 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
SUPPLEMENT ARTICLE |
* Departments of Pediatrics and Community and Preventive Medicine, University of Rochester School of Medicine and Dentistry, Rochester, New York
Institute for Child Health Policy, University of Florida, Gainesville, Florida
Center for Delivery, Organization, and Markets, Agency for Healthcare Research and Quality, Rockville, Maryland
|| Kansas Health Institute, Topeka, Kansas
¶ Departments of Pediatrics and Medicine, Indiana University School of Medicine, Indianapolis, Indiana
# The David and Lucile Packard Foundation, Los Altos, California
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Objectives. To use data from the Child Health Insurance Research Initiative to measure the prevalence of CSHCN in SCHIP in 4 states, describe their demographic and health care features at enrollment, and compare their sociodemographic characteristics, health status, prior health care experiences, and unmet needs versus children without special health care needs.
Methods. Children (0–18 years old) newly enrolled in SCHIP in 4 states were eligible for the study: New York, Florida (adolescents only), Kansas, and Indiana (CSHCN only). Telephone interviews were conducted shortly after enrollment and identified CSHCN by using the Child and Adolescent Health Measurement Initiative CSHCN screener. A common set of core questions assessed demographic characteristics, health status, special health care need status, insurance experience, access, use, quality of health care, and unmet needs during the year before enrollment. Bivariate and multivariate analyses were used to compare characteristics of CSHCN with characteristics of children without special needs.
Results. Interviews were completed for parents of 5296 children enrolled in SCHIP in the 4 states. By using the Child and Adolescent Health Measurement Initiative CSHCN screener, the prevalence of CSHCN among SCHIP enrollees was 17% (New York), 18% (Florida), and 25% (Kansas), higher than the prevalence of CSHCN reported in the general population in those states. More than half of CSHCN reported the use of a chronic medication. Demographic characteristics of CSHCN were similar to those of children without special needs, although CSHCN were more likely to reside in single-parent households. Although CSHCN had poorer health status than children without special needs, many CSHCN were reported to be in good health, suggesting a wide spectrum of severity of illnesses within the CSHCN group. Although CSHCN were more likely than children without special needs to have been insured before SCHIP, a large proportion of CSHCN were nevertheless uninsured for at least 12 months before SCHIP (New York, 56%; Florida, 68%; Kansas, 24%; Indiana, 25%). Although most SCHIP enrollees had a usual source of care (USC) before SCHIP and there was some variation across states, between 4% and 13% of CSHCN lacked a USC on enrollment, and 23% to 38% of CSHCN changed their USC after enrollment in SCHIP. The majority of all SCHIP enrollees (including CSHCN) had used some health care during the year before SCHIP including preventive, acute, or specialty care. A high proportion of all SCHIP enrollees, including >30% to 40% of CSHCN, were reported to have unmet health care needs at enrollment in SCHIP. A variety of unmet needs were reported by CSHCN including specialty care, mental health care, dental care, and prescription medications. Nevertheless, the vast majority of CSHCN as well as children without special needs rated the quality of their medical care before SCHIP highly on several specific quality measures. Findings from multivariate analyses were similar to bivariate results with CSHCN in several states having higher use of care and more unmet health care needs before enrollment.
Conclusions. SCHIP is enrolling many CSHCN, with the prevalence of these children occurring at least as high as the prevalence of CSHCN in the general population. CSHCN enrolled in SCHIP represent a heterogeneous population with a wide range of health status and health care needs. Although most CSHCN were already connected to the health care system with a USC and prior health care visits, many had unmet health care needs before enrolling in SCHIP.
Implications for Monitoring and Improving SCHIP for CSHCN Enrollees. 1) SCHIP benefit packages need to adequately cover services required by CSHCN such as prescription medications and specialty, mental health, developmental, and home services; 2) because utilization of care will be high among this large group of children, alternative methods of financing and managing care should be considered such as risk adjustment and special programs that involve case management and care coordination; 3) coordination of care across programs (such as between SCHIP and the state Title V Maternal and Child Health Services program, a component of which serves CSHCN) and ensuring adequate access to primary care and specialty providers might improve access to services for CSHCN; and 4) it is critical to monitor the quality of care for CSHCN enrolled in SCHIP, because these children are among the most vulnerable children covered by public health insurance programs and many of them are enrolling in SCHIP.
