Randomized Trials Stopped Early for Harm in HIV/AIDS: A Systematic Survey

PURPOSE

The decision to stop trials early because of the harmful effects of the intervention is complex and requires weighing statistical, logistical, and ethical considerations. We assessed the prevalence of randomized clinical trials (RCTs) stopped early for harm in HIV/AIDS and determined the quality of reporting of methods to inform the decision to stop the trial.

METHOD

We searched 11 electronic databases and major conference abstract databases, contacted trialist and advocacy groups, and searched the Internet. We selected RCTs stopped early for harm. We extracted data on journal and year of publication, reporting of methods and funding, planned sample size, number and planning of interim analyses, stopping rules, and effect size of the harm outcomes.

RESULTS

We found 10 RCTs stopped early for harm (median, n = 85; range, 7-1227). Most interventions (n = 9) were antiviral drugs; one trial studied vitamins to prevent vertical transmission of HIV. Five studies reported a priori defined adverse events, and only 1 trial reported planned stopping guidelines. The primary harm outcomes reported across trials included toxicity, death, and increased mother-to-child transmission. Two trials were stopped due to sudden unanticipated adverse events (Stevens-Johnson syndrome, death, and encephalopathy). Relative risk point estimates for harm ranged from 1 to 6.18. Six studies reported the presence of a data safety and monitoring board.

CONCLUSION

The reporting of methods to inform the decision to stop trials for harm in this population is deficient in a variety of ways, including lack of stopping guidelines. Clinicians should interpret RCTs stopped early for harm with caution and interpret the results in light of related evidence. Trialists should improve the transparency of their decision-making regarding early stopping for harmful effects.

Quality of reporting in systematic reviews of adverse events: systematic review

OBJECTIVES

To examine the quality of reporting of harms in systematic reviews, and to determine the need for a reporting guideline specific for reviews of harms.

DESIGN

Systematic review.

DATA SOURCES

Cochrane Database of Systematic Reviews (CDSR) and Database of Abstracts of Reviews of Effects (DARE).

REVIEW METHODS

Databases were searched for systematic reviews having an adverse event as the main outcome, published from January 2008 to April 2011. Adverse events included an adverse reaction, harms, or complications associated with any healthcare intervention. Articles with a primary aim to investigate the complete safety profile of an intervention were also included. We developed a list of 37 items to measure the quality of reporting on harms in each review; data were collected as dichotomous outcomes ("yes" or "no" for each item).

RESULTS

Of 4644 reviews identified, 309 were systematic reviews or meta-analyses primarily assessing harms (13 from CDSR; 296 from DARE). Despite a short time interval, the comparison between the years of 2008 and 2010-11 showed no difference on the quality of reporting over time (P=0.079). Titles in fewer than half the reviews (proportion of reviews 0.46 (95% confidence interval 0.40 to 0.52)) did not mention any harm related terms. Almost one third of DARE reviews (0.26 (0.22 to 0.31)) did not clearly define the adverse events reviewed, nor did they specify the study designs selected for inclusion in their methods section. Almost half of reviews (n=170) did not consider patient risk factors or length of follow-up when reviewing harms of an intervention. Of 67 reviews of complications related to surgery or other procedures, only four (0.05 (0.01 to 0.14)) reported professional qualifications of the individuals involved. The overall, unweighted, proportion of reviews with good reporting was 0.56 (0.55 to 0.57); corresponding proportions were 0.55 (0.53 to 0.57) in 2008, 0.55 (0.54 to 0.57) in 2009, and 0.57 (0.55 to 0.58) in 2010-11.

CONCLUSION

Systematic reviews compound the poor reporting of harms data in primary studies by failing to report on harms or doing so inadequately. Improving reporting of adverse events in systematic reviews is an important step towards a balanced assessment of an intervention.

Alemtuzumab

When the CAMMS 223 trial was formally reported in the New England Journal of Medicine in October of 2008, huge interest was generated in the media, among people with multiple sclerosis (MS) and their relatives and friends. The suggestion that a therapy Campath-1H (now known as alemtuzumab) was available which surpassed the effectiveness of current standard treatments, provided by infusions given at yearly intervals or even less frequently and which appeared to reduce the accumulation of disability was hailed in the lay press as virtually a 'cure'. The paper in this edition of the journal (Spotlight on Alemtuzumab, page 77) provides a more realistic account of alemtuzumab, describing the Cambridge experience, tracking the history of Campath-1H, one of the oldest humanized monoclonal antibodies, from its origin in the laboratory to its future licensing in the clinical management of people with MS.

