Modeling in-Hospital Patient Survival During the First 28 Days After Intensive Care Unit Admission: a Prognostic Model for Clinical Trials in General Critically Ill Patients

OBJECTIVE: The objective of the study was to develop a model for estimating patient 28-day in-hospital mortality using 2 different statistical approaches. DESIGN: The study was designed to develop an outcome prediction model for 28-day in-hospital mortality using (a) logistic regression with random effects and (b) a multilevel Cox proportional hazards model. SETTING: The study involved 305 intensive care units (ICUs) from the basic Simplified Acute Physiology Score (SAPS) 3 cohort. PATIENTS AND PARTICIPANTS: Patients (n = 17138) were from the SAPS 3 database with follow-up data pertaining to the first 28 days in hospital after ICU admission. INTERVENTIONS: None. MEASUREMENTS AND RESULTS: The database was divided randomly into 5 roughly equal-sized parts (at the ICU level). It was thus possible to run the model-building procedure 5 times, each time taking four fifths of the sample as a development set and the remaining fifth as the validation set. At 28 days after ICU admission, 19.98% of the patients were still in the hospital. Because of the different sampling space and outcome variables, both models presented a better fit in this sample than did the SAPS 3 admission score calibrated to vital status at hospital discharge, both on the general population and in major subgroups. CONCLUSIONS: Both statistical methods can be used to model the 28-day in-hospital mortality better than the SAPS 3 admission model. However, because the logistic regression approach is specifically designed to forecast 28-day mortality, and given the high uncertainty associated with the assumption of the proportionality of risks in the Cox model, the logistic regression approach proved to be superior.

Premature Discontinuation of Randomized Trials in Critical and Emergency Care: A Retrospective Cohort Study.

OBJECTIVES: Randomized clinical trials that enroll patients in critical or emergency care (acute care) setting are challenging because of narrow time windows for recruitment and the inability of many patients to provide informed consent. To assess the extent that recruitment challenges lead to randomized clinical trial discontinuation, we compared the discontinuation of acute care and nonacute care randomized clinical trials. DESIGN: Retrospective cohort of 894 randomized clinical trials approved by six institutional review boards in Switzerland, Germany, and Canada between 2000 and 2003. SETTING: Randomized clinical trials involving patients in an acute or nonacute care setting. SUBJECTS AND INTERVENTIONS: We recorded trial characteristics, self-reported trial discontinuation, and self-reported reasons for discontinuation from protocols, corresponding publications, institutional review board files, and a survey of investigators. MEASUREMENTS AND MAIN RESULTS: Of 894 randomized clinical trials, 64 (7%) were acute care randomized clinical trials (29 critical care and 35 emergency care). Compared with the 830 nonacute care randomized clinical trials, acute care randomized clinical trials were more frequently discontinued (28 of 64, 44% vs 221 of 830, 27%; p = 0.004). Slow recruitment was the most frequent reason for discontinuation, both in acute care (13 of 64, 20%) and in nonacute care randomized clinical trials (7 of 64, 11%). Logistic regression analyses suggested the acute care setting as an independent risk factor for randomized clinical trial discontinuation specifically as a result of slow recruitment (odds ratio, 4.00; 95% CI, 1.72-9.31) after adjusting for other established risk factors, including nonindustry sponsorship and small sample size. CONCLUSIONS: Acute care randomized clinical trials are more vulnerable to premature discontinuation than nonacute care randomized clinical trials and have an approximately four-fold higher risk of discontinuation due to slow recruitment. These results highlight the need for strategies to reliably prevent and resolve slow patient recruitment in randomized clinical trials conducted in the critical and emergency care setting.

