PEDIATRICS Vol. 105 No. 2 February 2000, pp. 363-371

Evaluation of a State Health Insurance Program for Low-Income Children: Implications for State Child Health Insurance Programs

Peter G. Szilagyi, MD, MPH^*, Jack Zwanziger, PhD, Lance E. Rodewald, MD^{*, #}, Jane L. Holl, MD, MPH^¶, Dana B. Mukamel, PhD, Sarah Trafton, JD, Laura Pollard Shone, MSW^*, Andrew W. Dick, PhD, Lynne Jarrell, MA, MPA^§, and Richard F. Raubertas, PhD

From the ^* Departments of Pediatrics,  Community and Preventive Medicine, ^§ Political Science, and  Biostatistics, University of Rochester School of Medicine and Dentistry, Rochester, New York; ^¶ Children's Memorial Hospital, Department of Pediatrics and Institute for Health Services Research and Policies Studies, Northwestern University School of Medicine, Chicago, Illinois; and the ^# National Immunization Program, Centers for Disease Control and Prevention, Atlanta, Georgia.

	ABSTRACT

Top Abstract Methods Results Discussion References

Background.  The State Child Health Insurance Program (SCHIP) is the largest public investment in child health care in 30 years, targeting 11 million uninsured children, yet little is known about the impact of health insurance on uninsured children. In 1991, New York State implemented Child Health Plus (CHPlus), a health insurance program that became a model for SCHIP.

Objective.  To examine changes in access to care, utilization of services, and quality of care among children enrolled in CHPlus.

Design.  A pre-post design was used to evaluate the health care experiences of children in the year before enrollment in CHPlus and during the year after CHPlus enrollment.

Setting.  New York State, stratified into 4 regions: New York City, urban counties around New York City, upstate urban counties, and upstate rural counties.

Participants.  A total of 2126 children (0-12.99 years of age) who enrolled in CHPlus in 1992-1993.

Data Collection.  Parents were interviewed by telephone, and primary care medical charts were reviewed for 694 children (0-3.99 years of age).

Analysis.  Access, utilization, and quality of care measures for each child were compared for the year before and the year after CHPlus enrollment, controlling for age, geographic region, previous insurance coverage, and CHPlus plan type (indemnity or managed care).

Results.  Enrollment in CHPlus was associated with fewer children lacking a medical home (5% before CHPlus vs 1% during CHPlus), with the greatest change occurring in New York City (11% vs 1%), where access before CHPlus was lowest. CHPlus was also associated with increased primary care visits: by 25% for preventive visits, by 52% for acute visits, and by 42% for total visits. The number of specialists seen during CHPlus was more than twice as high than before CHPlus. CHPlus was not associated with changes in emergency department utilization, although hospitalizations, which were not covered by CHPlus, were 36% lower during CHPlus coverage. Use of public health departments for immunizations declined by 64%, with more immunizations delivered in the medical home during CHPlus coverage. One third of parents reported improved quality of health care for their child as a result of CHPlus, and virtually none noted worse quality of care.

Conclusions.  This statewide health insurance program for low-income children was associated with improved access, utilization, and quality of care, suggesting that SCHIP has the potential to improve health care for low-income American children.  Key words:  Child Health Plus, uninsured, underinsured, health insurance, access, utilization, quality of care, State Child Health Insurance Program.

The State Child Health Insurance Program (SCHIP),¹ enacted in 1997 with passage of Title XXI of the Social Security Act, is the largest expansion in expenditures for child health care since Medicaid. SCHIP allocates $24 billion over 5 years, mostly as block grants to states (with required matching funds from states), to provide health insurance for children living below 200% of the federal poverty level who do not qualify for Medicaid.¹ SCHIP is a national response to one of the most vexing child health problems in the United Stateslack of adequate health insurance. More than 11 million children (14%) are uninsured for at least a year,^2-6 and even more have lengthy uninsured gaps^7-10 or are underinsured (>10% of household income spent on health care^9,10 or coverage lacking for preventive care¹¹). A critical question concerns the impact of SCHIP on access, use, and quality of care. This study is one of the first statewide evaluations of an SCHIP-like program and offers important lessons for the implementation and evaluation of SCHIP.¹²

