ArticlesComparison of outcomes in cancer patients treated within and outside clinical trials: conceptual framework and structured review
Introduction
The belief that clinical trials offer the best treatment for patients with cancer is widespread in the oncology community. This claim, motivated partly by aims to increase accrual1, 2, 3 and ensure third-party payment,4, 5 appears frequently in pronouncements by professional organisations and leaders. For example, the American Federation of Clinical Oncologic Societies maintains that “treatment in a clinical trial is often a cancer patient's best option”.6 Other people argue that “clinical trials are proven to offer children the best chance of survival”,7 and that trial access is one of the “basic requirements of quality cancer care.”5 Such claims suggest that trials are viewed not only as a way to improve future treatment, but also as the best treatment for current patients.
The view that trials lead to better outcomes, if correct, has important implications. First, that more than 95% of adults and perhaps 40% of children with cancer do not enrol in trials would constitute evidence of substandard care. Second, the suggestion that patients benefit directly by becoming research participants changes the traditional model of human experimentation. If so, clinicians arguably should advocate forcefully for enrolment on grounds of direct benefit, rather than presenting the risks and benefits for patients to weigh. In the conventional view, such advocacy might be criticised as misleading or coercive. Third, acceptance of this view might require substantial changes in trial financing and organisation, eligibility criteria, and patient selection. Anything that might constitute a barrier to participation (including considerations of scientific validity and integrity8) would be suspect. We must therefore be confident that trial participation improves outcomes before using the claim to inform practice or policy.
Ideally, the statement that trials are the best treatment option should rest on evidence that trial participants have better outcomes than similar patients treated off-protocol. Several studies9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22 have shown such a trial effect, also sometimes known as an inclusion benefit.23 However, showing a causal relation between trial participation and improved outcome is difficult.
Here, we seek to develop a conceptual framework for assessing the trial effect; describe the methodological challenges in studying this effect and the hierarchy of evidence that could be used to support its existence; and use these insights to assess systematically the quality, validity, and generalisability of the published work.
Section snippets
Methods
We sought to identify articles that presented primary data comparing outcomes between trial and non-trial patients with cancer. As others note,24 there is no obvious set of terms to capture all relevant reports. We therefore searched MEDLINE using the terms trial effect, inclusion benefit, population outcomes, community outcomes, trial benefit, patient preference trial, and comprehensive cohort trial, cross-referenced with cancer, oncology, neoplasms, and clinical trials. We also scanned an
Inclusion criteria
We identified 24 published articles that met our inclusion criteria.9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 29, 30, 31, 32, 33, 34, 35 Of these, seven were included in previous reviews.24, 49 Two articles22, 29 reported two comparisons each, thus, there was a total of 26 comparisons. Table 1 summarises these studies, arranged by population, study design, and dates of the primary data reported in the reports. Additional detail is available from the authors.
Study characteristics
table 2 presents
Discussion
In our review of the published work, we found little high-quality evidence to support the pervasive belief that cancer trial participation leads to improved outcomes. Although about half the studies provided some evidence for a trial effect, and none found trial participation to be harmful, methodological difficulties with most studies suggest the need for cautious interpretation.
There are four possible reasons that trial participants might be found to have improved outcomes when compared with
References (52)
- et al.
Retrospective assessment of quality of life and treatment outcome in patients with Hodgkin's disease from 1969 to 1994
Eur J Cancer
(1999) - et al.
Results of a breast-cancer-surgery trial compared with observational data from routine practice
Lancet
(1996) - et al.
Long-term survival in Hodgkin's disease patients: a comparison of relative survival in patients in trials and those recorded in population-based cancer registries
Eur J Cancer
(2000) The “inclusion benefit” in clinical trials
J Pediatr
(1999)- et al.
Are randomized clinical trials good for us (in the short term)? Evidence for a “trial effect”
JClinEpidemiol
(2001) - et al.
Prostate specific antigen response to mitoxantrone and prednisone in patients with refractory prostate cancer: prognostic factors and generalizability of a multicenter trial to clinical practice
J Urol
(2000) - et al.
The behavioral dynamics of clinical trials
Prev Med
(1991) - et al.
Large-scale randomized evidence: large, simple trials and overviews of trials
J Clin Epidemiol
(1995) - et al.
Publication bias and clinical trials
Control Clin Trials
(1987) - et al.
Can a clinical trial be the treatment of choice for patients with cancer?
J Natl Cancer Inst
(1988)