The Pennsylvania State University. The Graduate School. Department of Public Health Sciences

The Pennsylvania State University The Graduate School Department of Public Health Sciences THE IMPACT OF THE AFFORDABLE CARE ACT ON CONTRACEPTIVE USE AND COSTS AMONG PRIVATELY INSURED WOMEN A Thesis in Public Health Sciences by Ashley Hewlett Snyder 2017 Ashley Hewlett Snyder Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Science May 2017

ii The thesis of Ashley Hewlett Snyder was reviewed and approved* by the following: Cynthia H. Chuang Professor of Medicine and Public Health Sciences Thesis Advisor Douglas L. Leslie Professor of Public Health Sciences and Psychiatry Kristen H. Kjerulff Professor of Public Health Sciences Vernon M. Chinchilli Distinguished Professor Chair of Public Health Sciences *Signatures are on file in the Graduate School

iii ABSTRACT Objectives: The contraceptive coverage mandate in the Affordable Care Act (ACA) requires private health insurance plans to cover all Food and Drug Administrationapproved contraceptive methods without cost-sharing. We evaluate the impact of this policy on cost and use of long-acting reversible contraceptives (LARCs) and other prescription methods through 2014. Study Design: Data from Truven Health MarketScan were used to examine out-ofpocket costs and contraceptive use patterns for all reversible prescription contraceptives before and after implementation of the contraceptive mandate in August 2012. Study cohorts for calendar years 2005 to 2014 included women ages 13-45 with continuous medical and pharmacy coverage within each year. Costs were estimated by combining copayment, coinsurance and deductible payments for both contraception and insertion fees for LARCs. Multivariable logistic regression comparing LARC insertions pre- and post-aca was performed adjusting for year, age group, geographic region, and urban versus rural residence. Results: Rates of new LARC insertions increased over each study year. Rates of claims for LARC insertions decreased slightly post-aca when controlling for cohort year, age group, geographic region, and rural versus urban setting. Out-of-pocket costs for LARCs decreased sharply post-aca contraceptive mandate. Conclusions: While out-of-pocket costs for prescription contraceptives decreased post- ACA, there was not an increased uptake of LARCs beyond what would be expected based on the secular trend. The full impact of the contraceptive mandate on contraceptive use patterns may not be clear until more years of data are available.

iv TABLE OF CONTENTS List of Tables..v Acknowledgements...vi INTRODUCTION.. 1 MATERIALS AND METHODS...1 Data Source and Inclusion Criteria....1 Measures of Contraceptive Use..2 Measures of Contraceptive Cost.3 Statistical Analysis..3 RESULTS....3 DISCUSSION....6 REFERENCES......8

v LIST OF TABLES Table 1. Characteristics of reproductive age women by 3 year, 2005-2014. Table 2. Mean and median out-of-pocket costs by contraceptive 4 type, 2005-2014 (dollars). Table 3. LARC insertion rates and other contraceptive use by 5 year, 2005-2014 (percent of sample). Table 4. Adjusted odds of LARC insertion (IUD or implant) pre- 5 and post-aca implementation (n=57,027,745).

vi ACKNOWLEDGEMENTS This research was funded by the Robert E. Dye, M.D. Professorship at the Penn State College of Medicine and by the Penn State Center for Women s Health Research.

1 INTRODUCTION Long-acting reversible contraceptives (LARCs), which include the intrauterine device (IUD) and contraceptive implant, are highly effective forms of reversible prescription contraception. LARCs have become more affordable to insured women as a result of the contraceptive coverage mandate of the Affordable Care Act (ACA), which took effect in August 2012. The mandate requires most private health insurance plans to cover all Food and Drug Administration (FDA)-approved contraceptive methods without cost-sharing [1-3]. Prior to the ACA, the higher upfront out-of-pocket costs of LARCs likely discouraged women from choosing them over less effective prescription birth control methods with lower out-of-pocket costs [1, 2, 4]. A recent study by Pace and colleagues found higher rates of discontinuation and non-adherence with higher cost-sharing in women initiating generic oral contraceptives [5]. Furthermore a study by Carlin and colleagues found that reduced cost-sharing was associated with increased use of prescription contraceptives, including LARCs [4]. It has been described previously that there had been a background increase in LARC use among women using contraception from 2.4% of users in 2002 to 11.6% for 2011-2013 according to National Survey of Family Growth (NSFG) data [6-8]. Several prior studies have examined the effect of the ACA contraceptive provision on out-of-pocket costs for contraception [2, 3, 9, 10] and all show declining out-of-pocket costs to women after 2012. Two other studies have examined both out-of-pocket costs and types of contraception women use post-aca. Law and colleagues found a steep decline in out-of-pocket costs for LARCs following the ACA contraceptive provision and an increase in IUD claims from 1.2% in 2011, to 1.3% in 2012, to 1.6% in 2013. Pace and colleagues found that the proportion of claims without cost-sharing for IUDs and implants rose over time but found no significant increase in LARC uptake post-aca implementation as of 2013 [11, 12]. Our study reports contraceptive use patterns and out-of-pocket costs between 2005 and 2014 using a large national database of privately insured women. This is the first study, to our knowledge, with post-aca data through 2014. We aim to determine if the ACA has increased the proportion of women using prescription contraceptives, including LARC insertions, and reduced out-of-pocket contraceptive costs. MATERIALS AND METHODS Data Source and Inclusion Criteria Data are from the Truven Health Analytics MarketScan database which consists of reimbursed health care claims for employees, retirees, and their dependents from more