Previous studies have shown that uninsured children are more likely than insured children to lack a usual source of preventive or sick care,^4,7,8,13 to delay seeking care,^13,14 to use fewer ambulatory health services,^1315-18 and to have fewer visits for common pediatric conditions.^7,13 Uninsured children also have lower immunization rates,^19-21 are more likely to be perceived by their parents as being in poor or fair health,¹³ and are more likely to be hospitalized for potentially preventable conditions,²² to be discharged from the hospital early after birth,²³ and to have an increased risk of adverse outcomes after birth.²⁴

These studies compared uninsured and insured children, implying that provision of health insurance to uninsured children would result in outcomes equivalent to those in insured children. However, uninsured children can differ substantially from insured children on a range of important characteristics, some of which have been measured (eg, income, race, and employment)^3,8,13 and some of which may be difficult to measure (eg, propensity to seek health care). Therefore, it may not be appropriate to generalize the experience of insured children to the uninsured. The current study goes beyond previous studies by evaluating what actually happens when uninsured children are provided with health insurance.

The objectives of this study were to determine the association between the provision of health insurance for low-income children and their subsequent access to care, utilization of health services, and quality of care. We compared the experience of the same children during periods before enrollment and during enrollment in a state-funded health insurance program, thus providing information relevant to the implementation of SCHIP.

	METHODS

Top Abstract Methods Results Discussion References

Setting

This study evaluated the voluntary enrollment of uninsured and underinsured children into New York's statewide health insurance plan. In 1991, New York State introduced Child Health Plus (CHPlus), a program for low-income children that was one of the models for SCHIP. Children were eligible for CHPlus if they were state residents, were younger than 13 years of age, were not enrolled in Medicaid, and lacked equivalent coverage. Families that had gross incomes below 160% of the federal poverty level (86% of enrollees) were eligible for full subsidy; those who had incomes of 160% to 222% of the federal poverty level paid $25 per child up to a family maximum of $100 per year, and families who had incomes above 222% of the federal poverty level purchased CHPlus for $498 to $798 per child per year. The covered services included: preventive and illness visits, screening tests, immunizations, specialty care, emergency department (ED) care, ambulatory surgery, and prescription drugs (in some plans $1-$3 co-payment and limited to generic medications). Hospitalizations, mental health services, and durable medical equipment were not covered. Enrollment rose from 10 000 in 1992 to 66 000 in 1993, to 100 000 in 1996, and is currently over 200 000.²⁵ Statewide, at the time of this study, one third of eligible children were enrolled in CHPlus,²⁶ similar to the proportion expected to enroll in SCHIP.¹²

The study included children enrolled by the 14 private insurance plans that offered CHPlus during the study period, including 3 indemnity plans, 3 individual practice association-model plans, and 8 staff-model managed care plans. To account for demographic variations within the state,²⁷ subject sampling was stratified into 4 regions: New York City, urban counties around New York City, upstate urban (Standard Metropolitan Statistical Area²⁸) counties, and upstate rural (non-Standard Metropolitan Statistical Area²⁸) counties.

Study Design

Because a randomized trial of CHPlus was not feasible, assessment of the impact of CHPlus could be affected by confounding factors. To minimize the effects of child- or family-specific factors (such as the propensity to seek medical care), a pre-post study design was used. Health care access, utilization, and quality measures were compared for children for the year immediately before their enrollment in CHPlus to the year immediately after enrollment. Parent interviews were conducted to determine demographic characteristics, sources of health care, experience with CHPlus, and child health status. Medical charts were reviewed for a subset of children 0 to 3.99 years of age at enrollment to measure utilization and quality of care. Chart reviews were restricted to these younger children because they use primary care more frequently and uniform guidelines exist for their preventive care.²⁶

Sample Selection

Subjects were selected from the New York State CHPlus enrollment file that contained information on 70 000 children in 1993; names, contact information, and date and duration of enrollment were provided from both state-level and insurer-level administrative files and were used to construct a master sampling file. Children were sampled randomly from this file, stratifying by region. Children were eligible for the study if they: 1) enrolled in CHPlus between July 1, 1992 and June 30, 1993, 2) were 0 to 12.99 years of age at enrollment, 3) were continuously enrolled for 9 months or longer, and 4) had contact information including a telephone number or address (2431 children lacked contact information and were excluded). The 35% of children enrolled for fewer than 9 months were excluded because their length of enrollment was deemed to be insufficient for experience on CHPlus to be assessed. Only 1 child per family was included in the study. A total of 4342 sampled children (~1000-1200 per region) met these eligibility criteria, and telephone contact of their families was attempted.