Outcomes of patients who participate in randomized controlled trials compared to similar patients receiving similar interventions who do not participate

BACKGROUND

Some people believe that patients who take part in randomised controlled trials (RCTs) face risks that they would not face if they opted for non-trial treatment. Others think that trial participation is beneficial and the best way to ensure access to the most up-to-date physicians and treatments. This is an updated version of the original Cochrane review published in Issue 1, 2005.

OBJECTIVES

To assess the effects of patient participation in RCTs ('trial effects') independent both of the effects of the clinical treatments being compared ('treatment effects') and any differences between patients who participated in RCTs and those who did not. We aimed to compare similar patients receiving similar treatment inside and outside of RCTs.

SEARCH STRATEGY

In March 2007, we searched The Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, The Cochrane Methodology Register, SciSearch and PsycINFO for potentially relevant studies. Our search yielded 7586 new references. In addition, we reviewed the reference lists of relevant articles.

SELECTION CRITERIA

Randomized studies and cohort studies with data on clinical outcomes of RCT participants and similar patients who received similar treatment outside of RCTs.

DATA COLLECTION AND ANALYSIS

At least two review authors independently assessed studies for inclusion, assessed study quality and extracted data.

MAIN RESULTS

We identified 30 new non-randomized cohort studies (45 comparisons): no new RCTs were found. This update now includes five RCTs (yielding 6 comparisons) and 80 non-randomized cohort studies (130 comparisons), with 86,640 patients treated in RCTs and 57,205 patients treated outside RCTs. In the randomised studies, patients were invited to participate in an RCT or not; these comparisons provided limited information because of small sample sizes (a total of 412 patients) and the nature of the questions they addressed. When the results of RCTs and non-randomized cohorts that reported dichotomous outcomes were combined, there were 98 comparisons; there was also heterogeneity (P < 0.00001, I(2) = 42.2%) between studies. No statistical significant differences were found for 85 of the 98 comparisons. Eight comparisons reported statistically significant better outcomes for patients treated within RCTs, and five comparisons reported statistically significant worse outcomes for patients treated within RCTs. There was significant heterogeneity (P < 0.00001, I(2) = 58.2%) among the 38 continuous outcome comparisons. No statistically significant differences were found for 30 of the 38 comparisons. Three comparisons reported statistically significant better outcomes for patients treated within RCTs, and five comparisons reported statistically significant worse outcomes for patients treated within RCTs.

AUTHORS' CONCLUSIONS

This review indicates that participation in RCTs is associated with similar outcomes to receiving the same treatment outside RCTs. These results challenge the assertion that the results of RCTs are not applicable to usual practice.

Outcomes of patients who participate in randomised controlled trials compared to similar patients receiving similar interventions who do not participate

BACKGROUND

Some people believe that patients who take part in randomised controlled trials (RCTs) face risks that they would not face if they opted for non-trial treatment. Others think that trial participation is beneficial and the best way to ensure access to the most up to date physicians and treatments.

OBJECTIVES

To assess the effects of patient participation in RCTs ('trial effects') independent both of the effects of the clinical treatments being compared ('treatment effects') and any differences between patients who participated in RCTs and those who did not.

SEARCH STRATEGY

In May 2001, we searched The Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, The Cochrane Methodology Register, SciSearch and PsycINFO for potentially relevant studies. Our search yielded over 10,000 references. In addition, we reviewed the reference lists of relevant articles and wrote to over 250 investigators to try to obtain further information.

SELECTION CRITERIA

Randomised studies and cohort studies with data on clinical outcomes of RCT participants and similar patients who received similar treatment outside of RCTs.

DATA COLLECTION AND ANALYSIS

At least two reviewers independently assessed studies for inclusion, assessed study quality and extracted data. Study authors were contacted for additional information.

