A Research Agenda
In addition to taking the actions described in the previous chapters, improving the delivery of psychosocial health services will require targeted research. This research should aim to clarify the efficacy and effectiveness of new and existing services, including identifying subpopulations who benefit from specific services and the circumstances in which given services are most effective. Health services research also is needed to identify more effective and efficient ways of delivering these services to various populations in different geographic locations and with varying levels of resources. As discussed in Chapter 3, the economical production, interpretation, and application of research findings would be improved by a taxonomy and nomenclature for psychosocial health services that would be shared across disciplines.
A TAXONOMY AND NOMENCLATURE FOR PSYCHOSOCIAL HEALTH SERVICES
The committee reiterates the importance of the recommendation made in Chapter 3 for the development of a standardized, transdisciplinary taxonomy and nomenclature for psychosocial health services:
Recommendation: Standardized nomenclature. To facilitate research on and quality measurement of psychosocial interventions, the National Institutes of Health (NIH) and the Agency for Healthcare Research and Quality (AHRQ) should create and lead an initiative to
develop a standardized, transdisciplinary taxonomy and nomenclature for psychosocial health services. This initiative should aim to incorporate this taxonomy and nomenclature into such databases as the National Library of Medicine’s Medical Subject Headings (MeSH), PsycINFO, CINAHL (Cumulative Index to Nursing and Allied Health Literature), and EMBASE.
As discussed in Chapter 3, the absence of a commonly understood vocabulary to describe psychosocial health services and interventions hinders the identification, interpretation, analysis, and application of evidence of effective delivery of those services. Developing a language that can be used across professions and disciplines is critical to the production of better evidence to support the delivery of effective psychosocial health services, which are themselves multidisciplinary and multiprofessional.
EFFECTIVENESS AND HEALTH SERVICES RESEARCH
Although evidence described in Chapters 3, 4, and 5 supports the health care benefits of providing psychosocial health services and points to ways of doing so effectively, there are still many unanswered questions. Key questions remain about how to address certain psychosocial health problems most effectively, as well as how to deliver services most efficiently to the various individuals who need them.
As is the case in biomedical care, providing effective psychosocial services to all who need them is hindered in part by limitations of the knowledge base. For the past three decades, the National Cancer Institute (NCI) and other private organizations that fund cancer research have supported a wide range of psychosocial research studies involving cancer patients and their families. However, the challenge with cancer is that it is not a single disease (prostate cancer, for example, is different from lung cancer in its impact), and even for a particular cancer site, individuals’ specific psychosocial health care needs may vary (e.g., in early-stage versus advanced disease). Given that there are more than 100 specific cancer types, it is therefore difficult to generalize about the benefits of particular psychosocial interventions, as their efficacy may vary based on the cancer site or phase of disease. Increasingly, research studies have focused on homogeneous samples of patients, making interpretation of outcomes more salient. For these reasons, psychosocial research with cancer patients is more challenging than that focused on more homogeneous diseases, such as asthma, diabetes, or heart disease.
For some psychosocial problems, research has not yet identified effective services to resolve them. In other cases, evidence that a remedy works effectively for some populations does not necessarily mean that the same remedy is effective for (is generalizable to) all people in all situations. Moreover, evidence is frequently not as clear as one would like it to be. Findings can be mixed, with evidence of the effectiveness of a service being found in one study but not being replicated in others. Additionally, evidence that a given service is effective does not exclude the possibility that another service is more effective for the same problem or equally effective at lower cost. In such cases, research continues to be needed even for services and interventions whose efficacy is supported by research findings. All of these situations are found in the array of evidence pertaining to psychosocial health services.
Identification of Effective Interventions
For some psychosocial health problems faced by cancer patients, research has not yet identified efficacious remedies. For example, as discussed in Chapter 3, research does not well inform clinicians about how to address effectively continued tobacco use among cancer patients, cognitive impairment among adults treated for cancer, and difficulties with school reentry for children treated for cancer. Further, although cancer is recognized as having a large impact on family members, they are rarely the subject of or included in research on psychosocial health care (Helgeson, 2005). More commonly, research points to the effectiveness of specific psychosocial services, but offers limited evidence about whether a broad spectrum of patients (and family members) benefit equally from those services in all situations.
