Participants in research
http://www.100md.com
《英国医生杂志》
Neither guinea pigs nor sacrificial lambs, but pointers to better health care
Patients rightly worry that they risk being sacrificed in randomised trials whose results cannot be applied by clinicians in the front lines of real clinical practice. In a recent US survey, half the public considered trial participants to be "guinea pigs." Altruistic goals such as "making a contribution to science" were judged far more likely to be achieved than personal benefits such as "getting the best possible treatment" or "having access to the best physician."1 At the same time, many clinicians worry that the imposition of rigorous trial conditions on highly selected patients, especially in tertiary care settings, generates results that cannot be generalised to routine clinical practice.2
For the most part, these concerns originated from, and are perpetuated by, the examination of single trials or highly selective collections of them. Reports of the ghastly abuse of trial participants are sufficiently frequent and well publicised that they cause some to question whether randomised trials generally do more harm than good to their participants. Similarly, when certain subgroups have been under-represented in a trial (for example, women, older people, and those with comorbidity), some would conclude that it would be inappropriate to generalise from the trial's findings to guide their care. Finally, because explanatory trials control tightly the giving and taking of treatment in ways that are foreign to routine clinical practice, scepticism has arisen about whether the results obtained in research settings can ever be realised in the hurly-burly of frontline medical care, even among similar patients.
Individual cases and selective reviews are not the best ways to answer these important questions. Well over 20 000 reports of trials are published every year. If 999 out of every 1000 of them scrupulously protected their participants from abuse, this would still generate 2 scandals every month—more than enough to sustain criticism in the media and the resulting public concern. Similarly, showing that a particular subgroup of patients is under-represented in a trial cannot, by itself, provide convincing evidence that its results do not apply to them. Finally, the allegation that a treatment which works in a highly controlled trial will not work in routine practice deserves to be tested by something firmer than anecdote and opinion.
Recent systematic reviews are providing valuable, harder evidence for answering these questions. One group expanded, with a systematic literature search, the scope of an earlier review.3 They unearthed 11 articles in which the clinical outcomes of patients in trials were directly compared with those of concurrent eligible but non-recruited patients.4 Patients who participated in the trials fared better in nine of these 11 studies, and in six the difference in outcomes was big enough to reach significance. The other two studies found no evidence of any effect of being in a trial, and these investigators decided: "While the evidence is not conclusive, it is more likely that clinical trials have a positive rather than a negative effect on the outcome of patients." They were very cautious in attributing "trial advantage" to any specific causes, but their suggested reasons included better clinicians, better treatments, and better use of treatment protocols.4 A third review by a Boston based group was indeterminate.5 In these systematic reviews, trial and non-trial patients were not matched for treatment. Therefore, although these reviews provide important evidence on whether randomised trials do more good than harm to their participants, they do not answer the question of whether treatment effects within trials are generalisable to patients outside trials.
Researchers in this week's BMJ report how they used seven search strategies to identify 55 studies that compared the outcomes of patients inside and outside trials who were matched for the experimental and control treatments.6 7 Although they found occasional statistically significant differences in outcomes (10 favouring patients inside and four favouring patients outside trials), the 95% confidence interval around the overall relative effect of the same treatment for similar patients inside and outside trials excluded any important difference. Thus, although the results of individual comparisons are unpredictable, this systematic review provides important evidence, arguably for the first time, about the effects of the same treatment given inside and outside a randomised trial. The patients inside these trials seem unlikely to be harmed by their participation, and their outcomes with treatment seem to be generalisable to similar patients receiving the same treatment outside trials.
The extrapolation of the results of clinical trials to individual patients has become a basic clinical skill. Emerging evidence strengthens the scientific basis and rationale for these extrapolations.
David L Sackett, director
Trout Research and Education Centre at Irish Lake, RR 1, Markdale, ON, Canada N0C 1H0 (sackett@bmts.com)
Papers p 1175
I thank Mike Clarke, Brian Haynes, and Iain Chalmers for their thoughtful suggestions.
Competing interests: In addition to those at http://bmj.bmjjournals.com/cgi/content/full/324/7336/539/DC1, DS had encouraged Gunn Elisabeth Vist and her colleagues to carry out a systematic review of whether trials do more good than harm to their participants; they elected to do their current study instead.
References
Harris Interactive. News Room. Public awareness of clinical trials increases: new survey suggests those conducting trials are doing a better job of informing potential participants of opportunities. 11 June 2004. www.harrisinteractive.com/news/allnewsbydate.asp?NewsID=812 (accessed 12 May 2005).
Wei L, Ebrahim S, Bartlett C, Davey PG, Sullivan FM, MacDonald TM. Statin use in the secondary prevention of coronary heart disease in primary care: cohort study and comparison of inclusion and outcome with patients in randomised trials. BMJ 2005;330: 821.
Stiller CA. Centralised treatment, entry to trials, and survival. Br J Cancer 1994;70: 352-62.
Braunholtz DA, Edwards SJL, Lilford RJ. Are randomised clinical trials good for us (in the short term)? Evidence for a "trial effect." J Clin Epid 2001;54: 217-24.
Peppercorn JM, Weeks JC, Cook ECF, Joffe S. Comparison of outcomes in cancer patients treated within and outside clinical trials: conceptual framework and structured review. Lancet 2004;363: 263-70.
Vist GE, Hagen KB, Devereaux PJ, Bryant D, Kristoffersen DT, Oxman AD. Systematic review to determine whether participation in a trial influences outcome. BMJ 2005;330: 1175-9.
Vist GE, Hagen KB, Devereaux P, Bryant D, Kristoffersen DT, Oxman AD. Outcomes of patients who participate in randomised controlled trials compared to similar patients receiving similar interventions who do not participate. Cochrane Library 2005. Issue 2, 2005. Chichester: John Wiley.