Data Collection

Parent Interviews Telephone interviews were performed ~2 years after CHPlus enrollment. Contact was attempted during days, evenings, and weekends; up to 10 attempts at contact were made. Interviews were conducted in English, Spanish, and 3 Chinese dialects.

Medical Chart Reviews Medical charts were reviewed for children who were 0 to 3.99 years of age at CHPlus enrollment. Many children had multiple sites of primary care, and all primary care clinicians identified by parents on interview were contacted for permission for chart reviews. A child's medical chart review was considered complete if a chart was received from every source of primary care listed by the parent for the relevant 2-year interval.

Measures

Demographic characteristics and previous health insurance coverage were obtained (Table 1).

Child health status questions were adapted from the National Health Interview Survey and included overall health status, change over 2 years, and impact of CHPlus.

Access to care was measured by interview and included a routine source for preventive and sick care (National Health Interview Survey) and measures of accessibility.

Utilization of health services was measured by chart review rather than parent recall to minimize recall error and bias. Measures included primary care visits, ED visits, and hospitalizations. Thus, previous utilization was obtained by chart review to minimize error from recall on interview. Parent interviews also were used to assess hospitalizations (for which parent interviews have been shown to be accurate^29-31) and specialty care visits (which frequently were not documented in primary care charts).

Quality of care was measured in part by parental rating of quality. Medical chart reviews were used to measure immunization rates, compliance with recommended preventive care visits,³² and screening for anemia (8-13.9 months) and lead (8-13.9 months and 21-26.9 months).³² Immunization status was calculated for the day before CHPlus enrollment and 12 months later. Age-appropriate, series-complete immunization coverage was defined using the most permissive interpretation of established guidelines in effect at the time that children in the study were vaccinated.^33,34

Statistical Analysis

Each child contributed 2 observations to the analysis, 1 for the year before enrollment (pre-CHPlus period) and 1 for the year after enrollment (CHPlus period). To allow for possible correlations between observations for the same child and for the effect of aging, a mixed model analysis of variance was used to estimate CHPlus effects. The model included a random effect for each child and fixed effects for: period type (pre-CHPlus vs CHPlus); child age (in years) at the start of a period; region; pre-CHPlus insurance coverage; type of CHPlus plan (indemnity, staff-model health maintenance organization [HMO], or individual practice association); and interactions of period type with the other factors. Insurance coverage during the pre-CHPlus year was described by the most recent insurance type before CHPlus (uninsured throughout the year, underinsured, Medicaid, or fully insured), and the length of the uninsured gap immediately before CHPlus (<6 months vs 6 months). Age was treated as a categorical factor with 1 level for each year of life, so that there was no assumption of a linear age effect on outcome variables. SAS procedure MIXED was used (SAS Institute, Cary, NC).^35,36

Because pre-CHPlus insurance coverage was not relevant for infants, the data for infants were analyzed separately, with age and pre-CHPlus insurance coverage omitted from the statistical models. Thus, infants contributed 1 observation to the analysis.

Binary outcome variables were analyzed using the same models as quantitative variables. Logistic regression analysis with random effects³⁶ was also performed and generally gave similar results. For some variables, the observed proportions were close to 0 or 1. This made the linear model analysis inappropriate and prevented convergence of the logistic regression analysis. In these cases, unadjusted comparisons of proportions were made, applying McNemar's test to observations paired by child.

	RESULTS

Top Abstract Methods Results Discussion References

Patient Contact

Interviews were completed for parents of 2232 children (51%). Parents of 106 of these 2232 children stated that their children were enrolled for fewer than 9 months, leaving a study sample of 2126 children. Reasons for noncompletion of interviews included parent refusal (22%), inability to make telephone contact (14%), language barrier (6%), and other (7%).

The interview response rate differed substantially across the 4 regions: New York City (335 of 1008; 33%), the urban region surrounding New York City (520/1140; 46%), upstate urban (741/1213; 61%), and upstate rural (636/981; 65%). Language problems accounted for much of the regional differences in response rate (18%, 6.4%, 1.0%, and 1.0%). In light of the low response rates in the 2 downstate regions, analysis results were examined separately for each of the 4 regions as well as for all regions combined. The overall interview response rate was similar to that of many large-scale surveys of difficult-to-reach populations^37-39 or insurance plan enrollees.^40,41

There were 1291 children 0 to 3.99 years of age eligible for medical chart review, and charts were reviewed for 694 children (54%). Five physicians (8 subjects) refused to allow chart reviews; the remaining reviews were not completed because charts were not located at sites specified by parents.