MAIN RESULTS

We included five randomised studies (yielding 6 comparisons) and 50 non-randomised cohort studies (85 comparisons), with 31,140 patients treated in RCTs and 20,380 patients treated outside RCTs. In the randomised studies, patients were invited to participate in an RCT or not; these comparisons provided limited information because of small sample sizes (a total of 412 patients) and the nature of the questions they addressed. There was statistically significant heterogeneity (P < 0.002, I(2) = 36.2%) among the 73 dichotomous outcome comparisons; none of the potential explanatory factors we investigated helped to explain this heterogeneity. No statistically significant differences were found for 63 of the 73 comparisons. Eight comparisons reported statistically significant better outcomes for patients treated within RCTs, and two comparisons reported statistically significant worse outcomes for patients treated within RCTs. There were no statistically significant differences in heterogeneity (P = 0.53, I(2) = 0%) or in outcomes (SMD 0.01, 95% CI -0.10 to 0.12) of patients treated within and outside RCTs in the 18 comparisons which had used continuous outcomes.

AUTHORS' CONCLUSIONS

This review indicates that participation in RCTs is not associated with greater risks than receiving the same treatment outside RCTs. These results challenge the assertion that the results of RCTs are not applicable to usual practice.

Respiratory adverse event profiles in cystic fibrosis placebo subjects in short- and long-term inhaled therapy trials

The frequency and nature of adverse events (AEs) are important safety endpoints in clinical trials of therapies for cystic fibrosis (CF) subjects, yet published tables of background AE rates in the CF population are not readily available. Our objective in this study was to produce tables of respiratory AE rates for placebo subjects (pediatric and adult) for inhaled therapy trials in CF subjects. Respiratory AE rates in inhaled therapy trials were computed by combining data on placebo subjects from early-phase dosing studies and middle/late-phase studies, where placebo consisted of 4 or 5 mL of inhaled saline solution. AE rates were computed as number of events divided by number of placebo-subject days of observation, and 95% confidence intervals were computed based on a Poisson model. AEs were categorized as both broad (e.g., respiratory, reactive airway disease) and specific (e.g., cough, chest tightness, hemoptysis). In short-term studies, respiratory AE rates (95% confidence interval) were 1.1(0.7, 1.6)/person-week and 1.0(0.7, 1.4)/person-week in pediatric and adult subjects, respectively. In long-term studies, respiratory AE rates were 1.7(1.6, 1.8)/person-month and 2.2(2.1, 2.3)/person-month in pediatric and adult subjects, respectively. Stepwise Poisson models were fit to determine if baseline covariates were important in predicting AE rates. Forced expiratory volume in one second (FEV(1)) percent of predicted and age in short-term studies, and FEV(1) percent predicted and gender in long-term studies were statistically important in predicting respiratory AE rates. Although these variables were statistically significant, the models' predictive abilities were low, with adjusted R(2)'s of 0.06 and 0.12 in the short- and long-term studies, respectively. Combining placebo-subject AE data recorded from multiple CF clinical trials yields better estimates of true rates of occurrence in the CF population. The tables published from this study can be used to assist those charged with safety monitoring in CF clinical trials.

Brain lesions in newborns exposed to high-dose magnesium sulfate during preterm labor

High-dosage, tocolytic magnesium sulfate (MgSO4) administered to pregnant women during preterm labor can be toxic, and sometimes lethal, for their newborns (Cochrane Database of Systematic Reviews (relative mortality risk 2.82, 95% confidence interval 1.2-6.6)). Based on the results of the Magnesium and Neurologic Endpoints Trial and the work of many others, a unifying triangular concept is proposed to account for the increased prevalence of brain lesions, with their likely resultant mortality, in neonates and infants exposed to high-dose MgSO4 in the context of preterm labor. We review the evidence that: (1) elevated circulating levels of serum ionized magnesium occurring in mothers, and therefore in their babies, at the time of delivery are associated with subsequent neonatal intraventricular hemorrhage (IVH); (2) neonatal IVH is strongly associated with lenticulostriate vasculopathy (LSV), an unusual mineralizing lesion involving the thalami and basal ganglia of the neonate; and, (3) exposure to 50 g or more of tocolytic MgSO4 during preterm labor is associated with the development of LSV.

