CHLAMYDIA TRACHOMATIS INFECTION is the most commonly reported bacterial sexually transmitted infection in the United States,1 and prevalence is highest in young people.2 Most genital chlamydial infections are asymptomatic,3 and identification of infected individuals often relies on testing. The Centers for Disease Control and Prevention (CDC) and the U.S. Preventive Services Task Force (USPSTF) each currently recommends annual chlamydial screening for all sexually active women 25 years of age and younger as well as for older, at-risk women (e.g., with new or multiple sexual partners).4,5 The importance of annual chlamydial screening for young women is further emphasized in that such screening is an element in the Health Plan Employer Data and Information Set (HEDIS) measures widely used for comparing quality of healthcare provision.6 The benefits of chlamydial screening have been demonstrated in settings in which screening programs for women have reduced both the prevalence of infection7,8 and pelvic inflammatory disease (PID) rates.9,10
Despite screening recommendations and the availability of sensitive urine-based chlamydial diagnostic tests, which can simplify screening, C. trachomatis infections remain common. In a recent population-based study of young adults living in the United States, wave III of The National Longitudinal Survey of Adolescent Health (Add Health), the overall prevalence of chlamydial infection was more than 4%.11 There are likely many factors contributing to the sustained high chlamydial prevalence, and one such potential contributing factor is insufficient chlamydial screening12 perhaps either because persons for whom screening is recommended are not seeing providers or because providers are not testing them.
Among the many barriers to chlamydial screening in the U.S. population may be access to health care, which is, in turn, influenced by factors such as health insurance coverage and health care-seeking behaviors (e.g., primary site of health care and frequency of provider visits). In 2002, more than 43 million Americans, disproportionately including young adults, were without health insurance throughout the year, and millions more were uninsured for shorter periods.13 In particular, uninsured young people more often lack a usual source of health care, have unmet health needs, and go without provider contact during the course of a year.14 Moreover, health insurance coverage and health care-seeking behaviors may vary by gender and race/ethnicity, and select groups of individuals may have a disadvantage with regard to access to healthcare services. Although healthcare access does not assure that recommendations for chlamydial screening and management will be followed, access to care may be a component of efforts to accomplish these goals. Even among the insured, chlamydia screening efforts may be hampered by individuals not seeking provider care or receiving their health care at settings where chlamydial testing is not performed routinely.
To date, there have been no population-based studies that examined issues of healthcare access and chlamydial infection. The purpose of this study is to begin to investigate associations between health insurance status and health care-seeking behaviors on the likelihood of chlamydial infection in a national sample of young adults. For the first time, biomarker data are available as part of data from a representative national sample providing national prevalence estimates of chlamydial infection for specific racial and ethnic groups, separately by gender, as well as prevalence estimates by markers of healthcare access. We use multivariate techniques to estimate the effects of health insurance and health care-seeking behaviors on the odds of chlamydial infection net of age and race/ethnicity. Because we propose that the effects of healthcare access and race/ethnicity on the odds of infection differ for men and women, all analyses are conducted separately by gender.
Materials and Methods
Study Design and Study Population
The data used for this analysis are from wave III of the National Longitudinal Study of Adolescent Health (Add Health). Add Health is a longitudinal survey of youth attending U.S. schools in grades 7 through 12 during the 1994–1995 academic year. The Add Health study design has been described elsewhere (http://www.cpc.unc.edu/projects/addhealth/design.html). Wave I of the survey was collected in 1994–1995, wave II was collected in 1996, and wave III was collected in 2001–2002 when the respondents were ages 18 to 27 (i.e., young adults) and was comprised of 75.6% of the original eligible respondents interviewed in wave I. For wave III, 15,170 wave I respondents and 27 wave II respondents were located and reinterviewed (N = 15,197).
For our study, we used a subset of wave III respondents. Because of our interest in population-based estimates, we excluded respondents for whom wave III sampling weights were unavailable (n = 875). We also excluded those: 1) without chlamydia test results available (n = 1777); 2) who denied ever having had vaginal intercourse (n = 1617); 3) who reported antibiotic use in the prior 30 days (n = 1493); and 4) with missing information on health care-seeking behaviors (n = 88). This resulted in a final analytic sample of 9347 respondents (4473 men and 4874 women).
