DIRECTIONS FOR RESEARCH
Awful as it was, I was given a gift of experience. If I can use it [in research] to help someone else, it makes [my daughter’s] life mean something still.
Bereaved parent (Contro, 2002, p.15)
Among the most common phrases in this report are “research is limited” and “systematic data are not available.” Clinicians and parents must often make decisions about the care of children with little guidance from clinical or health services research that documents the potential burdens as well as the potential benefits of medical interventions. For example, parents of infants born more than 14 weeks prematurely are faced with urgent decisions about the extent of life support that they wish for their infant. They frequently must confront unanswerable questions and make their best guesses about what to do with the help and support of neonatolgists and nurses who must often, in turn, rely on their own experience and judgment with limited scientific knowledge to guide them. Neonatologists generally lack validated predictors of very premature infants’ risk of death or long-term morbidity, although both outcomes are frequent.
The knowledge base for organizational and policy decisions is likewise limited. For example, organizations considering the creation of a pediatric palliative care program have little research on which to base decisions about services, staffing, outreach, budgets, and similar matters. What information is available relies heavily on descriptive case studies of adult programs. Funding for comparative health services research to test different approaches to organizing adult or pediatric palliative care is minuscule
compared to the funding available for clinical research. Even when hospices and hospitals undertake data-intensive internal studies (e.g., for quality improvement projects), they often lack the resources and motivation to produce analyses that meet peer-review standards for publication and wider dissemination.
Researchers, clinicians, and policymakers have recognized shortfalls in clinical research involving palliative and end-of-life care for children and have taken steps to encourage and guide such research, some of which are described later in this chapter. The remainder of this chapter
briefly reviews initiatives to encourage pediatric research in general and in palliative and end-of-life care specifically;
describes directions for clinical, health services, and educational research to guide improvements in palliative, end-of-life, and bereavement care for children and their families;
summarizes the practical challenges of undertaking pediatric research; and
reviews ethical and legal questions raised by research involving children.
Lessons learned from the kinds of research recommended here should help inform and improve the care of children who survive as well as children who die. It should likewise help all families who experience a child’s serious or fatal medical problem.
INITIATIVES TO ENCOURAGE PEDIATRIC RESEARCH
Federal policymakers have taken steps to encourage certain kinds of pediatric research. In 1998, following directions from Congress in 1995, the National Institutes of Health (NIH) issued policies and guidelines for including children as research participants (NIH, 1998). Under the policy, children are to be included in all such research funded by the NIH unless their exclusion is justified on scientific or ethical grounds. Exclusion would, for example, be justified when a medical problem does not affect children. For some medical conditions, children’s developmental characteristics might suggest the need for a separate, child-only study.
Legislative conference language accompanying the legislation encouraged but did not require the NIH to establish pediatric research priorities (NIH, 1998). The organization has not developed an overall set of priorities, although some individual institutes have developed priorities for certain clinical problems or services including kidney disease (NIDDK, 2001),
HIV/AIDS (NIH, 2001a), and emergency medical services (NIH, 2001b). The priority-setting activity related to emergency services was stimulated in part by recognition that more needed to be done to implement the research recommendations in the 1993 Institute of Medicine (IOM) report on emergency medical services for children (IOM, 1993).
In 2000, the Children’s Health Act (P.L. 106-310), among its other provisions, created a pediatric research initiative to increase NIH support for research on diseases, disorders, and other conditions in children. The legislation also supported the training of more pediatric researchers to conduct basic and clinical research. Earlier, the Food and Drug Administration Modernization Act of 1997 (P.L. 105-115) provided incentives for pharmaceutical companies to test drugs in children, and Congress renewed that legislation with slight modifications late in 2001 (Best Pharmaceuticals for Children Act, P.L.107-109). Companies that undertake studies on their products’ effects on children get six months of exclusive marketing rights for the drugs in return.1 (A later section of this chapter describes provisions in this and other legislation and regulation related to protection of child participants in such studies.) In addition to the pediatric exclusivity provision, the Food and Drug Administration (FDA) issued regulations in 1998 that allow it to require drug companies to undertake pediatric testing for certain drugs likely to be used in children (63 Fed. Reg. 66632, December 2, 1998, effective April 1, 1999).2
In May 2001, the FDA published an updated list of priorities for pediatric drug research that included more than 425 drugs or drug uses (FDA,
Federal regulations requiring pediatric studies of certain drug and biologic products were proposed in 1997 and issued in final form in 1998 (http://www.fda.gov/ohrms/dockets/98fr/120298c.txt). In 2002, the government announced that it would suspend these rules for two years pending study of the need for them (Connolly, 2002); a month later, it reversed that decision (Kaufman and Connolly, 2002; Landa, 2002). Unlike the legislation, the regulations also cover biologics (e.g., vaccines, blood products, gene therapy products, HIV and hepatitis tests, innovative therapies for diseases such as cancer and arthritis).
In a set of questions and answers about the rule, the FDA has stated (emphasis in the original deleted): “Under the rule, FDA has the authority to require pediatric studies on a drug product for the product’s approved indications if there is substantial use in the pediatric population or the product would provide a meaningful therapeutic benefit—and the absence of adequate labeling could pose significant risk (see 21 CFR 201.23(a)). At this time, however, FDA will not require studies of approved drugs except if approved by the Center Director. Instead, FDA will seek to have manufacturers voluntarily submit studies for marketed drugs under the incentives provided by [the legislation] (see 63 FR 66634 Sec II). For those drugs in which voluntary measures fail to obtain necessary pediatric studies, FDA will consider requiring studies” (http://www.fda.gov/cder/pediatric/faqs.htm#the“Rule”; last update: March 8, 2001).
2001b). Of these, about two dozen uses might relate to symptoms of life-threatening illnesses.
In a report on the pediatric exclusivity provisions, the General Accounting Office noted that when the legislation was passed, approximately 70 to 80 percent of drugs were not adequately labeled for use with children. Since then, pediatric drug research has increased substantially and has provided “new and useful information about whether and how drugs work in children” (USGAO, 2001b, p. 6). According to a recent FDA report to Congress, “the pediatric exclusivity has done more to generate clinical studies and useful prescribing information for the pediatric population than any other regulatory or legislative process to date” (FDA, 2001c, p. ii). The FDA report also noted continuing problems in some areas including the irrelevance of exclusivity provisions to certain old, “off-patent” antibiotics and other drugs and to drugs with low volumes of sales (“orphan drugs”). Another problem has been inadequate incentives for studies in neonates and very young children that must follow studies conducted on older children. In addition, following the completion of studies, manufacturers have sometimes been slow to change drug labeling to reflect the results.
The 2001 legislation renewing “pediatric exclusivity” included provisions to encourage timely changes in drug labeling to reflect new research results. It also established an Office of Pediatric Therapeutics within the FDA to coordinate the agency’s activities related to children and pediatric practice and provided for the NIH-based Foundation for the National Institutes of Health (formerly the Foundation for Biomedical Research) to collect funds to support pediatric drug research.
Pediatric Palliative and End-of-Life Care
Research to support improvements in palliative, end-of-life, and bereavement care for children and their families constitutes only a tiny fraction of research involving children. Likewise, research involving children and their families occupies a small niche in the world of research on palliative and end-of-life care, which itself is small in comparison to other areas of clinical and health services research.
Some units of the National Institutes of Health are supporting or show a potential willingness to support relevant research (NINR et al., 1997). For example, in soliciting research proposals on emergency medical services for children, the National Institute of Child Health and Development (NICHD) identified the need for research on the biobehavioral aspects of pain, stress, and coping with illness or injury in situations of emergency care (NIH, 2000b). The National Institute of Mental Health suggested research on aspects of emergency medical services delivery (e.g., medical staff communication style, follow-up care) that could potentially identify
processes of emergency care that positively or negatively affect the experience of families whose children die.
In addition, as discussed elsewhere in this report and in Appendix H, the Center for Medicare and Medicaid Services (CMS) is funding several demonstration projects that are intended to provide information about the development, operation, effectiveness, and costs of comprehensive programs of palliative care for children and families from the time of diagnosis through bereavement. Although the evaluation designs do not involve controlled comparisons, the evaluations should provide considerable descriptive and analytic detail about the structure of palliative care programs, populations served, and costs.
Foundation-supported studies have been and continue to be particularly important in extending the knowledge base for adult and pediatric palliative, end-of-life, and bereavement care. Two pediatric programs are among 22 projects funded by the Robert Wood Johnson Foundation to test creative strategies for improving palliative care (Promoting Excellence, 2001). One, led by Children’s Hospital and Regional Medical Center of Seattle involves the state health department and regional Blue Cross Blue Shield plans in developing innovative ways to extend health plan coverage for palliative care during a child’s life-threatening illness. (See also Chapter 7.) The other, led by SSM Cardinal Glennon Children’s Hospital in St. Louis, involves support for a statewide network of health care providers ready to care for children at home. The hospital is developing a palliative care consult program to provide 24-hour consultative services, with a palliative care team to educate physicians, hospice providers, and community hospital staff.
