- Email: [email protected]
Improving recruitment to clinical trials for cancer in childhood
Personal View
Improving recruitment to clinical trials for cancer in childhood Kathy Pritchard-Jones, Mary Dixon-Woods, Marianne Naafs-Wilstra, Maria Grazia Valsecchi Lancet Oncol 2008; 9: 392–99 See Keynote Comment page 306 Section of Paediatric Oncology/ Children’s Unit, The Institute of Cancer Research and Royal Marsden NHS Foundation Trust, Sutton, Surrey, UK (Prof K Pritchard-Jones FRCPCH); Department of Health Sciences, University of Leicester, Leicester, UK (M Dixon-Woods DPhil); International Confederation of Childhood Cancer Parent Organisations (ICCCPO), Netherlands (M Naafs-Wilstra MA); and Department of Clinical Medicine and Prevention, University of Milano-Bicocca, Milan, Italy (M G Valsecchi PhD) Correspondence to: Prof Kathy Pritchard-Jones, Section of Paediatric Oncology/ Children’s Unit, The Institute of Cancer Research and Royal Marsden NHS Foundation Trust, Sutton, Surrey, UK kathy.pritchard-jones@icr. ac.uk
Clinical trials have underpinned progress in the treatment of cancer in childhood: about 75% of children newly diagnosed with cancer can expect to be long-term survivors. This success has been achieved because most children with cancer have participated in available clinical trials. This high level of engagement of the childhood-cancer community relies on the so-called therapeutic alliance that begins between doctors and families when a child is diagnosed with cancer. More research is needed to understand how to present most effectively the unfamiliar idea of a randomised clinical trial at a stressful time. High overall survival using current regimens presents challenges for future trial design—to secure further incremental increases in survival or to show survival equivalence from targeted agents that might have reduced side-effects. Small subgroups of children with cancers defined by molecular signatures mean that international recruitment is essential to do trials in a reasonable timeframe. Such collaboration across linguistic and cultural boundaries presents not only legal and regulatory hurdles, but also challenges the childhoodcancer research community to reappraise individual treatment preferences. The introduction of new paediatric regulations in Canada and the USA and Europe should encourage manufacturers of new anticancer drugs to lend support to clinical trials of cancer in childhood.
Introduction When a parent is faced with the diagnosis of a lifethreatening disease in their child, the normal reaction is shock and perhaps disbelief.1 Parents are desperate to give their child the best possible chance of cure and to understand what their child will have to endure. At this stressful time, the idea of entry into a clinical trial can be very challenging (figure 1). Importantly, clinical trials lie behind the story of remarkable scientific progress in the treatment of cancer in childhood. Sidney Farber’s 1948 study2 famously showed temporary remission of leukaemia in childhood— once almost invariably a fatal disease. Since then, clinical trials have addressed questions about the role of various intensities of chemotherapy, the need for radiotherapy, and the timing and extent of surgical treatment. For
Figure 1: The idea of randomisation in a clinical trial can be challenging for families of a child newly diagnosed with cancer
392
example, current survival from acute lymphoblastic leukaemia is associated with a successive series of collaborative clinical trials that spanned 35 years, during which 10-year survival increased from less than 10% to 80%.3 Treatment protocols for most cancers of childhood are highly complex, needing expert management by multidisciplinary teams. Overall, the results of highquality well-coordinated trials mean that about 75% of children in developed countries who are diagnosed with cancer will survive the illness.4 Despite the enviable success in the treatment of cancer in childhood, a pressing need for further research and development remains. Many children die from such illness, and reduction of mortality is an important goal. Moreover, survival is not the only objective in the study of cancer in childhood: the long-term and short-term iatrogenic effects of current cytotoxic drugs and radiotherapy present substantial hazards, including toxicity-related problems in later life. The need to identify and assess treatments that are safer and at least as efficacious as current treatments is ongoing. New therapeutic strategies that optimise existing treatment regimens and that use new agents are needed.5 Development of such strategies is likely to benefit from recent advances in cancer biology and from the new generation of molecularly targeted agents that are emerging in the treatment of adults with cancer. High rates of participation in clinical trials of patients and established networks of clinicians who care for children with cancer have made possible multicentre (and indeed national and international) collaborative clinical trials of rare diseases. Although more than 70% of children who are diagnosed with cancer currently enter national or international phase III clinical trials in most western European countries,6 recruitment and randomisation for different trials vary substantially. For example, UK studies of children with acute lymphoblastic http://oncology.thelancet.com Vol 9 April 2008
Personal View
leukaemia7 have recruited more than 85% of eligible children (with >90% accepting randomisation), whereas only 39% of eligible cases accepted randomisation in a concurrent trial in Wilms’ tumour.8 These variations raise important questions about how the quality of trial recruitment might be improved. Improvement does not necessarily mean maximisation of recruitment, but rather engaging families and patients as informed partners who can make appropriate choices. This Personal View focuses on the challenges associated with improving recruitment to phase III trials of cancer in childhood, which is the option generally offered to newly diagnosed patients. These trials usually compare current, best known treatment (ie, the standard group) against a modification of that treatment (ie, the experimental group), and are designed to test whether there is either improved efficacy or equal efficacy with reduced toxic effects. Phase III trials are designed to be low risk to participants in terms of not compromising the chance of cure, and the best interests of the child are prioritised above any research objectives. Although the interests of the child are of course prioritised in phase I and II trials, these early studies raise additional questions and require separate discussion.
Panel 1: Key issues for families of children with cancer in accepting a randomised clinical trial • • • • •
Understanding what is meant by a clinical trial Understanding that a trial means that their child is not being experimented on, but is being offered the best known treatment for their situation Understanding the questions that the trial is asking and what they mean for their child Understanding the relative risks of side-effects and keeping these in proportion Being asked to involve a competent child in decision-making when they might not wish to worry them about the possibility of non-cure or details of treatment sideeffects
Panel 2: Ways to improve recruitment to clinical trials of cancer in childhood • • • • • •
Ask a relevant question that is seen as important by families and physicians Keep the trial simple Provide easily understood information about clinical trials (eg, consider inclusion of graphics and a magazine-style layout) Reassure families that their child’s wellbeing is the doctor’s main aim Allow time for the child and family to digest the information and ask relevant questions Give feedback to the cancer community about trial results and future directions for research
Clinical trials of cancer in childhood: design and implementation It is essential for optimum recruitment that a clinical trial asks a question that is important to the eligible community, and that the design seems reasonable and acceptable both to physicians who will be recruiting and to the families of children who will be candidates (panels 1 and 2). The history of cancer in childhood highlights these issues.9 Throughout the 1960s, children with leukaemia in the UK were treated with antibiotics (and sometimes steroids) only, with invariably fatal outcomes. Despite encouraging results in the USA, until 1967 the UK Medical Research Council was reluctant to allow children to be treated with experimental drugs at high doses or with unpleasant toxic effects. This thinking was on the grounds that parents and nursing staff would object,9 while many paediatricians and haematologists thought that intensive multidrug chemotherapy was unwarranted.10 Although treatment efficacy (and the need for intensive treatment) has since been recognised in paediatric oncology, the general point that study designs must be acceptable to both recruiter and recruited remains salient. In adult trials, it has been recommended that qualitative research can help ensure relevance and acceptability to the target population. For example, a qualitative study of recruitment of men to a prostatecancer trial11 identified that trial information was being misinterpreted by potential participants; frequency of randomisation increased when clearer explanations were given. However, there is little activity of this kind in paediatric oncology. http://oncology.thelancet.com Vol 9 April 2008
Expected number of cases per year Acute lymphoblastic leukaemia
Expected number of deaths per year
372
71
Acute myeloid leukaemia
69
26
Hodgkin’s lymphoma
58
<2
Non-Hodgkin lymphoma
82
16
Medulloblastoma
70
31
155
33
Neuroblastoma
89
38
Retinoblastoma
43
2
Renal tumours
81
13
Astrocytoma
Hepatoblastoma
11
3
Osteosarcoma
31
10
Ewing’s sarcoma
22
6
Rhabdomyosarcoma
55
17
Malignant germ-cell tumours
31
2
Estimated from population-based registration and from mortality data for Great Britain12 and Europe.13
Table 1: Estimated annual number of new cases and number of deaths for the main types of cancer diagnosed before age 15 years in a European country with total population about 60 million
Clearly, trials must be well-designed and comply with principles of good clinical and scientific practice. When designing a trial, the number of children who can potentially be recruited to ensure sufficient statistical power must be anticipated. For most cancers of childhood, large numbers of participants are needed for recruitment into a randomised trial compared with the annual numbers of children that are diagnosed with cancer in an average-sized European country per year (tables 1 and 2).12–14 393
Personal View
Standard treatment
Total number of participants (events)†
5-year event-free survival (%)
Target difference in 5-year event-free survival (%)*
25
+20
167 (109)
25
+15
280 (189)
60
+10
711 (249)
60
+5
2897 (1086)
90
–4
2077 (249)
90
–3
3554 (409)
*Target difference in 5-year event-free survival to be detected based on log-rank test (one-sided test, type I error 0·025 and 0·80 power).14 †Number of participants to be enrolled in each treatment group is half the total number of participants.