Determination of Effectiveness in Different Populations and Scenarios
Questions about the effectiveness of many psychosocial services have evolved from addressing whether given services are effective to addressing for whom and under what circumstances specific services are needed and effective (Helgeson et al., 2000; Zebrack and Zeltzer, 2003; Cohen, 2004; Helgeson, 2005; Stanton, 2005). Effectiveness research on psychosocial health services has most often focused on women with breast cancer at the middle to upper middle socioeconomic levels without regard to the amount of psychosocial stress they are experiencing. Services need to be tested with men, in patients with sites of cancer other than breast, across different stages of cancer, with patients experiencing different types and levels of psychosocial needs and stress, and with those from different cultural and socioeconomic backgrounds.
Variables that moderate treatment effects need to be better understood as well (Helgeson, 2005). For example, many individuals diagnosed with cancer report manageable psychological distress that resolves over time without the need for formal services. Other research has found that patients with the highest levels of distress often show the greatest reduction in symptoms when provided psychosocial services (Andrykowski and Manne, 2006; Antoni et al., 2006). Research conducted with individuals with varying availability of social supports has found that the effectiveness of different types of psychosocial services can depend on the nature and extent of those supports (Cohen, 2004). Because of such findings, experts point to the need for a new generation of research on the effectiveness of psychosocial health services (Helgeson, 2005) involving “increasingly careful a priori consideration of the nature of the samples, interventions, and outcomes involved, as well as theory-guided examinations of mechanisms for the obtained effects” (Stanton, 2005:4819). Particular attention should be paid to socially disadvantaged populations, examining the effects of socioeconomic status and race/ethnicity on the risk for psychosocial problems and on the impact of interventions on these problems. Such work should also take into account developmental issues, particularly for children with cancer, and the effectiveness of interventions at different life stages (Zebrack and Zeltzer, 2003).
Use of More Robust Research Methods
The strength and generalizability of the evidence generated by research are increased by attention to several research design issues. First, the effectiveness of a service is often measured using dimensions of quality of life. Because measures of quality of life are numerous and variable, what one study finds effective may not be interpreted as such by others. The development of standard outcome measures by which the effectiveness of psychosocial services can be measured would increase the understanding and application of research results.
Research using more rigorous research designs is also needed, including use of longer follow-up periods (Helgeson, 2005). For example, with respect to the effects of psychotropic medications used to treat depression and other mental health conditions of patients with cancer, conclusions about effectiveness are limited (see Chapter 3) because of the few randomized controlled trials that have been done. More such trials are needed, using a larger patient cohort that is studied over a longer period of time to properly assess drug efficacy. These trials will likely require multicenter sites. The trials should be limited to patients with significant levels of depressive or anxious symptoms at baseline (e.g., severe adjustment disorder with depressive or anxious symptoms or anxiety, post-traumatic stress
disorder, or mood disorder) to ensure an adequate effect size. Studies also are needed that compare one drug with another and with the drug plus a psychosocial intervention.
More multisite research on services targeting children is needed as well. Because cancer in children is rare, most research involves small samples, limiting the conclusions that can be drawn (Patenaude and Kupst, 2005).
Research testing the effects of the receipt of psychosocial health services on physiological and clinical outcomes also could help build the conceptual framework underpinning those services, and point to new interventions and ways to target services and interventions to those who are most vulnerable (Patenaude and Kupst, 2005; Thacker et al., 2007). Such research should address, for example, the links between certain types of stress and immune system functioning and the effects of psychosocial supports on health, such as through changes in endocrine and immunological functioning and mediating physiological pathways.
Health Services Research
Health services research could help identify better ways of implementing some of the interventions necessary for the delivery of psychosocial health services. This research could be accomplished through the large-scale demonstration program recommended in Chapter 5 that would test various approaches to the effective provision of psychosocial health care in accordance with the standard of care set forth in this report. Health services research could also address how to implement components of the model described in Chapter 4 more efficiently and effectively, focusing in particular on methods for improving the patient–provider partnership, the development of better screening and needs assessment tools, comprehensive illness and wellness management interventions, approaches for effectively linking patients with services and coordinating care, and reimbursement arrangements that would support these interventions.
Methods for Improving the Patient–Provider Partnership
As discussed in Chapters 3 and 4, tools and approaches are needed to improve communication between patients and providers and to support patient decision making in the face of a large volume of complex information. Research is needed to develop such tools and approaches for populations at greatest risk (e.g., older adults; those of lower socioeconomic status; and those with comorbid conditions, including psychosocial distress and decreased cognition).
