HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) All Versions of this Article: AJPH.2004.059758v1 96/4/709    most recent Submit a response Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (3) Citing Articles via Google Scholar Google Scholar Articles by Johnson, T. J. Articles by Johnson, W. G. Search for Related Content PubMed PubMed Citation Articles by Johnson, T. J. Articles by Johnson, W. G. Related Collections Health Care Facilities/Services Insurance Access to Care Other Child and Adolescent Health Hispanics/Latinos Socioeconomic Factors

Tricia J. Johnson, PhD, Mary Rimsza, MD and William G. Johnson, PhD

Tricia J. Johnson is with the Department of Health Systems Management, Rush University, Chicago, Ill. Mary Rimsza and William G. Johnson are with the School of Health Management and Policy, W. P. Carey School of Business, Arizona State University, Tempe, Ariz.

Correspondence: Requests for reprints should be sent to Tricia J. Johnson, PhD, Dept of Health Systems Management, Rush University, TOB Suite 126B, Chicago, IL 60612 (e-mail: tricia_j_johnson{at}rush.edu).

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION References

 

Objectives. Many states are increasing the State Children’s Health Insurance Program (SCHIP) cost-sharing requirements to induce reductions in enrollment. We examined the effect of increasing SCHIP premiums on both health care use and cost to the public.

Methods. The net cost to the public of increased cost sharing for SCHIP-insured children in a border community was estimated with multivariate methods. The majority (88%) of children were of Mexican origin.

Results. We estimated that a $10 increase in monthly premiums would induce 10% of SCHIP children to disenroll, resulting in a 6% increase in public expenditures.

Conclusions. Families that disenroll from SCHIP and become uninsured typically turn to emergency departments for primary care, which increases total health care expenditures through the use of more expensive services.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION References

 
Federal rules have historically limited a family’s total cost sharing (i.e., premiums, enrollment fees, deductibles, coinsurance, copayments, and other fees) of State Children’s Health Insurance Program (SCHIP) expenditures to no more than 5% of the family’s income.¹ A weakening economy has increased the number of families that are eligible for Medicaid and SCHIP. At the same time, tax revenues have decreased. During fiscal year 2004, all 50 states either implemented or planned cost containment measures: 21 states planned to increase or add copayments, 18 states planned to restrict eligibility, and 17 states planned to reduce benefits.² Additionally, 6 states have frozen SCHIP enrollment, which has left uninsured children who qualify for SCHIP without insurance indefinitely.³

Enrollment in SCHIP dropped for the first time in the 6-year history of the program during the second half of 2003.⁴ Increased cost sharing makes medical services less affordable for low-income families and causes some beneficiaries to drop insurance coverage. An analysis of Washington’s Basic Health Plan, Minnesota’s Minnesota Care, and Hawaii’s Quest programs found that program participation declined from 57% to 18% as premiums rose from 1% to 5% of family income.⁵ Results from these 3 projects suggested that adding a premium to SCHIP that costs only 1% of family income for a family of 4 living at the federal poverty level would decrease enrollment by 16%.

Texas recently experienced a 29% reduction in SCHIP enrollment less than 1 year after (1) increasing premiums, (2) adding a 90-day waiting period for benefits, and (3) reducing the enrollment period from 12 months to 6 months.⁶ Reducing the number of children enrolled in SCHIP and reducing the program’s share of expenditures will reduce SCHIP expenditures. Reductions in expenditures by SCHIP, however, do not mean that public expenditures for the health care of affected children will be reduced. It is well known that children who lack health insurance are more likely than insured children to use emergency departments for primary care. It is reasonable to expect that children who lose SCHIP coverage and thus join the ranks of the uninsured will, on average, follow the health care use patterns of uninsured children. The majority of costs associated with uncompensated care are paid by the federal and state governments through Medicare and Medicaid (i.e., disproportionate share hospital adjustments and indirect medical education payments) and other federal programs (e.g., community health centers and the Maternal and Child Health Bureau). Federal and state funds have been estimated to cover 87% of the total costs of uncompensated care.⁷