Interviews and Specimen Collection
Methods for interviews and specimen collection have been previously described in detail (The Add Health Biomarker Team, 2003), and we briefly summarize relevant aspects here.11 After obtaining consent, interviewers conducted 90-minute, private interviews with each participant and entered responses directly into computers. Participants also used computer-assisted self-interview (CASI). After the interview and obtaining additional consent, interviewers instructed participants on urine collection techniques and a first-void urine sample was collected. Urine specimens were shipped on ice packs to arrive at the testing laboratory within 96 hours of collection. C. trachomatis ligase chain reaction (LCR) assays (Abbott Laboratories, Abbott Park, IL) were performed on the urine specimens according to the manufacturer’s instructions.
Of the 1777 respondents in wave III for whom valid chlamydial results were unavailable, 647 had no LCR assay results recorded and 1130 refused to provide urine specimens. We conducted additional analyses, which revealed no systematic differences between those who had valid urine LCR results compared with those who provided a urine specimen but for whom there were “no results.” There were, however, systematic differences between those who provided a urine sample and those who refused (e.g., refusers were significantly more likely to be older, single and not sexually active, female, college graduates, or to have missing information on age at first intercourse; data not shown). More comprehensive sensitivity analyses conducted by Miller and colleagues assessing the impact of nonresponse and test performance found that the substantive conclusions were robust regardless of the assumptions for nonresponders or test performance.11 We therefore present estimates unadjusted for nonresponders.
The outcome is C. trachomatis infection coded as a binary response (1 = positive test result and 0 = negative test result). The primary independent variables are health insurance coverage and health care-seeking behaviors. Respondents’ health insurance coverage was created based on the question, “Over the past 12 months, how many months did you have health insurance?” Conceptually, we are interested in examining whether continuity of health insurance coverage is associated with chlamydial infection because continued coverage could potentially provide greater opportunities for screening. We imputed 75 cases of missing data (“refused,” “don’t know,” or “not applicable” responses) using an ordinary least-squares regression that included age, race/ethnicity, gender, highest level of education completed, enrollment status, and union status.15 Responses were then categorized into no health insurance in the past year (0 months), insurance for 1 to 11 months, and insurance for 12 continuous months with no health insurance being the reference group. Because preliminary analyses found no difference by insurance product type, we combined all types into a single insurance variable in subsequent analyses.
The two health care-seeking behaviors examined are usual site of health care and time since last visit with a medical provider. For usual site of health care, respondents were asked, “Where do you usually go when you are sick or need health care?” Response categories included: never get sick or need health care; hospital-based clinic; hospital emergency room; community health center or clinic; health maintenance organization (HMO); private doctor’s office; school or college clinic; military hospital or clinic; clinic at work; or some other place. The categories of HMO, private doctor’s office, and clinic at work were combined because we felt that they represented primary care (PC) settings. The primary care setting category was the reference group. Time since last visit with a medical provider was based on information from the question, “How long ago did you last consult a doctor or a nurse?” We created a dummy variable measuring whether the respondent had consulted a doctor or nurse in the last 12 months.
Sociodemographic characteristics serving as control variables in our analyses include respondent’s gender, age, and race/ethnicity. Gender is a dichotomous variable, coded 1 for males and 0 for females. Age is measured in years and is included as a continuous variable ranging from 18 to 27 years. Respondents were asked whether they were of Hispanic or Latino origin. They were then asked their race and could select more than one option: white, black or black, American Indian or Native American, and Asian or Pacific Islander. Those who mentioned more than one race category were asked, “Which one category best describes your racial background?” Respondents who provided unusable information were coded to the interviewer’s perception of the respondent’s race. From these questions, a race/ethnicity composite measure was created giving priority to any mention of Hispanic or Latino origin. This resulted in the following race/ethnicity categories: non-Hispanic white; non-Hispanic black; Hispanic; non-Hispanic Asian; and non-Hispanic Native American. Non-Hispanic whites were the reference group.