Sufficiently Important Difference: Expanding the Framework of Clinical Significance

Background. It is generally agreed that randomized controlled trials should be powered to detect small but clinically significant treatment effects. Toward these ends, minimal important difference (MID) was proposed as a benchmark for designing trials and for interpreting health-related quality-of-life instrument scores. MID was defined in 1989 as “the smallest difference in score in the domain of interest which patients perceive as beneficial and which would mandate, in the absence of troubling side effects and excessive cost, a change in the patient’s management.” Objective. 1) To expand the idea of minimal clinically important difference so as to take into account harms as well as benefits. 2) To propose concepts and methods with which to do so. Summary. The authors define sufficiently important difference (SID) as the smallest amount of patient-valued benefit that an intervention would require to justify associated costs, risks, and other harms. As a means toward estimation of SID, the authors propose benefit-harm tradeoff methods, in which domains of benefit and harm are systematically traded off against each other and assessed in relation to the global decision of whether a treatment choice is worthwhile. Specific SID estimates can be used to power and interpret clinical trials or to inform health services research and/or public health policy. This article briefly describes the evolution of the important difference concept and outlines similarities and differences between MID and SID.

Detecting Treatment Emergent Adverse Events in Clinical Trials

BACKGROUND

The collection of adverse event data is an important component of clinical trials, but it is not clear whether solicited or unsolicited collection methods are better at distinguishing drug effects from the effects of placebo. The objective of this analysis is to compare the reporting rates and the ability to detect drug-placebo differences with spontaneous versus solicited adverse event collection methods.

METHODS

Adverse events were collected by spontaneous (unsolicited) reporting and by structured questionnaires in three randomised, double-blind clinical trials. For both spontaneous and solicited adverse event collection methods, a drug/placebo (D/P) reporting ratio was computed by dividing the reporting rate for the experimental drug by the reporting rate for placebo for each adverse event. An index (Sp-So index) was calculated by dividing the spontaneous D/P ratio by the solicited D/P ratio. A number >1.0 indicates that the spontaneous adverse event collection method is more effective in distinguishing the drug from placebo and a number <1.0 suggests that the solicited adverse event collection method is more effective in distinguishing the drug from placebo.

RESULTS

Reporting rates were greater when events were solicited than when the spontaneous reporting approach was used. The Sp-So index was >1.0 for 22 of the 29 (75.9%) events examined, suggesting that spontaneous collection of adverse events is more effective in distinguishing drug effect from placebo than the solicited approach. However, more statistically significant differences between drug and placebo were detected by the solicited method (nine events) than the spontaneous method (five events). This is due, in part, to the fact that differences in the percentages of adverse events between drug and placebo (rather than ratios of event rates) were more often greater when the solicited approach was used.

CONCLUSIONS

As expected, adverse events collected by solicitation leads to higher reporting rates. However, it is not clear that solicitation of events leads to greater ability to detect drug-placebo differences. By using a ratio to assess drug-placebo differences, spontaneous reporting provided larger drug-placebo differences more often than solicitation.

Availability of large-scale evidence on specific harms from systematic reviews of randomized trials

PURPOSE

To assess how frequently systematic reviews of randomized controlled trials convey large-scale evidence on specific, well-defined adverse events.

METHODS

We searched the Cochrane Database of Systematic Reviews for reviews containing quantitative data on specific, well-defined harms for at least 4000 randomized subjects, the minimum sample required for adequate power to detect an adverse event due to an intervention in 1% of subjects. Main outcome measures included the number of reviews with eligible large-scale data on adverse events, the number of ineligible reviews, and the magnitude of recorded harms (absolute risk, relative risk) based on large-scale evidence.

RESULTS

Of 1727 reviews, 138 included evidence on > or =4000 subjects. Only 25 (18%) had eligible data on adverse events, while 77 had no harms data, and 36 had data on harms that were nonspecific or pertained to <4000 subjects. Of 66 specific adverse events for which there were adequate data in the 25 eligible reviews, 25 showed statistically significant differences between comparison arms; most pertained to serious or severe adverse events and absolute risk differences <4%. In 29% (9/31) of a sample of large trials in reviews with poor reporting of harms, specific harms were presented adequately in the trial reports but were not included in the systematic reviews.

CONCLUSION

Systematic reviews can convey useful large-scale information on adverse events. Acknowledging the importance and difficulties of studying harms, reporting of adverse effects must be improved in both randomized trials and systematic reviews.

Safety reporting in randomized trials of mental health interventions

OBJECTIVE

The authors aimed to evaluate the adequacy of the reporting of safety information in publications of randomized trials of mental-health-related interventions.

