PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Matern Child Health J. Author manuscript; available in PMC 2017 May 1.
Published in final edited form as:
PMCID: PMC4826791
NIHMSID: NIHMS770544

Healthcare Utilization after a Children’s Health Insurance Program Expansion in Oregon

Steffani R Bailey, PhD,a Miguel Marino, PhD,a,b Megan Hoopes, MPH,c John Heintzman, MD MPH,a Rachel Gold, PhD MPH,c,d Heather Angier, MPH,a Jean P. O’Malley, MPH,b and Jennifer E DeVoe, MD DPhila,c

Abstract

Objective

The future of the Children’s Health Insurance Program (CHIP) is uncertain after 2017. Survey-based research shows positive associations between CHIP expansions and children’s healthcare utilization. To build on this prior work, we used electronic health record (EHR) data to assess temporal patterns of healthcare utilization after Oregon’s 2009–2010 CHIP expansion. We hypothesized increased post-expansion utilization among children who gained public insurance.

Methods

Using EHR data from 154 Oregon community health centers, we conducted a retrospective cohort study of pediatric patients (2–18 years old) who gained public insurance coverage during the Oregon expansion (n=3,054), compared to those who were continuously publicly insured (n=10,946) or continuously uninsured (n=10,307) during the 2-year study period. We compared pre-post rates of primary care visits, well-child visits, and dental visits within- and between-groups. We also conducted longitudinal analysis of monthly visit rates, comparing the three insurance groups.

Results

After Oregon’s 2009–2010 CHIP expansions, newly insured patients’ utilization rates were more than double their pre-expansion rates [adjusted rate ratios (95% confidence intervals); increases ranged from 2.10 (1.94–2.26) for primary care visits to 2.77 (2.56–2.99) for dental visits]. Utilization among the newly insured spiked shortly after coverage began, then leveled off, but remained higher than the uninsured group.

Conclusions

This study used EHR data to confirm that CHIP expansions are associated with increased utilization of essential pediatric primary and preventive care. These findings are timely to pending policy decisions that could impact children’s access to public health insurance in the United States.

Keywords: health insurance, children, utilization, electronic health records, Medicaid, CHIP

Introduction

The Children’s Health Insurance Program (CHIP) was informed by evidence that uninsured children have significant unmet healthcare needs.(13) CHIP was partially responsible for the uninsured rate among children in the United States (US) dropping from 14% in 1997, when CHIP began, to 7% in 2012.(4) CHIP provides public insurance for children otherwise not eligible for Medicaid; CHIP expansions have also been associated with increased enrollment in Medicaid via awareness of public coverage options.(5) Survey data suggest that newly insured children experience improved access to pediatric primary and dental care services,(614) but this self-reported data can be subject to certain biases (e.g., recall error, response fatigue).(4, 15, 16) Electronic health record (EHR) databases have the potential to overcome such biases. We know of just one prior study using electronic data to examine the association between CHIP expansions and care utilization.(17) That study assessed pre- and post-coverage utilization among children gaining CHIP coverage, and found higher rates of pediatric care visits in the post-period; however, this study did not include continuously uninsured or those already insured as comparators.

Funding for CHIP has been extended through 2017.(18) If CHIP is not reauthorized or similar funding is not available after that date, some children could face barriers to coverage.(4, 19) Given the uncertainty of CHIP’s future, the unknown and indirect effect of the Affordable Care Act on children’s insurance coverage, and the need for more evidence on potential impacts of these public insurance programs, we used EHR data to confirm and quantify the relationship between gaining public insurance coverage and children’s healthcare utilization. Using a retrospective cohort design, we identified 3 groups of established community health center (CHC) pediatric study participants: 1) patients who gained public insurance coverage during Oregon’s 2009–2010 CHIP expansion, 2) patients who were either continuously insured, or 3) continuously uninsured throughout the study period. Oregon expanded its CHIP program in 2009–2010 to offer coverage to children in families with incomes ≤200% of the federal poverty level (FPL); the eligibility limit in the pre-expansion period was ≤185% FPL. This expansion, referred to as “Healthy Kids,” was coupled with a large media campaign across Oregon and resulted in >100,000 newly insured children via Medicaid or CHIP,(20) and ultimately reduced the child uninsurance rate in the state by 50%.(21)