Another example of a privately funded initiative is the already-cited project supported by the Nathan Cummings Foundation and the Open Society Institute that is led by the Education Development Center of Boston (Solomon et al., 2001a). It involves eight children’s hospitals (Children’s Hospital/Dana-Farber Cancer Institute, Boston; Children’s Hospital of Philadelphia; Children’s Hospital and Health Center of San Diego; Johns Hopkins Children’s Center; Children’s Mercy Hospital, Kansas City; Lucile Packard Children’s Hospital, Stanford University Medical Center; University of California, San Francisco Children’s Hospital; and Vanderbilt Children’s Hospital). This project has examined clinicians’ and parents’ perspectives on palliative care and is developing quality improvement models and educational materials to support improvements in care for children with a range of life-threatening medical problems and their families.
Another project involving support from multiple private foundations (including Soros and the Charitable Leadership Foundation) is the Pediatric Advanced Illness Coordinated Care (PAICC) initiative that was discussed in Chapter 6. This effort, being led by the Center for Advanced Illness Coor-
dinated Care of Albany, New York, is now developing, evaluating, and refining a standardized model of communication and care coordination in five institutions (Himelstein et al., 2002; Hilden and Tobin, 2002).
DIRECTIONS FOR FUTURE RESEARCH
Throughout its work, the committee has been hampered by the lack of basic descriptive information about death in childhood as well as scant research testing the effectiveness of clinical interventions and organizational processes and structures in providing palliative, end-of-life, and bereavement care that meets the needs of seriously ill or injured children and their families. Studies reviewed by the committee in Chapter 3 describe deficits in care received by children who die and their families. Unfortunately, available research leaves much that is unclear about the extent and causes of shortfalls in care, the number and kinds of children and families who could benefit from palliative and end-of-life care, and the effectiveness of specific strategies to improve the delivery and financing of this care.
Recommendation: The National Center for Health Statistics, the National Institutes of Health, and other relevant public and private organizations, including philanthropic organizations, should collaborate to improve the collection of descriptive data—epidemiological, clinical, organizational, and financial—to guide the provision, funding, and evaluation of palliative, end-of-life, and bereavement care for children and families.
In the 2001 report Improving Palliative Care for Cancer (IOM, 2001c), the IOM’s National Cancer Policy Board included two recommendations aimed at stimulating palliative care research in designated “centers of excellence” and encouraging such centers to take a lead role as agents of national policy in promoting palliative care. The Board also recommended that the National Cancer Institute (NCI) should add the requirement of research in palliative care and symptom control for an institution’s designation as a “comprehensive cancer center.” The research activities suggested for such centers included the following:
formal testing and evaluation of new and existing practice guidelines for palliative and end-of-life care;
pilot testing of “quality indicators” for assessing end-of-life care at the level of the patient and the institution;
uncovering the determinants of disparities in access to care by
minority populations and developing specific programs and initiatives to increase access; and
providing clinical and research training fellowships in medical and surgical oncology in end-of-life care for adult and pediatric patients.
This general strategy should also prove productive in stimulating palliative, end-of-life, and bereavement care research involving children and their families. While by no means discouraging research in other institutions, the committee encourages initiatives that build on federally funded pediatric centers, networks, and similar structures.
Recommendation: Units of the National Institutes of Health and other organizations that fund pediatric oncology, neonatal, and similar clinical and research centers or networks should define priorities for research in pediatric palliative, end-of-life, and bereavement care. Research should focus on care for infants, children, adolescents, and their families, including siblings, and should cover care from the time of diagnosis through death and bereavement. Priorities for research include but are not limited to the effectiveness of
clinical interventions including symptom management;
methods for improving communication and decisionmaking;
innovative arrangements for delivering, coordinating, and evaluating care, including interdisciplinary care teams and quality improvement strategies, and
different approaches to bereavement care.
By organizing multiple sites to investigate a common problem using a common methodology, this strategy should increase the number of children involved in studies and increase the credibility of the findings. It should also stimulate the development of investigator expertise in pediatric palliative care research, encourage the formulation and successful completion of more high-quality research projects, and promote attention to palliative care, end-of-life, and bereavement issues in both pediatric clinical trials and regular patient care. By involving designated institutional participants in collaborative research, a “centers” strategy should also encourage tests of organizational interventions (e.g., random assignment of institutions to test innovations in information systems, training, staffing, or conflict management). The results should benefit children who survive, children who die, and the families of both groups.
The committee recognizes that the infrastructure for biomedical and clinical research varies considerably for different diseases, disorders, and other medical conditions affecting children. For example, the Children’s
Oncology Group (COG), supported by the NCI, involves nearly 240 centers (COG, 2001), whereas the Neonatal Research Network funded by the National Institute of Child Health and Development (NICHD) includes approximately a dozen centers (NICHD, 2000; see also http://neonatal.rti.org). Given differences in the research infrastructure and in the conditions that bring death to children, the specific incentives and mechanisms for a center- or network-focused research strategy will have to be flexible and creative. Although the recommendation focuses on NIH-funded centers and networks, philanthropic foundations and other private organizations can also participate in supporting creative projects that are based on these networks but are not limited to NIH priorities.
Flexibility and creativity on the part of both research sponsors and investigators will also be required to extend research earlier into the period following the diagnosis of a life-threatening problem and later into the period of bereavement experienced by the families of so many children with grave medical problems who participate in clinical research. If researchers focus on earlier stages in the trajectory of fatal and potentially fatal conditions, they may be able to increase the number of children participating in research on certain aspects of palliative care (e.g., effective symptom management). They may also be able to develop a fuller understanding of the symptoms and distress associated with serious illnesses (and their treatments) and the benefits and challenges of incorporating aspects of palliative care earlier in the course of a life-threatening condition.
In proposing directions for research to improve pediatric palliative and end-of-life care, the committee tried to focus on the gaps in the knowledge base needed to put the principles set forth in Chapter 1 into practice and to implement the recommendations listed in subsequent chapters. The discussion below focuses on several specific areas for further research including quality of life, symptom measurement and management, bereavement, edu-cation, and models for delivering and financing palliative, end-of-life, and bereavement care. At a more general level, a comprehensive research agenda for pediatric palliative, end-of-life, and bereavement care should consider
the needs of infants, children, and adolescents and developmentally appropriate care strategies;
a range of causes and trajectories of death including sudden, unexpected deaths, deaths from progressive chronic conditions, and deaths from conditions diagnosed prenatally;
the effects of uncertainty in diagnosis, prognosis, and treatment on communication with children and families, establishment of care goals and care plans, decisions about interventions, preparation for death, and family perspectives and emotions after death;
the needs of parents, siblings, and other family members;
the roles and relationships of different health care professionals and other personnel who are involved with children who may die or who have died and their families;
the range of care settings and organizations that are involved with children who may die or have died and their families (e.g., emergency first-response units, emergency departments, intensive care units, other inpatient units, hospices, home health agencies, and medical examiners’ offices);
the contribution of family, provider, and other factors to timely or delayed recognition that death is near and to differences in family and physician assessments of prognosis and care options;
the reports of children and families about their specific experiences with care (preferably concurrent with care rather than after the child’s death), not just their global assessments of satisfaction with care;
the experiences of children and families outside the health care system, including with schools;
the psychological effects on professionals of caring for children who die and the consequences for their ability to care for children and parents; and
methods and processes for improving communication and preventing or resolving conflicts among clinicians, patients, and family members.
Much of the research suggested in this chapter is descriptive and qualitative. It involves epidemiological, methodological, behavioral, organizational, and policy studies as well as clinical research. Controlled research strategies, including randomized clinical trials and careful case-control studies, should be encouraged. Such approaches will, however, often be difficult given the small numbers of children who die and the charged emotional circumstances surrounding a child’s life threatening illness or injury. Even for relatively narrow clinical questions (e.g., comparison of one pain management regimen versus another), randomized trials are relatively uncommon. Qualitative studies have an important role to play in describing the experiences, perspectives, and values of patients, family members, and caregivers (see, e.g., Sackett and Wennberg, 1997). In devising suitable research strategies, researchers will need to combine creativity, flexibility, and sensitivity both to patient and family burdens and anxieties.
Each research topic suggested here presents different methodological challenges. In general, researchers may encounter problems in defining and recruiting sufficient numbers of research participants, collecting information about subjective experiences from individuals who may be physically or emotionally limited in their ability or willingness to respond to questions, identifying necessary information in medical or other records, and defining measures relevant to patient and family experiences at different stages of life-threatening medical experiences. Epidemiologic and health
services research may have to rely on data that is years old by the time it is released. Researchers may also encounter difficulties in getting research approved by Institutional Review Boards, a topic that warrants investigation in its own right.
Research Directions: Quality of Life for Children and Families
To identify practices that affect the quality of life experienced by a child with a life-threatening medical problem requires measurement tools that can reliably and validly reflect the child’s experience, particularly when the problem has reached an advanced stage and death is expected or possible in the foreseeable future. Appendix C identifies some of the limitations of current measurement tools (especially those intended for well children) and the complexities of developing better ones. Although improvements in measures relevant for pediatric palliative and end-of-life care can build in a general way on measurement strategies used for adults, much of the work needed will have to be specific to infants, children, and adolescents. Box 10.1 summarizes some important directions for such work.