Cell therapy in dilated cardiomyopathy: from animal models to clinical trials

Dilated cardiomyopathy can be the end-stage form and common denominator of several cardiac disorders of known cause, such as hypertensive, ischemic, diabetic and Chagasic diseases. However, some individuals have clinical findings, such as an increase in ventricular chamber size and impaired contractility (classical manifestations of dilated cardiomyopathy) even in the absence of a diagnosed primary disease. In these patients, dilated cardiomyopathy is classified as idiopathic since its etiology is obscure. Nevertheless, regardless of all of the advances in medical, pharmacological and surgical procedures, the fate of patients with dilated cardiomyopathy (of idiopathic or of any other known cause) is linked to arrhythmic episodes, severe congestive heart failure and an increased risk of sudden cardiac death. In this review, we will summarize present data on the use of cell therapies in animal models of dilated cardiomyopathies and will discuss the few clinical trials that have been published so far involving patients affected by this disease. The animal models discussed here include those in which the cardiomyopathy is produced by genetic manipulation and those in which disease is induced by chemical or infectious agents. The specific model used clearly creates restrictions to translation of the proposed cell therapy to clinical practice, insofar as most of the clinical trials performed to date with cell therapy have used autologous cells. Thus, translation of genetic models of dilated cardiomyopathy may have to wait until the use of allogeneic cells becomes more widespread in clinical trials of cell therapies for cardiac diseases.

Stopping clinical trials because of treatment ineffectiveness: a comparison of a futility design with a method of stochastic curtailment

This paper introduces a simple futility design that allows a comparative clinical trial to be stopped due to lack of effect at any of a series of planned interim analyses. Stopping due to apparent benefit is not permitted. The design is for use when any positive claim should be based on the maximum sample size, for example to allow subgroup analyses or the evaluation of safety or secondary efficacy responses. A final frequentist analysis can be performed that is valid for the type of design employed. Here the design is described and its properties are presented. Its advantages and disadvantages relative to the use of stochastic curtailment are discussed. Copyright (C) 2003 John Wiley Sons, Ltd.

Sequential designs for phase III clinical trials incorporating treatment selection

Most statistical methodology for phase III clinical trials focuses on the comparison of a single experimental treatment with a control. An increasing desire to reduce the time before regulatory approval of a new drug is sought has led to development of two-stage or sequential designs for trials that combine the definitive analysis associated with phase III with the treatment selection element of a phase II study. In this paper we consider a trial in which the most promising of a number of experimental treatments is selected at the first interim analysis. This considerably reduces the computational load associated with the construction of stopping boundaries compared to the approach proposed by Follman, Proschan and Geller (Biometrics 1994; 50: 325-336). The computational requirement does not exceed that for the sequential comparison of a single experimental treatment with a control. Existing methods are extended in two ways. First, the use of the efficient score as a test statistic makes the analysis of binary, normal or failure-time data, as well as adjustment for covariates or stratification straightforward. Second, the question of trial power is also considered, enabling the determination of sample size required to give specified power. Copyright © 2003 John Wiley & Sons, Ltd.

A combined score test for binary and ordinal endpoints from clinical trials

There is growing interest, especially for trials in stroke, in combining multiple endpoints in a single clinical evaluation of an experimental treatment. The endpoints might be repeated evaluations of the same characteristic or alternative measures of progress on different scales. Often they will be binary or ordinal, and those are the cases studied here. In this paper we take a direct approach to combining the univariate score statistics for comparing treatments with respect to each endpoint. The correlations between the score statistics are derived and used to allow a valid combined score test to be applied. A sample size formula is deduced and application in sequential designs is discussed. The method is compared with an alternative approach based on generalized estimating equations in an illustrative analysis and replicated simulations, and the advantages and disadvantages of the two approaches are discussed.

The potential for bias in reporting of industry-sponsored clinical trials

Concerns about potentially misleading reporting of pharmaceutical industry research have surfaced many times. The potential for duality (and thereby conflict) of interest is only too clear when you consider the sums of money required for the discovery, development and commercialization of new medicines. As the ability of major, mid-size and small pharmaceutical companies to innovate has waned, as evidenced by the seemingly relentless decline in the numbers of new medicines approved by Food and Drug Administration and European Medicines Agency year-on-year, not only has the cost per new approved medicine risen: so too has the public and media concern about the extent to which the pharmaceutical industry is open and honest about the efficacy, safety and quality of the drugs we manufacture and sell. In 2005 an Editorial in Journal of the American Medical Association made clear that, so great was their concern about misleading reporting of industry-sponsored studies, henceforth no article would be published that was not also guaranteed by independent statistical analysis. We examine the precursors to this Editorial, as well as its immediate and lasting effects for statisticians, for the manner in which statistical analysis is carried out, and for the industry more generally.