Our study aim was to utilize EHR data to compare pre- and post-coverage rates of primary care and dental services utilization within and between these groups. We also longitudinally assessed the timing of post-coverage utilization at monthly intervals to identify temporal patterns. This study is unique in that it compares changes in healthcare utilization rates of newly insured patients to those of both continuously insured and continuously uninsured patients; it is the first to use EHR data to compare children’s temporal patterns with insurance group as the independent variable.

Methods

Data sources

We used EHR data from Oregon CHCs in the OCHIN community health information network with an EHR in place for practice management by 2008 (n=154 CHCs) (OCHIN was originally the Oregon Community Health Information Network but shortened to “OCHIN, Inc” when membership expanded beyond Oregon.) We used Medicaid and CHIP identification numbers to link individual-level EHR data with state administrative data to confirm public insurance coverage periods. All other patient-level data was extracted from the EHR.

Study setting and population

We identified 31,796 OCHIN patients who met the following criteria throughout the study period: aged 2–18; no record of insurance other than Medicaid/CHIP; and not pregnant, which independently impacts coverage and utilization. Medicaid/CHIP enrollment was restricted to eligibility based on poverty status (i.e., ‘categorically eligible’ children were excluded). Analyses were limited to Oregon patients with ≥1 billed encounter at an OCHIN CHC at any point before their ‘start date’ (defined below), plus ≥1 visit after their ‘start date,’ to establish a minimum of care continuity at OCHIN clinics. These encounters did not need to occur within the study period and encounters outside of the study period were not included in analyses.

Insurance coverage groups

The newly insured group included children continuously uninsured for ≥1 year prior to gaining public coverage during the Healthy Kids expansion period (09/01/09–12/31/10) (pre-period), then continuously insured by Medicaid or CHIP for ≥1 year after the coverage start date (post-period). Comparison groups were children who had continuous public coverage, or who were continuously uninsured, in an analogous period. To estimate these analogous periods, each comparison group child was randomly assigned a ‘start date’ based on the distribution of Medicaid/CHIP coverage start dates among newly insured children (e.g., if 2% of children in the group who gained insurance had an actual coverage start date of January 1, 2010, that start date was randomly assigned to 2% of those in each comparison group). We included children who gained either CHIP or Medicaid during this time period, as insurance expansions in one public program (i.e., CHIP) often increase awareness of other available public programs (i.e., Medicaid). Given the complexity and sporadic nature of insurance coverage among children with discontinuous insurance over the study period, children with public insurance for only part of the pre- or post-period were excluded from analyses (n=7,489).

Utilization outcomes

We examined the following visit outcomes: 1) primary care; 2) well-child; 3) acute care; 4) total dental; and 5) preventive dental. Primary care visits were defined as visits in the primary care setting (clinic types designated as primary care, public health, or early childhood in the EHR system) and included well-child visits. Well-child, acute care, and dental visits were identified via Current Procedural Terminology (CPT) codes, Healthcare Common Procedure Coding System (HCPCS) codes, and/or International Classification of Diseases, Ninth Revision (ICD-9) codes (see footnotes in Table 2 for specific codes). We assessed utilization rates 1 year prior to each individual’s ‘start date’ (pre-period) and 1 year after the ‘start date’ (post-period).

Table 2
Within and between group comparisons of pre-post coveragea CHC visit rates by insurance group

Analyses

We used Fisher’s exact, chi-square and t-tests to compare demographic characteristics of the study groups. Missing demographic data were coded as a separate category, thus no patients were dropped from the models due to missing data.