Key measures from parent interviews were compared for subjects with and without completed chart reviews (Table 1). The 2 groups did not differ significantly (P > .05) in age, gender, race, household composition, income, satisfaction with CHPlus, most reported utilization measures, and reported effects of CHPlus on quality of care or health status. Those without completed chart reviews were more likely to be from New York City or upstate urban areas and to have been uninsured before CHPlus; they also were less likely to be in managed care CHPlus plans. All 3 of these factors were controlled for in data analyses.

Demographic Characteristics and Parent Experience With Insurance (Table 1)

Demographic characteristics of the study population were similar to those of low-income families.²⁷ About one third of children had never had health insurance, half were insured (no gap in insurance) immediately before enrollment in CHPlus, and nearly half had an uninsured gap lasting 12 months or more immediately preceding CHPlus (insurance before CHPlus was controlled for in the analyses). Among children who had previous health insurance, reasons for loss of previous insurance included loss of a parent's job (36%), loss of Medicaid (30%), inability to afford the private insurance premiums (18%), change in family circumstances (6%), and change in benefit package or insurer cancellation of policy (5%).

Access to Care

The proportion of children who lacked any usual source of preventive care was significantly lower during CHPlus compared with before CHPlus (1% vs 5% before CHPlus; P < .001). This improvement was greatest in New York City (1% vs 11% before CHPlus), where access before CHPlus was poorest. The same was true for the proportion who lacked a source of sick care (<1% vs 3% before CHPlus; P < .001), again, with the greatest change occurring in New York City.

Accessibility of the medical home was higher during CHPlus than before CHPlus: 24-hour telephone coverage for sick care increased from 83% to 87% (P < .001), having a preventive care home <20 minutes away rose from 70% to 74% (P = .02), having 24-hour coverage for sick care increased for children 1 to 11.99 years of age from 83% to 87% (P < .001), and having the same usual source of preventive care and sick care rose from 95% to 98% (P < .001).

Tables 2 through 5 show the estimated effect of CHPlus on utilization and quality of care. Results are presented for all children in the statewide sample, and for each of the 4 regions. Measures are shown for the period before CHPlus (without CHPlus), and the difference between the 1-year period before CHPlus enrollment and the 1-year period after CHPlus enrollment.

Utilization of Health Services

Utilization of Primary Care (Table 2) Preventive, acute, and total visits to primary care practitioners were >25% higher during CHPlus than before CHPlus. For example, the estimated CHPlus effect for children 1 to 3.99 years of age was an additional 2.3 total primary care visits per year, including .38 additional preventive visits and 1.46 additional acute visits. Utilization was significantly higher during CHPlus in each region, with the greatest difference in New York City (additional 2.4 acute visits and 3.5 total visits per year; P < .001).

Subspecialty Care (Table 2) Significantly more specialists were seen during CHPlus, more than double the number before CHPlus for all children. This increase was noted in all regions. Subspecialty utilization was fivefold higher for those who were previously uninsured.

Emergency Care (Table 3) There was no significant difference in ED utilization associated with CHPlus. Approximately 10% of children made an ED visit each year. Except for infants in region 3, ED utilization was not significantly different after enrollment in CHPlus in any region.

Hospitalizations (Table 3) Hospitalization rates were significantly lower during CHPlus, whether measured by parent interview or by chart review. Overall, hospitalization rates were low and were similar to the national hospitalization rate of 4.1% reported by parents of uninsured children¹³ and to the New York State 1993 hospitalization rate for uninsured children of 3.6% per year for children 0 to 11.99 years of age (unpublished analysis of New York hospital discharge data). Within regions, there were few significant differences in hospitalizations after enrollment in CHPlus; however, the sample size was too small to detect anything but large changes in hospitalizations within regions.

Quality of Care

Parents reported improved quality of health care attributable to CHPlus for 33% of children, unchanged quality for 65%, and decreased quality for 2%. The reported effect of CHPlus on quality of care varied by geographic region (P < .001), with greatest improvements noted in New York City (43%); by CHPlus plan type (P = .015), with greatest improvements in staff model HMOs (37%); and by previous insurance (P < .001), with greatest improvements noted for previously uninsured children (40%).

Parents also stated that their child's health status was improved attributable to CHPlus for 29% of children, remained unchanged for 69%, and became worse for 2.8%. The extent of reported improvements in health status attributable to CHPlus varied by previous insurance (P = .006), with greatest improvements for children previously covered by Medicaid (36%); and by gap in insurance before CHPlus (P = .02), with greatest improvements for children who had uninsured gaps longer than 6 months (30%).

