Role of Partnerships: Second Annual Meeting of Child Health Services R
Adolescent Health Care and Health Services Research
Adolescent Health Care and Health Services Research
Elizabeth Shenkman, Ph.D.
Gail Slap, M.D.
Charles E. Irwin, Jr. M.D.
Jonathan D. Klein, M.D., M.P.H.
John Santelli, M.D., M.P.H.
Adolescent health care is seriously under-represented in child health services research. The goal of this session was to begin closing this gap by identifying a conceptual framework for thinking about health services research on adolescent health topics, and develop a research agenda:
- The session began with a profile of current adolescent health, an overview of adolescent demographic trends, and a discussion of issues related to access to care for adolescents provided by Charles Irwin, M.D.
- Jonathan Klein, M.D., M.P.H. followed with a discussion of issues in measuring content of care and validity and reliability of physician- reports versus adolescent self-reports.
- John Santelli, M.D., M.P.H. completed the session with a discussion of the progress achieved since the 1998 regulation regarding the inclusion of children in research, the increasing attention to safety in research, and issues related to rights and limits regarding confidentiality and IRB processes.
Profile and Trends in Adolescent Health
Adolescence represents a broad age range with significant changes in biological, psychological, and social systems. Dr. Irwin noted that in recent years there has been a shift in the way that adolescents are regarded. The long-term implications of behaviors begun in adolescence are beginning to be recognized in the context of how they relate to adult health status. Many major adult health problems have their origins in adolescence and often the negative outcomes that result from behaviors begun in adolescence have their onset in adulthood (such as cardiovascular disease, mental health disabilities, and habituation).
The adolescent population is also becoming the focus of increased research partly because of the increase in its size. In 1999 there were 39.4 million adolescents aged 10-19 in the United States. This number will increase by more than 2 million by the year 2005, at which time it will level off briefly before increasing again. The demographics of this population are also changing. The proportion of adolescents who are very poor is increasing, as well as the proportion who are raised in a single-parent household.
Although the media may lead the public to believe otherwise, indicated Dr. Irwin, overall mortality among adolescents is on the downswing. Death from homicide, motor vehicle accidents, and suicide are all decreasing. Adolescent males account for the majority of adolescent deaths, and minority groups (particularly African-American, Hispanic and Native American) also have disproportionately high death rates. Risky behaviors (substance abuse, unsafe sexual activity and violence, etc.) are responsible for more than 70 percent of adolescent morbidity/mortality.
Dr. Irwin pointed out an interesting juxtaposition in the changing adolescent demographics: although mortality is in a downward trend, adolescent participation in risky behaviors has not decreased for 20 years and nearly one in three adolescents engage in multiple risky behaviors. While the majority of the population (80 percent) report very good/excellent health, one in five have a mental/behavioral health disorder and 10 percent have a chronic disease/disability that limits daily functioning.
Despite the size of the population and their unique health needs, adolescents do not account for many medical visits and they have no group of providers who specifically care for them. Therefore, unlike adults and young children, adolescents tend to enter the health care system from multiple entry points and use a variety of care systems, often in fragmented ways. Barriers to utilization of health care among adolescents include lack of transportation, shortages of clinicians and cultural issues. The two most significant barriers that keep adolescents from seeking care are:
- Concern about confidentiality.
- Lack of health insurance coverage.
Currently, one in seven adolescents lacks health insurance coverage and older adolescents are the most likely population to be uninsured. This statistic has remained relatively constant in the last decade despite experiments such as Medicaid expansions, which are targeted at getting adolescents into care. There is a sense of uncertainty, Dr. Irwin indicated, over whether SCHIP will be able to change this. Although the proportion of insured adolescents has not changed, the type of coverage they have has changed: private coverage has decreased (from 74 to 69 percent) and public coverage has increased (from 10 percent to 15 percent). Dr. Irwin indicated that insurance trends such as higher employee contributions for dependents, lower family-income eligibility cut-offs for Medicaid, and overall increasing costs of insurance have accounted for this shift from private to public coverage.
Insurance coverage makes a significant difference in whether adolescents get into a physician's office. The average adolescent visits a physician's office nearly three times per year, while uninsured adolescents visit a physician's office just 1.5 times per year. This difference, however, disappears once a child has insurance coverage. Dr. Irwin cited a study conducted by Keane and associates (1999), which found that differences in unmet health care needs, as measured at baseline, nearly disappear in children of all age categories (6-10; 11-14; and 15-19 years old) when measured again one year after health insurance is obtained.
Even when adolescents visit a physician's office, it is unclear whether their health needs are being adequately met. Physicians spend approximately 18 minutes per visit with adolescents and there is not much information available on what goes on during that time. The current adolescent health guidelines have been developed based on the scarce data available, taking into account the cost of preventive services. Yet there remains a lack of solid data on what providers can accomplish during a visit, how effectively they can identify high risk behaviors, and what interventions are effective in the office visit setting.