As discussed in Appendix C, existing pediatric quality-of-life instruments may be generic in nature or they may be disease and condition specific. Either way, such instruments are designed primarily for well or chronically ill children and adolescents. They include items that measure function, problems with physical activities, emotional concerns, cognitive abilities to concentrate on and complete school tasks, and concerns with certain symptoms. Many of these items may not be relevant for a child with advanced illness. The instruments may also overlook some issues, for example, spiritual or existential concerns. In addition, completing an instrument intended for healthier children may be unduly burdensome for a seriously ill child.
Some instruments also exist to assess quality of life for parents of children who are seriously ill. These focus primarily on elements of caregiver burden and may only incompletely capture the quality of life of the parent of a dying child. The committee is aware of no instruments that assess quality of life for siblings of seriously ill children.
One task for methodologists is to identify specific domains (e.g., physical, spiritual) of quality of life that are relevant for a child with advanced illness. Another task is to determine whether the importance of a domain changes over time as an illness progresses. Methodologists will also have to consider the influence of development stage, culture, and context (e.g., sites of care, nature of the illness, family circumstances) both in the design of instruments and in their application. For example, an instrument may require different formats for young children and adolescents.
Interventions directed toward improving or at least maintaining a dying child’s quality of life could also improve the quality of life of family members who suffer with the child. Likewise, interventions to help parents, siblings, and other family members could benefit the ill child. Researchers designing interventions to protect and improve the quality of life of children with advanced illness and their families should consider how they might evaluate these “spillover” effects.
When possible, obtaining the dying child’s own report of his or her quality of life is preferred. It may not, however, be possible to identify definitively when a child is dying until illness is far advanced. Also, given the uncertainties associated with predicting time of death, death may come earlier than expected. Thus, some dying children will not be capable of providing self-reports, and the report of a parent or other proxy or surrogate may be necessary. Studies comparing children’s self-reports with reports from parents or professionals tend to show low to moderate agreement in ratings of symptoms or other aspects of quality of life (see Appendix C). Comparisons have not been completed for dying children and their parents or health care providers. Another question may arise when dying children experience changes in health care setting (e.g., home to hospital, distant medical center to community hospital) or health care provider (e.g., pediatric oncologist to generalist pediatrician or hospice professional). How such changes might affect proxy reports, especially those of health care professionals, is unknown.
Comparisons of child self-reports and proxy reports will be challenging in the context of a child’s advanced illness and will require great sensitivity. Nonetheless, it is important to get a better sense of the degree and nature of agreement or disagreement between the two and of the factors that might increase or diminish agreement (e.g., child’s development stage, family culture, location and continuity of end-of-life care). If parents or other proxies tend to overestimate, underestimate, or otherwise misperceive the quality of
life experienced by their children, their reports may misdirect efforts to improve care.
Research Directions: Physical and Psychological Symptoms of Serious Diseases and Their Treatments
In 1997, another IOM committee argued that pain research, which has achieved considerable success in illuminating the pathophysiology of pain and developing effective interventions, should be considered as a model for other areas of symptom research. It urged interaction between basic scientists and clinical investigators to direct new intellectual energy to research on significant end-of-life symptoms including fatigue, shortness of breath, nausea, confusion, anxiety, and depression.
This committee agrees with its predecessor that many features of the pain research strategy can be productively generalized to other symptom areas. These features include the following:
building the neuroscience base for understanding symptoms, for example, continuing to investigate the role of symptom-inhibiting neurotransmitters and receptor-specific opioids that avoid some unwanted opioid side effects;
encouraging the development and use of more precise descriptive terminology and classification schemes for symptoms to provide a “common language” for researchers;
continuing the development, testing, and refinement of standard tools for assessing symptoms including their “felt” burden;
studying the prevalence and severity of symptoms by physiological source or mechanism, medical condition, developmental stage, and other relevant characteristics; and
investigating potential therapies based on better understanding of both the pathophysiology of symptoms and the patient’s reported or observed experience of symptoms.
Research on symptoms has been constrained by numerous factors that have historically limited other pediatric research: small numbers, ethical constraints, complexities associated with outcomes measurement, and lack of financial incentives for the pharmaceutical industry (until recently). The limited research that has been performed has primarily concerned pain management, often in oncology patients only. Problems such as nausea, fatigue, sedation, shivering, anxiety and agitation, depression, seizures, spasticity, constipation and diarrhea, dyspnea, dysphagia, anorexia and cachexia, and dermatologic manifestations of disease and treatment (e.g., itching, mouth ulcers) require attention as well.
Further, although both children and adults have benefited from progress in pain research, many areas for investigation remain in this arena, including identifying the reasons for the underuse of effective therapies to prevent or relieve pain and devising interventions to increase appropriate use of these therapies. The American Academy of Pediatrics (AAP) recently noted the lack of research on the management of pain in newborns with conditions associated with extensive tissue damage or recurrent or chronic pain (e.g., necrotizing enterocolitis, meningitis) (AAP, 2000a). It also noted that the benefits and harms of sedatives and other agents to manage apparent anxiety in neonates have been little studied. Research directions for several areas related to the prevention and relief of pain and other symptoms are summarized in Box 10.2.
BOX 10.2 Directions for Research on Assessment and Management of Symptoms
Prevalence and Intensity of Symptoms in Children and Families
The prevalence and intensity of physical and psychological symptoms in children, especially symptoms other than pain, is inadequately documented. In particular, the extent and severity of psychological symptoms in children with life-threatening illness—and their parents and siblings—are unknown and probably significantly underappreciated. Depression, anxiety, posttraumatic stress disorder (PTSD), and other psychological and behavioral symptoms can significantly reduce the quality of life for children and their families. They can also compromise the ability of children and families to adhere to treatment regimens.
In addition to experiencing symptoms associated with their medical condition, many children with a life-threatening illness face a barrage of painful or uncomfortable procedures (e.g., aspiration of bone marrow or spinal fluid, drawing of blood, injections of diagnostic or therapeutic agents) that may not be necessary for adequate diagnosis and treatment. Painful or distressing procedures, including demanding chemotherapy or mechanical ventilation, often continue when children are nearing death. Stopping certain life-sustaining interventions for a dying child, for example, mechanical ventilation, can involve physical and emotional distress. Research should investigate the prevalence and intensity of treatment-related distress in children with different medical conditions as well as ways to prevent or relieve it.
Measuring symptoms in infants is a particular challenge. As discussed in Chapter 4 and early in this chapter, pain in infants was not recognized or taken seriously until fairly recently. Now, pediatric researchers are attempting to measure pain, stress, and more generally, distress or comfort in these populations using systematic clinical observations and physiological indicators (see, generally, Schecter et al., 1993; Taddio and Ohlson, 1997; Ballantyne et al., 1999; Stevens et al., 2001).
Measurement of Symptoms
Adequate measurement is the cornerstone of adequate treatment. Without the ability to quantify a clinical problem, it is often impossible to determine if interventions have been effective at ameliorating it. Although many developmentally appropriate instruments are available for most pediatric pain management situations, no uniformly accepted instruments exist for certain subpopulations such as children who are developmentally delayed or intubated. In addition, for symptoms other than pain, easy-to-use, reliable assessment tools are scarce. Treatment of fatigue, shortness of breath, drug-related sedation, spasticity, nausea and other more generalized suffering would all benefit from more uniform, quantified, and validated measurement strategies.
Management of Symptoms
In addition to better documenting the prevalence and intensity of psychological symptoms, research is also needed to refine and compare interventions appropriate for children of different ages and at different stages of a life-threatening medical condition. These interventions include individual psychotherapy, group support, and psychotropic medications.
Most of the pharmacologic agents used for physical and psychological symptoms have not been adequately studied in children. Despite recent incentives for drug research involving children, the developmental pharmacokinetic profile in children of many existing drugs is still unknown and must be established to allow and guide their “on-label” use. Then, evidence-based guidelines can be developed for selecting patients for different drug regimens, escalating or tapering off doses, and managing side effects. Although the negative long-term impact of unrelieved symptoms such as pain is well established, the long-term effects of opioids and other drugs used to ameliorate pain have yet to be established for children with serious, chronic problems that persist for months or years.
To reduce children’s discomfort during the administration of medications and to encourage cooperation with uncomfortable treatment regimens, novel routes of drug administration, including transdermal, transmucosal, and regional approaches, must be further developed in addition to better-tasting formulations of existing, orally administered drugs (AAP, 1997a). Development of new agents such as long-acting (e.g., days long) local anesthetics for children experiencing post-operative pain and better treatments for neuropathic and bone pain would also reduce the burden of suffering for many children.
In general, more research is needed both to develop better sedation and analgesic regimens and nonpharmacologic strategies for preventing and relieving procedural pain and to verify the safety and effectiveness of existing regimens. Although not an issue of procedure management per se, the efficacy of intensive, controlled sedation to relieve intractable pain, seizures, or other distress at the end of life also requires continued investigation. Such research is important regardless of one’s position on the use of such sedation when a side effect may be hastened death, although that result appears to be uncommon (see Chapters 5 and 8).
As discussed earlier in this report, pain management programs are now a requirement for hospital accreditation and an increasing focus of quality assessment initiatives (JCAHO, 2001). Careful documentation and analysis of sedation failures can improve the quality of care in an institution and also improve the evidence base overall for effective pain management.
