Changes in Premiums of Cancelled Nongroup Plans Under the Affordable Care Act

July 12, 2016
Jared Lane K. Maeda, PhD, MPH

,
Jersey Chen, MD, MPH

,
Brent R. Plemons, BS

Volume 22, Issue 7

Subscribers migrated to Affordable Care Act—compliant plans with modestly higher costs, but had higher levels of insurance coverage and stronger consumer protections.

ABSTRACT

Objectives: To examine the effect of the Affordable Care Act (ACA) on changes in premiums for subscribers of nongrandfathered, nongroup insurance plans that were “cancelled.”

Study Design: Retrospective multivariate analyses.

Methods: Changes in annual premiums post ACA were evaluated across subgroups of subscriber and health plan characteristics. Data was derived from databases containing information on premiums, plan benefit, and demographics for subscribers aged 18 to 64 years within Kaiser Permanente of the Mid-Atlantic States. A linear regression model was used to examine the independent association between subscriber and health plan characteristics on the relative change in premiums.

Results: In 2013, 4169 nongroup subscribers were enrolled in plans that were cancelled as a result of the ACA. The median pre-ACA premium was $3240 (range = $780-$39,492), which increased by a median of 21.3% (range = —77.4% to 193.6%), or $685 (range = –$27,464 to $8676), post ACA in 2014. Premiums increased more for high-deductible plans (median = 63.7%) than standard-deductible plans (median = 8.4%). Due to shifts in the age curve, premiums decreased for more than half of women aged 18 to 44 years, but increased by 35.2% for women aged 55 to 64 years. Premiums fell by 15.5% for subscribers who did not pass standard medical underwriting due to preexisting conditions.

Conclusions: Changes in premiums in the nongroup market post ACA, varied substantially across subgroups, primarily due to differences in the amount of coverage, changes in rating criteria, shifts in the age curve, and anticipated differences in risk selection and composition of the risk pool. Given the extent of this variation, it would be incorrect to conclude the ACA as being uniformly beneficial or detrimental to subscribers of these cancelled plans.

Am J Manag Care. 2016;22(7):e249-e257

Take-Away Points

This study examined how annual premiums for nongrandfathered, nongroup health plans in a large integrated delivery system changed after Affordable Care Act (ACA) implementation, by subscriber and health plan characteristics.

  • The changes in premiums varied substantially across subgroups, primarily due to differences in the amount of coverage, changes in rating criteria, shifts in the age curve, and anticipated differences in risk selection and composition of the risk pool.
  • On average, subscribers migrated to ACA-compliant plans with modestly higher costs, but had higher levels of insurance coverage and stronger consumer protections.
  • This study provides insight into how nongroup market reforms affect premiums.

The “cancellation” of health insurance plans because of noncompliance with the Affordable Care Act (ACA) is a contentious issue. Although the majority of Americans with employer-sponsored insurance were unaffected by this provision, nearly 18.6% of the 14 million individuals in the nongroup market received cancellation notices in 2013 because their plans were noncompliant.1,2 The ACA included new rules that reformed the nongroup market by prohibiting exclusions on preexisting conditions, providing guaranteed access to coverage, placing limits on insurers’ ability to price discriminate based on health status, and setting minimum standards on essential health benefits and actuarial value.3 Insurers had the option of modifying their existing policies to make them ACA-compliant or offering new policies.

At the center of this policy issue is the “grandfather” clause, where individuals had the option of keeping their plans that were in effect at the time the ACA was signed into law (March 23, 2010).4 Nongrandfathered plans offered or modified after that date must be made ACA-compliant, otherwise enrollees would need to purchase ACA-compliant plans either directly from insurers or from state health insurance exchanges.5 Although the Obama administration issued a transition policy that extended noncompliant plans until 2016,6 these plans will eventually be replaced. An estimated 2 million individuals purchased nongrandfathered policies in 2014, and the number of those who will continue to enroll in these plans is expected to be negligible by the end of 2016.7