Table 4 shows the estimated effect of CHPlus on immunization delivery. CHPlus was associated with greater use of the medical home for immunizations and fewer children using health departments for immunizations (by 64% statewide) after CHPlus enrollment (P < .001). Although immunization coverage rates for infants seemed to increase by 16% (similar in magnitude to results of a similar study³⁴), this change did not reach statistical significance.

Table 5 shows the results for recommended preventive visits (having at least 1 preventive visit per year and being up-to-date on recommended number of preventive visits) and screening tests. Although the proportion of children who completed all the recommended number of preventive visits³² did not improve significantly, the proportion who had at least 1 preventive care visit during the year was significantly higher after enrollment in CHPlus than before enrollment. This increase was noted in all regions. In addition, there was a 73% increase at staff-model HMOs. There was no statistically significant CHPlus effect on screening for anemia or lead toxicity, but power to detect differences was limited because fewer than 280 children were age eligible for screening.

	DISCUSSION

Top Abstract Methods Results Discussion References

This is one of the first statewide evaluations of an SCHIP-like program. Because CHPlus is a model for SCHIP and because key aspects of the 2 programs are similar (ie, voluntary enrollment, coverage by private insurance, and proportion of eligible children enrolled), findings from this study may predict the impact of SCHIP on children who enroll. Indeed, the results are similar to those from a study of a similar program in Western Pennsylvania, in which impact of insurance was assessed by interviews.⁴²

Enrollment in CHPlus was associated with an increase in the proportion of children having a medical home (reaching nearly 100%), greater accessibility to a primary care practitioner, and dramatically lower reliance on public health departments for immunizations. The changes were greatest in New York City, where baseline access to primary care was poorest, suggesting that SCHIP may have even larger effects for populations that have poorer access to medical homes.⁴ Findings from this study may underestimate the potential impact of SCHIP programs, especially for programs that are implemented in areas where children currently have poor access to primary care. However, having a primary care physician does not guarantee access to health care services^13-18; it is critical to measure actual use of services.

CHPlus was associated with markedly higher use of primary care services statewide and within all 4 regions of New York State. This suggests that previously described differences in use of services between uninsured and insured children^16-18 are, in large part, attributable to health insurance coverage rather than to some intrinsic characteristic of uninsured children that may limit their propensity to seek health care.

It is a widely held belief that uninsured children frequently visit EDs and that providing health insurance will reduce unnecessary ED visits by keeping children in their medical homes. This study found no evidence of an impact of CHPlus on ED utilization. The substantial increase in the number of acute illness visits after enrollment in CHPlus was borne entirely by primary care physicians. This increase in acute illness visits may represent previously unmet demand for illness care.

This study observed a twofold increase in the number of specialists seen after enrollment in CHPlus, suggesting improved access to specialty care attributable to CHPlus. Additional study is needed to determine the proportion of specialty visits that were medically necessary and the extent to which the increase was attributable to unmet previous demand.

CHPlus did not cover hospitalizations and was associated with reduced hospitalizations below an already low baseline level of 4% to 6% per year. The financial incentives of enhanced ambulatory coverage without concurrent inpatient coverage may have altered physician behavior and reduced hospitalizations. New York State recently expanded CHPlus to cover hospitalizations⁴³ (as do most SCHIP programs); thus the observed changes in hospitalization rates may not occur under SCHIP.

CHPlus was associated with improvements in quality of care, although these improvements were less uniform than the dramatic changes in primary care utilization. Children were more likely to stay in their medical homes for immunizations and to obtain some preventive services after enrollment in CHPlus, suggesting improved continuity of care. However, compliance with recommended preventive care visits and immunization coverage levels remained low among children enrolled in CHPlus. This suggests that barriers other than insurance coverage still prevented timely receipt of some preventive services and is consistent with findings from other studies of suboptimal levels of care despite comprehensive insurance coverage.⁴⁴ In addition to health insurance, more intensive, targeted efforts are needed to optimize children's quality of health care.

The estimated CHPlus effect in having a preventive visit was the only measure that varied significantly by the CHPlus plan being indemnity or a managed care plan. Perhaps the impact of receiving any insurance coverage overshadowed the effect of managed care on use of services.

