Effect of Recent State Policies on Child Health Health Care Use

The Effect of Welfare Reform and SCHIP/Medicaid Expansions on Child Health and Health Care Use Maki Ueyama, Kosali Simon, and Rachel Dunifon Cornell University Department of Policy Analysis and Management June 2007 Background • Welfare reform (AFDC waivers and TANF) – No stated goals on child health or health care use but it changed child’s environment that may have affected child health and health care use – Delinked Medicaid eligibility rules from those of cash assistance • Medicaid expansions and SCHIP implementation increased availability of public health insurance to children and families Research Questions • Research Questions – Did welfare reform and SCHIP/Medicaid expansions affect child’s health or use of formal health care? • How do family and child’s own public health insurance eligibility affect child’s health insurance coverage? • How do parents’ and child’s own public health insurance participation affect child’s health status and health care use? • Preliminary Findings – No effect of welfare reform on child health and health care use. – Both family and child’ own public health insurance eligibility decreases employer/private insurance coverage and increases public health insurance coverage. – No effect of parents’ and child’s own public health insurance participation on child’s health status and health care use. Relevant Prior Evidence • Welfare Reform – Small mixed impacts on child health and health care use • Medicaid/SCHIP Expansions – Mixed impacts on child health – Mixed impacts on doctor visits but an increase in dental visits – Parental health insurance coverage increases child’s coverage and health care use Contributions • Uses nationally representative data to study welfare reform and child health • Examines how parents’ public health insurance eligibility and participation affect child health and health care use Conceptual framework Welfare Reform SCHIP/Medicaid Expansions Mother’s Employment Family Income Quantity & Quality of Mother’s Time Health Insurance Child’s Health & Healthcare Use Empirical model: Welfare Reform Difference-in-difference model Yist   0  1 X ist   2 Z st   3 AWAVst   4TANFst   5Tist   6 AWAVst * Tist   7TANFst * Tist   8 s   9 t   ist AWAV: AFDC waiver dummy variable TANF: TANF dummy variable T: Treatment group dummy variable Treatment group: - children living in a family headed by a single mother who has less than a high school degree Control group: - children living in a family headed by a single mother who has high school education or more All regressions include child, mother, and state characteristics and fixed effects for: state, year, child’s age. Linear probability models are used for indicator dependent variables. Empirical model: Medicaid/SCHIP Instrumental variables model- 1st stages First stages: Health insurance eligibility or status (child, family, parents) OwnEligist   0  1 SimOwnElig   ist ist Fam Elig   0  1 SimFamElig   ist ist ist OwnIns   0  1 SimOwnElig   ist ist ist OwnIns   0  1 SimOwnElig   2 SimMomElig   3 SimDadElig   ist ist ist ist ist MomIns   0  1 SimOwnElig   2 SimMomElig   3 SimDadElig   ist ist ist ist ist DadIns   0  1 SimOwnElig   2 SimMomElig   3 SimDadElig   ist ist ist ist ist OwnIns   0  1 Sim Fam Elig   ist ist ist IV: fraction of children, families or parents eligible in the given state/year/month/age (for children) or state/year/month/sex/female head or not (for families and parents) created by using CPS All regressions include child, mother, and state characteristics and fixed effects for: state, year, child’s age. SEs are clustered at the state level. Linear probability models are used for indicator dependent variables. Empirical model: Medicaid/SCHIP Instrumental variables model- 2nd stages Second stages: Health insurance status, health status, health care use (child) ˆ OwnIns   0  1 X ist   2 Z st  3OwnEligist   4 s  5 t  ist ist ˆ OwnInsist   0  1 X ist   2 Z st  3 FamEligist   4 s  5 t  ist ˆ Yist   0  1 X ist   2 Z st  3OwnInsist   4 s  5 t  ist ˆ ˆ ˆ Yist  0  1 X ist   2 Z st  3OwnInsist   4 MomIns  5 DadIns  6 s  7 t  ist ist ist All regressions include child, mother, and state characteristics and fixed effects for: state, year, child’s age. SEs are clustered at the state level Linear probability models are used for indicator dependent variables. Child Health and Health Care Use Measures Health Insurance Status (years 1992-2002 only) - Medicaid/SCHIP - Employer/private insurance Child health - Dummy indicating whether the child’s health is excellent, very good or good - Dummy indicating whether the child has any limitation of activity - Number of school days lost to illness in past 2 weeks (ages 5-17 only) Health care use - Number of doctor visits in past 2 weeks Data: NHIS • 1990-2002 National Health Interview Survey • Sample: all children aged 0-17 • Outcomes self-reported (by adult member of the household) • Each observation at child-year level • N ranges across outcomes from approx. 