Key Words: children with special health care needs (CSHCN) • State Children’s Health Insurance Program (SCHIP)
Abbreviations: SCHIP, State Children’s Health Insurance Program • CSHCN, children with special health care needs • FPL, federal poverty level • CMS, Children’s Medical Services • USC, usual source of care • ED, emergency department
The State Children’s Health Insurance Program (SCHIP) was enacted by Congress to provide health insurance coverage to children in low-income families who are not eligible for Medicaid or covered by private health insurance. SCHIP legislation allows states to provide this coverage either by expanding Medicaid, developing free-standing separate insurance programs that are more like employer-sponsored private insurance, or a combination. Although Medicaid programs provide a rich set of services for children with special health care needs (CSHCN), free-standing SCHIP programs typically provide less extensive benefits. To design SCHIP to address the needs of CSHCN adequately, it is necessary to measure the degree to which children who enroll in SCHIP have special health care needs and to assess their health status and unmet health care needs at enrollment. As yet, little is known about the characteristics or preenrollment experience of CSHCN who enroll in SCHIP.1,2
There is no absolute consensus about how to define CSHCN.3 In the past, CSHCN were defined by using categorical approaches such as the presence of specific diagnoses such as cancer or eligibility for specific programs such as the state Title V Maternal and Child Health Services program. Recently, a group of child health experts sponsored by the federal Maternal and Child Health Bureau identified CSHCN as "those who have or are at increased risk for a chronic physical, developmental, behavioral, or emotional condition and who also require health and related services of a type or amount beyond that required by children generally."3 This broader definition is noncategorical and includes children with chronic physical, mental, developmental, and behavioral needs.
Some instruments have been developed that identify CSHCN by using a noncategorical approach.4–9 Two analyses that applied this noncategorical definition of CSHCN to the 1994 National Health Interview Survey estimated a national prevalence of CSHCN in the general population to be between 14.8%10 and 18.2%.11 A major new national survey, the National Survey of Children with Special Health Care Needs,12,13 soon will provide data on 
750 CSHCN per state. However, the number of CSHCN within SCHIP may be too low per state to assess how well SCHIP is meeting the health care needs of enrollees. Furthermore, national surveys tend not to distinguish SCHIP from other health insurance programs. Thus, state-based evaluations are needed.
Although noncategorical approaches for identifying CSHCN have been proposed as Health Plan Employer Data and Information Set measurements,14,15 they are not used widely by SCHIP or Medicaid managed care plans. State SCHIP strategies tend to follow the definitions of their state Medicaid programs,16,17 using available administrative data18 to identify children by the presence of specific diagnoses, a high level of services or costs, or enrollment in a special program such as the Title V Maternal and Child Health Services program.19 Because of the lack of uniformity in definition and the inability to even identify the CSHCN population within SCHIP, little is known about CSHCN enrolled in SCHIP.
Regardless of which criteria for CSHCN identification are used, many studies have noted that CSHCN have poorer health status, higher health care needs, higher use of outpatient and inpatient services, more frequent use of specialty services and ancillary services, and higher health care costs than the general population of children.20–30
This study is part of the Child Health Insurance Research Initiative, funded by the Agency for Health care Research and Quality, The David and Lucile Foundation, and the Health Resources and Services Administration. Study objectives are to 1) determine the prevalence of CSHCN among enrollees in SCHIP in several state programs, 2) describe sociodemographic and health characteristics of CSHCN, and 3) compare CSHCN versus children without special needs with respect to demographic characteristics, health status, prior insurance, prior health care experience, and prior unmet health care needs. We hypothesized that 1) SCHIP would enroll fewer CSHCN than their prevalence in the general population, because many might have been enrolled in other programs, 2) CSHCN who do enroll in SCHIP would already be higher users of health care than other children because of their chronic conditions, but 3) CSHCN would enter SCHIP with many unmet needs.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
With the exception of Florida, most CSHCN were enrolled into SCHIP in the same manner as other children. Florida has a carve-out program for CSHCN eligible for Medicaid or SCHIP called Children’s Medical Services (CMS), which is part of Florida’s state Title V Maternal and Child Health Services program. Florida children are identified through the SCHIP enrollment process as potential eligible CMS enrollees by using 3 screening questions modified from the CSHCN screener. More severely disabled CSHCN are referred from SCHIP to CMS.37 In all states, children with disabling special needs may be eligible for special programs such as the Supplemental Security Income program, Medicaid "medically needy programs," Katie Beckett Waivers, and Title V Maternal and Child Health Services programs,2 which may provide additional services or opportunity for enrollment in Medicaid.