than 250 employers and health plans from all 50 states and the District of Columbia. Individuals included in the database are covered under commercial (private) insurance plans. This large national database includes an annual population of over 50 million people and captures administrative claims with data from inpatient visits, outpatient visits, and pharmacy claims de-identified at the patient level. This study was approved by the Penn State College of Medicine Institutional Review Board. We conducted a retrospective cohort analysis to examine claims and out-of-pocket costs for prescription contraceptive methods used by women before and after implementation of the ACA contraceptive mandate in August 2012. Study cohorts were created for each calendar year between 2005 and 2014 (the most recent year for which data are available) that included women ages 13-45 who had continuous medical and pharmacy coverage during that year. We were unable to identify whether women belonged to employer groups who were exempt from the contraceptive mandate. Measures of Contraceptive Use Contraceptive claims were identified using Healthcare Common Procedure Coding System (HCPCS), International Classification of Diseases (ICD)-9, National Drug Code (NDC) and Current Procedural Terminology (CPT)-4 codes. IUD insertions were identified using ICD-9 codes V25.11 and 69.7, CPT-4 code 58300 or HCPCS codes J7300, J7301, J7302, S4981, and S4989. Implant insertions were identified using ICD-9 code V25.5, CPT-4 code 11981 and HCPCS codes J7306 and J7307. Because the CPT-4 code for implant insertion is not contraceptive specific, the CPT-4 code was combined with the contraceptive specific ICD-9 and HCPCS codes to ensure only the capture of contraceptive implant insertions. The LARC insertion rate was defined as the percent of women in each cohort year that had a LARC insertion claim. The LARC insertion rate does not represent the total proportion of LARC users during that year, as some LARC users will have had their LARC inserted in previous years. For non-larc methods, pharmacy claims were searched for oral contraceptive pills, patches, injection, and the contraceptive ring. Injections were additionally identified using procedure codes. Women with pharmacy claims for more than one type of contraceptive method in a calendar year were coded as using the method that was in use for the longest period of time in that year. Use rates of non-larc methods were defined as the percent of women using each of the contraceptive methods during each cohort year. Non-prescription contraceptive methods could not be accounted for because they do not generate claims. 2