To obtain pre-post utilization rate ratios and assess changes within and between groups, we fitted generalized Poisson mixed models with random effects to account for households nested within primary CHC (assuming a compound symmetry covariance structure).(22) Confidence intervals of rate ratios were used to assess significance in pre vs. post utilization rates within groups, and an interaction term between group and period was included to determine whether the pre-post rate ratios differed between groups.(23) To compare longitudinal patterns of primary care (well-child visits and acute care visits separately) and dental utilization across the two years, we computed monthly encounter rates using generalized estimating equations with a robust sandwich estimator(24) to account for within-patient temporal correlation.(25) All models were adjusted for potential confounders: child age on ‘start date’, gender, race/ethnicity, household income as percent of federal poverty level (FPL), and urban/rural classifications based on the 2004 Rural-Urban Commuting Area Zip Code file for the US. All statistical tests were two-sided and statistical significance was defined as a p-value<0.05. Statistical analyses were performed using SAS Enterprise Guide v.6.1. This study was approved by the Oregon Health & Science University’s Institutional Review Board and the data was collected under a waiver of authorization.

Results

Study population

We identified 24,307 patients meeting all study inclusion criteria. The newly insured group differed significantly from comparison groups on most demographic characteristics. Compared to the continuously insured, they were older, more commonly non-Hispanic white and less commonly Hispanic, less likely to have a household income under the FPL, and more commonly from a non-urban area. The only non-significant differences were gender and urban/rural classifications between the newly insured and the continuously uninsured group (Table 1).

Table 1
Patient demographic characteristics by insurance group

Utilization by insurance group

Among the newly insured group, post-period utilization rates were more than two times higher than in the pre-period for all outcomes (Table 2). Newly insured pre-post primary care visits increased from 0.54 to 1.28 per patient per year [adjusted rate ratio (aRR)=2.10, 95% confidence interval (95% CI)=1.95–2.26], well-child visits from 0.14 to 0.31 per patient per year (aRR=2.26, 95%CI=2.04–2.50), acute care visits from 0.18 to 0.55 per patient per year (aRR=3.12, 95%CI=2.80–3.47), total dental encounters from 0.35 to 0.98 per patient per year (aRR=2.77, 95%CI=2.56–2.99), and preventive dental encounters from 0.24 to 0.63 per patient per year (aRR=2.56, 95%CI=2.38–2.75). Comparison groups showed some significant pre-post utilization changes, though not of the same magnitude, direction, or consistency as observed in the newly insured. Between-group pre-post differences in rate ratios for all outcomes revealed that changes in utilization for the newly insured were significantly different from those of the continuously insured and continuously uninsured groups (p<0.001).

Temporal patterns of utilization

Well-child visit rates were higher among the newly insured compared to continuously insured immediately post-coverage with utilization rates leveling off over time. Primary care acute care visits also increased significantly in the post-period compared to pre-period for the newly insured group, but the rates for these visits were slightly lower than the continuously insured rates throughout the post-period. Utilization among the newly insured for both types of visits remained significantly higher than among the continuously uninsured (Figure 1). Adjusted monthly total dental visit rates showed a similar pattern as that of the well child visit rates, but spiked a few months later in the post-coverage period (Figure 2).

Figure 1
Adjusted1 monthly well child visit and acute care visit rates2 by insurance group, one year pre- through one year post-coverage ‘start date’3
Figure 2
Adjusted1 dental visit rates2 by insurance group, one year pre-through one year post-coverage ‘start date’3

Discussion

Our study uses EHR data to confirm findings from survey-based studies: children who gain public insurance coverage have increased utilization of important healthcare and dental services in the 12 months after gaining coverage. Among those who gained insurance, primary care visit rates doubled from an annual rate of 54 visits per 100 patients in the pre-period to an annual rate of 128 per 100 patients in the post-period; dental visit rates also doubled for this newly insured group. Our study adds substantially to the existing literature: we used EHR data to capture pediatric utilization rates during both insured and uninsured periods, and assessed monthly temporal patterns of utilization pre- and post-insurance coverage, including dental services.