METHOD

The authors randomly selected 200 entries from the PsiTri registry of mental-health-related controlled trials. This yielded 142 randomized trials that were analyzed for adequacy and relative emphasis of their content on safety issues. They examined drug trials as well as trials of other types of interventions.

RESULTS

Across the 142 eligible trials, 103 involved drugs. Twenty-five of the 142 trials had at least 100 randomly chosen subjects and at least 50 subjects in a study arm. Among drug trials, only 21.4% had adequate reporting of clinical adverse events, and only 16.5% had adequate reporting of laboratory-determined toxicity, while 32.0% reported both the numbers and the reasons for withdrawals due to toxicity in each arm. On average, drug trials devoted 1/10 of a page in their results sections to safety, and 58.3% devoted more space to the names and affiliations of authors than to safety. None of the trials of nondrug interventions had adequate or even partially adequate reporting of either clinical adverse events or laboratory-determined toxicity. In multivariate modeling, long-term trials and trials conducted in the United States devoted even less space to safety, while schizophrenia trials devoted more space to safety than did trials in other areas.

CONCLUSIONS

Safety reporting is largely neglected across trials of mental-health-related interventions, thus hindering the assessment of risk-benefit ratios for rational decision making in mental health care.

A potential bias in safety evaluation during open-label extensions of randomized clinical trials

PURPOSE

To describe a bias that can occur in the analysis of data from certain randomized trials.

METHODS AND RESULTS

Although randomized trials are effective at preventing confounding, a potentially strong confounding can arise in certain therapeutic drug trials in which follow-up is extended in an open-label phase that allows switching treatments. Many trials implement eligibility screening that excludes those at high risk of morbidity and mortality. In these trials, disease rates and death rates for the study population can rise rapidly during follow-up as the effect of screening wanes. During the open-label follow-up, a preponderance of patients may switch to the new therapy. If so, then any evaluation of the new therapy that includes follow-up from the open-label phase, as is often the case for safety evaluations, will be confounded. The confounding arises because the person-time experience of those on the new treatment will be more heavily weighted with the open-label phase experience, during which morbidity and mortality rates may be much greater than in the initial phase of follow-up. This confounding may be strong and will be in the direction of making the new treatment look worse, provided that the net switching is toward the new treatment during the open-label phase.

CONCLUSIONS

The confounding described here is not prevented by randomization because it develops in a non-randomized add-on analysis to the trial. The bias can be removed, however, by controlling for time since randomization in the analysis of the data.

Irreversible ototoxicity associated with difluoromethylornithine

Difluoromethylornithine (DFMO) is a potent, irreversible inhibitor of ornithine decarboxylase, the rate-limiting enzyme in the synthesis of polyamines that promote cellular proliferation. DFMO has been tested as a potential cancer therapeutic and chemopreventive agent in clinical trials. Reversible hearing loss is a recognized toxicity of DFMO that usually occurs at doses above 2 g/m(2)/d, and generally when the cumulative dose exceeds 250 g/m(2). In a recently completed Barrett's esophagus chemoprevention trial, a participant developed a 15-dB decrease in hearing at frequencies of 250, 2,000, and 3,000 Hz in the right ear and a > or =20-dB decrease in hearing at 4,000 to 6,000 Hz in the left ear after taking 0.5 g/m(2)/d DFMO for approximately 13 weeks (cumulative dose of 45 g/m(2)). The threshold shifts persisted 7 months after DFMO was discontinued. There was no obvious impact on the participant's clinical hearing, but these findings were consistent with irreversible hearing loss. This is the first case reported of irreversible ototoxicity in a clinical trial participant receiving DFMO and, thus, trial participants should be made aware of this small but important risk.

HABITS (hormonal replacement therapy after breast cancer--is it safe?), a randomised comparison: trial stopped

In the 1990s, two randomised clinical trials started in Scandinavia addressing whether hormone replacement therapy (HRT) is safe for women with previous breast cancer. We report the findings of the safety analysis in HABITS (hormonal replacement therapy after breast cancer--is it safe?), an open randomised clinical trial with allocation to either HRT or best treatment without hormones. The main endpoint was any new breast cancer event. All analyses were done according to intention-to-treat. Until September, 2003, 434 women were randomised; 345 had at least one follow-up report. After a median follow-up of 2.1 years, 26 women in the HRT group and seven in the non-HRT group had a new breast-cancer event. All women with an event in the HRT group and two of those in the non-HRT group were exposed to HRT and most women had their event when on treatment. We decided that these findings indicated an unacceptable risk for women exposed to HRT in the HABITS trial, and the trial was terminated on Dec 17, 2003.