Study Design
Comprehensive telephone interviews were performed 2 to 7 months after enrollment in SCHIP. The interviews addressed demographic and health care information during the year before enrollment in SCHIP. A core set of questions was developed across the 4 states to obtain similar information. The presence of special needs was assessed by using the CSHCN screener (see Table 2, described more completely below), a widely used 5-question screener identical to the one used in the National Survey of Children with Special Health Care Needs8,9,12,13 and one that is in the CSHCN module recently approved by the National Committee on Quality Assurance for optional use by Health Plan Employer Data and Information Set performance-measure assessments.14,15
|
Telephone Interviews
Study teams obtained the list of SCHIP enrollees from their states, and telephone calls were made at varying times of the day and week. Altogether, 5296 parent interviews were completed: 2620 in New York, 1824 in Florida, 776 in Kansas, and 76 in Indiana.
Socioeconomic, Demographic, and Health Status Measures (Table 3)
These measures included the child’s age group, gender, race and ethnicity, urbanicity, income, family structure, respondent education and employment, and 3 commonly used health status questions.
|
Access, Utilization, and Quality
The major comparison involved CSHCN versus the remainder of SCHIP enrollees (before enrollment in SCHIP) within each state for 1) prior insurance variables (the number of uninsured months during the year before SCHIP, the type of most recent insurance, and reasons for loss of prior insurance); 2) presence of a usual source of care (USC) before enrollment in SCHIP, type of source, reasons for lack of a source of care, and change in the USC after enrollment; 3) use of health care before SCHIP (any use and number of visits); 4) unmet needs before enrollment; and 5) quality of care (continuity, specific quality markers regarding provider’s care, and satisfaction with care).
Analyses
Descriptive statistics (
2 analyses or t tests) were used to tabulate proportions of SCHIP enrollees who were CSHCN and to compare characteristics of CSHCN versus children without special needs.
We used multivariate methods to generate "adjusted" rates for 5 key measures of health care access: 1) having a USC; 2) use of the emergency department (ED); 3) use of preventive care; 4) unmet needs; and 5) prior insurance. All sociodemographic measures from Table 3 were included in the multivariate analyses for the first 3 models; prior insurance was omitted as an independent variable in the fourth model because it was the dependent measure. For the first 4 dichotomous dependent variables, we estimated multivariate logistic regressions to examine the independent association with CSHCN, controlling for other sociodemographic characteristics. For prior health insurance, which has 3 levels (none, Medicaid, and private/other), we estimated a multivariate multinomial logistic regression model.
Adjusted estimates were calculated for the CSHCN group and for other children by predicting the outcomes for all children in the sample after setting the special needs variables to CSHCN first and then to "other children." The differences in the adjusted estimates between CSHCN and other children reflect outcomes that may be attributed to CSHCN. For ease of interpretation, all multivariate models were run with no interaction terms; otherwise, models across states would have had varying components and been difficult to compare. All analyses were weighted to be representative of the SCHIP populations by state using STATA 7.0, SUDAAN 8.0.1, or SAS 8.0.1 to adjust for the complex survey designs in each state.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The largest proportion of children screened on one dimension of special need alone (dependency on prescription medications, functional limitation, or service use above normal). However, 1.7% (New York) to 3.9% (Kansas) of children had all 3 dimensions of need.
Demographic and Health Characteristics of CSHCN Versus Other SCHIP Enrollees (Table 3)
Although CSHCN tended to be older and more likely to reside in a single-parent household, large demographic differences between CSHCN and other children were not noted. The majority of children, including those identified as CSHCN, were described by parents as having excellent, very good, or good health. However, there was a clear trend toward CSHCN having poorer health status than other children. A high proportion of parents reported worrying about their child’s health, and in Florida the level of worry was higher for CSHCN.