Impact of Symptom Control
Documentation of the impact of symptom control on children and families may promote a wider understanding and appreciation of its benefits among patients, families, and clinicians. Even today, some families may believe that severe pain and other physical distress are inevitable and must simply be endured because nothing can be done about them. Likewise, not all clinicians are aware of shortfalls in the management of pain and other symptoms or not all are knowledgeable about the impact on child and family quality of life of proven strategies for relieving symptoms.
Hospital, hospice, and insurance managers may be influenced if further research supports suggestions that better symptom control may shorten hospital stays and reduce rehospitalizations. Given the high cost of certain medications for pain and other symptoms, such evidence may also encourage rethinking of hospice reimbursements so that the appropriate use of highly effective but costly medications is not discouraged.
What are termed “complementary” or “alternative” therapeutic approaches3 such as acupuncture, massage, therapeutic touch, and aromatherapy may play a significant role in future symptom management. Most of these techniques are noninvasive, and some may be pleasurable for healthy and ill individuals alike. This makes them attractive to children and families facing pharmacological regimens with distressing side effects. The efficacy and acceptability of these techniques for children, especially those with life-threatening illnesses, has not been sufficiently studied to allow routine recommendation of their use. Research in this area could give children and their families, as well as providers, a new array of documented, effective treatment options. Credible research could also alert families to therapies that do not meet the claims made for them.
The benefits of cognitive and behavioral techniques are already well established in the area of pain management (see, e.g., Schechter et al.,
As described by the NIH’s National Center for Complementary and Alternative Medicine (NCCAM), complementary and alternative medicine (CAM) “covers a broad range of healing philosophies, approaches, and therapies. Generally, it is defined as those treatments and health care practices not taught widely in medical schools, not generally used in hospitals, and not usually reimbursed by medical insurance companies. . . . People use these treatments and therapies in a variety of ways. Therapies are used alone (often referred to as alternative), in combination with other alternative therapies, or in addition to conventional therapies (sometimes referred to as complementary)” (http://nccam.nih.gov/fcp/faq/index.html#what-is). As some of these strategies move into mainstream medicine, they may still continue to be described under the CAM label.
2002), but further research is necessary to demonstrate their efficacy in other symptoms such as fatigue, sedation, and nausea. The use of transcutaneous electrical nerve stimulation (TENs) units has received almost no formal research, yet based on anecdotal reports from patients and clinicians, such units may have a significant role to play in symptom control.
Research Directions: Perinatal Death
More children die in the neonatal period or immediately prior to birth than in any other period in childhood. Recent advances in prenatal diagnosis, prenatal treatment, and neonatal care have led to changes in the timing and nature of decisions that families make when confronted with a fetus or newborn infant with a lethal or potentially lethal condition. Little systematic descriptive information is available concerning the long-term impact on families of these decisions, the perceptions of parents of the benefits and harms of different life-sustaining and palliative interventions, or the best methods for communicating with families faced with these decisions. Likewise, little research has investigated the emotional impact of prenatal diagnosis on mothers and fathers, the effect of perinatal death on surviving or future siblings, and differences in effects on family members related to gender, culture, and socio-economic characteristics. More research in this area (Box 10.3) should help improve the care provided to a large group of bereaved families and should also contribute to a better understanding of decisionmaking about treatment choices (i.e., ending the pregnancy or trying to carry the fetus to term) and the consequences of different choices.
Despite the frequency of perinatal loss, population-based studies of the long-term impact of perinatal loss on parents are not available. Descriptive
research to assess parents’ quality of life and perceptions of their role and skills as parents should follow them for at least five years after a perinatal death.
Because of the differing physical and emotional impact of pregnancy on women compared to men, fathers and mothers may respond differently to information concerning fetal or infant diagnosis and prognosis. They may also differ in the rates at which they comprehend or assimilate information. Such differences may amplify already significant emotional stress or preexisting psychological disorders and may lead to confusion about the wishes of parents for interventions. Obstetricians, pediatricians, nurses, social workers, psychologists, and bereavement specialists would benefit from prospective studies that (1) investigate the coping mechanisms of parents who are confronted by an antenatal or postnatal fetal diagnosis that is potentially lethal and (2) explore parental perceptions of the benefits and burdens of different interventions (including palliative care) stratified by fetal or neonatal diagnosis and by timing of diagnosis.
Several perinatal interventions undertaken in the 1970s and 1980s were reported to benefit mothers during the first six months after a perinatal loss (Giles, 1970; Kennell et al., 1970; Parkes, 1980; Forrest et al., 1981, 1982). No data are, however, available on the long-term impact of these interventions, and no randomized trials have assessed the benefits of specific interventions (Chambers and Chan, 2000). Long-term (one to five years), prospective studies of parents who have received different interventions are needed. They should be stratified by fetal or neonatal diagnosis, reproductive history, and socioeconomic status and cultural characteristics. (See also the discussion below of directions for bereavement research.)
In addition to understanding parental experience with lethal or potentially lethal perinatal diagnosis or death, research attention also should extend to siblings subjected to the stress of perinatal loss and to the grief and parental chaos prompted by such loss. The committee located no longitudinal studies of parental perceptions of surviving siblings, self-perceptions of surviving siblings, or interactions between parents and surviving or later-born siblings.
Although the sensitivity and specificity of antenatal ultrasound diagnosis have been carefully defined (see, e.g., Sabbagha et al., 1985; Brocks and Bang, 1991; Chitty et al., 1991; Ewigman et al., 1993), little research (e.g., Cox et al., 1987; Michelacci et al., 1988; Cromie, 2001) documents the short-term and long-term emotional stress on parents of antenatal ultrasound, including when the procedure leads to a diagnosis of a lethal or potentially lethal condition. More research has focused on prenatal genetic testing and choices about pregnancy termination (see, e.g., Vintzileos and Egan, 1995; Bergsjo and Villar, 1997; Malone et al., 2000; Rillstone and Hutchison, 2001). Given the ubiquitous use of ultrasound and the known
impact of emotional stress on pregnancy outcome, prospective studies should evaluate the emotional cost of antenatal ultrasound and the emotional care provided by obstetricians following ultrasound diagnosis of a lethal condition.
Research Directions: Sudden and Unexpected Death
In addition to infant death, the other large contributor to death in childhood is injury, both unintentional and intentional. Despite the development of sophisticated emergency medical services capacities that each year benefit thousands of seriously ill or injured children and their families, little is known about key aspects of the care provided to children who die and their survivors or about the consequences for emergency care providers of their involvement in child deaths. Box 10.4 summarizes several directions for research on sudden and unexpected death, and Appendix F also discusses research needs.
When resuscitation should be terminated in children who have experienced cardiopulmonary arrest (CPA) is a continual subject of research and debate in emergency medicine. Continued efforts to establish predictors of resuscitation outcomes for out-of-hospital pediatric CPA are warranted to guide prehospital and emergency department care. One particular controversy involves the use of high-dose epinephrine (HDE) following cardiopulmonary arrest in children. The value of this intervention is not supported by
clinical research, yet HDE remains a pharmacologic option (Callaham et al., 1992; Dieckmann and Vardis, 1995; Patterson, 1999; Young and Seidel, 1999; AHA, 2000a).
Transient return of spontaneous circulation following prolonged resuscitation can trigger a tumultuous series of events, including admission to the pediatric intensive care unit (PICU) and subsequent decisions about withdrawal of life support, declaration of brain death, and organ donation. Although these events add stress to an already overwhelmingly stressful situation, it is possible that the time spent in intensive care benefits some families, for example, by helping them believe everything possible has been done to save their child. The consequences for families and for clinicians of predictably unsuccessful resuscitation efforts deserve systematic attention.
Several institutional studies have investigated issues related to family presence during attempted resuscitation (see, e.g., Meyer et al., 1998; Guzzetta et al., 2000; Clark et al., 2001). They suggest that families generally favor having the choice and that they tend to report afterwards that being present was helpful. These studies were, however, small, often retrospective, and limited in scope. Nonetheless, emergency medicine professionals are increasingly accepting and even encouraging parental presence, not only during resuscitation in the emergency department but also during other critical care procedures (AHA, 2000a). Given the vulnerability of parents during these high-intensity, often unsuccessful interventions, it seems prudent to undertake additional systematic research on the bereavement outcomes and other consequences of parental presence policies and their implementation. Prospective studies could follow families from the discussion of presence during resuscitation through their decisions, the procedures, and the aftermath, with follow-up during bereavement. The committee did not investigate differences in institutional protocols for family presence (among institutions that allowed or encouraged it), but if different approaches characterize different phases of family experience (e.g., discussion, presence, immediate and later bereavement support), it would be helpful to compare the effects on families and on health care personnel involved.
Sudden Death, Care Outside the Hospital, and Family Support
Each year, before or after a child’s sudden and expected death, thousands of families encounter emergency medical personnel, police, staff of medical examiners’ offices, and others who may positively support families in time of shock and grief—or whose behavior may unintentionally add to their pain and suffering. The committee identified little research on the aftermath of these encounters (see, e.g., Schmidt and Harrahill, 1995). Although it found examples of educational materials and policies intended
to prepare these personnel to support survivors and avoid harm, the actual use and effectiveness of such materials and policies appears little investigated (Jaslow et al., 1997).