We simulated the effect of increased cost sharing on both the use of health care and the cost to the federal and state governments (and therefore the taxpayers) of health care for SCHIP-insured children affected by an increase in the monthly SCHIP premium. Additionally, the added financial burden for these families may require them to seek additional public subsidies (e.g., food stamps) to pay for the premiums. Our simulation combined the estimated coefficients of multivariate models of uninsured children’s emergency department visits, physician visits, and inpatient hospitalizations with the characteristics of a set of children who were likely to be disenrolled from the program because of a modest ($10) increase in the costs of SCHIP. Our estimate of a $10 increase in costs is purposely conservative. Changes such as those that have occurred in Texas—a state with a number of border communities—would obviously have a much greater impact.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION References

 
We followed the model of access to medical care adopted by Aday et al. to distinguish the effects of health insurance coverage from other influences on emergency department use, inpatient hospitalizations, and physician services for nontraumatic illnesses and injuries (Figure 1).⁸ The model includes (1) characteristics that predispose children to use health care (e.g., age, gender, and race/ethnicity), (2) factors that enable children to obtain health care (e.g., availability of care from primary care providers, including office hours and transportation time), and (3) factors that represent the need for health services (e.g., nature of illness).

View larger version (15K): [in this window] [in a new window]   FIGURE 1—— Conceptual model of health care utilization. Note. ED = emergency department. Solid lines represent factors included in the study; the dotted line represents factors not included in the study; arrow represents causality.

The YHQ includes every child who used the emergency department or inpatient care or received care from a YHQ data partner or was insured by Medicaid or was immunized. The YHQ is a census of children aged 0 to 4 years and is an extremely large sample of children aged 5 to 19 years. The sample is not, however, necessarily representative of all the county’s children; therefore, an analysis weight was calculated for each child to compensate for potential underrepresentation.

Poststratification adjustments forced the distribution of the YHQ children to represent the distribution of children in Yuma County across age (0–4 years, 5–9 years, 10–14 years, and 15–19 years), gender, and race/ethnicity (African American, Hispanic, Native American, and White) in the 2000 US Census. The weights were generally quite small, which showed that the unweighted data were very close to a representative sample. The majority of the county’s children are of Mexican origin, and a small but significant number of Native American children also live in the county. The YHQ permits a uniquely detailed analysis of these vulnerable populations, whose access to health care is a national issue. Because of the unusually large, detailed, and comprehensive community database on children’s health and health care use, we used more sophisticated analytic techniques than have been used in previous studies of health insurance coverage and children’s use of health care services.^9–12

In 1998, the state of Arizona introduced SCHIP as a separate child health insurance program for children who were not eligible for Medicaid. Our study used YHQ data for 2001 to allow for 2 full years of SCHIP operation before the year of analysis and to allow the families time to gain experience with SCHIP enrollment and use processes. The analysis was conducted on the basis of a weighted sample of 41646 children who had health care episodes during 2001: 2734 (7%) were uninsured at some time during 2001 (1043 traumatic cases and 1691 non-traumatic cases), 21366 (51%) were insured by Medicaid, 4353 (10%) were insured by SCHIP, and the remaining 13193 children (32%) were covered by other sources of insurance (Figure 2).

View larger version (12K): [in this window] [in a new window]   FIGURE 2—— Breakdown of children, in the Yuma HealthQuery, by condition and insurance status. Note. SCHIP = State Children’s Health Insurance Program. Shaded boxes represent children included in the study.

In general, children’s use of an emergency department for primary care is influenced by travel time. To control for this potential effect, we included variables for the zip code of children’s homes. The emergency department, hospital, and largest pediatric practices are located in the city of Yuma,¹⁴ which is the largest community in the county, and 52% of the county’s children lived in the city during 2000. A single residence code for children who lived in Somerton or San Luis was used, because the availability of health care in the 2 communities was similar and they are both located on the same local road to Yuma. Both communities are served by community health centers.