Table 2: Number of participants needed in superiority trials and non-inferiority trials with two parallel groups of equal allocation and comparison of event-free survival
100
Survival (%)
80 1996–2000 (n=73) 1991–95 (n=65) 1986–90 (n=62) 1981–85 (n=51) 1976–80 (n=48) 1971–75 (n=43) 1966–70 (n=43)
60
40
20
0 0
5
10
15
20
25
30
35
40
Time since diagnosis (years)
Figure 2: Rapid increase in survival of children with liver cancer (hepatoblastoma and hepatocellular carcinoma; average 14 cases per year) in Great Britain after introduction of participation in SIOPEL clinical trials in 1991 Reproduced with permission from reference 12.
Optimum sample size in clinical trials depends strongly on the target difference the trial aims to identify: the smaller the difference, the greater the number of participants needed for enrolment. The required sample size can reach thousands of patients when the study is powered to detect less than a 10% absolute increase in event-free survival. Such relatively modest, but clinically important, improvements are to be expected with more efficient definition of therapeutic strategies that use existing chemotherapy drugs. Very large numbers of participants are needed for noninferiority trials. These studies are typically done when the aim is to reduce treatment burden, toxic effects, and sequelae for patients who have a favourable risk profile, accepting a small risk of slight inferiority in event-free 394
survival (although the few patients who relapse are expected to be treated successfully with more-intensive second-line treatment to maintain overall survival). By contrast, patients who have a poor outlook need new experimental therapeutic strategies and new drugs in hope of more marked improvements. The sample size needed (in the hundreds) is difficult to achieve because these high-risk cancers generally represent small subgroups. Since the 1970s, many countries have developed national childhood cancer study groups, which have stimulated the development of tumour-specific committees to design and run clinical trials. High frequency of participation in national clinical trials has been achieved through the active engagement of clinicians and patients in specialist treatment centres and through national consensus. However, even for the more common cancers of childhood, participation from several countries might be needed to achieve the numbers for a robustly powered trial in a reasonable timeframe (usually 5 years’ recruitment). This need adds to the time required to design and implement such trials and can mean national compromise in terms of accepting the definition of the standard-treatment group. Nonetheless, expansion of recruitment from national to international communities has been mainly successful, although it has sometimes needed an innovative approach. For example, a study design that used prospectively the principles of meta-analysis was adopted in an international trial of intensification of maintenance treatment for acute lymphoblastic leukaemia.15,16 For rare cancers of childhood or for specific subgroups, international cooperation is often the only way to produce sound evidence and to improve clinical practice. Thus, when the SIOPEL liver tumour group of the International Society of Paediatric Oncology (SIOP) commenced clinical trials in patients with the very rare tumour hepatoblastoma, survival increased in participating countries (figure 2).12 The issues raised by the precise characterisation of subgroups in paediatric oncology continue to grow in importance. Advances in the understanding of the molecular basis of many cancers of childhood and the development of more sensitive methods for assessment of response has led to the subdivision of common cancers into subgroups with different survival profiles. For example, acute lymphoblastic leukaemia can be subdivided into groups such as infants or patients who are positive for Philadelphia chromosome—each group has its own therapeutic questions.17 New statistical methods to trial design are needed to address important therapeutic questions, including that of how to increase efficacy for subgroups with poor survival. Optimisation of the design of phase I and II studies is also needed, in view of the limited number of patients available and the growing number of new drugs that must be assessed for toxic effects and efficacy. These questions have http://oncology.thelancet.com Vol 9 April 2008
Personal View
stimulated the development of adaptive designs, which might be most useful when drug activity is assessed in terms of early response to treatment, rather than in phase III studies where efficacy is measured by long-term assessment of event-free survival and overall survival.18 A Bayesian framework could be used to design phase II trials more efficiently by thinking of them as randomised selection trials that are comparing several treatments.19 Thus, in subgroups of patients with a rare cancer, it has been proposed to combine at the design stage the expected outcome of the new small trial being planned with the findings generated by earlier studies. Such a Bayesian approach estimates the potential contribution to the demonstration of treatment effect of the new trial which, if seen in isolation, would be underpowered.20 Medicines regulatory agencies have issued guidelines on how to run clinical trials in small populations, which show how non-standard designs are possible. These guidelines21 propose that there are no special methods for the design, conduct, or analysis of clinical trials in small populations, but note that the need for statistical efficiency should be weighed against the need for clinically relevant, interpretable results—the latter being more important. Even with innovative trial design and participation from major childhood-cancer research groups worldwide, as exemplified by the SIOPEL trials, asking randomised questions across different countries is very challenging. Multinational trials need additional time for design and implementation. Ensuring consistent data quality when working across national, cultural, legal, and linguistic boundaries adds further complexity.22 European Union directive 2001/20/EC23 introduced into the national law of each member state in May, 2004, aims to give a common set of regulations and definitions for standards of quality of clinical studies across member states. However, the directive has had several undesirable effects, including increased cost and bureaucracy for non-commercial trials done by universities and public research institutions. Specific requirements, such as the need for a single European trial sponsor, have created difficulties in the understanding of international trials—both within the European Union and for collaborations with cooperative groups worldwide.24
Trial recruitment Ethical issues An ongoing dilemma in clinical research is the desire for patients to benefit from progress in medical care realised by scientific research while avoiding the risk of harm from such research for any individual. The overarching ethical principle is that physicians must act in the best interests of every child. If there is substantial uncertainty—ie, clinical equipoise25—over which treatment is likely to be best for a child, it is not only defensible to offer the child the opportunity of a clinical trial, but it might also be the most ethically sound course of action.26 http://oncology.thelancet.com Vol 9 April 2008
The finding that physicians can assure families of equipoise in phase III randomised trials of cancer in childhood might be helpful,27 although the replacement of conventional cytotoxics that have proven effectiveness with new biologically targeted treatments that have fewer toxic effects is an emerging challenge. Current bioethics stresses the importance of informed consent to clinical trials. There are substantial challenges to securing informed consent in studies of cancer in childhood. Concern about children’s vulnerability underlies the bioethical standpoint on their participation in scientific research,28,29 and there are issues with the application of bioethical principles for adults to children. Ensuring that research participants have sufficient information to make a decision about whether to participate and that they have the capacity to understand the consequences of their decision is problematic in any area of research, but especially so when children are involved.
Bayesian framework A procedure used to improve a statistical model in the light of observed data
Physician perspective Physician-based factors are likely to affect trial recruitment, as exemplified by research in adult trials.30–32 Recruitment in trials of cancer in childhood raises distinct issues33 and additional complexities.34 Despite strong ethical arguments in favour of running clinical trials in situations of equipoise, doctors might have conflict and concern when assuming the dual role of caring clinician and researcher,35,36 particularly in earlyphase trials.37,38 Other possible factors that affect physicians’ willingness to recruit might include their level of comfort with randomisation and adherence to trial protocols; research of adult participants suggests that some physicians think these features of trials might undermine their ability to make individual treatment decisions.39,40 These issues might be especially important in multinational trials that involve national compromises in terms of what should be included in the standard group of the trial. Physicians face key challenges once they have made a decision to approach the family of a child with cancer about possible involvement in a trial. They must: introduce the possibility of a clinical trial to patients and parents, for whom it is probably an entirely unfamiliar idea or one that has unwelcome associations with passive experimentation; explain complex treatment protocols; and secure cooperation and trust under the time pressures imposed by the need to begin treatment.41 Consultations such as those during the diagnosis of a child with cancer or during treatment discussions might be understood as so-called focused encounters.42 Here, a specific matter (eg, how the child can be cured) is the focus of attention. The introduction of an idea such as research, which might seem irrelevant to families and a violation of the implicit rules that govern such encounters, can cause substantial discomfort.39 It is unfortunate that a controversial history in the conduct of medical research might provoke anxiety about 395
Personal View
the motives of doctors in recruitment to research or fears that the child will be treated as a guinea pig. Of pivotal importance is reassurance for families that their child’s wellbeing is the doctor’s main aim, that children are not viewed as a research resource, and that families retain the autonomy to make individualised treatment decisions if unexpected toxic effects occur. Many difficulties are compounded by linguistic, cultural, and religious differences among potential participants, particularly in the context of recent migration and poverty.43,44 Nevertheless, research is being informed by socialscience theory and methods that can help give an evidence-base for enhanced management of these processes by physicians,45 although much work remains to be done.