The committee recognizes that even retrospective, descriptive studies of the short-term and long-term effects on survivors of prehospital care may be costly and pose other practical and ethical challenges. Nonetheless, it encourages consideration of research on bereaved parents’ and siblings’ experiences and perceptions as a step toward identifying deficiencies in support and directions for improvement for personnel outside as well as inside the hospital. (Other questions about family experience following a child’s sudden death are discussed in the section on bereavement research.)
Critical Incident Stress Management
Professionals who care for children who die suddenly and unexpectedly may also experience serious emotional distress. Critical Incident Stress Management (CISM) is a widely used strategy for providing acute psychological support for emergency medical services (EMS) and other public safety personnel involved in traumatic incidents such as child deaths, multiple deaths, and other especially difficult situations (Everly and Boyle, 1999; Everly, 2000; Lipton and Everly, 2002). Recent reviews of the literature on this strategy concluded that research tends to support the short-term effectiveness of certain critical incident debriefing approaches with emergency and other personnel involved in stressful incidents (Everly and Boyle, 1999; Everly and Mitchell, 2000). Research on long-term effects is limited. More research would be useful in assessing or clarifying the timing of interventions, the criteria for selecting personnel and situations for intervention, the consequences of different levels or types of training for those conducting the debriefings, and the interactions among complementary interventions.
Research Directions: Bereavement Care
As in other areas of bereavement research, the scientific basis for current practices with parents, siblings, or others following a child’s death is limited (Stroebe et al., 1993, 2001b). Most studies are retrospective and relatively short term and suffer from the familiar limitations of late recall and abbreviated follow up. Research designs are also often inadequate in other ways (e.g., lack of control groups, small sample sizes), thus raising questions about the validity of the findings (Schut et al., 2001). One complication in intervention research is the relative (and fortunate) infrequency of child deaths in any single community. Possibly, Internet-based strategies can be explored as a means of involving larger numbers of family members.
The discussions above of perinatal and sudden death have already suggested directions for bereavement research in these areas. Box 10.5 lists additional areas for investigation, including continued refinement of tools for measuring grief and bereavement.
The committee sees a pressing need to develop prospective research that studies over many years the experiences of parents, siblings, and other family members during a child’s life-threatening illness and after a child’s death (whether sudden or anticipated). Such research should help increase understanding of variations in grief and bereavement experiences, including patterns of complicated bereavement and associated risk and protective factors. Prospective research likewise may increase understanding of parents’ willingness or reluctance to participate in bereavement support groups
and seek other mental health services for themselves or for their surviving children.
Research also should consider the impact of a child’s death as it affects not only parents and siblings individually but also as it affects the parents as a couple, the family as a unit, the extended family, and the family’s larger network of support. Measures applicable to individuals (e.g., psychological symptoms) will not necessarily apply to the family unit, and measures of functioning will appropriately vary for parents, young siblings, adolescents, grandparents, and other family members.
As part of a research strategy, the concept of complicated grief or bereavement needs further attention, as do criteria for identifying complicated grief in bereaved parents and siblings. Complicated bereavement has been proposed as a specific mental disorder but is not yet accepted as a diagnostic category (Prigerson and Jacobs, 2001).
The sudden, unexpected death of a child may create symptoms similar to posttraumatic stress disorder in perhaps one-third of surviving family members (Peebles-Kleiger, 2000).4 The experience of posttraumatic stress after a child’s death requires continued investigation, as does the relationship of this stress to complicated bereavement. A better understanding is necessary of the characteristics of individuals (e.g., gender) and situations (e.g., death by homicide, family dysfunction) that are associated with more severe distress or differences in responses. This will help in identifying high-
As set forth in the guide for diagnosing mental disorders (Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition [DSM-IV], American Psychological Association, 1994), the criteria for diagnosing PTSD include
Exposure to a traumatic event with both of the following elements: experiencing, witnessing, or being confronted with an event involving actual or threatened death or serious injury or a threat to the physical integrity of oneself or another and responding with intense fear, helplessness, or horror;
Persistent reexperience of the event, for example, in recurrent, intrusive, and upsetting recollections or dreams, in the sense that the event is happening again, or in intense reactions to cues that symbolize or are similar to the event;
Persistent avoidance of reminders of the traumatic event as demonstrated by three or more of the following: efforts to avoid thoughts, feelings, or conversations associated with the trauma; efforts to avoid activities, places, or people that arouse recollections of the trauma; inability to remember key aspects of the event; feelings of detachment or estrangement; significantly reduced interest or participation in activities; restricted range of emotions (e.g., to feel love); or sense of a limited future (e.g., loss of expectations related to job, family, life span); and
Persistent, new symptoms of increased arousal including at least two of the following: trouble sleeping; irritability or anger, difficulties concentrating; hypervigilance and exaggerated startle responses; duration of symptoms for more than one month; or clinically significant distress or impairment in social, occupational, or other important functioning.
risk individuals in advance for special attention and support and in tailoring strategies to individuals.
More generally, the development of effective interventions depends on better knowledge of the range of reactions to the sudden death of a child and of the factors related to similarities and differences in bereavement experiences and outcomes in parents, siblings, and other family members. Do the same risk and protective factors pertain to adults and children? Some commonly identified or hypothesized risk factors include deaths that are violent, self-inflicted, publicized, or linked to a broader catastrophic event. Hypothesized protective factors include stronger immediate and extended family relationships, relative economic security, religious faith, and absence of past traumatic losses. Research suggests men and women grieve differently. Do they differ otherwise in risk and protective factors?
When death is anticipated, are there factors that predict different outcomes, for example, continuity in relationships with the child’s physician or other caregivers, timely and accurate information, or access to psychosocial support services? The research by Wolfe and colleagues (2000a,b, 2001) discussed in Chapter 3 suggests some lines of inquiry. Ideally, these could be pursued using methods that do not depend entirely on after-death interviews with parents. In practice, as discussed elsewhere in this chapter, concurrent collection of information about end-of-life care for children is difficult because the timing of death is often uncertain and because interviews with children or parents during the child’s last days or weeks may be unacceptable or infeasible.
Research Directions: Models of Care Delivery
The committee identified a critical need for more systematic information on the organization, delivery, and outcomes of care for children with life-threatening conditions and their families (Box 10.6). After considering written statements from advocacy groups, discussions with parents, research and other literature, and professional experience, the committee identified several consistent themes: confusion about available resources and their quality, fragmentation of services with consequent lack of coordination and continuity of care, uneven and limited access to valued services, and deficits in communication, symptom management, and bereavement care in organizations that care for children with fatal or potentially fatal conditions. The very availability of key services is not well documented. For example, despite the increasing availability of useful information on the Internet, families cannot easily and reliably discover whether a children’s hospital has palliative care or pain management teams or which hospices have well-established pediatric programs—much less how that program performs in meeting child and family needs.
Developing the research base for organizational and system improvement is perhaps the most difficult research challenge identified by the committee. Such research tends to be practically and methodologically difficult, expensive, and not a priority for clinicians, health care managers, or research funders. Recognizing these obstacles, the committee proposes an incremental research strategy that includes both the identification and description of innovative programs and the evaluation of these strategies within and across institutions and settings.
Collecting data on the availability and characteristics of pediatric palliative and hospice care programs is one important step. Another is community-based case studies that describe child and family experiences with care coordination and continuity. Such case studies should help researchers to develop and refine hypotheses about factors that support or discourage coordination and continuity for families receiving services from multiple professionals and providers. Interviews with parents, patients, social workers, and others can generate useful information, but case studies will provide a fuller picture and help refine questions for further investigation.
Concurrently with descriptive research must come the most important step: the definition and testing of discrete, innovative care processes and broader models of palliative, end-of-life, and bereavement care. (For purposes of this discussion, processes of care include activities focused on specific tasks or goals such as symptom assessment and documentation of
discussions with parents; models of care include more comprehensive arrangements such as inpatient palliative care consultation services or inpatient hospice units.) Such evaluative research would, ideally, consider outcomes for children (e.g., levels of pain or anxiety, quality of life, continuity of care), families (e.g., satisfaction with care, understanding of diagnosis, prognosis, and treatment options), and providers (e.g., rates of specific problems such as unassessed symptoms).
Given the complexity and expense of organizing multi-site studies, especially in newer areas of investigation, evaluations will often focus on processes and models of care at single sites. Although generalization from single-site studies must be undertaken cautiously, controlled trials and other comparative studies within single institutions can help in the identification of more and less effective ways of providing care. One creative example is the small, randomized controlled trial, mentioned in Chapter 6, of a program providing parents of hospitalized premature infants with an interactive link with the neonatal intensive care unit and also Internet access to other resources.
Notwithstanding the challenges, multi-site, comparative studies are important to replicate and test the robustness of care models in different environments. If, as recommended earlier in this chapter, the NIH promoted research on palliative and end-of-life care based in pediatric cancer centers, neonatal network participants, and similar networks, this could facilitate comparative organizational as well as clinical studies.
Even multi-site studies that are not formally comparative may be a useful step in the development of systematic research initiatives. For example, although not comparative in design, the federally supported Medicaid demonstration projects described earlier in this report should provide individual evaluations of several home care and hospice models and suggest ideas for further investigation. Similarly, although focused primarily on adult care, the quality improvement initiative reported by Lynn and colleagues (2000) suggests opportunities for formal research to improve palliative and end-of-life care for children as well as adults (see Chapter 6).5 For
example, studies might compare the outcomes of different processes and structures for assessing, documenting, and managing pain and other symptoms, or they might test different procedures to improve the discussion, documentation, and implementation of care plans for children.