Statistical Analysis
We used a system of 3 logistic regression equations to estimate the probability of emergency department visits, inpatient hospitalizations, and physician visits, and we used a system of 3 nonlinear regression equations to estimate the quantity of services for each of the 3 types of services. Each system of equations was estimated separately for both SCHIP-insured and uninsured children. We estimated, therefore, a total of 12 equations to examine service use among these groups of children. Predisposing characteristics included age (0–4 years, 5–9 years, 10–14 years, and 15–19 years), gender, and race/ethnicity (White non-Hispanic; Hispanic; Asian; Native American; and other). Enabling characteristics included any care from a pediatrician and zip code location of residence to measure the effect of distance on access to health care. The need for care was represented by the presence of at least 1 ambulatory care sensitive condition, i.e., a condition that can be treated in a primary care setting if timely care is provided, but if left untreated, may result in 1 or more hospitalizations that could have been avoided. Examples include asthma, gastroenteritis, and kidney/urinary tract infections.¹⁵ Children who had an ambulatory care sensitive condition were twice as likely to have visited an emergency department¹⁶ and to have substantially larger hospital charges compared with children who did not have an ambulatory care sensitive condition.¹⁷

Our objective was to predict changes in health care use for nontraumatic medical care when SCHIP-insured children were disenrolled because of increases in cost sharing. Differences in use among uninsured and SCHIP-insured children were influenced by differences in demographic characteristics (e.g., because 1 group of children was older, the group was less likely to visit the emergency department) and by differences in insurance coverage (e.g., because uninsured children pay 100% of a physician’s fees for routine services, they are more likely to delay routine care). We used the Oaxaca decomposition—modified to fit health care comparisons—to separate differences in use between SCHIP-insured and uninsured children into differences caused by children’s characteristics and differences caused by insurance.^18–23 The Oaxaca decomposition separates the between-group difference in the dependent variable into the difference caused by observable characteristics (i.e., the portion of the difference explained by differences in the mean characteristics included in the model) and unobserved factors (i.e., the portion of the difference caused by differences in the coefficients between the 2 groups).

The final step of the analysis simulated the effect on the county’s health care delivery system of adding a $10 monthly premium increase to SCHIP coverage. On the basis of responses to cost-sharing increases in other states, we conservatively assumed that adding a $10 monthly premium increase would induce 10% of the SCHIP-insured population to drop coverage and thus become uninsured.^5,24 (In Arizona, a $10 increase changed the monthly premium to $25 for a family with 2 or more children.) A child who loses SCHIP coverage because of this increase could theoretically enroll in private insurance; however, this was unlikely if a $10 increase in the monthly premium was enough to dissuade a family from reenrolling in SCHIP. A study from the Kaiser Family Foundation found that the monthly premium for employer–provided health insurance was $222 per month for family coverage in 2004 ($149 in 2001), which is substantially more than the $25 monthly premium for SCHIP coverage.²⁵

Findings from recent studies of Medicaid also supported our assumption. For example, results from focus groups of adult Medicaid respondents in Oregon showed that all of those who lost coverage because of premium increases became uninsured. Other respondents who had incomes up to 170% of the federal poverty level stated that they could not afford private health insurance coverage without premium assistance from the state Medicaid program.²⁶ On the basis of Current Population Survey data, 54% of children who disenrolled from Medicaid between 1998 and 2001 became uninsured, despite being eligible for coverage.²⁷ While the cost of care incurred during spells without insurance may help a family previously enrolled in SCHIP to qualify for Medicaid in the future, Medicaid eligibility is not retroactive. Thus, the family’s ability to reenroll because of high medical debt as a result of hospitalizations or other health care bills does not reduce the cost of hospitalization. Therefore, we assumed that children who disenrolled from SCHIP would remain uninsured for the year.

To estimate the effect of the increase in cost sharing, 10% of the SCHIP-insured children were randomly selected and added to the pool of uninsured children. The pool of uninsured children increased from 1691 to 2050. The probabilities of use and the types and quantities of health care services consumed were reestimated for the new uninsured group of children by applying the previously estimated coefficients for the uninsured children to the average characteristics that now reflected the influx of previously insured SCHIP children.