This study builds on prior evidence suggesting that children utilize services differently when uninsured versus when insured, and that expansions in children’s public insurance coverage facilitates access to care for vulnerable children. Now, with CHIP’s future unclear, some experts warn that low- and middle-income children’s access to health insurance may be uncertain due to what has become known as the ‘family glitch.’(4, 26) If CHIP is not reauthorized or similar funding is unavailable, families with children who previously qualified for public health insurance would likely be eligible for federal subsidies through the ACA’s Insurance Exchange. However, if an employer offers affordable individual coverage for an employee based on his/her wages, the employee and his/her family are no longer eligible for subsidized insurance through the Exchange program. Thus, an employee might be able to afford individual insurance coverage but unable to afford it for the entire family. According to the Government Accountability Office, this ‘family glitch’ affects insurance coverage of almost half a million children.(27)

The comparison of monthly temporal patterns of utilization among the three groups has important implications. First, it demonstrates that ‘pent up demand’ for healthcare services observed among adults who gain coverage(28, 29) may also hold true for newly insured children. This finding has important workforce implications; specifically, whenever a group of children gain insurance coverage through expansions in a state or national program, the pediatric workforce should prepare for a ‘spike’ in utilization, to ensure capacity to meet this demand. Second, the increase in pre-post utilization among newly insured children suggests that insurance helped address unmet needs, which have been reported by parents of uninsured children.(30, 31) Similarly, the lower utilization rates among continuously uninsured children in this study, compared to newly insured or continuously insured children, might imply that these uninsured children continued to have significant unmet needs. Finally, after the initial increase, utilization rates of newly insured children fell below those of continuously insured children for both the well-child visits and dental visits. Rates of primary care acute visits for the newly insured did not surpass those for the continuously insured at any point in the post-period. Explanations for these interest findings include the possibility that families who are newly insured have different utilization patterns and/or less sustained need for services than those with more consistent coverage. These findings warrant further study as they could also be due to other unobserved differences between these groups.

Limitations

We did not include patients who were partially insured in the pre- or post-period because of the complex and sporadic nature of insurance coverage for this subgroup over the study period. Given our finding that gaining and maintaining insurance is, in fact, associated with increased utilization, future research is warranted to determine the graded effects of insurance coverage on utilization over time. We also excluded patients whose first visit was after the coverage start date, as we would not be able to assess pre- and post-utilization rates for these patients; thus, this paper only looked at established patients. EHR data was from patients at CHCs in Oregon; given differences across states in public insurance coverage eligibility standards and differences between CHCs and other primary care settings, results might not generalize to other healthcare settings. Our observational design prevents the ability to draw causal inferences; however, our finding that the temporal patterns of utilization among the comparison groups was relatively stable compared with utilization among the newly insured suggests that the post-period increase among those that gained insurance was not attributable to secular trends or other concurrent health policy initiatives. There were significant differences between the groups in measured covariates; while we controlled for available demographic variables, the groups may also have differed on unmeasured confounders. The monthly temporal correlation complicated the modeling approaches used. We were able to include households and clinics as nested random effects in the estimation of pre- and post-period rate ratios. However, the temporal modeling required random effects for individuals to account for within-patient correlation over time, and these models would not accommodate random effect for clinics and households. Finally, although access to, and appropriate utilization of, health care services are associated with improved health outcomes,(32, 33) and insurance coverage is shown to act interdependently with access to care to accomplish these outcomes,(34, 35) we did not examine the association between insurance and biomarker/health outcomes. Future studies are warranted to assess this relationship.

Conclusions

This study used EHR data to confirm previous evidence suggesting that public insurance coverage facilitates children’s access to primary and preventive care. It also suggests that healthcare utilization may spike in the immediate period after a child gains insurance, which has important pediatric workforce implications. These findings inform changes in the US health insurance landscape and how these changes might affect children’s access to primary care and dental care. If CHIP, Medicaid, or similar public funding becomes less available, unmet healthcare needs may again increase among low- and middle-income children in the US.

Significance

Survey-based research shows positive associations between CHIP expansions and children’s healthcare utilization. To build on this previous work, we used electronic health record (EHR) data to assess temporal patterns of healthcare utilization after Oregon’s 2009–2010 CHIP expansion. Using this novel data source, this study contributes to the body of evidence suggesting that insurance coverage facilitates children’s access to primary and preventive care. These findings are timely as the United States makes important policy decisions that could impact children’s access to public health insurance.