All analyses are conducted using wave III sampling weights and account for the complex survey design of Add Health using appropriate procedures (-svy- commands) available in Stata 8.216; the weighting technique adjusts for sampling procedure (e.g., addressing oversampling of certain demographics) and for nonresponse. All analyses are performed separately by gender. First, we use current study population characteristics. Then, chlamydial prevalence for each race/ethnic group is estimated separately by gender. We also examine racial/ethnic differences in health insurance status, usual site of care, and time since last provider visit. Chi-squared tests, corrected for survey design, are used for the biviariate analysis.17 Last, we investigate the associations between each of these covariates and chlamydia test results separately by gender, first unadjusted and then adjusted for age and race/ethnicity using weighted logistic regression. A P value of ≤0.05 is considered statistically significant.
Study Population and Prevalence of Chlamydial Infection
Of the 9347 Add Health wave III participants analyzed, 48.6% are women and the mean age is 21.9 years (standard deviation, 1.8 years). The majority are white (67.0%), followed by black (17.3%), Hispanic (11.6%), Asian (3.2%), and Native American (0.8%) (Table 1). Overall, 4.7% of young adults tested positive for chlamydia with lower percentages observed for men (4.1%) compared with women (5.4%) (P <0.05, Table 1). There are no significant differences by age within this population of young adults (data not shown).
Overall, chlamydial infection is highest among Native Americans (13.3%) followed by blacks (12.9%), Hispanics (6.9%), whites (2.3%), and Asians (1.8%) (P <0.001, Table 2). In young men, chlamydial infection is highest among blacks (11.5%) followed by Native Americans (10.3%), Hispanics (8.0%), whites (1.6%), and Asians (0.9%) (P <0.001). In young women, chlamydial infection is highest among Native American women (16.9%) followed by blacks (14.3%), Hispanics (5.6%), Asians (3.1%), and whites (3.0%) (P <0.001).
Health Insurance Coverage and Chlamydial Infection
Overall, 19.7% of participants were without health insurance during the 12 months preceding the interview and a greater percentage of men were without insurance compared with women (22.5% vs. 16.7%) (P <0.001, Table 3). Health insurance coverage differs significantly by race/ethnicity for both men and women (P <0.001 and P <0.01, respectively, Table 3). Among men, Hispanics have the lowest and Asians have the highest levels of coverage. Among women, whites and Asians have higher levels of coverage than blacks, Hispanics, or Native Americans.
Health insurance coverage is significantly associated with chlamydial prevalence among men with the lowest prevalence seen among men with continuous coverage (Table 4). Compared with men with no health insurance, the risk of chlamydial infection is significantly lower for men who were insured 1 to 11 months (P <0.05) or for 12 months (P <0.01). When we control for age and race/ethnicity, only men with continuous coverage have a significantly lower risk of infection than men with no coverage (P <0.05). Compared with women with no health insurance, women with continuous coverage have a significantly lower risk of infection (P <0.01); however, this difference is not statistically significant controlling for age and race/ethnicity (P <0.1) (Table 4).
Site of Usual Health Care and Chlamydial Infection
Overall, the most common sites of usual health care are primary care settings (47.1%) followed in descending frequency by hospital clinics, emergency rooms, community health centers, school clinics, and all other sites (Table 5). Approximately 7.0% of participants felt they did not need usual health care and denied having a site of usual health care. The reported sites of usual health care differ for men and women (P <0.001, Table 5). Compared with men, women more often sought care from primary care settings and less often in emergency rooms. Men also more often felt they did not need usual health care than did women. The sites of usual health care differ significantly by race/ethnicity for both men and women (P <0.001 for both, Table 5). White and Asian men more often sought services at primary care settings than black, Hispanic, and Native American men; and black men more often sought care at emergency rooms than all other racial/ethnic groups (P <0.001). White and Hispanic women more often sought care at primary care settings than black, Asian, and Native American women; and black and Native American women more often sought care at emergency rooms than the other racial/ethnic groups (P <0.001).