Family perspective The engagement of families as informed partners in clinical trials is a key goal of sound practice, but is far from straightforward in its achievement. A key issue is whether to involve a child with cancer in the decision about participating in a trial. The principle that children should be consulted, informed, and involved in decisions that affect them, but not have final authority over decisions, is stipulated in the UN Convention of the Rights of the Child (UNCRC). The guiding principle of the UNCRC is that the “best interests” of the child must be promoted, but differences of opinion between parents, children, and professionals about the “best interests” of the child will sometimes need to be resolved. Involvement of a child in a decision about participation in a cancer trial might be especially complex when parents who are struggling to accept the diagnosis prefer to defer telling the child the full diagnosis and its implications,46 perhaps until treatment is established and the parents are better able to cope with their emotions. Some parents might see a demand to involve the child in the decision—and by definition disclosure of the diagnosis—as interfering with their parental autonomy. Further complexity arises because of varying definitions of minors and of the legal age of consent between different countries; some jurisdictions rely on the assessment of competence rather than specification of a particular age of consent. How information should be communicated and how much understanding is necessary to qualify for informed consent poses further challenges. A US study47 found that although randomisation was explained in most consultations with families, 50% of parents did not subsequently show an understanding of it. Furthermore, longitudinal studies have identified that parents do not gain much further understanding over time.48 These findings are not unique to cancer in childhood: adult patients have a poor understanding of the idea and scientific benefits of randomisation.49 So-called misunderstanding by patients of ideas such as randomisation are commonly assumed to arise from 396
incomplete, badly written, or misleading information. Much attention, particularly from research ethics committees, tends to focus on the detail of patient information sheets. However, the role of written information in this context remains poorly understood, and parents might find discussion with staff more helpful than a consent document.50 Principles of honesty, clear writing, and presentation are of paramount importance. However, even the clearest writing is unlikely to give a full solution, and much of the current approach might rely on a defective and simplistic model of communication. This so-called deficit model assumes that misunderstanding is a technical issue to be resolved by leaflet design and full disclosure. A wide-ranging critique of this approach in the social sciences has shown how it fails to recognise the dynamic nature of the interaction between the text and individual readers, who bring many diverse competences, motivations, and knowledge to the task of reading.51 It is unsurprising that educational interventions to improve understanding of randomisation have proved disappointing.52 Current approaches to ethics need consultations about trial recruitment to take what sociologists would identify as a specific ritualised format of interactions between clinicians and families and to achieve certain recorded outcomes, such a signed consent form. However, clinical situations are commonly at odds with ethical or legal ideals. For example, the requirement that discussions take place at particular times and in a particular order poses substantial problems in practice. Families are frequently shocked and distressed at the diagnosis1 and might have limited ability to absorb or act on information in the first few days. The physician is left with a dilemma because trials need consent before treatment starts (for a day-1 randomisation or new treatment approach), or within a short defined period (for initial treatment that does not differ from standard practice). A more radical approach to resolving the issues associated with trial recruitment in this area is to argue that the current approach to information provision in clinical trials, where there is genuine uncertainty about which treatment is better, is simply wrong. Evans and colleagues53 in their influential book, Testing treatments, point to the discrepancy between the UK General Medical Council’s guidance on communicating with patients in routine practice, which recommends that the appropriate amount of information will vary depending on factors including the patient’s own wishes, and guidance on consent to research, which states that people who are asked to participate in clinical trials must be given “the fullest possible information”.53 Evans and colleagues53 propose the need for a flexible approach, which recognises that trust between doctor and patient is the bedrock. They suggest that people who are invited to join a trial should be given every opportunity of easy access to the fullest possible information, but that it would be up to them how much information they wanted before making a decision. http://oncology.thelancet.com Vol 9 April 2008
Personal View
Alongside such approaches to informing patients, which are the focus of ongoing debate, different trial designs should be assessed, including starting patients on standard treatment and introducing the idea of randomisation later. Any future research in this area must continue to acknowledge that the goal will never be maximisation of recruitment by persuading families to enroll children in clinical trials; indeed, provision of easily understood information might result in fewer enrolments in trials. Rather, parents and children should be engaged as partners, decision-making should be facilitated in the most supportive manner, and the best interests of every child secured.
Other factors Many organisational factors affect trial recruitment. The time and resource implications of trial management, strict eligibility criteria, and complex monitoring and follow-up can affect patient accrual to clinical trials in adult oncology,54 and there are similar issues in paediatrics. Compliance with all regulatory requirements for trials of medicines in a paediatric setting imposes a substantial overhead that many health services and individual treatment centres struggle to meet. Consequently, the decreasing number of open trials available for patients is an important concern.55 The bureaucracy and expense of pharmacovigilance in trials could be reduced if medicines regulators had a more flexible approach to the definition of “investigational medicinal products” in clinical trials. Medicines regulators should recognise that many drugs or treatment combinations that are used in phase III clinical trials of cancer in childhood are established with well-documented side-effects and safety profiles, even though individual drugs might be used outside the licensed indication because there is no, or very limited, paediatric data in their marketing authorisation. This issue is being addressed in an ongoing consultation between the European Commission and the noncommercial research community. Researchers also have some responsibility: they should define more clearly what is meant by “expected” toxicity, with the aim of reducing the number of episodes that need expedited reporting. Clinical trials of new drugs might be facilitated by the introduction of new regulations relevant to children in Canada and the USA (in 2003)56 and in Europe (in 2007).57 These regulations mandate submission of paediatric investigation plans for new drugs that have potential benefit for children’s diseases. This requirement might encourage pharmaceutical companies to engage with the academic community or public sector, which have traditionally led the way in paediatric oncology, to lend support to the clinical trials that will be necessary to incorporate the new generation of biologically targeted anticancer drugs and other therapies into treatment for cancer in childhood. For non-commercial partners, such http://oncology.thelancet.com Vol 9 April 2008
Search strategy and selection criteria This Personal View reflects the personal experience of the authors. Relevant research articles were identified from the authors’ knowledge of the subject and by searching PubMed by use of the keywords “clinical trials”, “recruitment”, “randomisation”, “cancer”, “children”, and “childhood”. Further papers were chosen from the reference lists of relevant articles.
dialogue is expected to increase access to new drugs and to the long established infrastructure for trial management in industry. The availability of clinical trials of cancer in childhood might improve, providing that the noncommercial sector has the necessary resources; however, the possible effects on trial recruitment are unknown.
Conclusion Clinical trials of cancer in childhood have substantially improved the survival and wellbeing of thousands of children worldwide. National groups who develop and apply treatment protocols, whether randomised or not, and evaluate and publish their results lend support to a culture of evidence-based treatments. Children who are treated in such settings are ensured a scientifically sound approach to care and are generally offered treatment under conditions of best clinical practice. The tendency to define clinical practice by working consistently over many years in national groups is not yet common practice in most adult cancers. This factor should be kept in mind when addressing the debate, still very open, on whether patients who are enrolled in oncology clinical trials have better outcomes than those who are not.58,59 Further research is needed: into how people respond to approaches by clinicians to consent a child to enrol in a trial; to identify how physicians can manage the informedconsent process in an optimum way; and to explore the role of written information. Such work should recognise that families might wish to engage in a therapeutic alliance of mutual trust and collaboration from the outset of their child’s treatment. There is little room for complacency: if progress is to be sustained, formidable challenges must be faced. As we have identified, these challenges include the scientific study of small populations, the complexities of the regulatory environment, the ethical and organisational constraints, the emotional context, and the attitudes of individual families and physicians. Some issues may be addressed by innovative trial design and implementation, but all will need engagement with the communities of families and physicians on whom trials depend. Clinical research must include a well-designed social science and ethics assessment of how best to inform and engage the patient community who need clinical trials to continue for progress to be made. Conflicts of interest The authors declared no conflicts of interest.