More generally, the research reviewed in Chapter 3 and the ideas and programs described in Chapters 4, 5, and 6 can suggest topics for organizational research into processes and models of care. For example, it would be useful to have additional research comparing outcomes for children and families associated with early versus late discussion of hospice or other options for care at the end-of-life.
Outcome and Performance Measures
One critical issue in organizational research is how to measure performance. For example, to what degree should investigators rely on assessments of patient or family satisfaction with care as indicators of the quality of care? Given the literature it reviewed and members’ own experiences, the committee is concerned that families may greatly appreciate the care provided to their children without realizing that more could have been done, for example, to treat a child’s pain effectively.
Thus, although it is appropriate to inquire about parents’ satisfaction with the care provided to their child and their family and about the help they had making decisions, it is also important to inquire about what the child and family experienced and to continue efforts to develop and use other measures. These measures include valid, reliable, and feasible instruments for assessing physical and psychological symptoms and quality of life. (See also Appendix C and Box 10.1.)
Information and Communication
This report, consistent with other discussions of care for patients with life-threatening medical conditions, stresses the importance of timely and accurate information and communication geared to child and family needs and circumstances. Most studies of communication in this area have focused on adult patients and their families, but many of the findings about barriers to effective communication have broad relevance for child patients and their families.
This committee has stressed the importance of involving children in discussions and decisions about their care, consistent with their development stage and preferences and with sensitivity to family culture and values.
To further such involvement, strategies for communicating with seriously ill children need further systematic investigation. As discussed in Chapter 4, most of what has been written about children’s understanding of death involves well children. Because children with advanced illness may have awareness and knowledge beyond their years, findings from studies of communication strategies involving well children need to be used with some caution. Rigorous research on communication with children made vulnerable by serious illness may raise ethical and practical challenges, but if carefully planned, reviewed, and conducted, it has the potential to help the research participants as well as children who follow them.
One persistent theme in statements of parents who have faced a child’s life-threatening illness is that shock, fear, and grief interfere with their ability to absorb and process information during initial and subsequent conversations with clinicians. Little or no systematic research is available to guide the timing, pace, or style of communication with these parents. One example of research underway in this area is the PAICC project mentioned earlier. It asks parents about the value and timeliness of information provided through a written manual for parents and other means. The plan is to have 20 parents evaluate the material, revise the manual with that feedback, and proceed again with another cohort of 20 parents.
The usefulness of specific modes of communication with children and parents may vary depending on the purpose of the communication (e.g., giving bad news, developing care objectives and plans), the type of information to be communicated (e.g., medical, administrative), the recipient (e.g., young versus older child), the opportunity for repeated reference to the information, and other factors. Structured, face-to-face communication appears essential for some purposes, but other strategies including written materials, graphics, videos, and Internet-based tools may reinforce such communication and, for some purposes, substitute for it. These expectations are reasonable but should be further investigated.
Interdisciplinary Care Teams
Despite their seeming ubiquity and this committee’s judgment of their importance, little rigorous research appears to have focused on the prevalence or functioning of different kinds of pediatric or adult palliative care teams or to have identified factors associated with more or less effective team performance in different contexts. Such systematic research as was identified by the committee focused primarily on primary care and chronic illness care (often as an element of primary care) (see e.g., Campbell et al.,
2001; Barrett et al., 2001; Roy-Byrne et al., 2001; Rost et al., 2000; but see also Shortell et al., 1994 on intensive care teams).6 Team care is, however, often a part of a multi-element intervention in which the individual structural elements or processes are not separately evaluated.7
Although problems in coordination and continuity often arise when multiple providers (e.g., hospitals, home care agencies) are involved in a child’s care, problems also arise within single organizations (e.g., as shifts change or clinicians from different disciplines interact with parents). As a starting point, it would be helpful to have more systematic information on the structure, composition, environment, and functioning of the lead care teams for children with life-threatening medical conditions and the characteristics of teams associated with better coordination and continuity for the dimensions of care identified in Chapters 4 and 5. In addition, an important question for pediatric palliative and end-of-life care is how care teams relate to each other during transitions in care (e.g., from inpatient to home hospice care) or when additional expertise (e.g., palliative care consultations) is needed.
As discussed in Chapter 6, access to expertise in pediatric palliative care is limited by geography. Some children who die will be cared for through the end of life at specialized regional centers, but families will take other children home to be cared for by family members and local pediatricians and community hospitals that have limited preparation for and experience in providing palliative and end-of-life care. In Chapter 6, the committee
Some relevant research may be categorized under related concepts or objectives such as care coordination or quality improvement. For example, team care is often associated with the use of formal clinical practice guidelines and quality improvement initiatives, and some studies suggest problems in the team care component of such initiatives, for example, staff turnover, physical separation, and responsibilities outside the team (Weissman et al., 1997; Hayward et al, 2000; see also Campbell E. et al., 1998).
Some research has focused on the contributions or acceptance of specific categories of professionals (e.g., nurse practitioners, social workers, or genetic counselors) in different environments (see, e.g., Burl et al., 1994; Aquilino et al., 1999; Inati et al., 1994). Research not specifically focused on the effectiveness of multidisciplinary teams has pointed to potential problems for team functioning arising from differences in the views of physicians and nurses about what constitutes clinically or ethically appropriate care at the end of life (see Chapter 3). In its call for grant applications on quality of life for individuals at the end of life, the National Institute of Nursing Research mentioned “multidisciplinary interventions” and noted that research is need to support care systems that include multidisciplinary teams (NIH, 2000).
recommended strategies for supporting these professionals and organizations, including written protocols, family guides, and various forms of telemedicine. The federal government has supported research to investigate the potential of telemedicine to improve care for rural populations, including people receiving care at home for serious chronic illness. (See, e.g., Office for the Advancement of Telehealth at http://telehealth.hrsa.gov.) Although limited, this research can provide a foundation for studies related to pediatric palliative and end-of-life care specifically.
Research Directions: Financing Pediatric Palliative and End-of-Life Care
Chapter 7 has made clear that few systematic data are available to describe the financing of palliative and end-of-life care for children, much less evaluate systematically the effects of different financing policies or practices on that care. Thus, the committee’s conclusions and recommendations about financing are based primarily on logic, experience, value judgments, and cautious generalization from studies involving adults. Although the committee believes it is appropriate to move to implement the recommendations, it also recognizes that further research and analysis are important to guide certain changes in financing policies and practices (Box 10.7).
More complete descriptive research on financing practices would help providers, managers, and policymakers better understand how care for children who die and their families is financed and how the situation for children differs from that for adults. The first steps should include a state-by-state inventory of variations in Medicaid coverage and reimbursement policies related to the elements of pediatric palliative and end-of-life care discussed in Chapter 7. Although a complete inventory is not feasible for private health plans given their large number, coverage and payment policies should be surveyed for a subset of representative plans or communities.
Next steps should also include some assessment of the extent to which written policies (including early periodic screening, detection, and treatment [EPSDT] policies) match implementation (e.g., whether claims administrators often deny claims for hospice and palliative services covered for children but not adult Medicaid beneficiaries). To be manageable, such an assessment might cover only a subset of states using a survey or case studies or both. Once the descriptive foundation is in place, the next step is to assess the impact of different coverage policies and reimbursement methods on cost, access, and quality of care for children who die and their families.
Payment, Coding, and Severity of Illness Systems or Measures
A number of Medicaid and private health plans have adopted or adapted payment and coding systems originally devised for Medicare (i.e., diagnosis-related groups [DRGs] for hospital care and the resource-based relative value scale [RBRVS] for physician care). Researchers and child health advocates have raised several concerns about the application of these schemes to children, as described in Chapter 7. Further studies are needed to examine whether state and private reimbursement systems based on
DRGs or the RBRVS properly categorize and weigh palliative and end-of-life services for children. Such research could also identify possible adjustments in these systems to correct any problems identified.
Chapter 7 made general suggestions about establishing children’s eligibility for certain types of palliative services using diagnosis and severity-of-illness criteria. Analytic work is needed to identify reliable and valid approaches for establishing such criteria for payment purposes and to assess their feasibility, cost implications, and likely consequences for child and family access to palliative care.
As discussed in Chapter 7, studies of hospice care for adults have identified serious limitations in Medicare’s per diem payments for that care. For example, studies have documented hospice access problems for adult patients with particularly high-cost needs (e.g., those requiring expensive pain medications or other palliative therapies). Research is needed to document the extent of hospice access problems for pediatric patients, examine hospice referral patterns, and identify characteristics of patients who are denied access. To explore these issues, prospective studies using existing clinical networks (e.g., the Children’s Oncology Group) for patient identification should be conducted.
In order to implement a payment outlier system for hospices (recommended by the committee in Chapter 7) and to facilitate research on utilization and costs of hospice care for adult and pediatric patients, a Medicaid hospice cost-reporting system should be created. Since many hospices do not currently maintain detailed data on number of hours or types of services provided by staff, Medicaid could begin by creating a cost-reporting system based on high-cost items that hospices are more likely to track (e.g., expensive medications, durable medical equipment).