There is little empirical data examining how premiums changed for subscribers of “cancelled plans” (ie, nongrandfathered, nongroup plans).8 Although much attention has focused on the aggregate number of cancelled policies, there has been less discussion about the extent to which certain subgroups might benefit from the ACA through lower premiums. Accordingly, we analyzed how premiums changed for subscribers of nongroup plans that were cancelled because of modification by the ACA within Kaiser Permanente of the Mid-Atlantic States (KPMAS), a large integrated delivery system with its own insurance plan and multi-specialty group practice that covers over 500,000 lives in Washington, DC (DC); Maryland; and Virginia.9

METHODS

Study Sample

In 2013, noncompliant KPMAS plans were cancelled and ACA-compliant plans were offered in their place for 2014. Subscribers of these cancelled (ie, nongrandfathered, nongroup) plans in 2013 were identified and linked to databases containing demographic and health plan benefit information. Demographic data included age, gender, race, family size, and insurance jurisdiction. Health plan data included premiums, deductibles, benefit offerings, and plan type. KPMAS offered 3 types of plans in the nongroup market: 1) health maintenance organization (HMO) plans with zero deductibles; 2) deductible-HMO (D-HMO) plans with deductibles for hospitalizations, outpatient surgery, and skilled nursing facilities; and 3) high-deductible HMO plans (HD-HMO) with deductibles for most services except preventative care.

In general, HMO, D-HMO, and HD-HMO plans had comparable sets of covered medical services (ie, hospital care, medications, and laboratory) but at different levels of cost sharing (ie, deductibles, co-payments, and coinsurance). A small number of nongroup subscribers (3.4% of the sample) previously enrolled in group plans that closed or dissolved, exercised an option to continue coverage; these “guaranteed-issue” subscribers did not undergo medical underwriting, but instead were subject to community-rated premiums typically higher than standard policies. A small number of “rate-up” policies (3.6% of the sample) were also offered where subscribers who fell outside standard underwriting risk by a certain margin could purchase insurance at rates 35% to 50% higher than standard policies.

For each subscriber in 2013 with a cancelled plan, we identified the most similar ACA-compliant plan in 2014 (ie, bronze, silver, and gold) based on the closest level of cost sharing and benefit offerings. We then calculated the difference in premiums, prior to any federal subsidies, based on premium data provided by KPMAS. Analyses were conducted at the subscriber level because prior to the ACA, each family policy was assigned a single premium inclusive of all members; less than 15% of subscribers in the sample were family policies. Due to limited generalizability, we excluded 6 subscribers with platinum-equivalent plans and 52 subscribers who were 65 years or older.

Analyses

χ

Changes in annual premiums were evaluated across subgroups of subscriber demographic and health plan characteristics. The Kruskal-Wallis test was used to assess for statistically significant differences in median premiums. We calculated the expected relative change (ie, percentage change) in premiums and examined how it varied across categories of subscriber and plan characteristics using the 2 test. We defined a threshold for “zero” premium change that accounted for increasing age (ie, all subscribers would be 1 year older) by estimating a log-linear regression model using 2013 data. To account for inflation, we used the annualized change in the 2013 Consumer Price Index (CPI) for medical care services.10 We used a multivariate linear regression model to examine the independent associations between subscriber and plan characteristics on relative change in premiums, using the ratio of 2014 to 2013 premiums as the dependent variable; heteroscedasticity-consistent standard errors were employed. Because of collinearity between the cost-sharing variables, multivariate models included only the plan type.

To understand the factors underlying the relationship between age and premiums before and after the ACA, we graphed the premium age, rating factors for the most common risk pool in the KPMAS nongroup market. The age curve represents the change in premiums due to age alone and was subject to approval by state insurance commissioners. For Maryland and Virginia, gender-specific age rating factors were used prior to the ACA, and the federal default age curve11 was used after the ACA. For DC, pre-ACA age rating factors were not gender-specific, and a DC-specific age curve was used post ACA.

Statistical analyses were performed using Stata version 13 (StataCorp LP, College Station, Texas). Significance testing was assessed at the P = .05 level. Approval for this study was obtained from the KPMAS Institutional Review Board.