397
Personal View
References 1 Dixon-Woods M, Findlay M, Young B, et al. Parents’ accounts of obtaining a diagnosis of childhood cancer. Lancet 2001; 357: 670–74. 2 Farber S, Diamond LK, Mercer RD. Temporary remissions in acute leukemia in children produced by folic acid antagonist, 4-aminopteroylglutamic acid (aminopterin). N Engl J Med 1948; 238: 787–93. 3 Reaman GH. Pediatric cancer research. From past success through collaboation to future transdisciplinary research. J Pediatr Oncol Nurs 2004; 21: 123–27. 4 Pritchard-Jones K, Kaatsch P, Steliarova-Foucher E, et al. Cancer in children and adolescents in Europe: developments over 20 years and future challenges. Eur J Cancer 2006; 42: 2183–90. 5 Estlin EJ. Novel targets for therapy in paediatric oncology. Curr Drug Targets Immune Endocr Metabol Disord 2002; 2: 141–50. 6 Ablett S, Pinkerton CR, United Kingdom Children’s Cancer Study Group. Recruiting children into cancer trials—role of the United Kingdom Children’s Cancer Study Group (UKCCSG). Br J Cancer 2003; 88: 1661–65. 7 Hann I, Vora A, Harrison G, et al. Determinants of outcome after intensified therapy of childhood lymphoblastic leukaemia: results from Medical Research Council United Kingdom acute lymphoblastic leukaemia XI protocol. Br J Haematol 2001; 113: 103–14. 8 Mitchell C, Pritchard-Jones K, Shannon R, et al. Immediate nephrectomy versus preoperative chemotherapy in the management of non-metastatic Wilms’ tumour: results of a randomised trial (UKW3) by the UK Children’s Cancer Study Group. Eur J Cancer 2006; 42: 2554–62. 9 Barnes E. Between remission and cure: patients, practitioners and the transformation of leukaemia in the late twentieth century. Chronic Illn 2007; 3: 253–64. 10 Anon. Acute leukaemia. Lancet 1965; 2: 939. 11 Donovan J, Mills N, Smith M, et al. Improving the design and conduct of trials by embedding them in qualitative research: ProctecT study. BMJ 2002; 325: 766–70. 12 Stiller C, ed. Childhood cancer in Britain. London: Oxford University Press, 2007. 13 Steliarova-Foucher E, Coebergh JW, Kaatsch P, et al, eds. Cancer in children and adolescents in Europe. Eur J Cancer (Special Issue) 2006; 42: 1913–2190. 14 Machin D, Campbell M, Fayers P, Pinol A. Sample size tables for clinical studies, 2nd edn. Malden: Blackwell Science, 1997. 15 Conter V, Valsecchi MG, Silvestri D, et al. Pulses of vincristine and dexamethasone in addition to intensive chemotherapy for children with intermediate-risk acute lymphoblastic leukaemia: a multicentre randomised trial. Lancet 2007; 369: 123–31. 16 Valsecchi MG, Masera G. A new challenge in clinical research in childhood ALL: the prospective meta-analysis strategy for intergroup collaboration. Ann Oncol 1996; 7: 1005–08. 17 Pieters R, Schrappe M, De Lorenzo P, et al. A treatment protocol for infants younger than 1 year with acute lymphoblastic leukaemia (Interfant-99): an observational study and a multicentre randomised trial. Lancet 2007; 370: 240–50. 18 Golub HL. The need for more efficient trial designs. Stat Med 2006; 25: 3231–35. 19 Estey EH, Thall PF. New designs for phase 2 clinical trials. Blood 2003; 102: 442–48. 20 Tan SB, Dear KB, Bruzzi P, Machin D. Strategy for randomised clinical trials in rare cancers. BMJ 2003; 327: 47–49. 21 European Medicines Agency (EMEA). Committee for medicinal products for human use. Guideline on clinical trials in small populations CHMP/EWP/83561/2005. http://www.emea.europa.eu/ pdfs/human/ewp/8356105en.pdf (accessed Jan 22, 2008). 22 Valsecchi MG, Silvestri D, Covezzoli A, De Lorenzo P. Web-based international studies in limited populations of pediatric leukemia. Pediatr Blood Cancer 2008; 50: 270–73. 23 The European Parliament and the Council of the European Union. Directive 2001/20/EC of the European Parliament and of the council of 4 April 2001 on the approximation of the laws, regulations and administrative provisions of the member states relating to the implementation of good clinical practice in the conduct of clinical trials on medicinal products for human use. http://eur-lex.europa.eu/LexUriServ/LexUriServ. do?uri=CELEX:32001L0020:EN:HTML (accessed Jan 22, 2008).
398
24 25
26 27
28 29
30
31
32
33
34
35
36
37
38
39
40 41 42 43
44
45
46
47 48
Hoey R. The EU clinical trial directive: 3 years on. Lancet 2007; 369: 1777–78. Djulbegovic B. Acknowledgment of uncertainty: a fundamental means to ensure scientific and ethical validity in clinical research. Curr Oncol Rep 2001; 3: 389–95. Chalmers I. Regulation of therapeutic research is compromising the interests of patients. Int J Pharm Med 2007; 21: 395–404. Kumar A, Soares H, Wells R, et al. Are experimental treatments for cancer in children superior to established treatments? Observational study of randomised controlled trials by the Children’s Oncology Group. BMJ 2005; 331: 1295. Field MJ, Behrman RE. Responsible research involving children. Ambul Pediatr 2005; 5: 47–49. Dixon-Woods M, Young B, Ross E. Researching chronic childhood illness: the example of childhood cancer. Chronic Illn 2006; 2: 165–77. Mansour EG. Barriers to clinical trials. Part III: knowledge and attitudes of health care providers. Cancer 1994; 74 (9 suppl): 2672–75. Mannel RS, Walker JL, Gould N, et al. Impact of individual physicians on enrolment of patients into clinical trials. Am J Clin Oncol 2003; 26: 171–73. Ellis PM. Attitudes towards and participation in randomised clinical trials in oncology: a review of the literature. Ann Oncol 2000; 11: 939–45. Simon CM, Siminoff LA, Kodish ED, Burant C. Comparison of the informed consent process for randomized clinical trials in pediatric and adult oncology. J Clin Oncol 2004; 22: 2708–17. Estlin EJ, Ablett S. Practicalities and ethics of running clinical trials in paediatric oncology—the UK experience. Eur J Cancer 2001; 37: 1394–98. Frayne SM, Mancuso M, Prout MN, Freund KM. Attitudes of primary care physicians toward cancer-prevention trials: a focus group analysis. J Natl Med Assoc 2001; 93: 450–57. Estlin EJ, Cotterill S, Pratt CB, et al. Phase I trials in pediatric oncology: perceptions of pediatricians from the United Kingdom Children’s Cancer Study Group and the Pediatric Oncology Group. J Clin Oncol 2000; 18: 1900–05. Mick R, Lane N, Daugherty C, Ratain MJ. Physician-determined patient risk of toxic effects: impact on enrollment and decision making in phase I cancer trials. J Natl Cancer Inst 1994; 86: 1685–93. Daugherty C, Ratain MJ, Grochowski E, et al. Perceptions of cancer patients and their physicians involved in phase I trials. J Clin Oncol 1995; 13: 1062–72. Taylor KM, Feldstein ML, Skeel RT, et al. Fundamental dilemmas of the randomized clinical trial process: results of a survey of the 1737 Eastern Cooperative Oncology Group investigators. J Clin Oncol 1994; 12: 1796–805. Fallowfield L, Ratcliffe D, Souhami R. Clinicians’ attitudes to clinical trials of cancer therapy. Eur J Cancer 1997; 33: 2221–29. Simon C, Eder M, Raiz P, et al. Informed consent for pediatric leukemia research: clinician perspectives. Cancer 2001; 92: 691–700. Goffman E. Encounters: two studies in the sociology of interaction. Indianapolis: Bobbs-Merill, 1961. Simon CM, Zyzanski SJ, Durand E, et al. Interpreter accuracy and informed consent among Spanish-speaking families with cancer. J Health Commun 2006; 11: 509–22. Abbe M, Simon C, Angiolillo A, et al. A survey of language barriers from the perspective of pediatric oncologists, interpreters, and parents. Pediatr Blood Cancer 2006; 47: 819–24. Eder ML, Yamokoski AD, Wittmann PW, Kodish ED. Improving informed consent: suggestions from parents of children with leukemia. Pediatrics 2007; 119: e849–59. Young B, Dixon-Woods M, Windridge K, Heney D. Managing communication with young people who have a potentially life threatening chronic illness: qualitative study of patients and parents. BMJ 2003; 326: 305–10. Kodish E, Eder M, Noll RB, et al. Communication of randomization in childhood leukemia trials. JAMA 2004; 291: 470–75. Greenley RN, Drotar D, Zyzanski SJ, Kodish E. Stability of parental understanding of random assignment in childhood leukemia trials: an empirical examination of informed consent. J Clin Oncol 2006; 24: 891–97.
http://oncology.thelancet.com Vol 9 April 2008
Personal View
49
50
51
52
53 54
55
Ellis PM, Dowsett SM, Butow PN, Tattersall MH. Attitudes to randomized clinical trials amongst out-patients attending a medical oncology clinic. Health Expect 1999; 2: 33–43. Kodish ED, Pentz RD, Noll RB, et al. Informed consent in the Childrens Cancer Group: results of preliminary research. Cancer 1998; 82: 2467–81. Dixon-Woods M. Writing wrongs? An analysis of published discourses about the use of patient information leaflets. Soc Sci Med 2001; 52: 1417–32. Robinson EJ, Kerr CE, Stevens AJ, et al. Lay public’s understanding of equipoise and randomisation in randomised controlled trials. Health Technol Assess 2005; 9: 1–192, iii–iv. Evans I, Thornton H, Chalmers I. Testing treatments: better research for better healthcare. London: British Library, 2006. Langley C, Gray S, Selley S, et al. Clinicians’ attitudes to recruitment to randomised trials in cancer care: a qualitative study. J Health Serv Res Policy 2000; 5: 164–69. Hearn J, Sullivan R. The impact of the ‘Clinical Trials’ directive on the cost and conduct of non-commercial cancer trials in the UK. Eur J Cancer 2007; 43: 8–13.
http://oncology.thelancet.com Vol 9 April 2008
56
57
58
59
Official Journal of the European Union. Regulation (EC) no 1901/2006 of the European Parliament and of the Council of 12 December 2006 on medicinal products for paediatric use and amending regulation (EEC) No 1768/92, Directive 2001/20/EC, Directive 2001/83/EC and Regulation (EC) No 726/2004. http:// ec.europa.eu/enterprise/pharmaceuticals/eudralex/vol-1/reg_2006_ 1901/reg_2006_1901_en.pdf (accessed Jan 22, 2008). US Food and Drug Administration. Best pharmaceuticals for Children Act. Jan 4, 2002. http://www.fda.gov/opacom/laws/ pharmkids/contents.html (accessed Jan 22, 2008). Braunholtz DA, Edwards SJ, Lilford RJ. Are randomized clinical trials good for us (in the short term)? Evidence for a “trial effect”. J Clin Epidemiol 2001; 54: 217–24. Peppercorn JM, Weeks JC, Cook EF, Joffe S. Comparison of outcomes in cancer patients treated within and outside clinical trials: conceptual framework and structured review. Lancet 2004; 363: 263.