Site of usual health care is associated with differences in chlamydial prevalence rates for men (Table 6). Among men, prevalence rates are highest for those using emergency room and other settings and lowest for those using community center clinics and primary care settings (P <0.001). Compared with men reporting usual care from primary care settings, the risk of chlamydial infection is higher for men who seek care in emergency rooms (P <0.001) or at some other place (P <0.05). When adjusting for age and race/ethnicity, men who seek care in the emergency room still have a higher risk of infection than those using primary care settings, but the odds ratio is reduced; in contrast, the odds ratio for other settings showed little change across the models. Site of usual care is also associated with variation of chlamydial prevalence rates among women (Table 6); prevalence is highest for women using emergency rooms and other settings and lowest for those using school clinics. The unadjusted regression results show that compared with women who sought services from primary care settings, those who use emergency rooms have a higher risk of infection and those who use school clinics have a lower risk of infection (P <0.01, respectively). When adjusting for age and race/ethnicity, only women who used school clinics have a significantly lower risk of infection compared with women who used primary care settings (P <0.01).
Time Since Last Provider Visit and Chlamydial Infection
Overall, 78.8% of study participants reported a healthcare visit in the preceding 12 months with women being more likely to report a visit than men (89.8% vs. 68.5%; P <0.001; Table 5). Although the majority of all racial/ethnic groups among both men and women reported a provider visit in the past 12 months, the differences are significant for both genders (P <0.05 for men, P <0.001 for women; Table 5). Hispanic and Asian men and women are less likely to have reported a recent provider visit compared with black, white, and Native American men and women.
Time since last provider visit is associated with differences in chlamydial infection in men but not women (Table 6). Men with provider visits in the preceding 12 months have a significantly lower risk of chlamydial infection than those who have not, even when adjusting for age and race/ethnicity (P <0.05).
Occurring in nearly one in 20 sexually active young adults in the United States, genital chlamydial infection prevalence remains unacceptably high.11 As an infection that is most often asymptomatic, C. trachomatis screening and treatment programs are critical elements of current control efforts. Therefore, access to health care (and thus screening opportunities) may well be among factors that affect chlamydial prevalence. To our knowledge, our study is the first to assess the relationship of indicators of healthcare access to variation in chlamydial infection among a nationally representative sample of sexually active young adults. We demonstrated that insurance status, site of usual health care, and having seen a provider in the last 12 months were each associated with variation in chlamydial infection independent of gender, age, and race/ethnicity, factors previously shown to be associated with variation in health care-seeking behavior.14,18–22
One interpretation of our findings is that young adults who have health insurance and seek usual care in primary care settings may be more likely to receive chlamydial screening and treatment, and consequently, the chlamydial infection is lower in this population. Alternatively, these healthcare access traits may be markers for populations who engage in overall healthier behaviors (including sexual behaviors) and for membership in lower-risk sexual networks or perhaps who have certain socioeconomic traits such as a higher income, which increases the likelihood they will have health insurance and access to health providers to get screened for chlamydia. The goal of our study was to examine some of the factors associated with chlamydial infection with a focus on healthcare access, a major potential barrier to chlamydial screening. However, we acknowledge there are likely other barriers to chlamydial screening (e.g., providers' willingness to screen and/or exclusion of screening in standard care procedures) and other factors influencing risk of chlamydial infection (e.g., sexual behaviors) that are important and deserve further study.
Women were insured more often than men; however, we were somewhat surprised to find that more than 75% of young men and women had health insurance for some amount of time during the year preceding the survey. Over the past decade, health insurance rates for young people may have increased as a result of the enactment of the State Children's Health Insurance Program in 1997 and Medicaid eligibility expansions, each providing mechanisms for adolescents and young adults to obtain insurance.23 The importance of health insurance expansion for the reproductive health of young people is highlighted in our finding of a significantly lower risk of chlamydial infection among insured individuals with these associations persisting even after controlling for race/ethnicity and age (more significant in men) (our study also confirmed prior reports of lower insurance rates for most minorities as compared with whites24). Furthermore, being insured continuously for 12 months was associated with even a lower risk of chlamydial infection when compared with not having insurance or being insured only part of the year. Many potential factors may contribute to our finding of lower chlamydial infection among the insured. Perhaps insured individuals are screened more often for chlamydia, are more often exposed to risk-reduction interventions, have greater access to care if they develop genitourinary symptoms, seek care from different healthcare provider types, or are more likely to be prescribed chlamydia-effective antibiotics at provider visits. Any or all or these explanations could contribute to this finding and are appropriate for further study.