Table 4 summarizes the prior health insurance experience of CSHCN versus other children. Four major findings are noted. First, there was a trend toward CSHCN being more likely than children without special needs to have been insured before SCHIP. Second, in most states there seem to be 2 predominant patterns of prior insurance: children covered for the entire 12 months (ie, enrolled immediately in SCHIP after their previous coverage ended) or children uninsured for the entire year before SCHIP. Florida differed in this regard in that fewer children were covered for the entire 12 months and more had private insurance before enrolling in SCHIP. Third, although there were variations by state, a substantial proportion of children had Medicaid as their most recent prior insurance. Fourth, the major reasons for loss of prior insurance involved a parent’s employment change, the high costs of the prior insurance, the perception that there would be better benefits from SCHIP compared with their prior insurance, or a life change (particularly loss of Medicaid because of rising income).
|
10% of CSHCN had no USC before SCHIP. The most common type of prior USC for all SCHIP enrollees was a doctor’s office outside of the hospital, followed by community health centers and hospital clinics. The type of USC differed across states, with Florida SCHIP enrollees being less likely than enrollees from other states to have received care from hospital-based clinics previously. A common reason for not having a USC was lack of health insurance. The reasons varied widely across states. In general, CSHCN were less likely than children without special needs to cite never being sick as a reason for not having a USC.
|
Many children, both with and without special needs, used health services during the year before SCHIP (Table 6). This includes a high proportion using the ED, >65% having a preventive visit, and a substantial proportion having an acute visit. Although there was some variability across states in the use of health care and not all visit types were assessed by all states, CSHCN were more likely than children without special needs to use the ED, mental health care, specialty care, and acute care.
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Other than being more likely to live in single-family households, demographic characteristics of CSHCN closely resembled those of children who did not screen positive for special needs. Confirming our second hypothesis, compared with children without special needs, CSHCN seemed to be somewhat more connected to the health care system before enrolling in SCHIP with respect to prior insurance, USC, and health care visits. This greater connection with the health care system may have been because of the poorer health status of CSHCN.
A higher proportion of SCHIP enrollees from New York and Florida were uninsured for the entire year before SCHIP, and a higher proportion in Kansas than the other study states had Medicaid as prior insurance. It is possible that the established nature of the New York and Florida programs accounts for the former finding in that the earlier SCHIP initiatives may have already enrolled children who were uninsured for short periods. The reasons for loss of prior insurance seem to reflect the type of prior insurance. For example, in Kansas where more children had Medicaid previously, fewer parents noted cost issues as reasons for loss of their prior insurance.
In support of the third hypothesis, the level of unmet health care needs of CSHCN on enrollment was high in all 4 states. This suggests that even as programs such as SCHIP mature, new enrollees will continue to have high levels of risk. In fact, both the CSHCN and children without special needs in the 2 states with mature programs were more likely than children in the 2 states with newer programs to have been uninsured for the entire year before SCHIP, underscoring the potential for unfulfilled health care needs among this population as well as the continuing need for long-term health insurance programs that cover all eligible children.
It is notable that the CSHCN population was generally similar to those without special needs on many access measures, and that the reported health status of most CSHCN was quite good. This suggests that the CSHCN screener identifies a broad spectrum of children and could not be used by SCHIP to focus exclusively on the most severely ill children. Also, many children who were not identified by the CSHCN screener still had unmet health care needs and fair or poor health. Thus, for the SCHIP population, the CSHCN screener does help to identify a group of at-risk children who merit extra attention from health care programs to ensure that their needs are met and to monitor their quality of care, but many other children are also potentially at risk for adverse health outcomes.
Two overarching trends emerged across New York, Florida, and Kansas in comparing CSHCN to children without special needs. CSHCN were more connected with the health care system (more likely to be insured, higher USC, and greater utilization) before SCHIP than children without special needs but nevertheless had more unmet health care needs (most commonly mental health, specialty care, and prescription medications). These trends could be detected despite the fact that the demographics of the children studied varied among the states (eg, the Florida study focused on adolescents, the New York sample had more black and Hispanic children, and the Kansas population was more rural). These findings suggest that SCHIP can work with existing health providers and networks to improve the health care delivery for CSHCN.
