Utilization of Long-Acting Reversible Contraceptives in the United States After vs Before the 2016 US Presidential Election

Abstract

This study examines the utilization rate of long-acting reversible contraceptives after the 2016 US presidential election compared with before the election and during the same period in 2015.

Soon after the US presidential election on November 8, 2016, media and industry reports described an increase in utilization of long-acting reversible contraceptive (LARC) methods (intrauterine devices and implants).¹ Proposed reasons included women’s concerns about contraceptive access should the Patient Protection and Affordable Care Act (ACA) be repealed during the Trump Administration. These reports, however, were descriptive and did not control for seasonal or secular trends. Using data from a large sample of commercially insured women, we sought to assess whether there was an increase in LARC utilization among commercially insured women during the 30 days after the election, compared with the 30 days before the election and the same period in 2015.

Methods

Using the IBM/Truven MarketScan Analytics Commercial Claims and Encounters Database, we studied women aged 18 to 45 years enrolled in commercial insurance during the 30 business days before and after November 8 in 2015 or 2016 who had at least 12 months of continuous enrollment. We used billing codes (eTable in the Supplement) to calculate daily LARC insertion rates during the 30 business days before (inclusive of November 8) and 30 days after November 8 in 2015 and 2016. To account for secular trends, we estimated changes in daily LARC insertions using a difference-in-differences generalized linear model with a Poisson distribution and log link function that compared the change in probability of LARC insertion during the 30 business days before vs after November 8, 2016, with the change in the comparable period in 2015. With person-day as the unit of analysis, we adjusted for age group, region, relationship to the insured individual, and plan type (Table) and accounted for clustering by individuals (because individuals contributed information for multiple time points). Wald tests were used to calculate the P values. We considered a 2-sided P < .05 to indicate statistical significance. The Harvard Medical School Office of Human Research Administration exempted the study from human subject review; therefore, patient consent was not required. All data were deidentified.

Table. Characteristics of the Study Sample^a.

Characteristic	Study Year
	2015 (n = 3 449 455)	2016 (n = 3 253 703)
Age, mean (SD), y	31.8 (8.3)	31.8 (8.4)
Age group, y
17-25	1 053 909 (30.6)	1 008 641 (31.0)
26-35	1 080 818 (31.3)	990 207 (30.4)
36-45	1 314 728 (38.1)	1 254 855 (38.6)
Region
Northeast	604 647 (17.5)	566 695 (17.4)
Midwest	719 795 (20.9)	671 995 (20.7)
South	1 488 945 (43.2)	1 399 966 (43.0)
West	625 392 (18.1)	599 055 (18.4)
Unknown	10 676 (0.3)	15 992 (0.5)
Relationship of patient to insured
Self (employee)	1 606 226 (46.6)	1 481 114 (45.5)
Spouse	926 129 (26.9)	854 588 (26.3)
Dependent	917 100 (26.6)	918 001 (28.2)
Health plan type
Comprehensive	76 937 (2.2)	84 111 (2.6)
Exclusive provider organization and health maintenance organization	365 457 (10.6)	347 518 (10.7)
Point of service	217 266 (6.3)	196 751 (6.1)
Preferred provider organization	1 797 425 (52.1)	1 657 299 (50.9)
Consumer-driven and high-deductible health plans	969 190 (28.1)	921 875 (28.3)
Missing	23 180 (0.7)	46 149 (1.4)

^aData are presented as number (percentage) of patients unless otherwise indicated.

Results

Among 3 449 455 women in 2015 (mean [SD] age, 31.8 [8.3] years) and 3 253 703 women in 2016 (mean [SD] age, 31.8 [8.4] years), demographic and health plan characteristics were similar (Table). In 2015, the mean adjusted daily LARC insertion rate during the 30 business days before and inclusive of November 8 was 12.9 per 100 000 women vs 13.7 per 100 000 women during the subsequent 30 business days. The comparable mean adjusted daily LARC insertion rates before and after the 2016 presidential election were 13.4 per 100 000 women and 16.3 per 100 000 women (Figure), an increase of 21.6%. The adjusted difference-in-difference rate was 2.1 insertions per 100 000 women per day (P < .001).

Figure. Long-Acting Reversible Contraceptive (LARC) Insertions per 100 000 Women Before and After November 8, 2015, and 2016.

Dashed line indicates midnight on November 8.

Discussion

We observed a significant increase in LARC insertions among commercially insured women during the 30 business days after the 2016 presidential election, adjusting for secular and seasonal trends and patient characteristics. If our findings were projected to the approximately 33 million women in the United States aged 18 to 45 years in 2016 with employer-sponsored health insurance, this rate would correspond to approximately 700 additional insertions per day in association with the 2016 election, above the approximately 4716 insertions per day that would have been otherwise predicted.

Long-acting reversible contraceptive methods are highly effective at preventing pregnancy for up to 12 years and are widely recommended as first-line contraception.² Long-acting reversible contraceptives remain underused in the United States.³ In 2012, the ACA eliminated cost-sharing for contraception for most women with private health insurance. Median out-of-pocket costs for an intrauterine device decreased from $40 to $0,⁴ and LARC insertions increased.⁵

Our study has limitations. We lacked information about important covariates, including race/ethnicity; we studied a short period after the election; and we only studied women with commercial insurance. Our findings may not generalize to women with public insurance or no health insurance. We could not determine whether the increase in LARC insertions was sustained or evaluate its public health significance. Although 2017 data were not yet available at the time of our study, we could repeat this analysis for that year.

The ACA’s contraceptive coverage mandate is an important strategy to reduce unintended pregnancies. The Trump Administration has weakened this mandate.⁶ Our findings could reflect a response to fears of losing contraceptive coverage because of President Trump's opposition to the ACA or an association of the 2016 election with reproductive intentions or LARC awareness. Our findings also suggest that women with commercial health insurance value contraceptive coverage and that concerns about potential reductions in access or coverage may affect their contraceptive choices.

Supplement.

eTable. Billing Codes Used to Identify Insertions of Long-acting Reversible Contraceptive Methods in 2015 and 2016