Having health insurance addresses one potential barrier to chlamydial screening, yet does not ensure young men and women will see providers for screening. There are sparse data on health care-seeking behaviors in young adults (18–27 years of age in wave III); however, in adolescents, it has been reported that their overall use of healthcare services is lower than for other age groups25 but that insured adolescents more often seek health care than uninsured.14 Using data from over 14,000 adolescents included in the 1995 National Health Interview Survey, Newacheck and colleagues reported uninsured adolescents were twice as likely to go without a physician visit during the course of a year.14 Most Add Health wave III participants, who were initially enrolled as adolescents in wave I and are now young adults, reported a provider visit in the preceding 12 months, possibly reflecting in part that most had health insurance. Racial/ethnic and gender disparities in time since last healthcare visit were noted for both men and women. Hispanics and Asians less often sought providers during the year. Women were more likely than men to have visited healthcare providers during the year, possibly related to visits for reproductive health (e.g., contraception, Pap smears, and so on). However, the risk of chlamydial infection was significantly lower for men, but not women, who reported having had a provider visit during the preceding year. This difference may reflect gender differences in reasons for seeking care as well as practitioner differences in screening based on gender, because national screening guidelines emphasize women. Women may tend to seek care more often for health maintenance. In contrast, recent healthcare provider visits for men may be more often problem-related (i.e., presenting with symptoms) and thus possibly be more likely to lead to antimicrobial therapy.
Chlamydial infection also varied by self-reported usual site of health care. Prevalence rates were higher among participants who reported usually using emergency rooms and other acute care settings than among those using more health maintenance-oriented care in primary care settings. Preventive healthcare activities such as chlamydial screening of asymptomatic persons is often beyond the usual scope of care provided in acute care settings where care is limited by time constraints and limited follow-up capacity.26–28 Men were more likely to report using emergency rooms for usual health care than women, an observation perhaps also potentially related to our finding that men also more often felt they did not need usual health care (and presumably would only be seen for acute problems such as genital symptoms). Interestingly, we also demonstrated that women seeking care in school clinics had even a lower risk of chlamydial infection than those seen in primary care settings, perhaps as a result of the provision of sexual health education and chlamydial screening or because of underlying differences in risk-related behaviors among women using these settings compared with those who use other settings.29,30
It is worth noting that health insurance status and health care-seeking behaviors are interrelated, an issue we investigated through a series of subanalyses (data available on request). Months of health insurance coverage in young men and women was positively associated with receipt of health care at primary care settings and negatively correlated with health care obtained at more acute care settings. This is a plausible finding given that persons seen in primary care settings would be expected to have more stable insurance coverage, access preventive services, including chlamydial screening, and to receive treatment for genital symptoms or a positive chlamydia test. An association between months of insurance and receipt of care in primary care settings has also been reported in adolescents. Newacheck and colleagues reported that uninsured adolescents were five times as likely as the insured to lack a usual source of care and, when receiving care, less often did so in physicians' offices or HMOs and more often in emergency rooms or urgent care settings.14 Another example of the interrelationships among health care-seeking behaviors in our study is that those who sought usual care at a primary care setting rather than an acute care setting also more often saw a healthcare provider in the last 12 months, which is also plausible because those with a primary care provider would probably be more likely to have routine yearly visits.
Our study has both strengths and limitations. The use of a nationally representative sample is a major strength of our study. This improves the generalizability and use of our results in considering the association between healthcare access and chlamydial prevalence in “real-world” clinical settings independent of differences in race/ethnicity, socioeconomic status, and risk behaviors. Furthermore, there are no other published data available, to our knowledge, that have evaluated the influence of healthcare access on chlamydial infection in a nationally representative sample; hence, our study provides information that significantly contributes to this important area of study and can serve as a basis for further investigation. Finally, the sensitivity of the chlamydial diagnostic assay used, chlamydial LCR, improves the accuracy of detecting chlamydia.