Development of Better Tools for Screening
The screening tools described in Chapter 4 all have somewhat different purposes. For example, the Distress Thermometer and its accompanying Problem List screen for generalized distress and identify emotional, spiritual, physical, family, and practical issues, such as problems with transportation, as sources of distress. PCM 2.0 focuses on role functioning and overall quality of life. Other types of psychosocial health needs not addressed by these instruments include, for example, social isolation/social support, difficulties in navigating the health system, and poor literacy. Health services research should focus on the development of psychosocial screening tools addressing a more comprehensive range of psychological and social stressors that can interfere with the ability of patients and families to manage cancer and its consequences.
Also needed are psychosocial screening tools and approaches for their use that consider the comorbidities frequently experienced by cancer survivors. Data from the 1998–2000 National Health Interview Survey, for example, show that 49 percent of cancer survivors are age 65 or older. Among cancer survivors aged 65 or older, 39 percent report having a diagnosis of heart disease, angina, heart attack, heart condition, or stroke; 13 percent report currently being treated for diabetes (Hewitt et al., 2003). These findings are significant because efforts are under way to understand psychosocial barriers to patients’ self-management of these conditions and to apply psychosocial screening instruments developed for individuals with these conditions (Glasgow et al., 2001; Whittemore et al., 2005). The Problem Areas in Diabetes (PAID) instrument, for example, is a 20-item self-administered measure of diabetes-specific emotional distress that has performed well in psychometric testing. It has been found to correlate strongly with “a wide range of theoretically related constructs such as general emotional distress, depression, diabetes self-care behaviors, diabetes coping, and health beliefs” and to be a statistically significant predictor of glycemic control (Polonsky et al., 2006). The Diabetes Self-Management Assessment Tool is another example.
The issue of comorbidity raises a number of questions. Hypothetically, should cancer patients be administered a separate psychosocial screening tool for each of their comorbid conditions? In addition to the burden that multiple screening instruments would likely impose on the patient, screening for psychosocial distress in cancer survivors might not be clinically feasible outside of oncology practices if medical providers had to implement multiple screens to address each patient’s unique multiple illnesses. Other questions arise as well. If a clinician cares for an individual with more than one chronic condition, which screening tool or tools should be used? Is there a minimum set of domains that should be included in all screening
tools for psychosocial health needs? Is there an ideal set? Could a tool be developed that would address all these domains? Should existing tools be improved to achieve greater utility?
Some of these questions would be addressed if there were a valid psychosocial screening instrument that could be used across multiple chronic conditions. One possibility would be to develop such an instrument along the lines of the NIH Patient-Reported Outcomes Measurement Information System (PROMIS) (http://www.nihpromis.org/default.asp). The goal of this initiative is to develop ways to measure patient-reported symptoms, such as pain and fatigue, and aspects of health-related quality of life across a wide variety of chronic diseases and conditions. The PROMIS initiative establishes a collaborative relationship between NIH and individual research teams in order to
Develop and test a large repository of items and questionnaires measuring patient-reported outcomes.
Build a Web-based resource for administering computerized adaptive tests, collecting self-report data, and reporting instant health assessments.
Evaluate the utility of PROMIS and promote widespread use of the instruments for clinical research and clinical care.
Sustain the repository and continued development of the PROMIS tools and system for clinical research and practice (NIH, undated).
The network will collaborate on the collection of self-reported data from diverse populations of individuals with a variety of chronic diseases using agreed-upon methods, modes, and questionnaires.
If a less sweeping initiative were desired with respect to screening instruments, research could address, for example, more testing of existing screening instruments, testing of the effectiveness of cancer-focused screening instruments for other chronic conditions and vice versa, and ways to incorporate quality-of-life measures into screening instruments. Use of screening instruments in conjunction with comprehensive needs assessment also should be addressed.