Prospective Payment for Home Health Care
Because Medicare policies often influence state and private policies, studies are needed to examine how new Medicare prospective payments systems for home health care and other services have diffused to state Medicaid programs and private health plans and how these systems are affecting care for children with life-threatening medical conditions. A first phase of research would involve surveying Medicaid and private plans to assess their adoption of Medicare reimbursement methods. A second phase would investigate the impact of these payment approaches on family burden and quality of care for Medicaid-covered children.
Cost Implications of Family Choices
For reasons discussed earlier in this report, families should have choices about the site of death for children with fatal medical problems. To assess the cost implications of these choices, prospective studies are needed to identify and compare the costs of different home and inpatient care options for children expected to die from congenital anomalies, cancers, progressive neurological disorders, and other serious conditions. Existing studies based on Medicare beneficiaries’ use of hospice care cannot be generalized to children and are, in any case, flawed by retrospective research designs that cannot control for selection bias (i.e., differences in preferences and other characteristics of those who choose hospice care).
Although the committee expects that the financing changes recommended in Chapter 7 will involve limited additional costs for the benefits expected, a simulation model would assist in estimating the cost and utilization effects of implementing different options. For example, the cost of eliminating the six-month prognosis requirement could be estimated.
Finally, as noted elsewhere in this chapter, long-term prospective research is needed to evaluate the effectiveness of bereavement interventions for parents and siblings. The estimated costs and long-term cost-effectiveness of alternative interventions should also be investigated if such studies are undertaken.
Research Directions: Educating Health Professionals
Despite several decades of research on methods for educating health care professionals, much uncertainty remains about methods that can be consistently applied to produce changes in knowledge and practice, especially changes that endure beyond the classroom or the continuing education program. One challenge for educators is that the knowledge, values, and actions they teach are not uniformly rewarded or reinforced in day-to-day practice.
Educational research is not always valued within academic departments or by research funders. Long-term follow-up of educational outcomes, in particular, tends to be expensive and logistically complicated. Nonetheless, given the huge investments in preparing health professionals for practice, research is essential to identify the effectiveness of different educational techniques and emphases in creating desired changes in knowledge, skills, attitudes, and especially, behaviors.
The committee expects that much research involving education in pediatric care, adult palliative care, and generally relevant attitudes or skills (e.g., empathy, interviewing patients) would be generalizable to education in pediatric palliative care. Nonetheless, in developing curricula or inter-
ventions for pediatric specialists in training or for established pediatric specialists, revisions or adjustments will have to be considered and evaluated to take into account the special emotional and other dimensions of caring for dying children and their families. For example, as discussed in Chapter 8, the use of standardized patients is now a common educational tool, but using child actors or volunteers to depict dying children might encounter practical and ethical obstacles. Standardized families can be used, and systematic research into the effectiveness of this strategy—and of different variants—would be helpful. The projects by Williams and colleagues (2001b, 2002) and Greenberg and colleagues (1993) discussed in Chapter 9 offer examples of how such standardized families might be used to improve skills in communicating bad news and informing family decisionmaking.
Box 10.8 lists several directions for educational research. The involvement of bereaved parents in educational programs should be investigated, both for its impact on trainees and for the consequences for the parents. Conversations with bereaved parents suggest that some have found consolation and strength in presenting their experiences and perspectives, answering questions, and otherwise helping to educate both clinicians-in-training and established clinicians. Self-selection of the most confident or educated parents is, however, a potential concern and should be taken into account in any research about the role of parents as educators.
Because it will take years for newly educated health professionals to dominate health care systems by numbers or other influence, it is important
to examine continuing education in pediatric palliative care for established clinicians. Further, research is needed to identify characteristics of pediatric care systems that promote and reinforce lessons learned about palliative and end-of-life care in all educational contexts. As noted above, without reinforcement, educational interventions may manage to create desired short-term results but fail to produce durable changes. Research questions include whether education programs are more effective when institutions have formal relationships with hospices, formal pediatric palliative care consultation programs, formal adult inpatient palliative care services (for institutions serving adults and children), or identified palliative care role models and educators.
This report has stressed the importance of team care to bring multiple clinical perspectives and skills to bear on the complex physical, emotional, spiritual, and practical needs of seriously ill or injured children and families. In addition to research on the performance of pediatric teams in different environments and contexts (including providing palliative, end-of-life, and bereavement care in hospitals or at home), research also should assess methods for successfully educating physicians, nurses, social workers, and others to function as effective members of health care teams.
Also stressed in this report is the importance of adequate and timely information for patients and families. As reviewed in Chapters 3 and 4, studies have suggested that clinicians may be reluctant to provide such information when the diagnosis and prognosis are grim. One study (Lamont and Christakis, 2001) suggested that clinicians may provide patients and families with misleadingly optimistic assessments and that younger adult patients were more likely to receive such misleading assessments than older adults. Another study involving physicians caring for adults found that they dreaded explicit discussion of patient prognosis, usually were not specific, and often waited until parents or families raised the issue (Christakis and Iwashyna, 1998). Other studies of adult and pediatric palliative care have identified discrepancies between physician assessments of care and the assessments of patients or family members (see, e.g., Wolfe et al., 2000b).
A variety of educational interventions—for example, role playing, standardized patients who provide feedback, and parents as teachers—could potentially promote greater clinician self-awareness and more explicit concern for the actual experiences and expectations of patients and family members. Studies of efforts to teach empathy that were conducted in the 1970s suggest strategies to assess and compare such educational interventions (Fine and Therrien, 1977; see also Spiro, 1992; Reiser, 1996; Halpern, 2001).
Finally, the development and evaluation of educational techniques to improve the use of proven symptom management strategies are important if their benefits are to reach children in need. In particular, although cognitive
and behavioral strategies have demonstrated effectiveness in reducing pain and improving coping, teaching of these techniques to clinicians generally appears to lack the rigor and uniformity that characterize other clinical domains.
CHALLENGES OF RESEARCH ON PALLIATIVE AND END-OF-LIFE CARE FOR CHILDREN AND THEIR FAMILES
The paucity of information about palliative and end-of-life care for children and their families has several explanations. Although over 50,000 children die each year, many of these deaths result from severe injuries or sudden infant death syndrome (SIDS). The medical care actually provided to or consciously experienced by these children is limited compared to that of children with chronic medical conditions. For surviving family members, interactions with emergency medical personnel (particularly first responders) or ICU personnel are often relatively brief and poorly documented, which makes it difficult for researchers to use medical records retrospectively to assess the extent and quality of the information and emotional support provided to survivors.
Particularly for deaths involving infants and very young children, a considerable fraction of deaths result from an array of relatively or very uncommon congenital conditions. Even over several years and across multiple medical centers, it may be difficult to identify enough cases for each condition to allow productive prospective or retrospective analyses of the patterns and quality of palliative and end-of-life care for these children and their families. Combining information about children with different conditions may sometimes be a plausible strategy if the symptoms and other characteristics of each condition and the consequences for the children and families are similar. Nonetheless, accumulating enough cases for analysis may still take years.
Even when numbers are sufficient for study, however, retrospective analyses of medical records may be constrained by failures to document relevant aspects of care, for example, whether or not pain and other symptoms were explicitly assessed. Such lack of documentation, in itself, is an important finding with implications for day-to-day clinical practice and quality improvement efforts.
Another constraint is that although Medicare-related databases offer much comprehensive information about the types and costs of care provided to elderly adults, no comparable information is available for younger individuals covered by the 50-plus state Medicaid programs and thousands of private insurance plans. Other information sources, such as government surveys of home health and hospice agencies, may not include enough children to allow meaningful analyses.
Despite many differences, researchers studying end-of-life care and researchers studying children share a common challenge: their subjects are often unable to report directly on the care they have received or on their physical or emotional distress or well-being. Even for children old enough to report reliably, the effects of a life-threatening medical problem or its treatment may render them unconscious, confused, drowsy, fatigued, or otherwise not able to answer questions. As a result, researchers frequently rely on surrogates—usually the parents—to describe their perceptions of the patient’s experience and quality of life (see Appendix C). Also, because the time of death is often unpredictable, researchers may be unable to assess patients’ status at comparable points before death and may have to rely on the recollections of surrogates about a patient’s status during the days, weeks, or months prior to death (see, e.g., Wolfe et al., 2000a,b). Nonetheless, surrogate reports and retrospective reports raise serious questions of validity and reliability, and researchers and research funders should consider alternatives whenever possible.
Another challenge is that research funders, who are accustomed to more stable research situations, measurement tools, clinical end points (e.g., remission, five-year survival), and already-conducted pilot testing, may be reluctant to support research that does not fit these patterns. In the committee’s experience, they are also reluctant to fund the kinds of pilot studies that provide an essential base for more rigorous research. This puts researchers in this new and difficult arena in a classic “Catch-22” situation.
In addition, the conduct of medical or health services research involving children faces more stringent ethical and legal constraints than apply to research involving most adults. These are discussed below.
ETHICAL AND LEGAL ISSUES IN RESEARCH INVOLVING CHILDREN
Policymakers, researchers, and ethicists have been working for decades to develop protections for people participating in research, especially “vulnerable” populations including children, prisoners, and those with serious cognitive or emotional problems. The first widely recognized principles for ethical research were the Nuremberg Code’s directives for human experimentation, which were developed in 1947 by judges presiding over the trials of Nazi physicians accused of research atrocities (Shuster, 1997). These judges developed 10 principles for research, the first of which stated that it is absolutely essential that human subjects voluntarily consent to participation. The statement did not consider the involvement of children in research.