It is important, however, to acknowledge study limitations not already discussed. First, the population enrolled may have missed several important subpopulations within the targeted age group. Although representative of youth in grades 7 to 12 in 1994–1995, young persons of these ages not enrolled in school may well have had different (higher) chlamydial prevalence and patterns of health care-seeking behaviors. Similarly, the construction of this multipurpose survey focused on heterosexual behaviors and thus makes assessment of chlamydial prevalence among males who only had sex with other males difficult to ascertain. Nonetheless, a large proportion of the nation's young people at the time the study was initiated are represented here. Second, we did not have detailed information on reasons for visits over the 12-month period of interest or specific details on the healthcare interactions that occurred (e.g., whether screening occurred and treatment was prescribed) during each visit. Third, data were not available on how frequently chlamydial screening was being routinely performed in the different healthcare settings we studied. Fourth, because this national study was a multipurpose study, not all healthcare access measures are as comprehensive as would have been useful for these analyses. For example, we excluded persons reporting antibiotic use in the prior 30 days, yet the type of antibiotic prescribed and the treatment indication was unknown. The majority of antibiotics prescribed in outpatients are for respiratory tract infections, and some of the antibiotics commonly prescribed (e.g., macrolides, fluoroquinolones, and tetracyclines) will have antichlamydial activity and may treat a chlamydial infection that was never diagnosed. This is why we excluded recent antibiotic use (if we included these treated subjects in our analysis, we may underestimate the overall chlamydial prevalence because of the unanticipated treatment effect), yet it is possible persons with insurance may more likely receive antibiotics and therefore have a lower chlamydial prevalence, and our exclusion of these individuals could bias our estimates of chlamydial prevalence. Finally, in our analyses, we treated race/ethnicity as a confounding variable. However, it seems reasonable that the effects of site of care, for example, on chlamydia infection, might depend on race/ethnicity; this is a question for future research. Although these limitations may hinder our ability to state with confidence that there is a causal link between healthcare access and chlamydial infection, we are able to still infer from our study findings that there is a relationship with the measures of healthcare access we studied and chlamydial infection, and thus provide some of the first published data on how measures of healthcare access may influence chlamydial infection.
In summary, our study demonstrates that health insurance and health care-seeking behaviors, both measures of healthcare access, are associated with chlamydial infection in a nationally representative sample of young adults. From our study findings, we postulate that efforts to further expand health insurance for young people may help to reduce C. trachomatis-associated reproductive health morbidity likely through improved provider access for chlamydial screening, diagnosis, treatment, and sex partner services. In addition, we believe at the same time that greater efforts are needed in educating young adults and their providers about the asymptomatic nature of chlamydia and the need for patients to see their providers for a yearly routine visit where chlamydia screening can occur. Because chlamydial screening is still not occurring in many settings routinely, more measures to encourage providers to follow current recommendations for screening are also likely to contribute to chlamydial control.
1. Cates W. Estimates of the incidence and prevalence of sexually transmitted diseases in the United States. Sex Transm Dis 1999; 26:S2–7.
2. Centers for Disease Control and Prevention. Sexually Transmitted Disease Surveillance 2002 Supplement, Chlamydia Prevalence Monitoring Project. Atlanta: US Department of Health and Human Services, Centers for Disease Control and Prevention, October 2003.
3. Stamm WE. Chlamydia trachomatis
—The persistent pathogen: Thomas Parran Award Lecture. Sex Transm Dis 2001; 28:684–689.
4. Centers for Disease Control and Prevention. Sexually transmitted disease treatment guidelines 2002. MMWR Morb Mortal Wkly Rep 2002; 51:1–80.
5. US Preventive Services Task Force. Screening for chlamydial infection: Recommendations and rationale. Am J Prev Med 2001; 20(suppl):90–94.
7. Mertz KJ, Levine WC, Mosure DJ, Berman SM, Dorian KJ, Hadgu A. Trends in the prevalence of chlamydial infections—The impact of community-wide testing. Sex Transm Dis 1997; 24:169–175.
