Table 3.

Summary of articles addressing the effects of public insurance on access to specialty care.

Author	Year	Sample Size	Data Source	Study Design	Access Measure	Statistic	Comparison	Findings	Endogeneity/Selection
Cabana et al. [21]	2002	3,163	Single MCO	Cross-sectional	Utilization	Logistic regression; likelihood of specialty care	Medicaid vs. non-Medicaid insured	Private with copay: OR = 2.52, p < 0.05 Private w/o copay: OR = 3.40, p = NS	Single MCO with Medicaid and private; patients do not choose
Damiano et al. [22]	2003	463	State SCHIP	Prospective cohort	Unmet need	McNemar; unmet need pre- vs. post-enrollment	SCHIP vs. prior coverage	40% vs. 13%; p < 0.05	No consideration of selection into program
Davidoff et al. [23]	2005	3413	National Health Interview Survey	Quasi-experimental	Utilization	Change in proportion with any visit	SCHIP ineligible vs. SCHIP eligible	+3.8, p = NS	Groups compared on eligibility, not enrollment
Forrest et al. [24]	1999	27,104	National practice-based research network	Prospective	Referral rates	t-test, percent referred; logistic regression, likelihood of referral to specialty	Medicaid vs. Private	4.46% vs. 2.61%, p < 0.001	No consideration of selection into coverage type
Holl et al [25]	2000	1,730	Single SCHIP	Quasi-experimental	Utilization	Change in proportion with any specialist visit	Prior to SCHIP enrollment vs. after enrollment	Age < 1 year: 15.5% vs. 16.1%, p = NS; Age 1–5 years: 19.7% vs. 19.4%, p = NS	No consideration of selection into program
Hwang et al. [26]	2005	54	Clinics in a single state	Cross-sectional	Appointment availability	t-test, proportion offering appointment	Private insurance vs. Medicaid	96% vs. 41%, p < 0.0001	Physician offices; no patient selection
Kempe et al. [27]	2000	596	Pediatric practices in a single state	Retrospective cohort	Referral rates	χ2; proportion with referral	Private insurance vs. Medicaid	11% vs. 20%, p = 0.09	No consideration of selection into coverage type
Kempe et al [28]	2005	480	Single SCHIP	Prospective cohort	Utilization	Logistic regression; saw specialist when needed; any specialist visit	Prior to enrollment vs. after enrollment	OR = 1.96, p < 0.05; OR = 1.22, p = NS	No consideration of selection into program
Mayer et al. [13]	2004	38,866	National Survey of CSHCN	Cross-sectional	Unmet need	Logistic regression; likelihood of unmet need for specialty care	Private insurance vs. Medicaid and SCHIP	Medicaid: OR = 1.26, p = NS; SCHIP: OR = 0.82, p = NS	No consideration of selection into coverage type
Ortega et al. [29]	2001	1,002	Multiple hospitals; single geographic region	Retrospective cohort	Utilization	χ2;percent seeing an asthma specialist	Private insurance vs. Medicaid	30% vs. 6%, p < 0.001	No consideration of selection into coverage type
Park et al. [14]	2002	1,985	National Health Interview Survey	Cross-sectional	Utilization	Proportion having seen a specialist	Private insurance vs. public insurance	Less likely vs. private insurance	No consideration of selection into coverage type
Perlstein et al. [15]	1997	544	Regional cardiac registry	Retrospective cohort	Time to referral	t-test; mean age at referral	Medicaid vs. "commercial"	168 days vs. 80 days, p < 0.05	No consideration of selection into coverage type
Price et al. [34]	1999	94	Single hospital	Cross-sectional	Utilization	t-test; number of specialist visits	Medicaid vs. fee-for-service	All: 3 vs. 6, p = NS; asthma-related: 2 vs.4, p < 0.05	No consideration of selection into coverage type
Szilagyi, et al. [31]	2000	187	Single SCHIP, children with asthma	Quasi-experimental	Utilization	χ2 and t-test; percent seeing specialist, number of visits	Prior to SCHIP enrollment vs. after enrollment	Any specialist: 30% vs. 40%, p = 0.02; Visits: 0.36 vs. 0.48, p = 0.02	No consideration of selection into program
Szilagyi, et al. [16]	2000	2,126	Single SCHIP	Quasi-experimental	Utilization	t-test, difference in number of specialist visits	Prior to SCHIP enrollment vs. after enrollment	0.174 more visits after enrollment, p < 0.001	No consideration of selection into program
Szilagyi et al. [30]	2004	2,644	Single SCHIP	Prospective cohort	Utilization and unmet need	Logistic regression, change in unmet needs pre- and post-enrollment	Prior to SCHIP enrollment vs. after enrollment	15.5 percentage point decrease after enrollment, p < 0.01	No consideration of selection into program
Wang et al. [32]	2004	100	Clinics in single state	Cross-sectional	Appointment availability	Percentage comparisons, no statistical test, percent offering an appointment	Private PPO vs. Medicaid	97% vs. 27%	Physician offices; no patient selection
Zwanziger, et al. [33]	2000	1,910	Single SCHIP	Quasi-experimental	Utilization	OLS, change in expenditures pre- and post-enrollment	Prior to SCHIP enrollment vs. after enrollment	$71.85 increase after enrollment	No consideration of selection into program