Publication and non-publication of clinical trials: longitudinal study of applications submitted to a research ethics committee

BACKGROUND: Not all clinical trials are published, which may distort the evidence that is available in the literature. We studied the publication rate of a cohort of clinical trials and identified factors associated with publication and nonpublication of results. METHODS: We analysed the protocols of randomized clinical trials of drug interventions submitted to the research ethics committee of University Hospital (Inselspital) Bern, Switzerland from 1988 to 1998. We identified full articles published up to 2006 by searching the Cochrane CENTRAL database (issue 02/2006) and by contacting investigators. We analyzed factors associated with the publication of trials using descriptive statistics and logistic regression models. RESULTS: 451 study protocols and 375 corresponding articles were analyzed. 233 protocols resulted in at least one publication, a publication rate of 52%. A total of 366 (81%) trials were commercially funded, 47 (10%) had non-commercial funding. 346 trials (77%) were multi-centre studies and 272 of these (79%) were international collaborations. In the adjusted logistic regression model non-commercial funding (Odds Ratio [OR] 2.42, 95% CI 1.14-5.17), multi-centre status (OR 2.09, 95% CI 1.03-4.24), international collaboration (OR 1.87, 95% CI 0.99-3.55) and a sample size above the median of 236 participants (OR 2.04, 95% CI 1.23-3.39) were associated with full publication. CONCLUSIONS: In this cohort of applications to an ethics committee in Switzerland, only about half of clinical drug trials were published. Large multi-centre trials with non-commercial funding were more likely to be published than other trials, but most trials were funded by industry.

Cluster randomized clinical trials in orthodontics: design, analysis and reporting issues

Cluster randomized trials (CRTs) use as the unit of randomization clusters, which are usually defined as a collection of individuals sharing some common characteristics. Common examples of clusters include entire dental practices, hospitals, schools, school classes, villages, and towns. Additionally, several measurements (repeated measurements) taken on the same individual at different time points are also considered to be clusters. In dentistry, CRTs are applicable as patients may be treated as clusters containing several individual teeth. CRTs require certain methodological procedures during sample calculation, randomization, data analysis, and reporting, which are often ignored in dental research publications. In general, due to similarity of the observations within clusters, each individual within a cluster provides less information compared with an individual in a non-clustered trial. Therefore, clustered designs require larger sample sizes compared with non-clustered randomized designs, and special statistical analyses that account for the fact that observations within clusters are correlated. It is the purpose of this article to highlight with relevant examples the important methodological characteristics of cluster randomized designs as they may be applied in orthodontics and to explain the problems that may arise if clustered observations are erroneously treated and analysed as independent (non-clustered).

Factorial designs: an overview with applications to orthodontic clinical trials.

Factorial designs for clinical trials are often encountered in medical, dental, and orthodontic research. Factorial designs assess two or more interventions simultaneously and the main advantage of this design is its efficiency in terms of sample size as more than one intervention may be assessed on the same participants. However, the factorial design is efficient only under the assumption of no interaction (no effect modification) between the treatments under investigation and, therefore, this should be considered at the design stage. Conversely, the factorial study design may also be used for the purpose of detecting an interaction between two interventions if the study is powered accordingly. However, a factorial design powered to detect an interaction has no advantage in terms of the required sample size compared to a multi-arm parallel trial for assessing more than one intervention. It is the purpose of this article to highlight the methodological issues that should be considered when planning, analysing, and reporting the simplest form of this design, which is the 2 × 2 factorial design. An example from the field of orthodontics using two parameters (bracket type and wire type) on maxillary incisor torque loss will be utilized in order to explain the design requirements, the advantages and disadvantages of this design, and its application in orthodontic research.