Policy Implications For SCHIP
These findings have policy implications for SCHIP directors, policy makers, and child health experts regarding 1) SCHIP benefit design and payment systems, 2) financing, 3) coordination of care, and 4) monitoring and improving the quality of care for CSHCN. Some policy implications depend on the state context, and there are some statewide differences in SCHIP with respect to CSHCN (as described previously). Despite these differences, the overall findings across states were similar for many measures such as the prevalence of CSHCN (17–25%), the portion of CSHCN in fair/poor health (14–21%), the income and employment-related demographic characteristics of CSHCN, the portion having a USC, and the use of health care before SCHIP.
Benefits and Program Design
In most states, the needs of CSHCN were not actively considered when policy makers debated the design of separate SCHIP models, although some states did make special provisions for CSHCN.42 Perhaps policy makers envisioned that SCHIP would enroll a relatively healthy population of children while children with special needs would enroll in Medicaid or special programs tailored to that population. In fact, an early qualitative study involving key informant interviews of SCHIP leaders in 5 states found "one issue that is particularly vexing to state officials and observers is the low number of children identified in their SCHIP as having special health care needs."43 The authors and key informants wondered whether causes for low enrollment of CSHCN included unsuccessful outreach to this population, the early stage of SCHIP, or perhaps that CSHCN were less likely to be uninsured. In contrast, SCHIP in our 4 states enrolled a population of children who, rather than being healthier than the general population, have more CSHCN than the general child population. The overall high level of unmet health care needs highlights the potential impact that SCHIP could have on the health of low-income children, especially CSHCN.
Separate, free-standing SCHIP programs including those in 3 of the 4 states evaluated in this study tend to offer more restricted benefit packages than Medicaid-expansion programs.44,45 This is important with respect to CSHCN. We found that CSHCN had more unmet health care needs than other children, and prior studies have noted that many CSHCN require more specialty and ancillary services than their healthy counterparts.21–30 For CSHCN, the benefit package design is critical.46 This is particularly important for states that do not have carve-out programs for CSHCN or aggressive referral systems to Title V Maternal and Child Health Services programs and other wrap-around services that enable CSHCN with the most severe disabilities to receive expanded services. It may also have relevance to the consideration of copayments as a means of reducing state costs for SCHIP, because copayments can inhibit the receipt of health care, particularly for low-income families.2,21
A continuing challenge for states when designing SCHIP programs and benefits is to consider and distinguish children who may have more mild conditions from those with more serious conditions and health needs. Some children with more mild conditions may not require the full array of services typically found in a Medicaid benefit package. Nonetheless, a lower proportion but overall large number of children who did not screen in as CSHCN also had unmet health care needs before SCHIP, suggesting that SCHIP has potential to improve health care for the general population as well as for CSHCN.
Financing
The high utilization rates and unmet needs before SCHIP suggest that utilization of services by CSHCN during SCHIP will be substantial, resulting in higher costs than for children without special needs. Programs may need to consider different financing mechanisms for at least some of the CSHCN with the most severe conditions. These include risk-adjustment strategies,47–49 high-risk pools,50,51 reinsurance mechanisms,52,53 and improved compensation of primary care practitioners (medical homes) for case management of CSHCN.
Coordination of Care
Coordination of care can occur at several levels, starting with the application process. Because many CSHCN enrolled in SCHIP and states have other programs that specifically serve CSHCN such as Title V Maternal and Child Health Services programs,54 Supplemental Security Income,55 or home- and community-based services waivers,56 stronger ties between SCHIP, Medicaid programs, and these other programs could enhance the identification of CSHCN and the monitoring of their level of care.57 Specific examples of coordination at this level include the use of joint application forms and single application processes to multiple programs, automatic referrals from one program to another, and joint care coordination.
