates of unintended pregnancy. (Poorer and uninsured students were much more likely to reduce usage, however, again demonstrating the aforementioned heightened impact of cost barriers among disadvantaged socioeconomic groups.)
The Obama Administration, for its part, has offered a 2012 ASPE Issue Brief to bolster its argument that the contraception mandate is cost-neutral for insurers. The two strongest citations therein are a memorandum stating that the 1999 contraceptive mandate for the Federal Employees Health Benefits Program did not change premium levels, and a 2001 report from Hawaii’s insurance commissioner concluding that there was no net effect on health insurance costs for four insurers following implementation of a state mandate. Other citations are somewhat less compelling—a2007 National Business Group on Health report concludes that “most [contraceptives] are cost-saving from the private-payer perspective”, but bases its findings on a PricewaterhouseCoopers actuarial report which assumes a deductible and coinsurance. Further, a paper from Trussell and colleagues (1997) estimates private-sector savings of $308-$946 for contraceptive use among adolescents over one year, but its analysis only compares contraceptives to “no method”, telling us little about net costs for insurers for a population (e.g., workers) likely to self-fund contraception.
The Administration also cites a study demonstrating millions in savings across six states following Medicaid contraception coverage expansions, as well as CBO-projected savings of $400 million if Medicaid family planning were to be expanded in all states. However, expanding contraceptive coverage in the Medicaid population is not the same as a private insurance expansion, since the price barrier of contraceptives is likely greater among Medicaid recipients.
There is also evidence to suggest that premium increases are a possible outcome of contraceptive mandates.
There are several studies that find evidence of premium increases caused by contraceptive mandates. Some are even mentioned by ASPE, such as a 1998 report using data from Buck Consultants estimating an increase of $21/year, and a 2003 PricewaterhouseCoopers report projecting an additional $41/year.
State actuarial reviews have also found slight but tangible premium increases attributable to their own contraception mandates: Texas found that oral contraceptives accounted for 0.18% of total claims in 2006; Connecticut estimated premium increases of $14.64 for individuals and $17.28 for groups in 2010; Maryland estimated that contraceptive coverage contributed to 0.6% of group premiums and 0.7% of individual premiums; Massachusetts estimated contraceptive costs accounted for 0.44% of premiums in 2008. However, many of these analyses focused on the up-front cost of contraception and did not consider long-term savings due to averted pregnancies.
Contraception-driven premium increases are small relative to overall health care costs, but savings are down the road.
It is important to bear in mind that the average “silver” premium on the exchanges is around $4,000 annually. Relative to that, a 15-40 dollar contraception bump, were it to manifest, is not exorbitant. We spend $750 Billion on unnecessary health services each year; contraception is certainly not one of them.
Still, to an insurer, that’s a cost, and one that might not be immediately recouped. As noted above and by the administration, actuarial studies like those cited above do not take always into account the longer-term savings from averted pregnancies, as well as the indirect cost savings from productivity that would otherwise be lost.
Some studies do indicate that most or all contraceptive technologies are cost-effective in the long run for payers (Trussell, 2009;Crespi, Kerrigan & Sood, 2013). Nevertheless, these studies stop short of predicting cost savings. Indeed, the actuarial studies described above were based on insurers’ experiences and pricing behavior, suggesting that from their point of view, contraceptive coverage bears an up-front cost that must be absorbed.
In the end, there isn’t enough evidence to conclude that contraceptive coverage will pay for itself from an insurer’s perspective, and certainly not in the short term. There is strong evidence that public contraceptive funding for underserved populations is cost-saving, and there is a chance that the cost-neutrality observed following the FEHB and Hawaii mandates will materialize for other insurers as well. There are undeniable economic benefits of contraception for society as a whole, as well as a multitude of social benefits that could fill many posts of their own. Focusing specifically on the economics of insurance, however, the literature on the subject is sparse. All told, we currently have little evidence to indicate the time frame needed for private insurers to realize cost offsets or savings, if there are indeed any to be had.
UPDATE: Be sure to read this follow-up, which considers the cost-effectiveness of different types of contraception.