33000 to 243000 Results: Welfare Reform Control Group: children in single mother families who has high school education or more Good Health Waiver TANF Treat 0.004 (0.004) 0.007 (0.006) dropped Activity Limitation 0.005 (0.006) -0.007 (0.010) -0.004 (0.006) Lost School Days 0.003 (0.027) -0.009 (0.052) -0.014 (0.028) Doctor Visits -0.013* (0.007) -0.021* (0.013) dropped Waiver*Tre at TANF* Treat N 0.005 (0.007) 0.003 (0.005) 65730 -0.025*** (0.009) 0.005 (0.007) 66090 0.001 (0.042) 0.130** (0.057) 38708 0.007 (0.009) -0.009 (0.008) 65925 Results: SCHIP – 2nd stage results Child’s Health Insurance Participation End. Var: Child’s pub insurance eligibilities; IV: Simulated child’s eligibilities Employer/Private Child’s SCHIP eligibility Child’s Medicaid eligibility -0.039 (0.038) -0.075* (0.042) Medicaid/SCHIP 0.060 (0.040) 0.173*** (0.052) N 146523 145857 End. Var: Family pub insurance eligibility; IV: Simulated family eligibility Employer/Private Family public insurance eligibility N -0.063 (0.134) 146523 Medicaid/SCHIP 0.256* (0.138) 145857 Results: SCHIP – 2nd stage results Child’s Health and Health Care Use End. Var: Child’s pub insurance participation; IV: Simulated child’s eligibility Good Health Activity Limitation Doctor Visits Child’s public insurance participation N -0.032 (0.114) 241967 -0.593 (0.687) 243140 -0.568 (0.679) 242517 End. Var: Child’s and parents’ pub insurance participations; IV: Simulated child’s and parents’ eligibilities Good Health Child’s public insurance participation -0.989 (3.768) Activity Limitation 0.033 (2.940) Lost School Days -0.415 (7.947) Doctor Visits -0.789 (5.420) Mother’s public insurance participation Father’s public insurance participation N -0.836 (3.559) 1.787 (6.212) 91281 0.744 (2.741) -1.203 (4.927) 91485 -3.439 (6.715) 6.521 (10.292) 34390 0.069 (4.554) 0.508 (8.783) 91309 Summary of Findings – Most effects of welfare reform on child health and health care use are statistically insignificant • No effect on general health status • Zero or a small ambiguous effect on activity limitation, the # of lost school days, and the # of doctor visits – Both family and child’ own public health insurance eligibilities: • Decreases employer/private insurance participation • Increases public health insurance participation – No effects of parents’ and child’s own public health insurance on health and health care use Further Studies • Examine the effect of EITC expansions on child health and health care use • Examine all three policies simultaneously • Separate the sample into age groups Descriptive Statistics: By Mother’s Characteristics Variable Child Age Female Hispanic Non-white, NonHispanic All Single Mother, HS dropout 8.392 0.498 0.410 0.406 Single Mother, HS or more 9.028 0.495 0.172 0.414 8.474 0.488 0.208 0.201 Mother Age 35.456 33.159 35.039 Child Health, Heath Care Use & Health Insurance: By Mother’s Characteristics Variable Good Health Any limitation of activity # of lost school days (past 2 wks) # of doctor visits (past 1 year) All 0.975 0.063 0.159 0.131 0.608 0.194 0.144 Single Mother, HS dropout 0.934 0.100 0.220 0.118 Single Mother, HS or more 0.964 0.093 0.218 0.142 Employer/Private Insurance Medicaid/SCHIP No Insurance Results: Welfare Reform Control Group: children in 2-parent families whose mother has less than high school degree Good Health Waiver TANF Treat -0.001 (0.005) -0.009 (0.007) 0.044* (0.027) Activity Limitation -0.014** (0.006) -0.012 (0.009) 0.021 (0.022) Lost School Days 0.016 (0.030) 0.111 (0.083) dropped Doctor Visits 0.012* (0.007) 0.004 (0.012) 0.038*** (0.010) Waiver*Tre at TANF* Treat N 0.009 (0.007) 0.004 (0.005) 58430 -0.0009 (0.008) 0.002 (0.007) 58862 0.024 (0.040) -0.024 (0.064) 32673 0.003 (0.009) -0.015* (0.008) 58691