Development of Better Needs Assessment Instruments
Chapter 4 presents the results of a systematic appraisal (Wen and Gustafson, 2004) of needs assessment instruments for cancer that identifies 17 patient and 7 family instruments for which information is available on their reliability, validity, burden, and psychometric properties. Across these instruments, many problems are noted, including wide variation in the
needs addressed,1 inconsistency in domains within the instrument and in items included in similarly named domains, a lack of evidence of sensitivity to change over time, failure to examine reading levels, and failure to address the period after initial treatment for cancer. The authors express their doubt as to whether any one instrument could be developed to address all areas of interest, but recommend that a common set of domain terms be adopted to form the core of needs assessment and that agreement be reached on some items to be placed in the domains. Research also is recommended to address obstacles to the practice of needs assessment, to identify characteristics of effective performance of needs assessment, and to establish the relative importance and significance of identified needs. As with the questions posed above with respect to screening, questions about how best to conduct needs assessment in the presence of comorbidities also require attention.
Comprehensive Illness and Wellness Management Interventions
As described in Chapters 3 and 4, comprehensive illness self-management programs have been found to be effective in improving patient knowledge, skills, and confidence in managing a number of chronic illnesses, such as diabetes, asthma, heart disease, lung disease, stroke, and arthritis. Some of these programs also have been found to be effective in improving health outcomes (Lorig et al., 2001; Bodenheimer et al., 2002; Lorig and Holman, 2003; Chodosh et al., 2005). Yet while particular interventions have been developed and found to be effective in helping cancer patients manage individual symptoms, such as pain and fatigue, comprehensive illness and health management programs similar to those that exist for individuals with other chronic illnesses have not been developed and tested in individuals living with the diagnosis and sequelae of cancer. Research to this end is needed.
Approaches for Effectively Linking Patients with Services and Coordinating Care
Also as discussed in Chapter 4, the various mechanisms used to link patients with services delivered by different health and human service providers (e.g., structured referral arrangements and formal agreements with external providers, case management, collocation and clinical integration of services, patient navigators, use of shared electronic health records) have
varying levels of empirical support. This support does not come from studies of the use of these mechanisms by oncology providers in ambulatory care settings. Because most oncology patients receive their cancer care in outpatient settings, research comparing the effectiveness and cost of using different mechanisms to link patients to psychosocial services and coordinate their care could help inform and redesign oncology practices. Such research also could evaluate the use of different types of personnel (e.g., nurses and social workers with varying levels of education and training and unlicensed, trained workers such as patient care navigators) to perform linkage and coordination activities.
Reimbursement Arrangements That Promote Psychosocial Care
As illustrated in Chapter 6, little information exists outside of Medicare about how group purchasers and health plans provide for psychosocial health care in their contracts with each other and with health care providers. Although there are anecdotal reports of best practices by some health plans and providers, qualitative and quantitative research could better illuminate the reimbursement and other mechanisms used by leading health care providers to address psychosocial health services and their effects on clinicians, work design, and patients and on the delivery of effective psychosocial health care. Such research could also address to what extent mechanisms have been developed for reimbursing for higher-than-average care coordination needs.
Consistent with the above discussion, the committee recommends the following research agenda.
Recommendation: Research priorities. Organizations sponsoring research in oncology care should include the following areas among their funding priorities:
Further development of reliable, valid, and efficient tools and strategies for use by clinical practices to ensure that all patients with cancer receive care that meets the standard of psychosocial care. These tools and strategies should include
approaches for improving patient–provider communication and providing decision support to cancer patients;
screening instruments that can be used to identify individuals with any of a comprehensive array of psychosocial health problems;
needs assessment instruments to assist in planning psychosocial services;
illness and wellness management interventions; and
approaches for effectively linking patients with services and coordinating care.
Identification of more effective psychosocial services to treat mental health problems and to assist patients in adopting and maintaining healthy behaviors, such as smoking cessation, exercise, and dietary change. This effort should include
identifying populations for whom specific psychosocial services are most effective, and psychosocial services most effective for specific populations; and
development of standard outcome measures for assessing the effectiveness of these services.
Creation and testing of reimbursement arrangements that will promote psychosocial care and reward its best performance.
Research on the use of these tools, strategies, and services should also focus on how best to ensure delivery of appropriate psychosocial services to vulnerable populations, such as those with low literacy, older adults, the socially isolated, and members of cultural minorities.
As part of this study, NIH requested that the Institute of Medicine make recommendations for how the impact of this report could be evaluated. The committee believes that evaluation activities could be useful in promoting action on all the recommendations made in this report and in designing future studies. Accordingly, the committee makes the following recommendation.