Acknowledgments

This study was supported by the Agency for Healthcare Research and Quality (AHRQ) grant number (R01 HS018569), AHRQ grant number (K08 HS021522), and the National Institute on Drug Abuse grant number (K23 DA037453). The funding agencies had no role in the study design; collection, analysis and interpretation of the data; in the writing of the report; or in the decision to submit the article for publication.

The authors are grateful to Carrie Tillotson, MPH, for statistical consultation, and to Christine Nelson, PhD, RN, for editing and reviewing the final version of this manuscript. The authors also gratefully acknowledge the OCHIN network and the OCHIN PBRN.

Footnotes

Conflict of Interest

The authors declare that they have no conflict of interest.

References

1. DeVoe JE, Tillotson CJ, Wallace LS. Children's receipt of health care services and family health insurance patterns. Ann. Fam. Med. 2009;7(5):406–413. [PubMed]
2. Newacheck PW, Stoddard JJ, Hughes DC, et al. Health insurance and access to primary care for children. N. Engl. J. Med. 1998;338(8):513–519. [PubMed]
3. Szilagyi PG, Schuster MA, Cheng TL. The scientific evidence for child health insurance. Acad. Pediatr. 2009;9(1):4–6. [PMC free article] [PubMed]
4. Kaiser Family Foundation. Children’s Health Coverage: Medicaid, CHIP and the ACA. [Accessed December 15, 2015];2014 http://kff.org/health-reform/issue-brief/childrens-health-coverage-medicaid-chip-and-the-aca/
5. Hess C, Basini LO, Plaza CI. Keeping Children’s Coverage Strong in the Context of the Affordable Care Act: Perspectives from State Children’s Health Insurance Leaders. Portland, ME: National Academy for State Health Policy; 2012.
6. Yu J, Harman JS, Hall AG, et al. Impact of Medicaid/ SCHIP disenrollment on health care utilization and expenditures among children: a longitudinal analysis. Med. Care. Res. Rev. 2011;68(1):56–74. [PubMed]
7. Szilagyi PG, Dick AW, Klein JD, et al. Improved access and quality of care after enrollment in the New York State Children's Health Insurance Program (SCHIP) Pediatrics. 2004;113(5):e395–e404. [PubMed]
8. Kenney G, Rubenstein J, Sommers A, et al. Medicaid and SCHIP coverage: findings from California and North Carolina. Health Care Financ. Rev. 2007;29(1):71–85. [PMC free article] [PubMed]
9. Kempe A, Beaty BL, Crane LA, et al. Changes in access, utilization, and quality of care after enrollment into a state child health insurance plan. Pediatrics. 2005;115(2):364–371. [PubMed]
10. Damiano PC, Willard JC, Momany ET, et al. The impact of the Iowa S-SCHIP program on access, health status, and the family environment. Ambul. Pediatr. 2003;3(5):263–269. [PubMed]
11. Howell EM, Kenney GM. The impact of the Medicaid/CHIP expansions on children: a synthesis of the evidence. Med. Care Res. Rev. 2012;69(4):372–396. [PubMed]
12. Brickhouse TH, Rozier RG, Slade GD. The effect of two publicly funded insurance programs on use of dental services for young children. Health Serv. Res. 2006;41(6):2033–2053. [PMC free article] [PubMed]
13. Duderstadt KG, Hughes DC, Soobader M-J, et al. The impact of public insurance expansions on children's access and use of care. Pediatrics. 2006;118(4):1676–1682. [PubMed]
14. Kenney GM, Coyer C. National findings on access to health care and service use for children enrolled in Medicaid or CHIP. MACPAC Contractor Report No. 1. 2012 Mar
15. Choi BC, Pak AW. A catalog of biases in questionnaires. Prev. Chronic Dis. 2005;2(1):A13. [PMC free article] [PubMed]