3 Measures of Contraceptive Cost Individual out-of-pocket costs for each type of contraception were estimated by combining copayment, coinsurance and deductible payments for both contraception and insertion fees (in the case of LARCs). Costs for LARCs are reported as out-of-pocket cost for insertion (including device and insertion fees). Oral contraceptives, patches and rings are reported as cost per 28-day supply obtained (e.g. a pack of contraceptive pills). Injection is reported as cost per injection. All costs were adjusted for inflation to 2014 dollars using the Consumer Price Index. Statistical Analysis For each study year, the distribution of the study sample by age, U.S. region, and rural versus urban residence was determined. The IUD and implant insertion rates and percent of women using oral contraceptives, injections, ring, and patch are reported. The mean and median out-of-pocket costs for each method type in each study year are reported. To estimate likelihood of LARC insertion post-aca implementation (compared with pre- ACA implementation), a multivariable logistic regression was performed adjusting for year (to account for secular trends), age group (to account for reproductive life stage, noting that the ACA extends coverage to dependents under 26), geographic region (to account for variation in prescribing patterns), and urban vs. rural residence (to account for patient access to providers). Age groups were defined as 13-17, 18-25, 26-35 and 36-45. Geographic region (northeast, north central, south and west) and an indicator of urban versus rural residence were available from the MarketScan database. Statistical analyses were performed using SAS version 9.4. RESULTS Table 1 describes the characteristics of the study sample by cohort year. For all years, the smallest proportion of women was in the youngest age group and the largest proportion of women in the oldest age group. The smallest proportion of women live in the northeast region, and the largest proportion live in the south region. Most women were living in urban areas. Table 1. Characteristics of reproductive age women by year, 2005-2014. Year N (millions) Age (%) U.S. Region (%) Rural/Urban (%) 13-18- 26-36- Northeast North South West Unknown Rural Urban 17 25 35 45 Central 2005 3.68 16.73 17.37 27.50 38.40 10.17 21.89 42.88 24.24 0.82 15.32 84.68 2006 3.88 16.99 17.73 27.48 37.79 12.19 22.57 47.52 17.03 0.68 16.06 83.94 2007 4.44 16.90 17.99 27.45 37.66 11.15 23.60 47.13 17.63 0.48 17.08 82.92 2008 5.68 16.60 18.20 28.04 37.16 13.79 26.20 43.12 16.36 0.54 15.09 84.91 2009 5.86 16.08 18.55 28.74 36.63 11.45 28.08 43.37 16.86 0.25 14.49 85.51

2010 6.31 16.12 18.76 28.63 36.49 13.53 25.88 40.34 19.89 0.36 14.06 85.94 2011 7.13 15.50 21.80 27.68 35.02 15.63 24.04 39.79 18.56 1.98 15.44 84.56 2012 7.32 15.42 23.02 27.51 34.05 16.14 24.32 37.82 20.41 1.30 14.87 85.13 2013 6.26 15.17 23.96 27.03 33.84 16.75 21.69 36.16 22.48 2.92 14.83 85.17 2014 6.47 14.92 23.98 27.20 33.90 19.24 20.32 40.08 17.77 2.59 14.77 85.23 4 Table 2 shows the mean and median out-of-pocket costs for each prescription contraceptive method in each study year. Pre-ACA (2005-2012), the mean out-of-pocket cost for IUD, implant, ring, and patch was increasing while remaining stable to slightly decreasing for the pill, and stable for the injection. During this same time, the median out-of-pocket cost remained stable for IUD (around $20), was variable for the implant, increasing for the ring, was stable to slightly decreasing for the pill, decreasing for the injection, and stable to slightly increasing for the patch. Post-ACA contraceptive mandate (2013-2014), the mean out-of-pocket cost for all types of contraception decreased. Similarly, the median out-of-pocket cost for all types of prescription contraception, with the exception of the injection, decreased to $0. Cost data for the contraceptive implant is not shown for 2005 as the contraceptive implant did not become available during this study period until 2006. Table 2. Mean and median out-of-pocket costs by contraceptive type, 2005-2014 (dollars). Year IUD Implant* Oral Contraceptive Injection Ring Patch Mean Median Mean Median Mean Median Mean Median Mean Median Mean Median 2005 47.63 18.18 --- --- 114.47 78.79 19.17 11.95 23.67 0 27.45 0 2006 59.45 17.61 38.23 35.34 118.27 82.20 22.29 11.50 49.85 17.73 44.24 13.70 2007 63.06 22.84 85.29 57.10 117.68 79.94 22.01 11.41 52.38 27.86 57.98 22.84 2008 61.24 21.99 71.63 21.99 110.69 71.47 19.19 7.70 57.06 26.39 51.84 22.34 2009 68.32 22.07 77.79 22.07 109.63 69.52 19.22 7.13 65.57 27.59 53.53 24.46 2010 83.12 21.71 86.78 21.71 109.52 65.14 20.22 6.90 87.86 39.74 71.72 28.56 2011 92.37 21.05 93.91 21.05 110.56 63.78 21.84 7.37 82.17 37.89 68.59 28.54 2012 107.18 20.62 131.21 36.09 103.70 61.87 21.81 7.13 105.26 37.12 70.46 28.55 2013 21.03 0 29.67 0 38.45 0 10.00 2.69 58.63 0 53.61 0 2014 16.79 0 23.84 0 24.05 0 7.89 1.51 41.23 0 38.87 0 NOTE: Dollars adjusted for inflation to 2014 dollars using the Consumer Price Index. IUD and implant cost presented as out-of-pocket cost in dollars per insertion. Injection cost presented as cost per injection. Cost for other methods presented as out-of-pocket cost per 28-day supply obtained. *Cost data for the contraceptive implant is not shown for 2005 as the contraceptive implant did not become available during this study period until 2006. Table 3 shows the trend in prescription contraceptive use over each successive cohort year. The IUD insertion rate was 0.42% in 2005 and increased over time until 2014 when it was 1.95%. The contraceptive implant insertion rate was 0.0004% in 2006 and also increased every year to 0.42% in 2014. The greatest proportion of women in each year were oral contraceptive users (26% in 2014), while only 1.9%, 1.7%, and <1% were