Recommendation. Promoting uptake and monitoring progress. The National Cancer Institute/NIH should monitor progress toward improved delivery of psychosocial services in cancer care and report its findings on at least a biannual basis to oncology providers, consumer organizations, group purchasers and health plans, quality oversight organizations, and other stakeholders. These findings could be used to inform an evaluation of the impact of this report and each of its recommendations. Monitoring activities should make maximal use of existing data collection tools and activities.
This recommendation could be implemented using a variety of approaches. For example, to determine the extent to which patients with cancer receive
Conduct an annual, patient-level, process-of-care evaluation using a national sample and validated, reliable instruments, such as the Consumer Assessment of Healthcare Providers and Systems (CAHPS) instruments.
Add measures of the quality of psychosocial health care for patients (and families as feasible) to existing surveys, such as the Centers for Disease Control and Prevention’s Behavioral Risk Factor Surveillance System (BRFSS).
Conduct annual practice surveys to determine compliance with the standard of care.
Monitor and document the emergence of performance reward initiatives (e.g., content on psychosocial care in requests for proposals [RFPs] and pay-for-performance initiatives that specifically include incentives for psychosocial care).
For the committee’s recommendation on patient and family education (see Chapter 5), NCI could
Routinely query patient education and advocacy organizations about their efforts to educate patients with cancer and their family caregivers that they should expect, and request when necessary, cancer care that meets the standard of care recommended in this report.
Assess whether patients and caregivers show greater knowledge of how oncology providers should address their psychosocial needs (the standard of care) and whether they report more receipt of psychosocial health services as part of their cancer care. Surveys could be used to gather this information and would indicate the extent to which cancer care is meeting the standard of care.
Use an annual patient-level, process-of-care evaluation (such as CAHPS) to identify patient education experiences.
For the committee’s recommendation on dissemination and uptake of the standard of care (see Chapter 5), DHHS could determine whether NCI, the Centers for Medicare & Medicaid Services (CMS), and AHRQ had conducted demonstration projects and how they had disseminated the findings from those demonstrations.
For the committee’s recommendation on support from payers (see Chapter 6), NCI and/or advocacy, provider, or other interest groups could
Survey national organizations (e.g., America’s Health Insurance Plans, the National Business Group on Health) about their awareness of and/or advocacy activities related to the recommendations in this report and the initiation of appropriate reimbursement strategies/activities.
Monitor and document the emergence of performance reward initiatives (e.g., RFP content on psychosocial care, pay-for-performance that specifically includes incentives for psychosocial care).
Evaluate health plan contracts and state insurance policies for coverage, copayments, and carve-outs for psychosocial services.
Assess coverage for psychosocial services for Medicare beneficiaries.
For the committee’s recommendation on quality oversight (see Chapter 6), DHHS could
Examine the funding portfolios of NIH, CMS, AHRQ, and other public and private sponsors of quality-of-care research to evaluate the funding of quality measurement for psychosocial health care as part of cancer care.
Query organizations that set standards for cancer care (e.g., National Comprehensive Cancer Network, American Society of Clinical Oncology, American College of Surgeons’ Commission on Cancer, Oncology Nursing Society, American Psychosocial Oncology Society) and other standards-setting organizations (e.g., National Quality Forum, National Committee for Quality Assurance, URAC, the Joint Commission) to determine the extent to which they have
created oversight mechanisms used to measure and report on the quality of ambulatory cancer care (including psychosocial health care);
incorporated requirements for identifying and responding to psychosocial health care needs into their protocols, policies, and standards in accordance with the standard of care put forth in this report; and
used performance measures of psychosocial health care in their quality oversight activities.
For the committee’s recommendation on workforce competencies (see Chapter 7), DHHS could
Monitor and report on actions taken by Congress and federal agencies to support and fund the establishment of a Workforce
Development Collaborative on Psychosocial Care during Chronic Medical Illness.
Review board exams for oncologists and primary care providers to identify questions relevant to psychosocial care.
Review accreditation standards for educational programs used to train health care personnel to identify content requirements relevant to psychosocial care.
Review certification requirements for clinicians to identify those requirements relevant to psychosocial care.
Examine the funding portfolios of NIH, CMS, AHRQ, and other public and private sponsors of quality-of-care research to quantify the funding of initiatives aimed at assessing the incorporation of workforce competencies into education, training, and clinical practice and their impact on achieving the standard for psychosocial care.