[Safety assessment in radomized controlled clinical trials]

OBJECTIVE

To introduce some methods of safety assessment in randomized controlled clinical trials.

METHODS

Recent advances and current parctice in normalized safety assessment were reviewed and relevant data analyzed. RESTULTS: The statistical issues including analysis and presentation of adverse events data and laboratory data were involved and summed up.

CONCLUSION

With the progressive development of randomized controlled clinical trials in China, the methods introduced in this paper are sure to prove of consultative value for the safety assessment.

How can we improve the assessment of safety in child and adolescent psychopharmacology?

OBJECTIVE

To identify approaches to improving methods for assessing tolerability and safety of psychotropic medications in children and adolescents.

METHOD

Strengths and limitations of current methodology were reviewed and possible alternatives examined.

RESULTS

Research on the validity of safety evaluation has been extremely limited. No evidence-based "gold standard" exists. Clinical trials remain the best design to establish causality, but sample size limitations prevent the detection of infrequent, though serious, adverse events. Other designs, such as cohort and case-control studies, and approaches, such as mining of large databases, must be considered.

CONCLUSION

The current lack of methodological standardization across studies prevents generalizations and meta-analyses. Because the issues relevant to drug safety are diverse, a variety of methodological approaches and instruments are needed. It is, however, possible to adopt standard basic definitions of adverse events, degree of severity, ascertainment methods, and recording procedures, as a common "core," to which more specific assessment instruments can be added. Systematic empirical testing and validation of safety methodology is needed.

Suicide risk in placebo vs active treatment in placebo-controlled trials for schizophrenia

BACKGROUND

If there is an increased risk of suicide in the placebo arms of placebo-controlled studies in patients with schizophrenia, it would be a strong ethical argument against the conduct of placebo-controlled studies in this patient population. We tested whether the risk of suicide and attempted suicide in the placebo arms of placebo-controlled studies among patients with schizophrenia is higher than in the active treatment arms of such studies.

METHODS

All placebo-controlled double-blind studies that were part of a registration dossier for the indication schizophrenia, and that were submitted to the regulatory authority of the Netherlands from January 1, 1992, through December 31, 2002, were reviewed for suicide and attempted suicide.

RESULTS

In 31 studies, 7152 patients were included: 1888 in placebo groups (398.2 person-years) and 5264 in active compound groups (981.3 person-years). One suicide occurred in the placebo groups (0.05%, or an incidence rate of 251 per 100,000 years of exposure) and 1 in the active compound groups (0.02%, or an incidence rate of 102 per 100,000 years of exposure). This difference was not statistically significant. Two attempted suicides occurred in the placebo groups (0.11%, or an incidence rate of 502 per 100,000 years of exposure) and 11 in the active compound groups (0.21%, or an incidence rate of 1121 per 100,000 years of exposure). This difference was also not statistically significant.

CONCLUSION

Concern about increased risk of suicide or attempted suicide in the placebo group should not be an argument against the conduct of placebo-controlled trials in schizophrenia, provided that appropriate precautions are taken.

Status of fetal tissue transplantation for the treatment of advanced Parkinson disease

In the first double-blind, placebo-controlled randomized study of fetal tissue transplantation for the treatment of patients with advanced Parkinson disease (PD), investigators found that implanted dopaminergic tissue can produce measurable improvement in young PD in the absence of medication (that is, the "off" state). The results of the study, however, also highlighted several serious limitations of transplantation. In the group of older patients in the study (in the typical age range of individuals afflicted with PD) no improvement was derived from the implant despite positron emission tomography-documented scan evidence that the graft survived and produced dopamine. Patients in the study were selected because they experienced motor fluctuations, and the transplant did not improve dyskinesias or the time required to remain "on" medication for any subgroup of patients, including young patients. Five of 33 implant-treated patients developed involuntary movements (dyskinesias or dystonia) that could not be eliminated by reducing antiparkinsonian medications. These included four patients with the best responses to transplantation. Finally, some sham-operated patients experienced a dramatic placebo effect lasting at least 1 year, which justified the controversial sham surgery. The authors believe that these problems must be solved before fetal tissue transplantation can be considered a therapeutic option for PD.