RESULTS

Description of Study Cohort

Table 1

eAppendices

A total of 7842 subscribers were enrolled in nongroup plans in 2013; 4169 of these subscribers were enrolled in cancelled plans that required modification by the ACA, representing 1.8% of all KPMAS subscribers aged 18 to 64 years. Subscribers of cancelled plans had a median age of 39.6 years; 41.3% of subscribers were female, 56.5% were white, and 20% were black (). Most subscribers resided in Maryland (43.2%) or Virginia (40.2%) and were either enrolled in D-HMO plans (44.6%) or HD-HMOs (38.8%). Bronze-equivalent plans consisted entirely of HD-HMOs with annual deductibles of $4500 or $8000. Gold-equivalent plans consisted of HMO plans or D-HMO plans with deductibles of $2000 or less (eAppendix Table 1 [ available at www.ajmc.com]). Only 3.4% and 3.6% of subscribers, respectively, were enrolled in guaranteed-issue or rate-up policies.

Changes in Plan Benefits

Table 2

In 2014, all cancelled nongroup plans at KPMAS incorporated the following general provisions of the ACA: 1) annual limits on out-of-pocket expenses, 2) elimination of lifetime limits on expenses for essential medical benefits, 3) guaranteed issue and renewal, and 4) guaranteed ability for children to be covered under a parent’s health plan until age 26. KPMAS plans already offered comprehensive “essential health benefits”; there were only minor changes required after the ACA, consisting primarily of adding pediatric dental coverage, mental health parity with medical services, additional coverage for state-specific selected habilitative services, and elimination of cost sharing for some maternity/newborn services ().

Changes in Premiums After ACA

Table 3

Figure 1

Figure 2

The median annual premium for cancelled nongroup plans was $3240 (range = $780-$39,492) in 2013 and $3803 (range = $1355-$26,513) in 2014. Premiums increased with older subscriber age and larger family size (). Overall variation in premiums decreased in the post-ACA period when examining subscribers with no spousal or dependent coverage (). The median change in premiums after the ACA was $685 (range = —$27,464 to $8676). In relative terms, premiums increased by a median of 21.3% and a mean of 31.5% (SD = 50.4%; range = –77.4% to 193.6%). The distribution of percentage premium change was bimodal, with peaks at approximately 15% and 115% (). The peak with the higher percent increase in premiums consisted almost entirely of subscribers who were either male or aged 55 to 64 years and enrolled in HD-HMOs. Of the 560 subscribers whose premiums at least doubled after the ACA, 97% who were either male or aged 55 to 64 years were enrolled in HD-HMOs.

In the overall cohort, subscribers aged 55 to 64 years had the highest unadjusted relative increase in premiums (median = 34.7%) compared with other age groups (P <.001) (Table 3). Overall, men had higher relative increases in premiums compared with women (median = 29.1% vs 9.9%; P <.001). However, when stratified by both age and gender, women aged 18 to 25 years, 26 to 34 years, and 35 to 44 years had median decreases in premiums of 15.2%, 3.0%, and 3.3%, respectively; in contrast, women aged 55 to 64 years had a median increase in premiums of 35.2%. Families with 2, 3, or 4 members had smaller median increases in premiums compared with individual subscribers (3.5% vs 0.6% vs 12.1%, vs 23.4%, respectively; P <.001). Maryland residents had higher premium increases compared with those in Virginia or DC (31.8% vs 14.1% vs 12.6%, respectively; P <.001). HD-HMO enrollees had a median increase in premiums of 62.7%, while premiums for the non—high-deductible plans rose by 8.4% overall; specifically, D-HMO and zero-deductible HMO premiums changed by 13.9% and –5.5%, respectively. Subscribers of guaranteed-issue and rate-up policies had a median reduction in premiums of 15.6% and 15.5%, respectively, while premiums for standard medically underwritten policies increased by a median of 23.4%.