This study examined the experience of children who voluntarily enrolled in CHPlus. Although findings may not be generalizable to all uninsured children, results are relevant to children who will voluntarily enroll in SCHIP.¹² Of course these enrollees may differ in important characteristics from children who are eligible but are not enrolled. It is possible that the enrollees had greater unmet needs than eligible children who were not enrolled and that the CHPlus effects would, therefore, be lower for those children not enrolled. Different study designs involving population-based monitoring⁴⁵ will be required to evaluate the impact of SCHIP on the entire population of uninsured children, and those studies will be greatly influenced by the proportion who enroll.

A concern is the response rate of interviews (51%) and the completion rate of chart reviews (54%) in this study. Many eligible children were not included because their families could not be contacted or refused to participate. The 2126 index children who had completed interviews were similar in gender to the entire statewide CHPlus population but were slightly younger (mean age: 5.2 years vs 5.7 years of age; P < .05) and were less likely to be enrolled in managed care plans (70% vs 80%; P < .05); however, age and plan type were controlled for in data analyses. Children for whom chart reviews were completed had similar characteristics based on parent interviews to those for whom chart reviews were not completed (Table 1). This suggests that any nonresponse bias may have been small.

The interview response rate in the New York City region was particularly low (33%). A major reason was language barriers, despite interviews being conducted in 5 languages. Clearly an important subset of the CHPlus population in New York City consists of children of immigrant families, and it is likely that the barriers to health care faced by these families were not measured by this study.

To more accurately assess the potential bias resulting from the low response rate in New York City, a separate substudy was performed. An outreach organization was contracted to attempt to contact 100 children from the New York City region sample whom the interviewers had been unable to reach after 10 telephone attempts. Three outreach workers made home visits and contacted neighbors, physician offices, and community organizations to try to locate these patients. Of 100 eligible patients, contact was made with 31 families. Two parents refused an interview, and 29 were interviewed. Although the sample size was small, there were no substantial differences between this group and the sample interviewed in the primary study in terms of the demographic characteristics in Table 1, the health status of children, or reported utilization of health services. Nevertheless, we believe that results for the New York City region should be interpreted with care because of the low response rate.

Another potential concern involves recall errors for interview items. Interviewers attempted to minimize recall errors by using cues to help parents recall dates or child ages. The longer recall required for the pre-CHPlus year may have led to omissions in reported services or health care sites. In contrast, because parents may not have been able to limit their recall to the 12-month period before CHPlus, there may have been some over-reporting of services for the pre-CHPlus year. Thus, the extent of recall bias is unknown. However, recall error is unlikely to affect chart reviews, which were used to assess many of the key utilization and quality measures. The completion rate for chart reviews was similar for the period before and after enrollment in CHPlus.

This was an observational study, and the estimated CHPlus effects may have been influenced by secular trends. However, study enrollment covered a single year and pre- and post-CHPlus observations for each child covered consecutive 12-month periods, so the large differences observed are unlikely to be attributable solely to secular trends. In addition, because the enrollments were distributed throughout a 12-month period, and because pre- and post-CHPlus observations also covered 12-month periods, there is a great deal of overlap in the pre- and post-CHPlus groups, limiting the threat that secular trends in outcomes could be confounding the results. Furthermore, statewide hospitalization rates were unchanged between 1993 and 1994 (unpublished data), and annual CHPlus reports²⁵ do not identify any specific trends in characteristics of enrollees since the implementation of CHPlus. Lastly, a recent national study⁴ found that, since the 1980s, access to primary care for uninsured children had not improved.¹³ Thus, secular trends are unlikely to account for the improvements noted in this study.

As states begin to enroll children in SCHIP,⁴⁶ it is critical to evaluate the impact on children's health care and the limits to the benefits associated with health insurance coverage. This study found that provision of health insurance to low-income children was associated with substantially improved access to care, increased primary care utilization, and improvements in many measures of quality of health care. If similar changes also are realized with SCHIP, millions of uninsured and underinsured children in the United States are likely to benefit from expansions in health insurance for low-income children.

	ACKNOWLEDGMENTS

This study was supported by Grant CO11956 from the New York State Department of Health and Grant 92-5391 from the David and Lucile Packard Foundation.

We appreciate the expert computer programming of Richard Barth, the medical chart review expertise of Jacqueline Jennings, and the many individuals who contributed to the data collection and data management. We thank Michael Weitzman, MD, and Eugene Lewit, PhD, for their thoughtful review of the manuscript.

	FOOTNOTES

This study was approved by the Research Subjects Review Board of the University of Rochester.