There are differing views regarding adolescent development that affect how health care to this population is provided and measured. Those that emphasize the immaturity of the adolescent favor more restricted approaches to care that may ultimately cause decreased access to care by advocating for more parental involvement and less adolescent decision-making, while others who over-emphasize independence and competence of this developmental period may ultimately push toward a system that is set up to address adult health care issues. The balance needed to effectively care for the unique needs of the adolescent population is difficult to achieve, but with increasing scientific information on how the adolescent brain develops and functions as well as new funding streams created by programs such as SCHIP, new opportunities are being created to positively impact and improve the health status of adolescents.
The greatest adolescent risk factors for morbidity and mortality, as identified in the Youth Risk Behavior Surveillance System (YRBSS) are related to:
- Mental health.
- Tobacco, alcohol and other drugs.
- Nutrition and exercise.
Dr. Klein pointed out that most health problems related to these factors are preventable. Regarding access to and use of health care, most adolescents report having a usual source of care and having used some preventive care. Although many adolescents who have had a clinical visit think that they have had a "well-care" or "preventive visit," this has not been confirmed in the research. This discrepancy, Dr. Klein asserts, raises concerns about how we measure what happens in a clinical visit and what must be included in a "preventive-care" visit according to current guidelines.
To assess the quality of adolescent care, accurate surveillance tools are needed. Current surveillance methods for health behaviors rely primarily on adolescent self-reporting and current methods for health services rely primarily on parent-reporting, chart review, or administrative data. Dr. Klein and his colleagues have been involved in research to develop measures for public health surveillance, managed care quality assessment, and future research to measure the receipt of adolescent preventive services by assessing adolescent self-report validity and reliability and physician-report validity.
One recent study assessed the validity of the adolescent self-report. Using a convenience sample of 15 community-based primary care practices with 14-21 year-olds being seen for preventive care, the prevalence of physicians' discussions of preventive care topics was assessed by adolescents' self-reports and compared with audiotapes of the visit and chart reviews. Adolescent patients were interviewed either "early" (two to four weeks following their visit), or "late" (five to seven months following the physician visit). The preventive care topics were coded according to the 33 areas recommended by the Guidelines for Adolescent Preventive Service. The preventive topic screening tools included specific discussion items under larger topic headings such as tobacco, alcohol and drugs, sexuality and STDs, and psychosocial issues and health maintenance behaviors.
Adolescent self-report in the late interview group had reasonable sensitivity, and specificity and reliability was comparable to the early interview group, which suggests that reliability of the self-report is somewhat robust over time. In general, however, the specificity of the self-report was not good when compared with the audiotape. In some cases, particularly in relation to specific sensitive topics such as sexual orientation, the audiotapes under-represented physician discussion when compared to the chart and self-reports. When a topic had a trigger question in the chart, a greater proportion of physicians reported discussing that topic, whereas a smaller proportion of physicians using patient charts without the trigger question reported discussion of the topic.
The findings suggested that chart reviews may over-report service delivery and that despite being audiotaped, many providers do not deliver recommended services. The most salient conclusions drawn from the study were that:
- Surveying adolescents about their own care is feasible.
- Adolescent self-report is a valid method for assessing the content and quality of some preventive care delivered to youth.
Dr. Klein also discussed the findings of his recent study to assess the test-retest reliability of adolescents' self-report about their receipt of preventive health screening/counseling services and their health behaviors. A written survey similar to the CDC's Youth Risk Behavior Survey (YRBS) was administered to students of one New York high school (grades 9-12). The survey assessed recall/behaviors for a range of time periods from "lifetime" to "at last visit to doctor" to "current," and agreement between the time one (T1) and the time two (T2) surveys was assessed. Neither age nor ethnicity significantly impacted reliability and gender had only weak effects. Sentence complexity and time frame were strongly associated with reliability.
However, an important limitation of the study, Dr. Klein noted, was that it did not test multiple ways to ask about specific items. The conclusions he drew were that adolescents are most reliable in reporting lifetime or current behaviors; they demonstrate a reasonably high reliability in self-reporting survey items of most health behaviors, counseling services and utilization; and sentence complexity and time frame of recall have the greatest effect on the reliability of adolescent self-reports.
Dr. Klein then discussed the use of physician surveys regarding adolescent health care. He indicated that surveying physicians is cost effective but may be unreliable and inaccurate. In one recent study he assessed the validity of provider self-reporting by surveying a random sample of primary care pediatricians and family physicians and comparing the results to the charts of their 14-19 year-old adolescent patients. The findings of the study indicated that physicians were most likely to over-report discussing topics such as birth control, sexual activity and cigarette use, while under-reporting discussions of chewing tobacco, sexual orientation, and people smoking at home. Overall, physicians who used screening questionnaires had lower overall disagreement with adolescent patients than those who did not use screening tools.