We used a threshold of a 5% premium increase to account for incremental subscriber age and inflation, based on a mean 2.5% premium increase in premiums for each 1-year increment in age and a 2.5% change in the CPI for medical services. Approximately one-third of subscribers (32.2%) would have paid less for nongroup coverage in 2014 compared with 2013, after adjusting for age and inflation (ie, a difference of 5% or less), while 41.6% would expect premium increases of 5% to 50%, 12.7% would have increases of 51% to 100%, and the remaining 13.4% would have increases greater than 100% (eAppendix Table 2). One-fourth of subscribers aged 55 to 64 years—and one-fourth of women aged 55 to 64 years—could expect premium increases of more than 100% compared with 7.1% to 16.9% of subscribers in other age categories. In general, 54.9% of women would have higher premiums after the ACA, whereas 76.7% of men would have premium increases. Additionally, one-third of HD-HMO subscribers had a more than 100% increase in premiums after the ACA, while two-thirds of subscribers in zero-deductible HMO plans would have lower premiums. Of guaranteed-issue and rate-up subscribers, 62% and 75%, respectively, would have lower premiums after the ACA.

The multivariate regression model explained a substantial proportion of the relative change in premiums (R2 = 0.64), with findings consistent with the unadjusted comparisons (eAppendix Table 3). Compared with men aged 18 to 25 years, on average, women aged 18 to 25, 26 to 34, and 35 to 44 years had significantly lower premiums post ACA (—60.7%, –63.0%, and –50.2%, respectively; P <.001), but not women aged 55 to 64 years (P =.80). No statistically significant differences in premiums were observed across subscriber race. Virginia and DC residents had premium decreases (—15.7% and –6.8%, respectively; P <.001) compared with Maryland residents. Families with 2 to 4 members had 23.2% to 24.9% (P <.001) relative decreases in premiums compared with individual subscribers—this difference was smaller for families with 5 or more members (—10.7%; P = .02). Guaranteed-issue plans were associated with a 19.3% decrease in premiums, while rate-up policies were associated with a 45.5% decrease in premiums (P <.001). Compared with zero-deductible HMO plans, HD-HMOs and D-HMO plans were associated with a 78.2% and 18.5% increase in premiums, respectively (P <.001).

To understand the relationship between age and gender on premiums, we examined changes in the age curve before and after the ACA. In Maryland and Virginia, where gender rating was permitted prior to the ACA, the age curve shifted upwards substantially for women aged 55 to 64 years after the ACA (eAppendix Figure 1). In DC, where gender rating had already been eliminated in 2011, the age curve shifted upwards for men and women 40 years or older (eAppendix Figure 2).

DISCUSSION

Our study provides insight into how reforms in the nongroup market under the ACA affected premiums prior to federal subsidies within a large integrated healthcare delivery system. Fewer than half of nongroup subscribers were enrolled in nongrandfathered plans subject to change under the ACA; affected subscribers with cancelled plans made up only a small portion (1.8%) of the overall KPMAS population. Subscribers of cancelled plans had an overall median increase of 21.3% in premiums before subsidies. On average, enrollees of HD-HMO plans face substantially higher premiums after the ACA, whereas women younger than 45 years and enrollees of guaranteed-issue and rate-up policies can expect premium decreases.

Understanding how the insurance rating system, benefit generosity, risk selection, and composition of the risk pool changed after the ACA clarifies why particular subgroups would experience higher or lower premiums. Prior to ACA, premiums in the KPMAS nongroup market were determined according to the age of the oldest enrollee (either subscriber or spouse), gender, number of dependents, and a factor representing level of cost sharing (ie, plan type and other deductibles). All enrollees had to undergo medical underwriting, except for those with guaranteed-issue policies. Post ACA, gender was not permitted as a rating factor; each family member was assigned a premium independently and medical underwriting also ceased. Health status was also no longer allowed as an explicit rating factor and was accounted for in the overall post-ACA base rate. KPMAS also considered the scenario where the composition of the risk pool post ACA would be of worse health status. Because of the ACA premium stabilization program, however, KPMAS anticipated that it would receive payments, and adjusted their rates to account for this. In addition, KPMAS assumed that members’ utilization would be sensitive to the level of cost sharing. Insurance rates for each age level were assigned according to the new federal default age curve11 in Maryland and Virginia; the DC insurance commission implemented its own age curve.