399
Improving recruitment to clinical trials for cancer in childhood Kathy Pritchard-Jones, Mary Dixon-Woods, Marianne Naafs-Wilstra, Maria Grazia Valsecchi Lancet Oncol 2008; 9: 392–99 See Keynote Comment page 306 Section of Paediatric Oncology/ Children’s Unit, The Institute of Cancer Research and Royal Marsden NHS Foundation Trust, Sutton, Surrey, UK (Prof K Pritchard-Jones FRCPCH); Department of Health Sciences, University of Leicester, Leicester, UK (M Dixon-Woods DPhil); International Confederation of Childhood Cancer Parent Organisations (ICCCPO), Netherlands (M Naafs-Wilstra MA); and Department of Clinical Medicine and Prevention, University of Milano-Bicocca, Milan, Italy (M G Valsecchi PhD) Correspondence to: Prof Kathy Pritchard-Jones, Section of Paediatric Oncology/ Children’s Unit, The Institute of Cancer Research and Royal Marsden NHS Foundation Trust, Sutton, Surrey, UK kathy.pritchard-jones@icr. ac.uk
Clinical trials have underpinned progress in the treatment of cancer in childhood: about 75% of children newly diagnosed with cancer can expect to be long-term survivors. This success has been achieved because most children with cancer have participated in available clinical trials. This high level of engagement of the childhood-cancer community relies on the so-called therapeutic alliance that begins between doctors and families when a child is diagnosed with cancer. More research is needed to understand how to present most effectively the unfamiliar idea of a randomised clinical trial at a stressful time. High overall survival using current regimens presents challenges for future trial design—to secure further incremental increases in survival or to show survival equivalence from targeted agents that might have reduced side-effects. Small subgroups of children with cancers defined by molecular signatures mean that international recruitment is essential to do trials in a reasonable timeframe. Such collaboration across linguistic and cultural boundaries presents not only legal and regulatory hurdles, but also challenges the childhoodcancer research community to reappraise individual treatment preferences. The introduction of new paediatric regulations in Canada and the USA and Europe should encourage manufacturers of new anticancer drugs to lend support to clinical trials of cancer in childhood.
Introduction When a parent is faced with the diagnosis of a lifethreatening disease in their child, the normal reaction is shock and perhaps disbelief.1 Parents are desperate to give their child the best possible chance of cure and to understand what their child will have to endure. At this stressful time, the idea of entry into a clinical trial can be very challenging (figure 1). Importantly, clinical trials lie behind the story of remarkable scientific progress in the treatment of cancer in childhood. Sidney Farber’s 1948 study2 famously showed temporary remission of leukaemia in childhood— once almost invariably a fatal disease. Since then, clinical trials have addressed questions about the role of various intensities of chemotherapy, the need for radiotherapy, and the timing and extent of surgical treatment. For
Figure 1: The idea of randomisation in a clinical trial can be challenging for families of a child newly diagnosed with cancer
392
example, current survival from acute lymphoblastic leukaemia is associated with a successive series of collaborative clinical trials that spanned 35 years, during which 10-year survival increased from less than 10% to 80%.3 Treatment protocols for most cancers of childhood are highly complex, needing expert management by multidisciplinary teams. Overall, the results of highquality well-coordinated trials mean that about 75% of children in developed countries who are diagnosed with cancer will survive the illness.4 Despite the enviable success in the treatment of cancer in childhood, a pressing need for further research and development remains. Many children die from such illness, and reduction of mortality is an important goal. Moreover, survival is not the only objective in the study of cancer in childhood: the long-term and short-term iatrogenic effects of current cytotoxic drugs and radiotherapy present substantial hazards, including toxicity-related problems in later life. The need to identify and assess treatments that are safer and at least as efficacious as current treatments is ongoing. New therapeutic strategies that optimise existing treatment regimens and that use new agents are needed.5 Development of such strategies is likely to benefit from recent advances in cancer biology and from the new generation of molecularly targeted agents that are emerging in the treatment of adults with cancer. High rates of participation in clinical trials of patients and established networks of clinicians who care for children with cancer have made possible multicentre (and indeed national and international) collaborative clinical trials of rare diseases. Although more than 70% of children who are diagnosed with cancer currently enter national or international phase III clinical trials in most western European countries,6 recruitment and randomisation for different trials vary substantially. For example, UK studies of children with acute lymphoblastic http://oncology.thelancet.com Vol 9 April 2008
Personal View
leukaemia7 have recruited more than 85% of eligible children (with >90% accepting randomisation), whereas only 39% of eligible cases accepted randomisation in a concurrent trial in Wilms’ tumour.8 These variations raise important questions about how the quality of trial recruitment might be improved. Improvement does not necessarily mean maximisation of recruitment, but rather engaging families and patients as informed partners who can make appropriate choices. This Personal View focuses on the challenges associated with improving recruitment to phase III trials of cancer in childhood, which is the option generally offered to newly diagnosed patients. These trials usually compare current, best known treatment (ie, the standard group) against a modification of that treatment (ie, the experimental group), and are designed to test whether there is either improved efficacy or equal efficacy with reduced toxic effects. Phase III trials are designed to be low risk to participants in terms of not compromising the chance of cure, and the best interests of the child are prioritised above any research objectives. Although the interests of the child are of course prioritised in phase I and II trials, these early studies raise additional questions and require separate discussion.
Panel 1: Key issues for families of children with cancer in accepting a randomised clinical trial • • • • •
Understanding what is meant by a clinical trial Understanding that a trial means that their child is not being experimented on, but is being offered the best known treatment for their situation Understanding the questions that the trial is asking and what they mean for their child Understanding the relative risks of side-effects and keeping these in proportion Being asked to involve a competent child in decision-making when they might not wish to worry them about the possibility of non-cure or details of treatment sideeffects
Panel 2: Ways to improve recruitment to clinical trials of cancer in childhood • • • • • •
Ask a relevant question that is seen as important by families and physicians Keep the trial simple Provide easily understood information about clinical trials (eg, consider inclusion of graphics and a magazine-style layout) Reassure families that their child’s wellbeing is the doctor’s main aim Allow time for the child and family to digest the information and ask relevant questions Give feedback to the cancer community about trial results and future directions for research
Clinical trials of cancer in childhood: design and implementation It is essential for optimum recruitment that a clinical trial asks a question that is important to the eligible community, and that the design seems reasonable and acceptable both to physicians who will be recruiting and to the families of children who will be candidates (panels 1 and 2). The history of cancer in childhood highlights these issues.9 Throughout the 1960s, children with leukaemia in the UK were treated with antibiotics (and sometimes steroids) only, with invariably fatal outcomes. Despite encouraging results in the USA, until 1967 the UK Medical Research Council was reluctant to allow children to be treated with experimental drugs at high doses or with unpleasant toxic effects. This thinking was on the grounds that parents and nursing staff would object,9 while many paediatricians and haematologists thought that intensive multidrug chemotherapy was unwarranted.10 Although treatment efficacy (and the need for intensive treatment) has since been recognised in paediatric oncology, the general point that study designs must be acceptable to both recruiter and recruited remains salient. In adult trials, it has been recommended that qualitative research can help ensure relevance and acceptability to the target population. For example, a qualitative study of recruitment of men to a prostatecancer trial11 identified that trial information was being misinterpreted by potential participants; frequency of randomisation increased when clearer explanations were given. However, there is little activity of this kind in paediatric oncology. http://oncology.thelancet.com Vol 9 April 2008
Expected number of cases per year Acute lymphoblastic leukaemia
Expected number of deaths per year
372
71
Acute myeloid leukaemia
69
26
Hodgkin’s lymphoma
58
<2
Non-Hodgkin lymphoma
82
16
Medulloblastoma
70
31
155
33
Neuroblastoma
89
38
Retinoblastoma
43
2
Renal tumours
81
13
Astrocytoma
Hepatoblastoma
11
3
Osteosarcoma
31
10
Ewing’s sarcoma
22
6
Rhabdomyosarcoma
55
17
Malignant germ-cell tumours
31
2
Estimated from population-based registration and from mortality data for Great Britain12 and Europe.13
Table 1: Estimated annual number of new cases and number of deaths for the main types of cancer diagnosed before age 15 years in a European country with total population about 60 million
Clearly, trials must be well-designed and comply with principles of good clinical and scientific practice. When designing a trial, the number of children who can potentially be recruited to ensure sufficient statistical power must be anticipated. For most cancers of childhood, large numbers of participants are needed for recruitment into a randomised trial compared with the annual numbers of children that are diagnosed with cancer in an average-sized European country per year (tables 1 and 2).12–14 393
Personal View
Standard treatment
Total number of participants (events)†
5-year event-free survival (%)
Target difference in 5-year event-free survival (%)*
25
+20
167 (109)
25
+15
280 (189)
60
+10
711 (249)
60
+5
2897 (1086)
90
–4
2077 (249)
90
–3
3554 (409)
*Target difference in 5-year event-free survival to be detected based on log-rank test (one-sided test, type I error 0·025 and 0·80 power).14 †Number of participants to be enrolled in each treatment group is half the total number of participants.