The conclusions of the study were that physician self-report of services provided are not particularly accurate and that adolescents' self-report should be the standard used in health services research regarding what services are provided to youth. Dr. Klein also suggested that providing physicians with screening instruments may be helpful in improving quality of services provided to adolescents. It was suggested that having trigger questions in the patient chart could help improve quality of care by encouraging physicians to include preventive care discussions in the patient visit. He also indicated that other young adult health surveys have only found consistent adolescent and physician agreement about care regarding highly salient events, but physicians who used "trigger" mechanisms (such as screening instruments) had higher rates of agreement with their adolescent patients. Dr. Klein concluded his discussion by emphasizing that the bottom-line and most critical implication of these studies is that researchers cannot begin to improve care for adolescents if there is not an accurate understanding of what is currently being delivered.
The inclusion of children and adolescents in research has been at the forefront of ethical debates in recent years, and is being reflected in policy decisions such as the 1998 National Institutes of Health (NIH) policy on inclusion of children. Dr. Santelli noted that this debate has been propelled by the evolution in biological and legal understanding of children, which has resulted in the development of laws regarding children's rights to treatment and privacy.
Changes in the understanding of research risks and benefits have also contributed to greater inclusion of adolescents in research. Historically, research has been viewed as inherently dangerous and the history of research abuse in the United States has helped to maintain that view. In recent years, however, that view has been changing: research is increasingly seen as beneficial and there is a growing awareness that if a population is not included in research they do not benefit from it.
With that in mind, in 1999 the NIH established a policy regarding the inclusion of children in research so that adequate data will be developed to support the treatment modalities for disorders and conditions that may affect children. The NIH policy was developed because "medical treatments applied to children are often based upon testing done only in adults and scientifically evaluated treatments are less available to children due to barriers in their inclusion in research studies." Although there is a growing body of guidelines on the inclusion of children, researchers and Institutional Review Boards (IRBs) continue to have difficulty with the appropriate inclusion of adolescents.
Dr. Santelli discussed how the basic ethical principles of respect for persons, beneficence and justice are driving the current notion of inclusion in research. Federal law mandates basic requirements for human subject protection and special protections for vulnerable populations such as children. The Federal regulations on research with children mandate the following:
- Permission of the parent/guardian.
- Assent (affirmative agreement to participate) of the child/adolescent.
- Alternative mechanisms for protection when parental permission is not advisable.
- Hierarchy of risk.
Most research for the adolescent population is minimal risk, defined by the CDC as having an anticipated probability and magnitude of harm that is not greater than those risks ordinarily encountered in daily life or during the performance of routine examinations or tests.
Consent is often a central concern in obtaining adolescent participation in research. Capacity and confidentiality are two key issues in the consent debate. "Capacity" to provide informed consent is dependent upon cognitive abilities/maturity and life experience. Dr. Santelli indicated that adolescents over age 13 have cognitive capacities similar to adults, while younger adolescents (ages 9-13) can understand risk/benefits and make decisions similar to adults, but they are less able to consider multiple conflicting points.
Confidentiality is a key issue because of the requirement of parental permission by most IRBs. Provisions have been made in the Federal regulations to circumvent this barrier. Waiver of parental permission is allowed for conditions in which adolescents may legally receive treatment without parental consent, for mature minors, and when no more than minimal risk is involved, and in cases in which the child's parents are legally or functionally incompetent or other circumstances in which it is not "reasonable" to get parental permission. Factors which determine whether it is "reasonable" to waive parental permission include:
- Risks of requiring permission.
- Burden of obtaining permission.
- Level of risk/benefit of study.
- Developmental capacity.
- Ethical principles.
Despite the guidelines, mandates, and provisions established at both the Federal and professional association levels, children and adolescents are still systematically excluded from research at major research institutions throughout the United States. The issue of inclusion of children and adolescents and its essential role in improving the health of this population is still widely debated and continues to pose significant challenges for investigators and IRBs across the nation.
Discussion among participants identified the concern that although some aspects of adolescent health and research on adolescent health have changed and evolved over time, a conceptual model for adolescent research is still lacking. Following the panelists' presentations, participants discussed whether a model for adolescent research is needed, given the information presented in today's session.
Much of the knowledge about adolescents that currently exists has been gleaned from large surveys or "fishing expeditions." The large sample sizes of these studies allow for easy identification of differences, but make it difficult to determine why those differences exist. IRBs have historically been critical of studies which include sensitive items that are trying to collect data on the "why" of trends identified. One participant pointed out that in order to "move to the next level" in knowledge about adolescent health and begin to uncover why mortality rates are changing, a conceptual framework with specific hypotheses is needed. Other participants agreed that much knowledge about adolescents' health care, including where they receive care and what kind of care they receive from physicians, is still limited.
As the panelists responded to participants' comments, they made a key closing observation: in today's health system, while quality of care improvements can be made at the provider level, standards of care and accountability are increasingly being set at the health plan level. Therefore, in order to truly impact and improve the quality of health care for adolescents, knowledge from adolescent health research must be translated into recommendations made at the plan level, not just targeted toward changing physician practices.
Current as of June 2000
Adolescent Health Care and Health Services Research. Role of Partnerships: Second Annual Meeting of Child Health Services Researchers. June 27, 2000. Agency for Healthcare Research and Quality, Rockville, MD. http://www.ahrq.gov/research/chsr2ado.htm