For the committee’s recommendation on standardized nomenclature and research priorities (see Chapter 3 and this chapter, respectively), DHHS could
Report on NIH/AHRQ actions to develop a taxonomy and nomenclature for psychosocial health services.
Examine the funding portfolios of public and private research sponsors to assess whether funding priorities included the recommended areas.
Andrykowski, M. A., and S. L. Manne. 2006. Are psychological interventions effective and accepted by cancer patients? I. Standards and levels of evidence. Annals of Behavioral Medicine 32(2):93–97.
Antoni, M. H., S. K. Lutgendorf, S. W. Cole, F. S. Dhabhar, S. E. Sephton, P. G. McDonald, M. Stefanek, and A. K. Sood. 2006. The influence of bio-behavioral factors on tumor biology: Pathways and mechanisms. Nature Reviews. Cancer 6(3):240–248.
Bodenheimer, T., K. Lorig, H. Holman, and K. Grumbach. 2002. Patient self-management of chronic disease in primary care. Journal of the American Medical Association 288(19): 2469–2475.
Chodosh, J., S. C. Morton, W. Mojica, M. Maglione, M. J. Suttorp, L. Hilton, S. Rhodes, and P. Shekelle. 2005. Meta-analysis: Chronic disease self-management programs for older adults. Annals of Internal Medicine 143(6):427–438.
Cohen, S. 2004. Social relationships and health. American Psychologist 59(8):676–684.
Glasgow, R. E., D. J. Toobert, and C. D. Gillette. 2001. Psychosocial barriers to diabetes self-management and quality of life. Diabetes Spectrum 14(1):33–41.
Helgeson, V. S. 2005. Recent advances in psychosocial oncology. Journal of Consulting and Clinical Psychology 73(2):268–271.
Helgeson, V. S., S. Cohen, R. Schultz, and J. Yasko. 2000. Group support interventions for women with breast cancer: Who benefits from what? Health Psychology 19(2): 107–114.
Hewitt, M., J. H. Rowland, and R. Yancik. 2003. Cancer survivors in the United States: Age, health, and disability. Journal of Gerontology 58(1):82–91.
Lorig, K. R., and H. Holman. 2003. Self-management education: History, definition, outcomes, and mechanisms. Annals of Behavioral Medicine 26(1):1–7.
Lorig, K. R., P. Ritter, A. L. Stewart, D. S. Sobel, B. W. Brown, A. Bandura, V. M. Gonzalez, D. D. Laurent, and H. R. Holman. 2001. Chronic disease self-management program: 2-year health status and health care utilization outcomes. Medical Care 39(11): 1217–1223.
NIH (National Institutes of Health). undated. Primary goals of PROMIS. http://www.nihpromis.org/Web%20Pages/Goals.aspx?PageView=Shared (accessed September 12, 2007).
Patenaude, A. F., and M. J. Kupst. 2005. Psychosocial functioning in pediatric cancer. Journal of Pediatric Psychology 30(1):9–27.
Polonsky, W. H., B. J. Anderson, P. A. Lohrer, G. Welch, A. M. Jacobson, J. E. Aponte, and C. E. Schwartz. 2006. RCMAR measurement tools: Problem areas in diabetes (PAID). Resource Centers for Minority Aging Research (RCMAR). http://www.musc.edu/dfm/RCMAR/PAID.html (accessed January 9, 2007).
Stanton, A. L. 2005. How and for whom? Asking questions about the utility of psychosocial interventions for individuals diagnosed with cancer. Journal of Clinical Oncology 23(22):4818–4820.
Thacker, P. H., S. K. Lutgendorf, and A. K. Sood. 2007. The neuroendocrine impact of chronic stress on cancer. Cell Cycle 6(4):430–433.
Wen, K.-Y., and D. H. Gustafson. 2004. Needs assessment for cancer patients and their families. Health and Quality of Life Outcomes 2:11.
Whittemore, R., G. D. E. Melkus, and M. Grey. 2005. Metabolic control, self-management and psychosocial adjustment in women with type 2 diabetes. Journal of Clinical Nursing 14(2):195–203.
Zebrack, B., and L. Zeltzer. 2003. Quality of life issues and cancer survivorship. Current Problems in Cancer 27(4):198–211.