5 injection, ring, and patch users, respectively. Insertion rate for the contraceptive implant is not shown for 2005 as the contraceptive implant did not become available during this study period until 2006. Table 3. LARC insertion rates and other contraceptive use by year, 2005-2014 (percent of sample). LARC insertion rates (%) Non-LARC method use (%) Year IUD Implant* Oral contraceptive Injection Ring Patch 2005 0.42 --- 23.66 2.16 0.78 0.004 2006 0.60 0.0004 24.76 2.25 1.32 0.004 2007 0.84 0.003 24.27 2.26 1.51 0.005 2008 1.19 0.06 25.15 2.27 1.67 0.004 2009 1.31 0.09 26.06 1.38 2.34 0.005 2010 1.26 0.11 25.30 1.41 2.36 0.003 2011 1.37 0.16 25.70 2.41 1.82 0.003 2012 1.53 0.20 25.59 2.48 1.77 0.004 2013 1.78 0.32 25.47 1.76 1.71 0.003 2014 1.95 0.42 26.06 1.88 1.68 0.002 *Insertion rate for the contraceptive implant is not shown for 2005 as the contraceptive implant did not become available during this study period until 2006. Table 4 shows the results of a multivariable logistic regression analysis modeling the adjusted odds of LARC insertion for the years before and after implementation of the contraceptive mandate. There was a statistically significant 3.2% reduced odds of LARC insertion after the contraceptive mandate was implemented, after adjusting for covariates. There was a statistically significant 16% increased odds of LARC insertion with each subsequent year. Compared to the oldest age group, girls 13-17 years old were significantly less likely to have a LARC insertion while women aged 18-26 and 27-35 had increased odds of LARC insertions. Women living in the Northeast had decreased odds of LARC insertions, while women in the South and West had increased odds of LARC insertions compared with women living in the North central region. There was no statistically significant difference in LARC insertion for women in rural versus urban areas. Table 4. Adjusted odds of LARC insertion (IUD or implant) pre- and post-aca implementation (n=57,027,745). Adjusted odds ratio (95% confidence interval) Post-ACA vs. Pre-ACA 0.968 (0.961, 0.974) Cohort year (1 year increments, 2005-2014) Age group 13-17 18-25 26-35 36-45 U.S. region 1.161 (1.159, 1.162) 0.365 (0.361, 0.369) 1.620 (1.610, 1.630) 2.250 (2.238, 2.262) reference

6 Northeast North central South West 0.900 (0.893, 0.907) reference 1.048 (1.1042, 1.054) 1.200 (1.192, 1.208) Rural vs. urban 1.006 (1.000, 1.013) NOTE: Hosmer and Lemeshow Goodness-of-Fit Test Chi-Square statistic 4045.5 (p<0.0001). DISCUSSION The absolute rate of claims for new LARC insertions increased over time, including in the two years post- implementation of the ACA contraceptive mandate. However, when we controlled for year to account for secular trends in the multivariable logistic regression analysis, the rate of LARC insertions showed a slight decline post-aca through 2014. The 3.2% reduced odds of LARC insertion after the contraceptive mandate was statistically significant given the very large sample size, but is unlikely to be clinically significant due to the small effect size. Because cost is often a leading indicator, we may see a greater increase in LARC uptake in coming years. Consistent with prior studies, we observed a dramatic reduction in the average out-ofpocket cost for most contraceptive methods, including LARCs, after the ACA contraceptive mandate was implemented in 2012. This study extends this finding through 2014. While most women had no out-of-pocket costs for LARCs after 2012, the mean cost for an IUD was still between $16 and $21. These post-2012 costs may be attributable to grandfathered plans, employers with religious exemptions to the contraceptive mandate, or noncompliance with the ACA contraceptive mandate [13]. While reduction in out-of-pocket contraceptive costs to women is an important finding, the potential to improve public health and decrease rates of unintended pregnancy will not be fully realized unless those reduced costs translate into increased use of more effective contraceptive methods. We aimed to fill a gap in the current literature by examining how rates of use of various forms of contraception have been affected by the ACA contraceptive coverage mandate. Our findings suggest that LARC uptake did not increase following implementation of ACA and that more years of data post-aca are needed to fully understand trends in LARC use. A limitation of this study is that the MarketScan database does not include all private insurers, and it does not include those covered by Medicaid. In addition, because this is a claims database, we cannot account for use of non-prescription contraceptive methods or for prescription methods obtained by the woman for which a claim was not generated (e.g. family planning or school-based clinics). In addition, we could not account for LARC use by women who obtained the method in a year outside of our period of observation. Finally, it was not possible to follow individual women over time and observe changes in contraceptive use. Strengths of this database include its large size and