Regulatory perspectives on data monitoring

Data monitoring is a critical component of the conduct of clinical trials that provide the evidence of efficacy and safety of investigational drugs. These trials may be conducted either by a pharmaceutical sponsor or by the government, especially those large trials that assess the impact of therapies on serious morbidity and/or mortality. While not extensive, I will review a regulatory history of FDA's evolving concerns and positions on data monitoring. I will review the key aspects of data monitoring and interim analysis of clinical trials contained in the recently published International Conference on Harmonization's statistical guidance as well as some other issues being considered for a draft guidance on data monitoring. Finally, some suggestions for improving and enhancing tools and statistical methods for monitoring clinical trials for safety assessment will be offered. This latter area deserves more consideration by statisticians than it has received to date.

Role of the Data and Safety Monitoring Committee (DSMC)

The Data and Safety Monitoring Committee (DSMC) is a committee independent of both the Steering Committee (SC) and any sponsor. It advises these bodies on continuation or stopping based upon safety and efficacy considerations. The primary objective is to assure safety for the patients in the trial. Blinding of the DSMC regarding type of treatment is usually avoided. The DSMC is composed of three to five members and should include one or two clinicians knowledgeable in the field of the trial, one or two statisticians and in some situations a pharmacologist and an ethicist. It is important that the members have experience from other trials and have high integrity. Rules for the statistical analyses have to be set up in the beginning of the trial. However, the DSMC cannot rely only upon statistical rules. Information from other sources may cause stopping before the scheduled end of the trial. The DSMC has to check that the overall quality of the data is good. Of special importance is that the Endpoint Committee is current with classification of endpoints. Timing of the DSMC meetings is dependent on several factors: (i) the incidence of events; (ii) the power of the study design; (iii) the effectiveness of the intervention. Legal requirements to report serious adverse events to legal authorities may be taken over by the DSMC, which ideally is the only body that has access to unblinded data.

Using the CABLES model to assess and minimize risk in research: control group hazards

CABLES is both an acronym and metaphor for conceptualizing research participation risk by considering 6 distinct domains in which risks of harm to research participants may exist: cognitive, affective, biological, legal, economic, and social/cultural. These domains are described and illustrated, along with suggestions for minimizing or eliminating the potential hazards to human participants in biomedical and behavioral science research. Adoption of a thoughtful ethical analysis addressing all 6 CABLES strands in designing research provides a strong protective step toward safeguarding and promoting the well-being of study participants.

Improving safety reporting from randomised trials

Randomised clinical trials offer a unique opportunity for capturing safety information under a controlled setting that minimises biases in the comparison of different therapeutic options. Nevertheless, empirical evidence across diverse medical fields suggests that the reporting of safety information in clinical trials is largely neglected and receives less attention compared with efficacy outcomes. An analysis of 192 randomised trials has shown that reasons for withdrawals due to toxicity were specified per study arm in only 46% of the trial reports. Adequate reporting of clinical adverse effects and laboratory-determined toxicity occurred in only 39 and 29% of the trials, respectively, even with lenient definitions of what constitutes adequate reporting. The use of standardised scales for adverse effects is a prerequisite for improved reporting on safety in randomised trials. Safety data need to be collected and analysed in a systematic fashion and active surveillance for toxicity during the conduct of a randomised trial is preferable to passive surveillance. Standardised reporting of safety data does not necessarily require extensive space to accomplish. It is essential to provide numerical data per study arm on each type of adverse effect along with a categorisation of the severity of the adverse effects with an emphasis on severe and life-threatening reactions. The severity grading must be referred to well-known standardised scales and new scales need to be carefully defined. Information on withdrawals due to toxicity is also important to report, along with the specific reasons leading to discontinuation. Tabulation of information may be helpful and rare or not previously reported adverse effects should be described in detail. The availability of newer options such as electronic publication, publication of raw databases, large database research, meta-analytic approaches, and prospective registration of clinical trials and of their databases may further improve the safety insights we can gain from randomised clinical trials.