16. Bhandari A, Wagner T. Self-reported utilization of health care services: improving measurement and accuracy. Med Care Res. Rev. 2006;63(2):217–235. [PubMed]
17. Eisert S, Gabow P. Effect of Child Health Insurance Plan enrollment on the utilization of health care services by children using a public safety net system. Pediatrics. 2002;110(5):940–945. [PubMed]
18. Hudson JL, Hill SC, Selden TM. If rollbacks go forward, up to 14 million children could become ineligible for public or subsidized coverage by 2019. Health Aff (Millwood) 2015;34(5):864–870. [PubMed]
19. Racine AD, Long TF, Helm ME, et al. Children's Health Insurance Program (CHIP): accomplishments, challenges, and policy recommendations. Pediatrics. 2014;133(3):e784–e793. [PubMed]
20. Oregon Health Authority. [Accessed December 15, 2015]; http://www.oregonhealthykids.gov/healthykids/history.html.
21. Oregon Health Authority. Oregon Health Plan Annual Report, 2010–2011. [Accessed December 15, 2015];2011 http://www.oregon.gov/oha/healthplan/DataReportsDocs/FFY%202011%20Annual%20Report.pdf.
22. McCulloch CE, Neuhaus JM. Encyclopedia of Biostatistics. John Wiley & Sons, Ltd; 2005. Generalized Linear Mixed Models.
23. Puhani PA. The treatment effect, the cross difference, and the interaction term in nonlinear “difference-in-differences” models. Economics Letters. 2012;115(1):85–87.
24. White H. Maximum likelihood estimation of misspecified models. Econometrica. 1982;50(1):1–25.
25. Ziegler A. Generalized estimating equations. New York, New York: Springer; 2011.
26. Health Policy Brief: The Family Glitch. [Accessed December 15, 2015];Health Aff (Millwood) 2014 Nov 10; http://www.healthaffairs.org/healthpolicybriefs/brief.php?brief_id=129.
27. Gorvernment Accountabiity Office. Washington, DC: GAO; 2012. Jun, [Accessed December 15, 2015]. GAO-12-648 Children's Health Insurance: Opportunities Exist for Improved Access to Affordable Insurance. http://www.gao.gov/assets/600/591797.pdf.
28. Ku L, MacTaggart P, Pervez F, et al. Assocation for Community Affiliated Plans. Washington, DC: 2009. [Accessed December 15, 2015]. Improving Medicaid's continuity and quality of care. http://www.communityplans.net/Portals/0/ACAP%20Docs/ACAP%20MCQA%20Report.pdf.
29. Gold R, Bailey SR, O'Malley JP, et al. Estimating demand for care after a medicaid expansion: lessons from Oregon. J. Ambul. Care Manage. 2014;37(4):282–292. [PMC free article] [PubMed]
30. Federico SG, Steiner JF, Beaty B, et al. Disruptions in insurance coverage: patterns and relationship to health care access, unmet need, and utilization before enrollment in the State Children's Health Insurance Program. Pediatrics. 2007;120(4):e1009–e1016. [PubMed]
31. DeVoe J, Graham A, Angier H, et al. Obtaining healthcare services for low-income children: a hierachy of needs. Journal of Healthcare for the Poor and Underserved. 2008;19(4):1192–1211. [PMC free article] [PubMed]
32. Shi L. The impact of primary care: a focused review. Scientifica (Cairo) 2012;2012:432892. [PMC free article] [PubMed]
33. Starfield B, Shi L, Macinko J. Contribution of primary care to health systems and health. [Review] [158 refs] Milbank Q. 2005;83(3):457–502. [PubMed]
34. De Voe JE, Tillotson CJ, Lesko SE, et al. The case for synergy between a usual source of care and health insurance coverage. J. Gen. Intern. Med. 2011;26:1059–1066. [PMC free article] [PubMed]
35. Sommers BD, Baicker K, Epstein AM. Mortality and access to care among adults after state Medicaid expansions. N. Engl. J. Med. 2012;367(11):1025–1034. [PubMed]