Reprints not available. Address correspondence to Peter G. Szilagyi, MD, MPH, Department of Pediatrics, University of Rochester School of Medicine, Box 632, Strong Memorial Hospital, 601 Elmwood Ave, Rochester, NY 14642. E-mail: peter_szilagyi{at}urmc.rochester.edu

Received for publication Mar 22, 1999; accepted Jun 22, 1999.

This work was presented in part at the 1996 Annual Meeting of the Pediatric Academic Societies; May 8, 1996; Washington, DC.

	ABBREVIATIONS

SCHIP, State Child Health Insurance Program; CHPlus, Child Health Plus, ED, emergency department; HMO, health maintenance organization.

	REFERENCES

Top Abstract Methods Results Discussion References

1. Children's Health Insurance Program. Public Law 105-34. The Revenue Reconciliation Act of 1997
2. Lewit EM, Baker LS Health insurance coverage. Future Child 1995; 5:192-204 [CrossRef][Medline]
3. Fronstein P Children without health insurance coverage: an analysis of the increase in uninsured children between 1992 and 1993. Inquiry 1995; 32:353-359 [Medline]
4. Newacheck PW, Stoddard JJ, Hughes DC, Pearl M Health insurance and access to primary care for children. N Engl J Med 1998; 338:513-519 [Abstract/Free Full Text]
5. US General Accounting Office. Health Insurance for Children: Private Insurance Coverage Continues to Deteriorate. Washington, DC: US General Accounting Office; 1996. GAO/HEHS-96-129
6. Monheit AC, Cunningham PJ. Children Without Health Insurance. Washington, DC: US Department of Health and Human Services; 1992. Publication 93-0025
7. Kogan MD, Alexander GR, Teitelbaum MA, Jack BW, Kotelchuck M, Pappas G The effects of gaps in health insurance on continuity of a regular source of care among preschool-aged children in the United States. JAMA 1995; 274:1429-1435 [Abstract]
8. Newacheck P, Hughes D, Cisternas M Children and health insurance: an overview of recent trends. Health Aff (Millwood) 1995; 14:244-254 [Abstract]
9. Pepper Commission. Bipartisan Commission on Comprehensive Health Care: A Call for Action. Washington, DC: Government Printing Office; 1990:101-114
10. Short PF, Banthin JS New estimates of the underinsured younger than 65 years. JAMA 1995; 274:1302-1306 [Abstract]
11. Farley PJ Who are the uninsured? Milbank Memorial Fund Q Health Soc 1985; 63:476-503
12. Congressional Budget Office. Budgetary Implications of the Balanced Budget Act of 1997, Table 11. Washington, DC: Congressional Budget Office; 1997
13. Holl JL, Szilagyi PG, Rodewald LE, Byrd RS, Weitzman ML Profile of uninsured children in the United States. Arch Pediatr Adolesc Med 1995; 149:398-406 [Abstract]
14. Weissman JS, Stern R, Fielding SL, Epstein AM Delayed access to health care: risk factors, reasons, and consequences. Ann Intern Med 1991; 114:325-331
15. Newacheck, PW, Stoddard JJ, Hughes DC, Pearl M. Children's access to health care: the role of social and economic factors. In: Stein REK, ed. Health Care for Children. New York, NY: United Hospital Fund; 1997
16. Patrick DL, Madden CW, Diehr P, Martin DP, Cheadle A, Skillman SM. Health status and use of services among families with and without health insurance. Med Care. 1992;30-941-949
17. Cornelius LJ Barriers to medical care for white, black, and Hispanic children. J Natl Med Assoc 1993; 85:281-288 [Medline]
18. Stoddard JJ, St Peter RF, Newacheck PW Health insurance status and ambulatory care for children. N Engl J Med 1994; 330:1421-1425 [Abstract/Free Full Text]
19. Himmelstein DU, Woolhandler S Care denied: US residents who are unable to obtain needed medical services. A m J Public Health 1995; 85:341-344
20. Rodewald LE, Shiuh T, Zill E, Dietz V, Szilagyi PG Health insurance and underimmunization: lessons from the 1991 National Health Interview Survey. Pediatr Res 1995; 37:144
21. Zimmerman R, Janoski J Immunization barriers in Minnesota private practices: the influence of economics and training in vaccine timing. Fam Pract Res J 1993; 13:213-214 [Medline]
22. Weissman JS, Gatsonis C, Epstein AM Rates of avoidable hospitalization by insurance status in Massachusetts and Maryland. JAMA 1992; 268:2388-2394 [Abstract]