The net changes in use among SCHIP-insured and uninsured children were calculated by multiplying the probabilities of use and quantities of services by the numbers of children in each group before and after the simulated premium change. The product was the net effect on the number of children who used services and quantities of services. For example, if there were 100 uninsured children who had 1.1 emergency department visits on average per year, and if there were 100 SCHIP-insured children who had 0.5 emergency department visits on average per year, the total number of emergency department visits would have been 160. After a $10 premium increase that resulted in 10% of the SCHIP-insured children dropping coverage, there would have been 110 uninsured children who had 1.1 emergency department visits on average per year and 90 SCHIP-insured children who had 0.5 emergency department visits on average per year. The total number of emergency department visits would have been 166, and the net increase in use would have been 6 emergency department visits.

Total health care expenditures were calculated on the basis of actual 2001 payments by Medicaid and SCHIP for physician visits, inpatient hospitalizations, and emergency department visits. The average amount paid by Medicaid and SCHIP for each service was multiplied by the quantity of each service before and after the simulated premium change to estimate the aggregate change in health care expenditures. For example, if the average SCHIP payment per emergency department visit was $225, total expenditures for emergency department visits would have been $36 000 before the $10 premium increase and $37 350 after the premium increase. The net increase in spending would have been $1350.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION References

 
We used estimates from the regression models to examine the effect of reductions in insurance coverage for children on the use and costs of emergency department visits, inpatient hospitalizations, and physician visits for nontraumatic health conditions. The emphasis on SCHIP reflects the assumption that children who disenroll from SCHIP will be uninsured.

Differences in Use Between Uninsured and SCHIP-Insured Children
Uninsured children were more than 4 times more likely to use emergency department services for nontraumatic medical care compared with SCHIP-insured children (Table 1). The probability of visiting the emergency department at least once during 2001 was 54% among uninsured children and 12% among SCHIP-insured children. Uninsured children also were 10 times more likely to use inpatient hospital care. The probability of experiencing at least 1 inpatient hospitalization was 23% among uninsured children and only 2% among SCHIP-insured children. Conversely, uninsured children were less than half as likely to use physician services. The probability of at least 1 physician visit was 36% among uninsured children and 80% among SCHIP-insured children. Among children who used some form of health care, uninsured children had substantially more emergency department visits and inpatient hospitalizations (0.53 more emergency department visits and 0.21 more inpatient hospitalizations) and had 1.4 fewer physician visits on average (2.0 visits among uninsured children vs 3.4 among SCHIP-insured children) (Table 2).

A smaller proportion of the difference in the number of inpatient hospitalizations was caused by the insurance effect (67%), and 33% of the difference was caused by differences in the children’s characteristics (Table 2). As children are disenrolled from SCHIP after a premium increase, the percentage change in emergency department visits is expected to be larger than the percentage change in inpatient hospitalizations, because the insurance effect was larger for emergency department visits. The 1.4 difference in physician visits was completely explained by the insurance effect. If the uninsured and SCHIP-insured children’s characteristics were identical, SCHIP-insured children would have had 1.9 fewer physician visits, but differences in children’s characteristics somewhat mitigated the insurance effect. These results show that the number of emergency department and physician visits by children who were disenrolled from SCHIP changed more dramatically than inpatient hospitalizations, because a larger portion of the difference in inpatient hospitalization use was attributed to differences in the children’s characteristics.

Total Health Care Expenditures
We estimated that adding a $10 monthly premium increase to SCHIP coverage would increase the number of uninsured by 21%, or 359 children (from 1691 to 2050 children). The addition of 359 children to the uninsured would have added 159 emergency department visits and 54 inpatient hospitalizations during 2001, and the number of physician visits would have decreased by 654 visits (Table 3). These changes would have increased total health care expenditures for SCHIP-insured and uninsured children by $167000, or a 6% increase in the $2940000 currently spent on health care for these 2 groups of children in 2001 (Table 4). This amounts to $464 per child who drops SCHIP coverage and becomes uninsured.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION References

Our data are from a semirural border community in Arizona; thus, our estimates are not nationally representative. Our results do, nevertheless, provide insights into the effects of cost sharing on a large number of children who are of Mexican origin and whose access to care is a question of national interest. Additionally, the increase in health care costs caused by disenrollment may be the responsibility of the state in some instances, and, in other communities, it may be the responsibility of private or public health systems, depending on state SCHIP policies about reenrollment and other factors.