The 1964 Declaration of Helsinki from the World Medical Association called for parents of child participants in research to give their consent and
for children’s “assent” to be obtained when possible (WMA, 1964). In the United States, a national commission on protection of human subjects in research likewise proposed in 1977 that parental permission be required as well as assent in some form from children age 7 and over (USDHEW, 1977). Subsequent books, articles, and conferences have continued to explore ethical issues involved in research on children, the boundaries of permissible research, and strategies for protecting child research subjects (see, e.g., Stanley and Sieber, 1992; Grodin and Glantz, 1994; AAP, 1995a; Levine, 1995; Rosato, 2000).
If conducting research presents ethical questions, so does the failure to conduct research. As noted by the American Academy of Pediatrics in a statement on drug research, the “lack of studies in children presents the treating physician with an ethical dilemma. The physician must frequently either not treat children with potentially beneficial medications or treat them with medications based on adult studies or anecdotal empirical experience in children” (AAP, 1995a, p. 286).
Federally Funded Research
As set forth in Box 10.9, federal regulations adopted in 1983 and revised in 1991 set forth requirements related to children’s participation as subjects of federally funded research (45 CFR 46; USDHSS, 1993). The regulations also provided for institutional review boards (IRBs) to be established by each research institution to review federally funded research that involves human subjects and determine whether proposed research meets federal standards.
Although the regulations and accompanying federal guidance encourage respect for a child’s views about participating in research, they make clear that a research project may be approved without requiring that each child provide assent before enrollment (45 CFR 46.408[a]). The IRB must conclude in such cases that the research has the potential to produce an important direct benefit to the child, that such benefit is possible only in the research context, and that provisions are adequate for obtaining informed permission from the child’s parents for the research.
The regulations also require that research conform to state laws and judicial decisions including those related to the age of consent for medical treatment. If an adolescent is considered legally competent under state law to consent to treatment, he or she can usually consent to research presenting equivalent risk.
In some situations (e.g., for research on children who have been abused), researchers and the IRB may seek alternatives to or reviews of parental permission for a child’s participation in research. If a child is a ward of the state, additional limitations are imposed on his or her participation in
Research Involving No Greater Than Minimal Risk (46.404)
“The Institutional Review Board (IRB) must find that the research involves no greater than minimal risk to children and that adequate provisions are made for soliciting the assent of the children and the permission of their parents or guardians.”
Research Involving Greater Than Minimal Risk That Has Potential to Benefit the Research Subjects Directly (46.405)
“The IRB must find that more than minimal risk to children is presented by an intervention and then must also find that (a) the research may directly benefit the individual research subjects, (b) the anticipated benefit is sufficient to justify the risk, (c) the relation of the benefit to the risk is equally or more favorable as that presented by available nonresearch alternatives, and (d) adequate provisions are made for soliciting the assent of the children and the permission of their parents or guardians.”
Research Involving Greater Than Minimal Risk That Has No Prospect of Direct Benefit to Individual Subjects but Is Likely to Yield Generalizable Knowledge (46.406)
“The IRB must find that more than minimal risk to children is presented by an intervention that is not expected to benefit the individual subject directly but that (a) the increased risk is minor, (b) the experiences associated with the intervention are reasonably similar to those inherent in the subject’s actual or expected medical situation, (c) the intervention is likely to yield generalizable knowledge about the subject’s condition that is vitally important to understand or ameliorate the condition, and (d) adequate provisions are made for soliciting assent of the children and the permission of their parents or guardians.”
Research That Is Not Otherwise Approvable (46.407)
“The IRB must conclude that the research does not meet the above requirements but does present a reasonable opportunity to further the understanding, prevention, or alleviation of a serious problem affecting the health or welfare of children. The research must then be judged by the Secretary of the Department of Health and Human Services (after consultation with appropriate experts and opportunity for public review and comment) to fit one of the categories described above or to meet the following conditions of (a) presenting a reasonable opportunity for better understanding, preventing, or alleviating a serious problem affecting the health or welfare of children, (b) following sound ethical principles for research, and (c) including adequate provisions for soliciting the assent of children and the permission of their parents or guardians.”
SOURCE: 45 CFR 46 Protection of Human Subjects, Additional Protections for Children Involved as Subjects in Research, Subpart D, Sections 401–409. 56 FR 28032, June 18, 1991 (http://www.med.umich.edu/irbmed/FederalDocuments/hhs/HHS45CFR46.html).
research when the research involves more than minimal risk with no prospect of direct benefit to the child. The research must either be related to children’s status as wards or be conducted in settings such as hospitals or schools where the majority of child research participants are not wards of the state (45 CFR 46.409). Also, a special advocate (other than the guardian) must be appointed for a child participating in such research. The committee understands that some states refuse to permit wards of the state to participate in research, even if no alternative treatment exists for the child’s life-threatening illness.
Explanations of the proposed research are to be appropriate for the specific child’s age, maturity, experience, and medical condition. The National Cancer Institute provides guidance for parents and guardians, including advice that parents not withhold information about the discomforts or risks involved in research and that they monitor whether the child really seems to understand the explanations being provided (NCI, 2001c).
Based on members’ experiences and their conversations with other researchers, the committee believes that many IRBs, as presently constituted, may lack the expertise and background to evaluate proposals for research on pediatric palliative and end-of-life care. They may have unwarranted concerns about the potential of such research to cause harm to children and families. As described below, Congress has asked for a study that will investigate the competence of IRBs to assess research involving children.
Some research should soon be available to document positive responses by families to studies investigating their experiences and perspectives (Wolfe, 2001). As noted in a recent report on family perspectives on the quality of pediatric palliative care, “The families who participated in our assessment unanimously expressed gratitude for the opportunity to ‘tell their stories’ ” (Contro et al., 2002, p. 18). Researchers in that study reported, however, that 20 of 63 families they reached declined to participate because it would be too painful.
Privately Funded Research
The regulations described in the preceding section apply to federally funded research involving children. Pediatric drug research promoted by the Food and Drug Administration is generally funded by pharmaceutical companies. As directed by Congress in the Children’s Health Act of 2000, the FDA must apply Department of Health and Human Services requirements for federally funded research that involves children to privately funded studies of products regulated by the FDA. (The agency had long required IRB review of drug research submitted to it by private companies [Levine, 1988].)
In April 1, 2001, the FDA issued an interim rule to change its regulations accordingly (21 CFR 50 and 56; FDA, 2001a). The agency invited comment on several issues including children’s participation in placebocontrolled trials, the definition of what constitutes a “minor increase over minimal risk” (and ways of measuring minor risk and determining when a minimal risk becomes a major risk), and ways to provide age-appropriate explanations of research to children.
The interim rule differs from the regulations governing federally funded research in certain respects. For research in the “not otherwise approvable” category (see Box 10.8) that is referred to the commissioner of the FDA for consideration, the agency may not be able to offer public review and comment if the research sponsor is unwilling to make public necessary information that is privileged under other FDA requirements.
Research that is neither federally funded nor conducted under the FDA provisions described earlier may be covered by state regulations and institutional policies requiring review of all research undertaken by its employees or students. In general, IRBs apply the same criteria to research proposals without reference to the source of funding.
Continuing Concerns about Children’s Participation in Research
Policies about children’s participation in research continue to prompt debate. The major issues involve
the appropriateness of current regulations for children of different ages or intellectual and emotional maturity,
the ethics of including children and other vulnerable subjects in research that is not expected to benefit these subjects directly,
the definition of minimal risk,
the extent of children’s and adolescents’ comprehension of information about the risks and benefits of treatments and participation in research,
the potential for conflict between parent and child and the meaningfulness of “assent” or discussions about assent with a child when parents can override the child’s wishes, and
the performance of IRBs in evaluating research involving children.
The legislation reauthorizing the incentives for pediatric drug testing provided for an Institute of Medicine study to examine such issues and review federal regulations relating to research involving children. The results of this report are to be given to Congress by the close of 2003.
This chapter has proposed a range of research efforts to strengthen the very limited base of knowledge now available to guide those providing, organizing, financing, and monitoring palliative, end-of-life, and bereavement care for children and families. It has recommended strategies to promote relevant research in federally funded pediatric centers, networks, and similar structures. The intent is to increase the numbers of children and families involved in studies, encourage the development and use of common research methods, and increase the credibility and acceptance of the research findings.
The research directions proposed here focus on children who have conditions that are certain or likely to prove fatal, but investigation of some of the suggested research questions may involve participation by children who survive and their families. In some cases, such participation will reflect the unpredictability of certain life-threatening conditions and the requirements of prospective research designs. In other cases, it will reflect a focus on questions that affect the well being of children and families facing a life-threatening problem, whether or not that problem actually ends in death. In either case, when it is appropriate, including a larger group of children usually has methodological advantages.
Much of the research proposed here thus should provide knowledge that will inform and improve the care of children who survive as well as those who do not—and likewise will help every family that suffers with a seriously ill or injured child. Indeed, all of the recommendations in this report, if implemented, should help create a care system that all children and families can trust to provide capable, compassionate, and reliable care when they are in need.