Enrollees of HD-HMO plans faced substantially higher premiums (median = 62.7%) because the post-ACA bronze plans were of higher actuarial value. For example, HD-HMO plans with $8000 deductibles were no longer permitted, and were replaced by higher actuarial value bronze plans with $5000 deductibles. HD-HMO enrollees would pay more in premiums, but would benefit from higher levels of coverage. Our findings are consistent with an analysis by the Congressional Budget Office that projected additional insurance coverage would explain the majority of premium increases in the nongroup market.12

Changes in the age curve explained why older women were less likely to have lower premiums. In general, women paid more for health insurance prior to ACA due to higher overall healthcare expenditures compared with men.13,14 As such, elimination of gender as a rating factor in Maryland and Virginia would lead to a 1-time decrease in premiums for women residing in these states; DC had already ceased gender rating in 2011. However, when stratified by age, we found that these savings largely accrued to women aged under 45 years. Changes in the shape of the age curve explain this phenomenon. In Maryland and Virginia, the age curve shifted downwards after ACA for women aged 21 to 40 years, but shifted considerably upwards for women 55 years or older; for subscribers in DC, the age curve shifted upwards for both women and men 40 years or older. As a result, the changes in premiums post ACA for women aged 55 to 64 years were not significantly different compared with similarly aged men.

Prior to the ACA, premiums for families with children were based on the rate of the oldest parent multiplied by a fixed family factor, regardless of the number of children aged under 26 years, yielding a relative discount for large families. After the ACA, each adult family member and up to 3 children aged under 21 years were assigned separate premiums based on their individual ages; children aged 21 to 25 years were assigned adult premiums. For many families with 2 to 4 members, these changes resulted in lower aggregate premiums since the effect of the age of the oldest adult became diminished. For families with 5 or more members, the savings were less pronounced as the relative discount prior to ACA disappeared with individual premium assignment; furthermore, larger families were more likely to have children aged 21 to 25 years who were newly assigned adult premiums. Because we were only able to examine premiums at the subscriber level, this might have increased the absolute difference in the average premiums rather than if we were able to determine premiums for each individual member covered by the nongroup policy in both time periods. However, we believe this impact to be small as less than 15% of subscribers had family policies. Although our results are more representative of individual subscribers, this finding reflects the actual impact on this portion of KPMAS’ market segment.

A number of factors likely explain why premiums changed differently across jurisdictions: 1) definitions of “essential health benefits” depended on benchmark plans in each jurisdiction, 2) different levels of competition in the health insurance market,15 or 3) differences in the use of ACA rating criteria (eg, DC did not permit tobacco use as a rating factor post ACA).16

Patients with preexisting conditions had lower premiums as a result of the ACA, as seen in the savings observed for subscribers of guaranteed-issue and rate-up policies. Rate-up subscribers were considered too high-risk to pass standard medical underwriting, while guaranteed-issue subscribers typically had preexisting conditions since they otherwise would have chosen to undergo medical underwriting to purchase insurance at lower standard rates. After the ACA, these higher-risk subscribers were able to obtain less costly plans at rates comparable to healthy individuals, thereby shrinking the variation in premiums. The characteristics of subscribers in guaranteed-issue policies, however, differed significantly from standard policy subscribers and are not representative of the overall nongroup market segment with cancelled policies.

Because KPMAS is an integrated delivery system providing a full range of health services, only minor changes were required to comply with the ACA rules for essential health benefits; the addition of pediatric dental benefits was an exception, but affected a small number of dependents. However, all cancelled plans were modified with respect to guaranteed issue and renewal and elimination of lifetime limits on essential health expenses. Quantifying the dollar amount of these benefits is complex, but presumably would be more valuable for higher-risk individuals. Additional economic models are required to quantify the value of these consumer protections and how it translates into premium increases observed for particular subgroups in the nongroup market.

Overall, subscribers of cancelled plans had considerable variation in premium changes, ranging from a 77.4% decrease to a 193.6% increase. Nearly one-third of subscribers would pay less in the subsequent year—predominately individuals of presumably worse health status (ie, guaranteed issue and rate-up policy holders) and women aged under 45 years, as medical underwriting and gender were no longer allowed as rating factors. Subscribers of high-deductible plans would have the greatest increase in premiums, but would be required to purchase higher levels of coverage. Our findings demonstrate that the ACA had a substantially varied effect on holders of cancelled policies and that it would be incorrect to generalize the ACA as being uniformly beneficial or detrimental to these subscribers.

