Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
The use of composite end points to assess harm from treatment
- M Justin S Zaman (4 April 2007)
-
Composite Approach to Solving the Composite Trial Endpoint Problem
- Bruce Ovbiagele (5 April 2007)
-
Composite end points and clinical practice
- Margaret E Cupples, Susan M Smith, Mary C Byrne, Andrew W Murphy (11 April 2007)
-
Composite end points
- Philipp Conradi (15 April 2007)
-
Multilevel end points
- Grazyna T Adamiak (9 May 2007)
-
Decision Analysis: A Potential Tool for Composite End Point Interpretation
- Paul S. Chan, Brahmajee K. Nallamothu, Rodney A Hayward, Sandeep Vijan (25 May 2007)
|
|
|||
|
M Justin S Zaman, Clinical Research Fellow in Epidemiology University College London, WC1E 6BT
Send response to journal:
|
The use of composite end points to increase event rates and statistical power is rarely used in the opposite way - to assess harm from treatment. Ferreira-González et al should be applauded for pointing out this 'statistical tap-dancing'. Competing interests: None declared |
|||
|
|
|||
|
Bruce Ovbiagele, Assistant Professor of Neurology UCLA Medical Center, Los Angeles, CA 90095
Send response to journal:
|
Ferreira-González and colleagues are to be congratulated on their study that underscores the challenges with the use of composite endpoints in clinical trials. One cannot but agree with their conclusion that clinical trialists need to better report single component endpoints to enhance optimal interpretation of study results. However, an additional approach towards assisting healthcare providers to determine whether a particular intervention has an effect on endpoints of premier importance to their patients, would be to incorporate these clinically relevant distinctions in the national and international treatment guidelines written by various subject experts. Consensus guidelines can be useful in fostering quality care and reducing disease burden [1], but currently the evidence grading in most guidelines reifies the significance of the scientific evidence, but often does not systematically accommodate the importance of patient preferred outcomes. Beyond just making recommendations based on levels and class of evidence [2], for the increasing number of clinical trials that have composite primary endpoints, distinguishing the influence of the therapy on individual hard clinical outcomes could better inform the perspectives of clinicians and patients about the merits of prescribing or receiving a treatment. It could of course be argued that that these studies were not specifically designed to assess the individual components independently, but it would seem that with the current guideline development approach the risk of inadvertently attributing a patient-important benefit to a therapy that does not indeed provide such a benefit, is high. Furthermore, there appear to be viable strategies for interpreting the results of clinical trials with composite end points of varying importance [3], that could be applied in developing guideline recommendations based on aggregate clinical outcome data. References: 1. Deutsch SC, Denton M, Borenstein J. Clinical practice guidelines: a tool to help provide quality care. Geriatrics. 1998;53:57, 61-4, 70, 73- 4. 2. The U.S. Preventive Services Task Force Guide to Clinical Preventive Services. Second Edition 1996:15-38. 3. Montori VM, Permanyer-Miralda G, Ferreira-Gonzalez I, Busse JW, Pacheco-Huergo V, Bryant D, Alonso J, Akl EA, Domingo-Salvany A, Mills E, Wu P, Schunemann HJ, Jaeschke R, Guyatt GH. Validity of composite end points in clinical trials. BMJ. 2005;330:594-596. Competing interests: None declared |
|||
|
|
|||
|
Margaret E Cupples, Senior Lecturer, General Practitioner Dept of General Practice, Queen's University, 1 Dunluce Avenue, Belfast BT9 7HR, Susan M Smith, Mary C Byrne, Andrew W Murphy
Send response to journal:
|
Ferreria-Gonzalez et al 1 illustrate how mortality, as a component of composite end points in cardiovascular trials may contribute very little to the composite evaluation. Death is however, an outcome of high importance to patients. On reflection, quality of life and activities of daily living are also important to most patients, yet reports of trials include such data infrequently, particularly as primary outcomes. Recognition of individuals’ needs and priorities is relevant when translating research findings into practice. Healthcare outcomes may be prioritised differently by patients and practitioners and by these groups in different healthcare settings.2 For example, patients balance costs of treatment against the other expenses which impact on their lives: this balance varies in different healthcare systems and social circumstances.2 Also, various outcomes, such as blood pressure and cholesterol targets, have different administrative and financial implications for