Patients rightly worry that they risk being sacrificed in randomised trials whose results cannot be applied by clinicians in the front lines of real clinical practice. In a recent US survey, half the public considered trial participants to be "guinea pigs." Altruistic goals such as "making a contribution to science" were judged far more likely to be achieved than personal benefits such as "getting the best possible treatment" or "having access to the best physician."1 At the same time, many clinicians worry that the imposition of rigorous trial conditions on highly selected patients, especially in tertiary care settings, generates results that cannot be generalised to routine clinical practice.2
For the most part, these concerns originated from, and are perpetuated by, the examination of single trials or highly selective collections of them. Reports of the ghastly abuse of trial participants are sufficiently frequent and well publicised that they cause some to question whether randomised trials generally do more harm than good to their participants. Similarly, when certain subgroups have been under-represented in a trial (for example, women, older people, and those with comorbidity), some would conclude that it would be inappropriate to generalise from the trial's findings to guide their care. Finally, because explanatory trials control tightly the giving and taking of treatment in ways that are foreign to routine clinical practice, scepticism has arisen about whether the results obtained in research settings can ever be realised in the hurly-burly of frontline medical care, even among similar patients.
Individual cases and selective reviews are not the best ways to answer these important questions. Well over 20 000 reports of trials are published every year. If 999 out of every 1000 of them scrupulously protected their participants from abuse, this would still generate 2 scandals every month—more than enough to sustain criticism in the media and the resulting public concern. Similarly, showing that a particular subgroup of patients is under-represented in a trial cannot, by itself, provide convincing evidence that its results do not apply to them. Finally, the allegation that a treatment which works in a highly controlled trial will not work in routine practice deserves to be tested by something firmer than anecdote and opinion.
Recent systematic reviews are providing valuable, harder evidence for answering these questions. One group expanded, with a systematic literature search, the scope of an earlier review.3 They unearthed 11 articles in which the clinical outcomes of patients in trials were directly compared with those of concurrent eligible but non-recruited patients.4 Patients who participated in the trials fared better in nine of these 11 studies, and in six the difference in outcomes was big enough to reach significance. The other two studies found no evidence of any effect of being in a trial, and these investigators decided: "While the evidence is not conclusive, it is more likely that clinical trials have a positive rather than a negative effect on the outcome of patients." They were very cautious in attributing "trial advantage" to any specific causes, but their suggested reasons included better clinicians, better treatments, and better use of treatment protocols.4 A third review by a Boston based group was indeterminate.5 In these systematic reviews, trial and non-trial patients were not matched for treatment. Therefore, although these reviews provide important evidence on whether randomised trials do more good than harm to their participants, they do not answer the question of whether treatment effects within trials are generalisable to patients outside trials.
Researchers in this week's BMJ report how they used seven search strategies to identify 55 studies that compared the outcomes of patients inside and outside trials who were matched for the experimental and control treatments.6 7 Although they found occasional statistically significant differences in outcomes (10 favouring patients inside and four favouring patients outside trials), the 95% confidence interval around the overall relative effect of the same treatment for similar patients inside and outside trials excluded any important difference. Thus, although the results of individual comparisons are unpredictable, this systematic review provides important evidence, arguably for the first time, about the effects of the same treatment given inside and outside a randomised trial. The patients inside these trials seem unlikely to be harmed by their participation, and their outcomes with treatment seem to be generalisable to similar patients receiving the same treatment outside trials.
The extrapolation of the results of clinical trials to individual patients has become a basic clinical skill. Emerging evidence strengthens the scientific basis and rationale for these extrapolations.
David L Sackett, director
Trout Research and Education Centre at Irish Lake, RR 1, Markdale, ON, Canada N0C 1H0 (sackett@bmts.com)
Papers p 1175
I thank Mike Clarke, Brian Haynes, and Iain Chalmers for their thoughtful suggestions.
Competing interests: In addition to those at http://bmj.bmjjournals.com/cgi/content/full/324/7336/539/DC1, DS had encouraged Gunn Elisabeth Vist and her colleagues to carry out a systematic review of whether trials do more good than harm to their participants; they elected to do their current study instead.
References
Harris Interactive. News Room. Public awareness of clinical trials increases: new survey suggests those conducting trials are doing a better job of informing potential participants of opportunities. 11 June 2004. www.harrisinteractive.com/news/allnewsbydate.asp?NewsID=812 (accessed 12 May 2005).
Wei L, Ebrahim S, Bartlett C, Davey PG, Sullivan FM, MacDonald TM. Statin use in the secondary prevention of coronary heart disease in primary care: cohort study and comparison of inclusion and outcome with patients in randomised trials. BMJ 2005;330: 821.
Stiller CA. Centralised treatment, entry to trials, and survival. Br J Cancer 1994;70: 352-62.
Braunholtz DA, Edwards SJL, Lilford RJ. Are randomised clinical trials good for us (in the short term)? Evidence for a "trial effect." J Clin Epid 2001;54: 217-24.
Peppercorn JM, Weeks JC, Cook ECF, Joffe S. Comparison of outcomes in cancer patients treated within and outside clinical trials: conceptual framework and structured review. Lancet 2004;363: 263-70.
Vist GE, Hagen KB, Devereaux PJ, Bryant D, Kristoffersen DT, Oxman AD. Systematic review to determine whether participation in a trial influences outcome. BMJ 2005;330: 1175-9.
Vist GE, Hagen KB, Devereaux P, Bryant D, Kristoffersen DT, Oxman AD. Outcomes of patients who participate in randomised controlled trials compared to similar patients receiving similar interventions who do not participate. Cochrane Library 2005. Issue 2, 2005. Chichester: John Wiley.