Table 2: Number of participants needed in superiority trials and non-inferiority trials with two parallel groups of equal allocation and comparison of event-free survival
100
Survival (%)
80 1996–2000 (n=73) 1991–95 (n=65) 1986–90 (n=62) 1981–85 (n=51) 1976–80 (n=48) 1971–75 (n=43) 1966–70 (n=43)
60
40
20
0 0
5
10
15
20
25
30
35
40
Time since diagnosis (years)
Figure 2: Rapid increase in survival of children with liver cancer (hepatoblastoma and hepatocellular carcinoma; average 14 cases per year) in Great Britain after introduction of participation in SIOPEL clinical trials in 1991 Reproduced with permission from reference 12.
Optimum sample size in clinical trials depends strongly on the target difference the trial aims to identify: the smaller the difference, the greater the number of participants needed for enrolment. The required sample size can reach thousands of patients when the study is powered to detect less than a 10% absolute increase in event-free survival. Such relatively modest, but clinically important, improvements are to be expected with more efficient definition of therapeutic strategies that use existing chemotherapy drugs. Very large numbers of participants are needed for noninferiority trials. These studies are typically done when the aim is to reduce treatment burden, toxic effects, and sequelae for patients who have a favourable risk profile, accepting a small risk of slight inferiority in event-free 394
survival (although the few patients who relapse are expected to be treated successfully with more-intensive second-line treatment to maintain overall survival). By contrast, patients who have a poor outlook need new experimental therapeutic strategies and new drugs in hope of more marked improvements. The sample size needed (in the hundreds) is difficult to achieve because these high-risk cancers generally represent small subgroups. Since the 1970s, many countries have developed national childhood cancer study groups, which have stimulated the development of tumour-specific committees to design and run clinical trials. High frequency of participation in national clinical trials has been achieved through the active engagement of clinicians and patients in specialist treatment centres and through national consensus. However, even for the more common cancers of childhood, participation from several countries might be needed to achieve the numbers for a robustly powered trial in a reasonable timeframe (usually 5 years’ recruitment). This need adds to the time required to design and implement such trials and can mean national compromise in terms of accepting the definition of the standard-treatment group. Nonetheless, expansion of recruitment from national to international communities has been mainly successful, although it has sometimes needed an innovative approach. For example, a study design that used prospectively the principles of meta-analysis was adopted in an international trial of intensification of maintenance treatment for acute lymphoblastic leukaemia.15,16 For rare cancers of childhood or for specific subgroups, international cooperation is often the only way to produce sound evidence and to improve clinical practice. Thus, when the SIOPEL liver tumour group of the International Society of Paediatric Oncology (SIOP) commenced clinical trials in patients with the very rare tumour hepatoblastoma, survival increased in participating countries (figure 2).12 The issues raised by the precise characterisation of subgroups in paediatric oncology continue to grow in importance. Advances in the understanding of the molecular basis of many cancers of childhood and the development of more sensitive methods for assessment of response has led to the subdivision of common cancers into subgroups with different survival profiles. For example, acute lymphoblastic leukaemia can be subdivided into groups such as infants or patients who are positive for Philadelphia chromosome—each group has its own therapeutic questions.17 New statistical methods to trial design are needed to address important therapeutic questions, including that of how to increase efficacy for subgroups with poor survival. Optimisation of the design of phase I and II studies is also needed, in view of the limited number of patients available and the growing number of new drugs that must be assessed for toxic effects and efficacy. These questions have http://oncology.thelancet.com Vol 9 April 2008
Personal View
stimulated the development of adaptive designs, which might be most useful when drug activity is assessed in terms of early response to treatment, rather than in phase III studies where efficacy is measured by long-term assessment of event-free survival and overall survival.18 A Bayesian framework could be used to design phase II trials more efficiently by thinking of them as randomised selection trials that are comparing several treatments.19 Thus, in subgroups of patients with a rare cancer, it has been proposed to combine at the design stage the expected outcome of the new small trial being planned with the findings generated by earlier studies. Such a Bayesian approach estimates the potential contribution to the demonstration of treatment effect of the new trial which, if seen in isolation, would be underpowered.20 Medicines regulatory agencies have issued guidelines on how to run clinical trials in small populations, which show how non-standard designs are possible. These guidelines21 propose that there are no special methods for the design, conduct, or analysis of clinical trials in small populations, but note that the need for statistical efficiency should be weighed against the need for clinically relevant, interpretable results—the latter being more important. Even with innovative trial design and participation from major childhood-cancer research groups worldwide, as exemplified by the SIOPEL trials, asking randomised questions across different countries is very challenging. Multinational trials need additional time for design and implementation. Ensuring consistent data quality when working across national, cultural, legal, and linguistic boundaries adds further complexity.22 European Union directive 2001/20/EC23 introduced into the national law of each member state in May, 2004, aims to give a common set of regulations and definitions for standards of quality of clinical studies across member states. However, the directive has had several undesirable effects, including increased cost and bureaucracy for non-commercial trials done by universities and public research institutions. Specific requirements, such as the need for a single European trial sponsor, have created difficulties in the understanding of international trials—both within the European Union and for collaborations with cooperative groups worldwide.24
Trial recruitment Ethical issues An ongoing dilemma in clinical research is the desire for patients to benefit from progress in medical care realised by scientific research while avoiding the risk of harm from such research for any individual. The overarching ethical principle is that physicians must act in the best interests of every child. If there is substantial uncertainty—ie, clinical equipoise25—over which treatment is likely to be best for a child, it is not only defensible to offer the child the opportunity of a clinical trial, but it might also be the most ethically sound course of action.26 http://oncology.thelancet.com Vol 9 April 2008
The finding that physicians can assure families of equipoise in phase III randomised trials of cancer in childhood might be helpful,27 although the replacement of conventional cytotoxics that have proven effectiveness with new biologically targeted treatments that have fewer toxic effects is an emerging challenge. Current bioethics stresses the importance of informed consent to clinical trials. There are substantial challenges to securing informed consent in studies of cancer in childhood. Concern about children’s vulnerability underlies the bioethical standpoint on their participation in scientific research,28,29 and there are issues with the application of bioethical principles for adults to children. Ensuring that research participants have sufficient information to make a decision about whether to participate and that they have the capacity to understand the consequences of their decision is problematic in any area of research, but especially so when children are involved.
Bayesian framework A procedure used to improve a statistical model in the light of observed data
Physician perspective Physician-based factors are likely to affect trial recruitment, as exemplified by research in adult trials.30–32 Recruitment in trials of cancer in childhood raises distinct issues33 and additional complexities.34 Despite strong ethical arguments in favour of running clinical trials in situations of equipoise, doctors might have conflict and concern when assuming the dual role of caring clinician and researcher,35,36 particularly in earlyphase trials.37,38 Other possible factors that affect physicians’ willingness to recruit might include their level of comfort with randomisation and adherence to trial protocols; research of adult participants suggests that some physicians think these features of trials might undermine their ability to make individual treatment decisions.39,40 These issues might be especially important in multinational trials that involve national compromises in terms of what should be included in the standard group of the trial. Physicians face key challenges once they have made a decision to approach the family of a child with cancer about possible involvement in a trial. They must: introduce the possibility of a clinical trial to patients and parents, for whom it is probably an entirely unfamiliar idea or one that has unwelcome associations with passive experimentation; explain complex treatment protocols; and secure cooperation and trust under the time pressures imposed by the need to begin treatment.41 Consultations such as those during the diagnosis of a child with cancer or during treatment discussions might be understood as so-called focused encounters.42 Here, a specific matter (eg, how the child can be cured) is the focus of attention. The introduction of an idea such as research, which might seem irrelevant to families and a violation of the implicit rules that govern such encounters, can cause substantial discomfort.39 It is unfortunate that a controversial history in the conduct of medical research might provoke anxiety about 395
Personal View
the motives of doctors in recruitment to research or fears that the child will be treated as a guinea pig. Of pivotal importance is reassurance for families that their child’s wellbeing is the doctor’s main aim, that children are not viewed as a research resource, and that families retain the autonomy to make individualised treatment decisions if unexpected toxic effects occur. Many difficulties are compounded by linguistic, cultural, and religious differences among potential participants, particularly in the context of recent migration and poverty.43,44 Nevertheless, research is being informed by socialscience theory and methods that can help give an evidence-base for enhanced management of these processes by physicians,45 although much work remains to be done.