Limitations

Our study is subject to a few limitations. We were unable to assess the effect of federal subsidies in our analysis.17 More than half of enrollees in the nongroup market are expected to be eligible for subsidies, which would reduce premium expenditures by 56% to 59%.12 Additionally, our study only evaluated premiums and not total out-of-pocket costs, and we only examined subscribers already enrolled in full-service insurance; individuals with insurance that lacked major essential health benefits will likely have higher premiums post ACA before subsides.18,19 Subscribers could have chosen to enroll in ACA-compliant plans that were more or less generous than what was offered at the time of renewal. Participation in the nongroup insurance market is often transient,20 and subscribers may not have remained insured for the entire year as we assumed. We acknowledge that health status and individual underwriting factors were not considered, which limits interpretation of the analysis. Our study also evaluated the impact of the ACA for subscribers residing in 3 jurisdictions at a single large integrated delivery system, which may not be applicable to other regions.15,17 Previous studies, however, have found KP membership to be generally representative of the broader community.21,22

CONCLUSIONS

Within a large integrated delivery system, the changes in premiums for cancelled plans in the nongroup market varied substantially across subgroups, primarily due to differences in the amount of coverage, changes in rating criteria, shifts in the age curve, and anticipated differences in risk selection and composition of the risk pool. On average, subscribers migrated to ACA-compliant plans with modestly higher costs, but had higher levels of coverage and stronger consumer protections.

Author Affiliations: Mid-Atlantic Permanente Research Institute, Kaiser Permanente Mid-Atlantic States (JLKM, JC), Rockville MD; Kaiser Foundation Health Plan of the Mid-Atlantic States, Inc (BRP), Rockville, MD.

Source of Funding: None.

Author Disclosures: Drs Maeda and Chen and Mr Plemons are former employees of Kaiser Permanente, a health insurer.

Authorship Information: Concept and design (JLKM, JC); acquisition of data (JLKM, JC, BRP); analysis and interpretation of data (JLKM, JC, BRP); drafting of the manuscript (JLKM); critical revision of the manuscript for important intellectual content (JLKM, JC, BRP); statistical analysis (JC); administrative, technical, or logistic support (JLKM, BRP); and supervision (JLKM).

Address correspondence to: Jared Lane K. Maeda, PhD, MPH, formerly of Mid-Atlantic Permanente Research Institute, Kaiser Permanente Mid-Atlantic States, Rockville, MD. E-mail: jared.maeda@gmail.com.

REFERENCES

1. Clemans-Cope L, Anderson N. How many nongroup policies were canceled? estimates from December 2013. Health Affairs blog website. http://healthaffairs.org/blog/2014/03/03/how-many-nongroup-policies-were-canceled-estimates-from-december-2013/. Published March 3, 2014. Accessed October 18, 2015.

2. Appleby J, Gorman A. Thousands of consumers get insurance cancellation notices due to health law changes. Kaiser Health News website. http://www.kaiserhealthnews.org/stories/2013/october/21/cancellation-notices-health-insurance.aspx. Published October 21, 2013. Accessed January 16, 2014.

3. Gruber J. The impacts of the Affordable Care Act: how reasonable are the projections? National Tax Journal. 2011;64(3):893-908. doi: 10.3386/w17168.

4. Blumenthal D, Collins SR. Health care coverage under the Affordable Care Act—a progress report. New Engl J Med. 2014;371(3):275-281. doi: 10.1056/NEJMhpr1405667.

5. Public Law 111-148. HR 3590. Patient Protection and Affordable Care Act [111th US Congress]. US Government Printing Office website. https://www.gpo.gov/fdsys/pkg/PLAW-111publ148/pdf/PLAW-111publ148.pdf. Published March 23, 2010. Accessed June 21, 2016.

6. Cohen G. Insurance Standards Bulletin Series—extension of transitional policies through October 1, 2016. CMS website. https://www.cms.gov/cciio/resources/regulations-and-guidance/downloads/transition-to-compliant-policies-03-06-2015.pdf. Published March 5, 2014. Accessed October 18, 2015.