Are randomized clinical trials good for us (in the short term)? Evidence for a "trial effect"

OBJECTIVE

To assess whether there is evidence that randomized controlled trials are systematically beneficial, or harmful, for patients. In other words, is there a "trial effect"? If so, to examine whether the evidence sheds light on the likely sources of the difference in outcomes.

METHODS

Systematic review of the literature.

RESULTS

We set out in some detail potential sources of a "trial effect" and potential biases. We found only 14 research articles (covering more than 21 trials) with relevant primary data. We extracted, with difficulty, quantitative data-sets from the articles, and classified these according to likely source of any apparent trial effect. The categories used were: differences in prognosis; superior treatment in the trial; and "protocol/Hawthorne effect" (benefit from improved routine care within a trial).

ANALYSIS

The evidence available is limited in breadth (coming largely from cancer trials) and quality, as well as quantity. There is weak evidence to suggest that clinical trials have a positive effect on the outcome of participants. This does not appear to depend strongly on the trial demonstrating that an experimental treatment is superior. However, benefit to participants is less evident where scope for a "protocol/Hawthorne effect" was apparently limited (because there was no effective routine treatment or because the comparison group also received protocol care). A form of bias, arising if clinicians who tend to recruit to trials also tend to be better clinicians, could also explain these results.

CONCLUSION

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. In the limited data available, the effect seems to be larger in trials where an effective treatment already exists and is included in the trial protocol.

RECOMMENDATION

That carefully researched treatment protocols, and monitoring of outcomes, be used for all patients, not just those in trials.

Clinical practice guidelines in nutrition support: can they be based on randomized clinical trials?

As rationing of health care services becomes an increasing reality, the pressure to justify interventions such as nutrition support will intensify. The establishment of clinical practice guidelines is one means of providing practitioners with such justification, but clinical practice guidelines for nutrition support cannot be based primarily on prospective randomized trials. This situation arises as the result of limitations specific to nutrition support whereby the most malnourished patients-those who appear most likely to show a benefit from the treatment-cannot be randomized to a no feeding group and are therefore excluded from participation in the study. As the result of this limitation, marginal candidates for nutrition support have been included in some trials, potentially masking the benefits of this treatment. An additional problem limiting present interpretation of published reports of randomized trials in nutrition support is the fact that ongoing research continues to yield improvements in the clinical practice of nutrition support. Thus the nutrition support group in such trials may not have received this treatment according to current practice. The A.S.P.E.N. Guidelines, based on both randomized prospective trials and other types of evidence, represent an important contribution to the practice of nutrition support. Testing of the performance of these and other guidelines in clinical practice and further outcomes research will be important steps toward revision and improvement of nutrition support, but may be difficult to achieve in the near future.

Treatment-related deaths in small cell lung cancer trials: can patients at risk be identified? Medical Research Council Lung Cancer Working Party

OBJECTIVES

This paper investigates the problem of treatment-related deaths in small cell lung cancer (SCLC).

DESIGN

To observe and define increased hazard levels, and to identify factors relating to these excess deaths.

SETTING

The United Kingdom.

SUBJECTS

A total of 2196 patients entered into the series of six randomised clinical trials in SCLC conducted by the Medical Research Council (MRC) Lung Cancer Working Party (LCWP).

RESULTS

In this large series of patients an increased risk of death in the second week after commencing the first cycle of chemotherapy was observed, suggesting that of the 10% of patients who died within 3 weeks of starting chemotherapy, half may have been treatment-related. Much less additional risk was associated with subsequent cycles of chemotherapy, and no additional risk with either initial surgery or radiotherapy. Radford et al. [Eur J Cancer 1993; 29A: 81-86] suggested that the risk factors for death from sepsis were a Karnofsky Performance (KP) score of < or = 50 (translated as a WHO performance grade (PS) > or = 3), age > 50 years and three or more drugs in the chemotherapy regimen utilised. Starting with this model we found that our data suggest it can be refined by omitting age and including a white blood cell count > or = 10,000/mm3 (this variable was not tested by Radford), and changing the other categories to WHO PS > or = 2 (KP < or = 70), and four or more drugs. Within our data this revised model identified a high risk group of patients with an excess death rate of more than 15% in the second week after starting chemotherapy. Radford et als' suggestion that high risk patients be given half doses of drugs at the first cycle should be tested in a randomised clinical trial.