practitioners in different healthcare settings: this can influence their priorities in providing care. Clinicians and patients can decide better whether or not interventions are applicable to them if they have knowledge of treatment effects relating to individual components of composite end points. They are also more able to determine the relevance of research results to their own situations if researchers describe trial contexts. Campbell et al3 recently emphasised the importance of context in relation to trials of complex health service interventions. Ferreria-Gonzalez et al1 provide a salutary reminder of the need for careful interpretation of research findings. Their application of science in explaining how a composite end point may mislead if detail of its components and the context of measurements are not reported is welcomed. Consideration of the relative importance of various endpoints to clinicians and patients also reminds us that clinicians would be unwise to forget the ‘art’ component which sits alongside science in good medical practice. Understanding of individual situations is required in order to provide best quality care. References 1. Ferreira-Gonzalez I, Busses JW, Heels-Ansell D, Montori VM, Akl EA, Bryant DM, et al. Problems with use of composite endpoints in cardiovascular trials: systematic review of randomised controlled trials. BMJ,doi:10.1146/bmj.39136.682083AE (published 2 April 2007) 2. Corrigan M, Cupples ME, Smith SM, Byrne M, Leathem CS, Clerkin P, Murphy AW. The contribution of qualitative research in designing a complex intervention for secondary prevention of coronary heart disease in two different healthcare systems BMC Health Services Research 2006, 6:90 doi:10.1186/1472-6963-6-90 3. Campbell NC, Murray E, Darbyshire J, Emery J, Farmer A, Griffiths F, et al. Designing and evaluating complex interventions to improve health care. BMJ 2007;334:455-459 Competing interests: None declared |
|||
|
|
|||
|
Philipp Conradi, Allgemeinarzt Otto-Dix-Ring 98 01219 Dresden, Germany
Send response to journal:
|
composite end points ( or aggregate end points - as in UKPDS) are a theoretical construct with little relevance for the individual patient. The authors have to be applauded for making this clear. The Heart Protection Study is a prime example of manufactured science. Its main outcome ,,Major First Vascular Events'' lump together a whole variety of different end points : several coronary and non coronary vascular interventions including amputations and also fatal and non-fatal strokes and fatal and non-fatal coronary events. This is obviously a logical contradiction and stands for undisciplined thinking. Unfortunately this scientific malaise proliferates by finding its way into meta-analyses ( like the Poly pill ) and influences experts, advisory boards, gouverment advisers, think tanks and so on. Clinicians have the difficult task explaining their patients the possible benefits of treatment. British GPs face a much harder challenge: with the New Contract they are paid handsomely for putting large swaithes of patients on different drugs for medium blood pressure, cholesterol, renal disease... If patients were fully informed they might object these treatments and family doctors would lose ,Quality Points' and - money. How to stop a rolling band wagon? Competing interests: I posed this question to the GMC in 2003 |
|||
|
|
|||
|
Grazyna T Adamiak, PhD, MA, MPH Unemployed
Send response to journal:
|
Performing the present review the Authors categorized a number of components of composite end points according to their importance for patients. The importance to patients is defined by the clinicians participating in the studies, obviously according to professional criteria and not as defined by patients them selves even if the mortality as an end point is a generally agreed adverse outcome. There’s a problem proposing the of notion of “importance to patients” since it rests on assumption that patients define conditions and outcomes in similarity to clinicians. In reality, this kind of schemes impose the view of clinicians or drug industries on what matters for patients, pretending to patients’ or general public’s point of view. A classification by diagnoses is a professional matter while patients experience symptoms and signs, anxiety and functional difficulties in daily living etc. In addition, most trials chosen for the review as well the review it self are targeting professional audiences and authorities licensing drugs. The social distance and differences in cognizance are large between the above mentioned interest groups and it is questionable if it is desirable to define end points in generic terms rather than narrow clinical in this particular context. Thus, the notion “importance to patients” could have been replaced by “importance for outcome of intervention ” without changing the meaning