While demand-side cost sharing is seen as a way to reduce unnecessary services, it also reduces necessary service use by low-income individuals and thus results in delaying needed care and increased severity of potentially avoidable illnesses. The increase in severity can ultimately translate into large and significant increases in both the quantity and the intensity of services. As our study shows, a $10 increase in the monthly premium for SCHIP coverage in Yuma County, Ariz, would have increased the number of uninsured children by 21% and resulted in an overall increase of $167000 in total health care expenditures because of a shift from physician visits to more expensive emergency department visits and inpatient hospitalizations. The apparent savings from increased cost sharing are, therefore, illusory.

	Acknowledgments

 
This research was supported by the Flinn Foundation and the Arizona State University Office of Research. Helpful comments were provided by Richard Butler, Brigham Young University.

Human Participant Protection
This study was approved by the institutional review board of Arizona State University.

	Footnotes

 
Peer Reviewed

Contributors
M. Rimsza and W.G. Johnson originated the study. T.J. Johnson was responsible for statistical analysis. All the authors originated ideas, interpreted results, and edited drafts of the article.

Accepted for publication June 12, 2005.

	References

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION References

 
1. State Child Health: Implementing Regulations for the State Children’s Health Insurance Program, Final Rule. Federal Register. 2001;66(Pt 431, 433, 435, 436 and 437):2490–2538.[Medline]

2. Smith VK, Ramesh R, Gifford K, Ellis E, Wachino V, O’Malley M. States Respond to Fiscal Pressure: A 50-State Update of State Medicaid Spending Growth and Cost Containment Actions. Washington, DC: Kaiser Commission on Medicaid and the Uninsured; 2004. Publication No. 7001.

3. Ross DC, Cox L. Preserving Recent Progress on Health Coverage for Children and Families: New Tensions Emerge. A 50 State Update on Eligibility, Enrollment, Renewal and Cost-Sharing Practices in Medicaid and SCHIP. Washington, DC: Kaiser Commission on Medicaid and the Uninsured; 2003. Publication No. 4125.

4. Smith VK, Rousseau DM, O’Malley M. SCHIP Program Enrollment: December 2003 Update. Washington, DC: Kaiser Commission on Medicaid and the Uninsured; 2004. Publication No. 7134.

5. Ku L, Coughlin TA. Sliding scale premium health insurance programs: four state experiences. Inquiry. 1999–00;36:471–480.

6. Dunkelberg A, O’Malley M. Children’s Medicaid and SCHIP in Texas: Tracking the Impact of Budget Cuts. Washington, DC: Kaiser Commission on Medicaid and the Uninsured; 2004. Publication No. 7132.

7. Hadley J, Holahan J. Who Pays and How Much? The Cost of Caring for the Uninsured. Washington, DC: Kaiser Commission on Medicaid and the Uninsured; 2003. Publication No. 4088.

8. Aday L, Andersen R, Fleming G. Health Care in the US: Equitable for Whom? Beverly Hills, Calif: Sage Publications; 1980.

9. Halfon N, Newacheck PW, Wood DL, St. Peter RF. Routine emergency department use for sick care by children in the United States. Pediatrics. 1996;98: 28–34.[Abstract/Free Full Text]

10. Petersen LA, Burstin HR, O’Neil AC, Orav EJ, Brennan TA. Nonurgent emergency department visits: the effect of having a regular doctor. Med Care. 1998; 36:1249–1255.[CrossRef][ISI][Medline]

11. Phelps K, Taylor C, Kimmel S, Nagel R, Klein W, Puczynski S. Factors associated with emergency department utilization for nonurgent pediatric problems. Arch Fam Med. 2000;9:1086–1092.[Abstract/Free Full Text]

12. Sharma V, Simon SD, Backwell JM, Ellerbeck EF, Fox MH, Wallace DD. Factors influencing infant visits to emergency departments. Pediatrics. 2000;106: 1031–1039.[Abstract/Free Full Text]

13. International Classification of Diseases, Ninth Revision. Geneva, Switzerland: World Health Organization; 1980.

14. Arizona Department of Health Services. Yuma community health profile 2000. Available at: http://www.azdhs.gov/hsd/chpweb/2003/places/85540.pdf. Accessed December 23, 2003.