Family perspective The engagement of families as informed partners in clinical trials is a key goal of sound practice, but is far from straightforward in its achievement. A key issue is whether to involve a child with cancer in the decision about participating in a trial. The principle that children should be consulted, informed, and involved in decisions that affect them, but not have final authority over decisions, is stipulated in the UN Convention of the Rights of the Child (UNCRC). The guiding principle of the UNCRC is that the “best interests” of the child must be promoted, but differences of opinion between parents, children, and professionals about the “best interests” of the child will sometimes need to be resolved. Involvement of a child in a decision about participation in a cancer trial might be especially complex when parents who are struggling to accept the diagnosis prefer to defer telling the child the full diagnosis and its implications,46 perhaps until treatment is established and the parents are better able to cope with their emotions. Some parents might see a demand to involve the child in the decision—and by definition disclosure of the diagnosis—as interfering with their parental autonomy. Further complexity arises because of varying definitions of minors and of the legal age of consent between different countries; some jurisdictions rely on the assessment of competence rather than specification of a particular age of consent. How information should be communicated and how much understanding is necessary to qualify for informed consent poses further challenges. A US study47 found that although randomisation was explained in most consultations with families, 50% of parents did not subsequently show an understanding of it. Furthermore, longitudinal studies have identified that parents do not gain much further understanding over time.48 These findings are not unique to cancer in childhood: adult patients have a poor understanding of the idea and scientific benefits of randomisation.49 So-called misunderstanding by patients of ideas such as randomisation are commonly assumed to arise from 396
incomplete, badly written, or misleading information. Much attention, particularly from research ethics committees, tends to focus on the detail of patient information sheets. However, the role of written information in this context remains poorly understood, and parents might find discussion with staff more helpful than a consent document.50 Principles of honesty, clear writing, and presentation are of paramount importance. However, even the clearest writing is unlikely to give a full solution, and much of the current approach might rely on a defective and simplistic model of communication. This so-called deficit model assumes that misunderstanding is a technical issue to be resolved by leaflet design and full disclosure. A wide-ranging critique of this approach in the social sciences has shown how it fails to recognise the dynamic nature of the interaction between the text and individual readers, who bring many diverse competences, motivations, and knowledge to the task of reading.51 It is unsurprising that educational interventions to improve understanding of randomisation have proved disappointing.52 Current approaches to ethics need consultations about trial recruitment to take what sociologists would identify as a specific ritualised format of interactions between clinicians and families and to achieve certain recorded outcomes, such a signed consent form. However, clinical situations are commonly at odds with ethical or legal ideals. For example, the requirement that discussions take place at particular times and in a particular order poses substantial problems in practice. Families are frequently shocked and distressed at the diagnosis1 and might have limited ability to absorb or act on information in the first few days. The physician is left with a dilemma because trials need consent before treatment starts (for a day-1 randomisation or new treatment approach), or within a short defined period (for initial treatment that does not differ from standard practice). A more radical approach to resolving the issues associated with trial recruitment in this area is to argue that the current approach to information provision in clinical trials, where there is genuine uncertainty about which treatment is better, is simply wrong. Evans and colleagues53 in their influential book, Testing treatments, point to the discrepancy between the UK General Medical Council’s guidance on communicating with patients in routine practice, which recommends that the appropriate amount of information will vary depending on factors including the patient’s own wishes, and guidance on consent to research, which states that people who are asked to participate in clinical trials must be given “the fullest possible information”.53 Evans and colleagues53 propose the need for a flexible approach, which recognises that trust between doctor and patient is the bedrock. They suggest that people who are invited to join a trial should be given every opportunity of easy access to the fullest possible information, but that it would be up to them how much information they wanted before making a decision. http://oncology.thelancet.com Vol 9 April 2008
Personal View
Alongside such approaches to informing patients, which are the focus of ongoing debate, different trial designs should be assessed, including starting patients on standard treatment and introducing the idea of randomisation later. Any future research in this area must continue to acknowledge that the goal will never be maximisation of recruitment by persuading families to enroll children in clinical trials; indeed, provision of easily understood information might result in fewer enrolments in trials. Rather, parents and children should be engaged as partners, decision-making should be facilitated in the most supportive manner, and the best interests of every child secured.
Other factors Many organisational factors affect trial recruitment. The time and resource implications of trial management, strict eligibility criteria, and complex monitoring and follow-up can affect patient accrual to clinical trials in adult oncology,54 and there are similar issues in paediatrics. Compliance with all regulatory requirements for trials of medicines in a paediatric setting imposes a substantial overhead that many health services and individual treatment centres struggle to meet. Consequently, the decreasing number of open trials available for patients is an important concern.55 The bureaucracy and expense of pharmacovigilance in trials could be reduced if medicines regulators had a more flexible approach to the definition of “investigational medicinal products” in clinical trials. Medicines regulators should recognise that many drugs or treatment combinations that are used in phase III clinical trials of cancer in childhood are established with well-documented side-effects and safety profiles, even though individual drugs might be used outside the licensed indication because there is no, or very limited, paediatric data in their marketing authorisation. This issue is being addressed in an ongoing consultation between the European Commission and the noncommercial research community. Researchers also have some responsibility: they should define more clearly what is meant by “expected” toxicity, with the aim of reducing the number of episodes that need expedited reporting. Clinical trials of new drugs might be facilitated by the introduction of new regulations relevant to children in Canada and the USA (in 2003)56 and in Europe (in 2007).57 These regulations mandate submission of paediatric investigation plans for new drugs that have potential benefit for children’s diseases. This requirement might encourage pharmaceutical companies to engage with the academic community or public sector, which have traditionally led the way in paediatric oncology, to lend support to the clinical trials that will be necessary to incorporate the new generation of biologically targeted anticancer drugs and other therapies into treatment for cancer in childhood. For non-commercial partners, such http://oncology.thelancet.com Vol 9 April 2008
Search strategy and selection criteria This Personal View reflects the personal experience of the authors. Relevant research articles were identified from the authors’ knowledge of the subject and by searching PubMed by use of the keywords “clinical trials”, “recruitment”, “randomisation”, “cancer”, “children”, and “childhood”. Further papers were chosen from the reference lists of relevant articles.
dialogue is expected to increase access to new drugs and to the long established infrastructure for trial management in industry. The availability of clinical trials of cancer in childhood might improve, providing that the noncommercial sector has the necessary resources; however, the possible effects on trial recruitment are unknown.
Conclusion Clinical trials of cancer in childhood have substantially improved the survival and wellbeing of thousands of children worldwide. National groups who develop and apply treatment protocols, whether randomised or not, and evaluate and publish their results lend support to a culture of evidence-based treatments. Children who are treated in such settings are ensured a scientifically sound approach to care and are generally offered treatment under conditions of best clinical practice. The tendency to define clinical practice by working consistently over many years in national groups is not yet common practice in most adult cancers. This factor should be kept in mind when addressing the debate, still very open, on whether patients who are enrolled in oncology clinical trials have better outcomes than those who are not.58,59 Further research is needed: into how people respond to approaches by clinicians to consent a child to enrol in a trial; to identify how physicians can manage the informedconsent process in an optimum way; and to explore the role of written information. Such work should recognise that families might wish to engage in a therapeutic alliance of mutual trust and collaboration from the outset of their child’s treatment. There is little room for complacency: if progress is to be sustained, formidable challenges must be faced. As we have identified, these challenges include the scientific study of small populations, the complexities of the regulatory environment, the ethical and organisational constraints, the emotional context, and the attitudes of individual families and physicians. Some issues may be addressed by innovative trial design and implementation, but all will need engagement with the communities of families and physicians on whom trials depend. Clinical research must include a well-designed social science and ethics assessment of how best to inform and engage the patient community who need clinical trials to continue for progress to be made. Conflicts of interest The authors declared no conflicts of interest.