In addition, health plans could expand their specialty provider network based on providers’ participation in the other programs and offer resources for coordination of care by health care providers. Furthermore, because the USC for many children enrolled in SCHIP was a community health center, it is important to promote connections between these providers and specialty providers, particularly for CSHCN. Our finding that many CSHCN changed their primary care providers after enrollment into SCHIP highlights the importance of provider networks. Continuity of care is critical for CSHCN because of their chronic conditions.58,59 Additional study is needed to determine why so many children changed providers, to assess whether limited provider panels may have accounted for these findings, and to evaluate the impact of changing providers on quality of care for CSHCN. Our study suggests that the adequacy of specialty provider networks including mental health care also needs to be evaluated within SCHIP. Although studies have found that access to specialists has been one of the more common barriers to CSHCN in managed care settings,25,60–64 little is known about access to specialty care within SCHIP.
Monitoring Quality
The high prevalence of CSHCN in SCHIP and the high rates of change in USC after enrollment in SCHIP both point to the importance of monitoring the quality of care for this vulnerable population. An important first step is the use of a systematic means to identify the population of CSHCN within SCHIP. One potential strategy is for SCHIP to adopt the CSHCN and/or other screening instruments for families to complete at enrollment and recertification. The next step would be to monitor their level of care by using routine assessment-of-health measures that are critical to CSHCN65–67 or children with specific diseases.68 Our results of quality of care show high marks for the quality of providers’ care even before SCHIP. These quality measures did not distinguish CSHCN from other children, and it is possible that quality measures more specific to CSHCN would be useful for this subgroup. For example, performance measures focused on mental health, specialty, and dental care may be useful in particular for CSHCN, and these were all areas of high unmet needs in the current study.
Strengths and Limitations
This study has several strengths. The results represent early findings from SCHIP in several states representing >25% of all SCHIP enrollees in the nation. Although there are some variations in aspects of these programs, the similarities across states in prevalence of CSHCN, demographic characteristics, and prior health care experiences of CSHCN suggests that the findings may be applicable beyond single states. These results will be critical to determining the impact of SCHIP on the health care of CSHCN, because the quality of health care during SCHIP represents a combination of the performance of plans and providers as well as sociodemographic features and prior unmet needs of patients.
Several study limitations must also be noted. Large-scale, multistate comparisons would be needed to assess whether findings from the study’s 4 states are applicable to other states. Unfortunately, national surveys including the National Survey of Children with Special Needs often cannot distinguish the SCHIP population or else have a limited number of SCHIP subjects. However, comparisons of the study findings to aggregate results from national-level studies that can identify the SCHIP population would be helpful.
Second, several questions varied somewhat across states, and the study populations varied in that the Florida sample included only adolescents and the Indiana sample included only CSHCN. The major study objective was to compare CSHCN versus other children within the same states and differences in questions and populations-limit comparisons across the states. Third, some of the differences between CSHCN and children without special needs may not be because of the special needs but to other confounding factors that were not accounted for by multivariate analyses. Fourth, no screening tool is perfect, and the CSHCN screener may have misclassified some children into or not into the CSHCN group.13,18 Fifth, these findings are from parent reports and subject to inaccuracies. The questions on utilization of care may be most subject to errors in reporting, although some evidence exists that parents accurately recall key visits to health care professionals such as ED visits and specialty visits.69,70 Moreover, because the same measures were used to compare CSHCN versus children without special needs, inaccuracies in parental recall may not translate to a bias toward either higher or lower utilization for CSHCN versus children without special needs.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
We gratefully acknowledge collaboration of Robert St. Peter (Child Health Insurance Research Initiative Principal Investigator in Kansas) and Karen VanLandeghem. Thanks are also due to Cindy Mann and Steve Berman, who conscientiously commented on an earlier draft of this article.
|
FOOTNOTES |
|---|
Reprint requests to AHRQ Publications Clearinghouse, P.O. Box 8547, Silver Spring, MD 20907-8547. E-mail: ahrqpubs{at}ahrq.gov
This article represents the views of the authors and not necessarily those of the funding agencies.
This study was approved by the human investigations boards of each institution.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S. A. Fisher-Owens, J. C. Barker, S. Adams, L. H. Chung, S. A. Gansky, S. Hyde, and J. A. Weintraub Giving Policy Some Teeth: Routes To Reducing Disparities In Oral Health Health Aff., March 1, 2008; 27(2): 404 - 412. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