of the studies included or the present article. The notion “importance to patients” is a rather rhetoric measure in this context. Among the end points listed according to the criteria “importance to patients” there are, from the clinical point of view, some less specific and consistently defined end points, e.g. myocardial infarctions and strokes according to their scope or size and resting severity, respectively. The lack of consistency between the above categories does not make them mutually exclusive. In addition, what is surprising is the use of hospital admission as a measure of clinical outcome in cardiological studies. This kind of end point could indicate availability of hospital beds rather than clinically specific outcomes. While AMI is a condition that justifies an admission into inpatient care, there are many less well defined and diffuse conditions within internal medicine and cardiology, such as chest pain, that also could result in hospital admissions despite inconclusive etiology. From this point of view, the notion of “importance to patients” is ambiguous. Clinicians decide on admissions and many patients never reach the hospital with their symptoms. Less knowledgeable readers or less experienced physicians could get impression that a drug reduces rates of hospital admissions and thus represents a desirable intervention. It could sound well as a marketing argument. My ending point is, that the claims on the superiority of clinical trials according to the clinical hierarchy of evidence might be questioned if the RCT:s and others introduce and reproduce, and thus maintain, end points that are clinically deficient what regards specificity and appropriateness, which are more context dependent and multilevel than a RCT is designed to manage. Competing interests: None declared |
|||
|
|
|||
|
Paul S. Chan, cardiovascular researcher University of Michigan Medical Center. Zipcode 48103, Brahmajee K. Nallamothu, Rodney A Hayward, Sandeep Vijan
Send response to journal:
|
To the Editor: We applaud the work by Ferreira-Gonzalez et al. in addressing a crucial issue common to clinical trials that use composite endpoints.1 The authors demonstrate that cardiovascular interventions can exhibit significant gradients in their impact on individual components of a composite endpoint, with ‘softer’ endpoints often providing the power for statistical significance. With increasingly lower event rates common to contemporary studies, composite endpoints allow for the conduct of clinical trials within rational time and financial constraints. We agree though that clinical trials should be cautiously interpreted if individual components of a composite endpoint are highly dissimilar.2 3 We believe a potential solution to this concern is to use decision analysis as a mechanism to provide a common denominator – quality-adjusted life-years – to compare the benefits of interventions that have composite endpoints. Quality- adjusted life-years provide differential weights to outcomes based on the impact on both quality and quantity of life, and can be combined with costs to evaluate economic outcomes (such as avoiding hospitalizations). In a recent study, we demonstrated the value of this approach by quantifying the incremental benefit (in quality-adjusted life-years) achieved by high-dose statin therapy over conventional statin dosing from four clinical trials whose composite endpoints ranged from death to rehospitalization.4 Ultimately, we live in an era where it is increasingly becoming difficult for clinical trials to demonstrate reductions in mortality. This is a testament to how far we have come in treating disease. We recognize that investigators need to conduct clinical trials which are feasible in terms of cost and sample size, and that composite endpoints are a necessity. However, we urge organizations that guide clinical trial design and reporting, such as CONSORT and major medical journals, to provide guidance in developing clinically meaningful and transparent means of reporting composite endpoints in future trials. REFERENCES 1. Ferreira-Gonzalez I, Busse JW, Heels-Ansdell D, Montori VM, Akl EA, Bryant DM, et al. Problems with use of composite end points in cardiovascular trials: systematic review of randomised controlled trials. BMJ 2007;334(7597):786. 2. Chan PS, Nallamothu BK, Hayward RA. High-Dose Statins and the IDEAL Trial. JAMA 2006;295(21):2477. 3. Chan PS, Nallamothu BK, Hayward RA. Keeping apples and oranges separate: reassessing clinical trials that use composite end points as their primary outcome. J Am Coll Cardiol 2006;48(4):850; author reply 851- 2. 4. Chan PS, Nallamothu BK, Gurm HS, Hayward RA, Vijan S. Incremental benefit and cost-effectiveness of high-dose statin therapy in high-risk patients with coronary artery disease. Circulation 2007;115(18):2398-409. Competing interests: None declared |
|||