15. Falik M, Needleman J, Wells B, Korb J. Ambulatory care sensitive hospitalizations and emergency visits: experiences of medicaid patients using federally qualified health centers. Med Care. 2001;39:551–561.[CrossRef][ISI][Medline]

16. Rimsza M, Johnson WG, Johnson TJ. Children who use the emergency department as their "medical home." Working paper; 2004.

17. Shi LH, Samuels ME, Pease M, Bailey W, Corley E. Patient characteristics associated with hospitalizations for ambulatory sensitive conditions in South Carolina. South Med J. 1999;92:989–998.[ISI][Medline]

18. Oaxaca RL. Male-female wage differentials in urban labor markets. Int Econ Rev. 1973;14:693–709.[CrossRef]

19. Cotton J. On the decomposition of wage differentials. Rev Econ Stat. 1988;70:236–243.[CrossRef]

20. Johnson WG, Baldwin ML, Burton JF, Jr. Why is the treatment of work-related injuries so costly? New evidence from California. Inquiry. 1996;33:53–65.[ISI][Medline]

21. Baldwin ML, Butler RJ, Johnson WG. A hierarchical theory of occupational segregation and wage discrimination. Econ Inq. 2001;39:94–110.[CrossRef][ISI]

22. Baldwin ML, Johnson WG. Labor market discrimination against men with disabilities. J Hum Resour. 1994; 29:1–19.

23. Baldwin ML, Johnson WG. Labor market discrimination against women with disabilities. Ind Relations. 1995;34:555–577.

24. Madden C, Cheadle A, Diehr P, Martin D, Patrick D, Skillman S. Voluntary public insurance for low-income families: the decision to enroll. J Health Polit Policy Law. 1995;20:955–972.[ISI][Medline]

25. Kaiser Family Foundation Health Research and Educational Trust. Employer Health Benefits 2004 Annual Survey. Menlo Park, Calif: Kaiser Family Foundation; 2004.

26. LeCouteur G, Perry M, Artiga S, Rousseau DM. The Impact of Medicaid Reductions in Oregon: Focus Group Insights. Washington, DC: Kaiser Commission on Medicaid and the Uninsured; 2004. Publication No. 7233.

27. Sommers BD. From Medicaid to uninsured: dropout among children in public insurance programs. Health Serv Res. 2005;40:59–78.[CrossRef][ISI][Medline]

28. Dept of Health and Human Services, Office of the Secretary. Annual update of the HHS poverty guidelines. Federal Register. 2003;68:6456–6458.

29. Hudman J, O’Malley M. Health Insurance Premiums and Cost-Sharing: Findings from the Research on Low-Income Populations. Washington, DC: Kaiser Commission on Medicaid and the Uninsured; 2003. Publication No. 4071.

This article has been cited by other articles:

				M. E. Rimsza, R. J. Butler, and W. G. Johnson Impact of Medicaid Disenrollment on Health Care Use and Cost Pediatrics, May 1, 2007; 119(5): e1026 - e1032. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) All Versions of this Article: AJPH.2004.059758v1 96/4/709    most recent Submit a response Purchase Article View Shopping Cart Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (3) Citing Articles via Google Scholar Google Scholar Articles by Johnson, T. J. Articles by Johnson, W. G. Search for Related Content PubMed PubMed Citation Articles by Johnson, T. J. Articles by Johnson, W. G. Related Collections Health Care Facilities/Services Insurance Access to Care Other Child and Adolescent Health Hispanics/Latinos Socioeconomic Factors