8. Centers for Disease Control and Prevention. Chlamydia trachomatis
genital infections United States, 1995. MMWR Morb Mortal Wkly Rep 1997; 46:193–198.
9. Scholes D, Stergachis A, Heidrch FE, Andrilla H, Holmes KK, Stamm WE. Prevention of pelvic inflammatory disease by screening for cervical chlamydial infection. N Engl J Med 1996; 334:1362–1366.
10. Kamwendo F, Forslin L, Bodin L, Danielsson D. Decreasing incidences of gonorrhea- and chlamydia-associated acute pelvic inflammatory disease: A 25-year study from an urban area of central Sweden. Sex Transm Dis 1996; 23:384–391.
11. Miller WC, Ford CA, Morris M, et al. Prevalence of chlamydial and gonococcal infections among young adults in the United States. JAMA 2004; 291:2229–2236.
12. Centers for Disease Control and Prevention. Chlamydia screening among sexually active young female enrollees of health plans—United States, 1999–2001. MMWR Morb Mortal Wkly Rep 2004; 53:983–985.
14. Newacheck PW, Brindis CD, Cart CU, Marchi K, Irwin CE. Adolescent health insurance coverage: Recent changes and access to care. Pediatrics 1999; 104:195–202.
15. Paul C, Mason WM, McCaffrey D, Fox SA. What should we do about missing data? (A case study using logistic regression with missing data on a single covariate.) California Center for Population Research On-Line Working Paper Series, 2003.
16. Stata Statistical Software: Release 8.0. College Station, TX: Stata Corp, 2003.
17. Rao JNK, Scott AJ. On chi square tests for multiway contingency tables with cell proportions estimated from survey data. Ann Stat 1984; 12:46–60.
19. Liao Y, Tucker P, Okoro CA, Giles WH, Mokdad AH, Harris VB; Division of Adult and Community Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention (CDC). REACH 2010 Surveillance for Health Status in Minority Communities—United States, 2001–2002. MMWR Surveill Summ 2004; 53:1–36.
20. Shenkman E, Youngblade L, Nackashi J. Adolescents' preventive care experiences before entry into the State Children's Health Insurance Program (SCHIP). Pediatrics 2003; 112:e533.
21. Shone LP, Dick AW, Brach C, et al. The role of race and ethnicity in the State Children's Health Insurance Program (SCHIP) in four states: Are there baseline disparities, and what do they mean for SCHIP? Pediatrics 2003; 112:e521.
23. American Academy of Pediatrics, Section on Adolescent Health. Improving the implementation of State Children's Health Insurance Programs for Adolescents. Pediatrics 2000; 105:906–912.
24. Johnson RL, Saha S, Arbelaez JJ, Beach MC, Cooper LA. Racial and ethnic differences in patient perceptions of bias and cultural competence in health care. J Gen Intern Med 2004; 19:101–110.
25. Klein JD. Adolescents, the health care delivery system, and health care reform. In: Irwin CE, Brindis C, Holt K, Langlykke K, eds. Health Care Reform: Opportunities for Improving Adolescent Health. Arlington, VA: National Adolescent Health Information Center, National Center for Education for Maternal and Child Health, 1994:17–28.
26. Todd CS, Haase C, Stoner BP. Emergency department screening for asymptomatic sexually transmitted infections. Am J Public Health 2001; 91:461–464.
27. Yealy DM, Greene TJ, Hobbs GD. Underrecognition of cervical Neisseria gonorrhoeae
and Chlamydia trachomatis
infections in the emergency department. Acad Emerg Med 1997; 4:962–967.
28. Kuhn GJ, Campbell A, Merline J, O'Neil BJ. Diagnosis and follow-up of Chlamydia trachomatis
infections in the ED. Am J Emerg Med 1998; 16:157–159.
29. Kirby D. Effective approaches to reducing adolescent unprotected sex, pregnancy, and childbearing. J Sex Res 2002; 39:51–57.
30. Kirby D, Short L, Collins J, et al. School-based programs to reduce sexual risk behaviors: a review of effectiveness. Public Health Rep 1994; 109:339–360.