7 nationally representative sample over many years, including two full years post-aca mandate implantation. While it is clear that the ACA contraceptive mandate has dramatically reduced the out-ofpocket cost for prescription contraceptives including LARCs, our findings suggest that LARC uptake did not increase post-aca. The full impact of the ACA contraceptive mandate on contraceptive use patterns may not be clear until more years of data are available. Additional studies will be needed to help inform future policy.

8 REFERENCES [1] Chuang CH, Mitchell JL, Velott DL, Legro RS, Lehman EB, Confer L, et al. Women's Awareness of Their Contraceptive Benefits Under the Patient Protection and Affordable Care Act. American Journal of Public Health. 2015;105 Supplement 5:713-5. [2] Becker NV, Polsky D. Women Saw Large Decrease In Out-Of-Pocket Spending For Contraceptives After ACA Mandate Removed Cost Sharing. Health Affairs. 2015;34:1204-11. [3] Bearak JM, Finer LB, Jerman J, Kavanaugh ML. Changes in out-of-pocket costs for hormonal IUDs after implementation of the Affordable Care Act: an analysis of insurance benefit inquiries. Contraception. 2016;93:139-44. [4] Carlin CS, Fertig AR, Dowd BE. Affordable Care Act's Mandate Eliminating Contraceptive Cost Sharing Influenced Choices Of Women With Employer Coverage. Health Affairs. 2016;35:1608-15. [5] Pace LE, Dusetzina SB, Keating NL. Early Impact Of The Affordable Care Act On Oral Contraceptive Cost Sharing, Discontinuation, And Nonadherence. Health Affairs. 2016;35:1616-24. [6] Xu X, Macaluso M, Ouyang L, Kulczycki A, Grosse SD. Revival of the intrauterine device: increased insertions among US women with employer-sponsored insurance, 2002-2008. Contraception. 2012;85:155-9. [7] Guttmacher Institute. Use of Highly Effective Contraceptives in the U.S. Continues to Rise, with Likely Implications for Declines in Unintended Pregnancy and Abortion, https://www.guttmacher.org/article/2014/12/use-highly-effective-contraceptives-uscontinues-rise-likely-implications-declines; 2014 [accessed 4/7/16]. [8] Daniels K, Daugherty J, Jones J, Mosher W. Current Contraceptive Use and Variation by Selected Characteristics Among Women Aged 15-44: United States, 2011-2013. National Health Statistics Reports. Hyattsville, MD. National Center for Health Statistics; 2015. [9] Finer LB, Sonfield A, Jones RK. Changes in out-of-pocket payments for contraception by privately insured women during implementation of the federal contraceptive coverage requirement. Contraception. 2014;89:97-102.

9 [10] Sonfield A, Tapales A, Jones RK, Finer LB. Impact of the federal contraceptive coverage guarantee on out-of-pocket payments for contraceptives: 2014 update. Contraception. 2015;91:44-8. [11] Law A, Wen L, Lin J, Tangirala M, Schwartz JS, Zampaglione E. Are women benefiting from the Affordable Care Act? A real-world evaluation of the impact of the Affordable Care Act on out-of-pocket costs for contraceptives. Contraception. 2016;93:392-7. [12] Pace LE, Dusetzina SB, Keating NL. Early Impact of the Affordable Care Act on Uptake of Long-acting Reversible Contraceptive Methods. Medical Care. 2016;54:811-7. [13] Tschann M, Soon R. Contraceptive Coverage and the Affordable Care Act. Obstetrics and Gynecology Clinics of North America. 2015;42:605-17.