397
Personal View
References 1 Dixon-Woods M, Findlay M, Young B, et al. Parents’ accounts of obtaining a diagnosis of childhood cancer. Lancet 2001; 357: 670–74. 2 Farber S, Diamond LK, Mercer RD. Temporary remissions in acute leukemia in children produced by folic acid antagonist, 4-aminopteroylglutamic acid (aminopterin). N Engl J Med 1948; 238: 787–93. 3 Reaman GH. Pediatric cancer research. From past success through collaboation to future transdisciplinary research. J Pediatr Oncol Nurs 2004; 21: 123–27. 4 Pritchard-Jones K, Kaatsch P, Steliarova-Foucher E, et al. Cancer in children and adolescents in Europe: developments over 20 years and future challenges. Eur J Cancer 2006; 42: 2183–90. 5 Estlin EJ. Novel targets for therapy in paediatric oncology. Curr Drug Targets Immune Endocr Metabol Disord 2002; 2: 141–50. 6 Ablett S, Pinkerton CR, United Kingdom Children’s Cancer Study Group. Recruiting children into cancer trials—role of the United Kingdom Children’s Cancer Study Group (UKCCSG). Br J Cancer 2003; 88: 1661–65. 7 Hann I, Vora A, Harrison G, et al. Determinants of outcome after intensified therapy of childhood lymphoblastic leukaemia: results from Medical Research Council United Kingdom acute lymphoblastic leukaemia XI protocol. Br J Haematol 2001; 113: 103–14. 8 Mitchell C, Pritchard-Jones K, Shannon R, et al. Immediate nephrectomy versus preoperative chemotherapy in the management of non-metastatic Wilms’ tumour: results of a randomised trial (UKW3) by the UK Children’s Cancer Study Group. Eur J Cancer 2006; 42: 2554–62. 9 Barnes E. Between remission and cure: patients, practitioners and the transformation of leukaemia in the late twentieth century. Chronic Illn 2007; 3: 253–64. 10 Anon. Acute leukaemia. Lancet 1965; 2: 939. 11 Donovan J, Mills N, Smith M, et al. Improving the design and conduct of trials by embedding them in qualitative research: ProctecT study. BMJ 2002; 325: 766–70. 12 Stiller C, ed. Childhood cancer in Britain. London: Oxford University Press, 2007. 13 Steliarova-Foucher E, Coebergh JW, Kaatsch P, et al, eds. Cancer in children and adolescents in Europe. Eur J Cancer (Special Issue) 2006; 42: 1913–2190. 14 Machin D, Campbell M, Fayers P, Pinol A. Sample size tables for clinical studies, 2nd edn. Malden: Blackwell Science, 1997. 15 Conter V, Valsecchi MG, Silvestri D, et al. Pulses of vincristine and dexamethasone in addition to intensive chemotherapy for children with intermediate-risk acute lymphoblastic leukaemia: a multicentre randomised trial. Lancet 2007; 369: 123–31. 16 Valsecchi MG, Masera G. A new challenge in clinical research in childhood ALL: the prospective meta-analysis strategy for intergroup collaboration. Ann Oncol 1996; 7: 1005–08. 17 Pieters R, Schrappe M, De Lorenzo P, et al. A treatment protocol for infants younger than 1 year with acute lymphoblastic leukaemia (Interfant-99): an observational study and a multicentre randomised trial. Lancet 2007; 370: 240–50. 18 Golub HL. The need for more efficient trial designs. Stat Med 2006; 25: 3231–35. 19 Estey EH, Thall PF. New designs for phase 2 clinical trials. Blood 2003; 102: 442–48. 20 Tan SB, Dear KB, Bruzzi P, Machin D. Strategy for randomised clinical trials in rare cancers. BMJ 2003; 327: 47–49. 21 European Medicines Agency (EMEA). Committee for medicinal products for human use. Guideline on clinical trials in small populations CHMP/EWP/83561/2005. http://www.emea.europa.eu/ pdfs/human/ewp/8356105en.pdf (accessed Jan 22, 2008). 22 Valsecchi MG, Silvestri D, Covezzoli A, De Lorenzo P. Web-based international studies in limited populations of pediatric leukemia. Pediatr Blood Cancer 2008; 50: 270–73. 23 The European Parliament and the Council of the European Union. Directive 2001/20/EC of the European Parliament and of the council of 4 April 2001 on the approximation of the laws, regulations and administrative provisions of the member states relating to the implementation of good clinical practice in the conduct of clinical trials on medicinal products for human use. http://eur-lex.europa.eu/LexUriServ/LexUriServ. do?uri=CELEX:32001L0020:EN:HTML (accessed Jan 22, 2008).
398
24 25
26 27
28 29
30
31
32
33
34
35
36
37
38
39
40 41 42 43
44
45
46
47 48
Hoey R. The EU clinical trial directive: 3 years on. Lancet 2007; 369: 1777–78. Djulbegovic B. Acknowledgment of uncertainty: a fundamental means to ensure scientific and ethical validity in clinical research. Curr Oncol Rep 2001; 3: 389–95. Chalmers I. Regulation of therapeutic research is compromising the interests of patients. Int J Pharm Med 2007; 21: 395–404. Kumar A, Soares H, Wells R, et al. Are experimental treatments for cancer in children superior to established treatments? Observational study of randomised controlled trials by the Children’s Oncology Group. BMJ 2005; 331: 1295. Field MJ, Behrman RE. Responsible research involving children. Ambul Pediatr 2005; 5: 47–49. Dixon-Woods M, Young B, Ross E. Researching chronic childhood illness: the example of childhood cancer. Chronic Illn 2006; 2: 165–77. Mansour EG. Barriers to clinical trials. Part III: knowledge and attitudes of health care providers. Cancer 1994; 74 (9 suppl): 2672–75. Mannel RS, Walker JL, Gould N, et al. Impact of individual physicians on enrolment of patients into clinical trials. Am J Clin Oncol 2003; 26: 171–73. Ellis PM. Attitudes towards and participation in randomised clinical trials in oncology: a review of the literature. Ann Oncol 2000; 11: 939–45. Simon CM, Siminoff LA, Kodish ED, Burant C. Comparison of the informed consent process for randomized clinical trials in pediatric and adult oncology. J Clin Oncol 2004; 22: 2708–17. Estlin EJ, Ablett S. Practicalities and ethics of running clinical trials in paediatric oncology—the UK experience. Eur J Cancer 2001; 37: 1394–98. Frayne SM, Mancuso M, Prout MN, Freund KM. Attitudes of primary care physicians toward cancer-prevention trials: a focus group analysis. J Natl Med Assoc 2001; 93: 450–57. Estlin EJ, Cotterill S, Pratt CB, et al. Phase I trials in pediatric oncology: perceptions of pediatricians from the United Kingdom Children’s Cancer Study Group and the Pediatric Oncology Group. J Clin Oncol 2000; 18: 1900–05. Mick R, Lane N, Daugherty C, Ratain MJ. Physician-determined patient risk of toxic effects: impact on enrollment and decision making in phase I cancer trials. J Natl Cancer Inst 1994; 86: 1685–93. Daugherty C, Ratain MJ, Grochowski E, et al. Perceptions of cancer patients and their physicians involved in phase I trials. J Clin Oncol 1995; 13: 1062–72. Taylor KM, Feldstein ML, Skeel RT, et al. Fundamental dilemmas of the randomized clinical trial process: results of a survey of the 1737 Eastern Cooperative Oncology Group investigators. J Clin Oncol 1994; 12: 1796–805. Fallowfield L, Ratcliffe D, Souhami R. Clinicians’ attitudes to clinical trials of cancer therapy. Eur J Cancer 1997; 33: 2221–29. Simon C, Eder M, Raiz P, et al. Informed consent for pediatric leukemia research: clinician perspectives. Cancer 2001; 92: 691–700. Goffman E. Encounters: two studies in the sociology of interaction. Indianapolis: Bobbs-Merill, 1961. Simon CM, Zyzanski SJ, Durand E, et al. Interpreter accuracy and informed consent among Spanish-speaking families with cancer. J Health Commun 2006; 11: 509–22. Abbe M, Simon C, Angiolillo A, et al. A survey of language barriers from the perspective of pediatric oncologists, interpreters, and parents. Pediatr Blood Cancer 2006; 47: 819–24. Eder ML, Yamokoski AD, Wittmann PW, Kodish ED. Improving informed consent: suggestions from parents of children with leukemia. Pediatrics 2007; 119: e849–59. Young B, Dixon-Woods M, Windridge K, Heney D. Managing communication with young people who have a potentially life threatening chronic illness: qualitative study of patients and parents. BMJ 2003; 326: 305–10. Kodish E, Eder M, Noll RB, et al. Communication of randomization in childhood leukemia trials. JAMA 2004; 291: 470–75. Greenley RN, Drotar D, Zyzanski SJ, Kodish E. Stability of parental understanding of random assignment in childhood leukemia trials: an empirical examination of informed consent. J Clin Oncol 2006; 24: 891–97.
http://oncology.thelancet.com Vol 9 April 2008
Personal View
49
50
51
52
53 54
55
Ellis PM, Dowsett SM, Butow PN, Tattersall MH. Attitudes to randomized clinical trials amongst out-patients attending a medical oncology clinic. Health Expect 1999; 2: 33–43. Kodish ED, Pentz RD, Noll RB, et al. Informed consent in the Childrens Cancer Group: results of preliminary research. Cancer 1998; 82: 2467–81. Dixon-Woods M. Writing wrongs? An analysis of published discourses about the use of patient information leaflets. Soc Sci Med 2001; 52: 1417–32. Robinson EJ, Kerr CE, Stevens AJ, et al. Lay public’s understanding of equipoise and randomisation in randomised controlled trials. Health Technol Assess 2005; 9: 1–192, iii–iv. Evans I, Thornton H, Chalmers I. Testing treatments: better research for better healthcare. London: British Library, 2006. Langley C, Gray S, Selley S, et al. Clinicians’ attitudes to recruitment to randomised trials in cancer care: a qualitative study. J Health Serv Res Policy 2000; 5: 164–69. Hearn J, Sullivan R. The impact of the ‘Clinical Trials’ directive on the cost and conduct of non-commercial cancer trials in the UK. Eur J Cancer 2007; 43: 8–13.
http://oncology.thelancet.com Vol 9 April 2008
56
57
58
59
Official Journal of the European Union. Regulation (EC) no 1901/2006 of the European Parliament and of the Council of 12 December 2006 on medicinal products for paediatric use and amending regulation (EEC) No 1768/92, Directive 2001/20/EC, Directive 2001/83/EC and Regulation (EC) No 726/2004. http:// ec.europa.eu/enterprise/pharmaceuticals/eudralex/vol-1/reg_2006_ 1901/reg_2006_1901_en.pdf (accessed Jan 22, 2008). US Food and Drug Administration. Best pharmaceuticals for Children Act. Jan 4, 2002. http://www.fda.gov/opacom/laws/ pharmkids/contents.html (accessed Jan 22, 2008). Braunholtz DA, Edwards SJ, Lilford RJ. Are randomized clinical trials good for us (in the short term)? Evidence for a “trial effect”. J Clin Epidemiol 2001; 54: 217–24. Peppercorn JM, Weeks JC, Cook EF, Joffe S. Comparison of outcomes in cancer patients treated within and outside clinical trials: conceptual framework and structured review. Lancet 2004; 363: 263.
399