23. Margolis LH, Kotelchuck M, Chang HY Factors associated with early maternal postpartum discharge from the hospital. Arch Pediatr Adolesc Med 1997; 151:466-472 [Abstract]
24. Braveman P, Oliva G, Grisham M, Reiter R, Egerter S Adverse outcomes and lack of health insurance among newborns in an eight-county area of California, 1982 to 1986. N Engl J Med 1989; 321:508-513 [Abstract]
25. New York State Department of Health. Child Health Plus Health Plan for Kids: Annual Reports (1992, 1993, 1994, 1995, 1996, 1997, and 1998). Albany, NY: New York State Department of Health; 1992-1998
26. Szilagyi PG, Zwanziger J, Rodewald LE, et al. Evaluation of Child Health Plus in New York State: Final Report to the New York State Department of Health. Albany, NY: New York State Department of Health; 1996
27. New York State Department of Health, Bureau of Child and Adolescent Health. Maternal, Child and Adolescent HealthProfile 1991. Albany, NY: New York State Department of Health; 1991
28. New York State Department of Labor. New York State Standard Metropolitan Statistical Areas (New Area Definitions). Albany, NY: New York State Department of Labor; 1997
29. Harel Y, Overpeck MD, Jones DH, et al The effects of recall on estimating annual nonfatal injury rates for children and adolescents Am J Public Health 1994; 84:599-605 [Abstract/Free Full Text]
30. Pless CE, Pless IB How well they remember: the accuracy of parent reports. Arch Pediatr Adolesc Med 1995; 149:553-558 [Abstract]
31. Stephenson H Evaluation of self-report by adults of childhood otitis media histories. Audiology 1995; 34:124-134 [Medline]
32. American Academy of Pediatrics. Guidelines for Health Supervision II. Elk Grove Village, IL: American Academy of Pediatrics; 1988
33. American Academy of Pediatrics, Committee on Infectious Diseases. Red Book: Report of the Committee on Infectious Diseases. Elk Grove Village, IL: American Academy of Pediatrics; 1994
34. Rodewald LE, Szilagyi PG, Holl J, Shone LR, Zwanziger J, Raubertas RF Health insurance for low-income working families: effect on the provision of immunizations to preschool-aged children. Arch Pediatr Adolesc Med 1997; 151:798-803 [Abstract]
35. SAS Institute. SAS/STAT Software: Changes and Enhancements Through Release 6.11. Cary, NC: SAS Institute; 1996
36. Wolfinger R, O'Connell M Generalized linear mixed models: a pseudo-likelihood approach. J Stat Comput Simulation 1993; 48:233-243
37. Sisk JE, Gorman SA, Reisinger AL, Glied SA, DuMouchel WH, Hynes MM Evaluation of Medicaid managed care: satisfaction, access, and use. JAMA 1996; 276:50-55 [Abstract]
38. Gold M, Wooldridge J. Plan-Based Surveys of Satisfaction With Access and Quality of Care: Review and Critique. Washington, DC: US Agency for Health Care Policy and Research; 1994:1-32
39. Weiss B, Senf J Patient satisfaction survey instrument for use in health maintenance organizations. Med Care 1990; 28:434-445 [CrossRef][Medline]
40. Davis K, Collins KS, Schoen C, Morris C Choice matters: enrollees' views of their health plans. Health Aff 1995; 14:99-112 [Medline]
41. Misha SI, Dooley D, Catalano R, Serxner S Telephone health surveys: potential bias from non-completion. Am J Public Health 1993; 83:94-99 [Abstract/Free Full Text]
42. Lave JR, Keane CR, Lin CJ, Ricci EM, Amersbach G, LaValle CP Impact of a children's health insurance program on newly enrolled children. JAMA 1998; 279:1820-1825 [Abstract/Free Full Text]
43. New York Health Care Reform Act (HCRA) of 1996
44. Fielding J, Cumberland W, Petit L Immunization status of children of employees in a large corporation. JAMA 1994; 271:525-530 [Abstract]
45. Weinick RM, Weigers ME, Cohen ME Children's health insurance, access to care, and health status: new findings. Health Aff 1998; 17:127-136 [Abstract]
46. Gauthier AK, Schrodel SP. State initiatives in health care reform: expanding children's coverage, lessons from state initiatives in health care reform. May 1997. Unpublished report