7. Updated estimates of the effects of the insurance coverage provisions of the Affordable Care Act, April 2014. Congressional Budget Office website. http://www.cbo.gov/sites/default/files/cbofiles/attachments/45231-ACA_Estimates.pdf. Published April 2014. Accessed October 18, 2015.

8. Gruber J. Growth and variability in health plan premiums in the individual insurance market before the Affordable Care Act. The Commonwealth Fund website. http://www.commonwealthfund.org/publications/issue-briefs/2014/jun/health-insurance-premiums. Published June 5, 2014. Accessed June 12, 2014.

9. Cutting CC, Collen MF. A historical review of the Kaiser Permanente medical care program. J Soc Health Syst. 1992;3(4):25-30.

10. Table 1. Consumer price index for all urban consumers (CPI-U): U.S. city average, by expenditure category and commodity and service group. Bureau of Labor Statistics website. http://www.bls.gov/news.release/cpi.t01.htm. Accessed April 2, 2014.

11. Cohen G. Sub-regulatory guidance regarding age curves, geographical rating areas and state reporting. CMS website. https://www.cms.gov/CCIIO/Resources/Files/Downloads/market-reforms-guidance-2-25-2013.pdf. Published February 25, 2013. Accessed June 12, 2014.

12. An analysis of health insurance premiums under the Patient Protection and Affordable Care Act. Congressional Budget Office website. https://www.cbo.gov/sites/default/files/cbofiles/ftpdocs/107xx/doc10781/11-30-premiums.pdf. Published November 30, 2009. Accessed April 18, 2014.

13. Pear R. Gender gap persists in cost of health insurance. The New York Times website. http://www.nytimes.com/2012/03/19/health/policy/women-still-pay-more-for-health-insurance-data-shows.html?_r=0. Published March 19, 2012. Accessed April 28, 2014.

14. Women’s health USA 2011. Health Resources and Services Administration website. http://www.mchb.hrsa.gov/whusa11/more/downloads/pdf/w11.pdf. Published October 2011. Accessed June 21, 2016.

15. Health insurance marketplace premiums for 2014. Office of the Assistant Secretary for Planning and Evaluation website. http://aspe.hhs.gov/health/reports/2013/MarketplacePremiums/ib_marketplace_premiums.cfm. Published September 2013. Accessed April 28, 2014.

16. The Center for Consumer Information and Insurance Oversight—market rating reforms: state specific rating variations. CMS website. http://www.cms.gov/CCIIO/Programs-and-Initiatives/Health-Insurance-Market-Reforms/state-rating.html. Updated February 2016. Accessed June 21, 2016.

17. Kaiser Family Foundation. What Americans pay for health insurance under the ACA. JAMA. 2014;311(11):1100. doi:10.1001/jama.2014.1211.

18. Gabel JR, Lore R, McDevitt RD, et al. More than half of individual health plans offer coverage that falls short of what can be sold through exchanges as of 2014. Health Aff (Millwood). 2012;31(6):1339-1348. doi: 10.1377/hlthaff.2011.1082.

19. Whitmore H, Gabel JR, Pickreign J, McDevitt R. The individual insurance market before reform: low premiums and low benefits. Med Care Res Rev. 2011;68(5):594-606. doi: 10.1177/1077558711399767.

20. Sommers BD. Insurance cancellations in context: stability of coverage in the nongroup market prior to health reform. Health Aff (Millwood). 2014;33(5):887-894. doi: 10.1377/hlthaff.2014.0005.

21. Gordon NP. A comparison of sociodemographic and health characteristics of the Kaiser Permanente Northern California membership derived from two data sources: the 2008 Member Health Survey and the 2007 California Health Interview Survey. Kaiser Permanente Division of Research website. https://www.dor.kaiser.org/external/chis_mhs_comparison_2008/. Published January 4, 2012. Accessed June 21, 2016.

22. Koebnick C, Langer-Gould AM, Gould MK, et al. Sociodemographic characteristics of members of a large, integrated health care system: comparison with US Census Bureau data. Perm J. 2012;16(3):37-41.