Premature Discontinuation of Randomized Trials in Critical and Emergency Care: A Retrospective Cohort Study.

OBJECTIVES Randomized clinical trials that enroll patients in critical or emergency care (acute care) setting are challenging because of narrow time windows for recruitment and the inability of many patients to provide informed consent. To assess the extent that recruitment challenges lead to randomized clinical trial discontinuation, we compared the discontinuation of acute care and nonacute care randomized clinical trials. DESIGN Retrospective cohort of 894 randomized clinical trials approved by six institutional review boards in Switzerland, Germany, and Canada between 2000 and 2003. SETTING Randomized clinical trials involving patients in an acute or nonacute care setting. SUBJECTS AND INTERVENTIONS We recorded trial characteristics, self-reported trial discontinuation, and self-reported reasons for discontinuation from protocols, corresponding publications, institutional review board files, and a survey of investigators. MEASUREMENTS AND MAIN RESULTS Of 894 randomized clinical trials, 64 (7%) were acute care randomized clinical trials (29 critical care and 35 emergency care). Compared with the 830 nonacute care randomized clinical trials, acute care randomized clinical trials were more frequently discontinued (28 of 64, 44% vs 221 of 830, 27%; p = 0.004). Slow recruitment was the most frequent reason for discontinuation, both in acute care (13 of 64, 20%) and in nonacute care randomized clinical trials (7 of 64, 11%). Logistic regression analyses suggested the acute care setting as an independent risk factor for randomized clinical trial discontinuation specifically as a result of slow recruitment (odds ratio, 4.00; 95% CI, 1.72-9.31) after adjusting for other established risk factors, including nonindustry sponsorship and small sample size. CONCLUSIONS Acute care randomized clinical trials are more vulnerable to premature discontinuation than nonacute care randomized clinical trials and have an approximately four-fold higher risk of discontinuation due to slow recruitment. These results highlight the need for strategies to reliably prevent and resolve slow patient recruitment in randomized clinical trials conducted in the critical and emergency care setting.

A simulation study of the effects of small center enrollment in multi -center clinical trials

Multi-center clinical trials are very common in the development of new drugs and devices. One concern in such trials, is the effect of individual investigational sites enrolling small numbers of patients on the overall result. Can the presence of small centers cause an ineffective treatment to appear effective when treatment-by-center interaction is not statistically significant?^ In this research, simulations are used to study the effect that centers enrolling few patients may have on the analysis of clinical trial data. A multi-center clinical trial with 20 sites is simulated to investigate the effect of a new treatment in comparison to a placebo treatment. Twelve of these 20 investigational sites are considered small, each enrolling less than four patients per treatment group. Three clinical trials are simulated with sample sizes of 100, 170 and 300. The simulated data is generated with various characteristics, one in which treatment should be considered effective and another where treatment is not effective. Qualitative interactions are also produced within the small sites to further investigate the effect of small centers under various conditions.^ Standard analysis of variance methods and the "sometimes-pool" testing procedure are applied to the simulated data. One model investigates treatment and center effect and treatment-by-center interaction. Another model investigates treatment effect alone. These analyses are used to determine the power to detect treatment-by-center interactions, and the probability of type I error.^ We find it is difficult to detect treatment-by-center interactions when only a few investigational sites enrolling a limited number of patients participate in the interaction. However, we find no increased risk of type I error in these situations. In a pooled analysis, when the treatment is not effective, the probability of finding a significant treatment effect in the absence of significant treatment-by-center interaction is well within standard limits of type I error. ^

OMERACT-OARSI Initiative: Osteoarthritis Research Society International set of responder criteria for osteoarthritis clinical trials revisited.

Background: The OARSI Standing Committee for Clinical Trials Response Criteria Initiative had developed two sets of responder criteria to present the results of changes after treatment in three symptomatic domains (pain, function, and patient's global assessment) as a single variable for clinical trials (1). For each domain, a response was defined by both a relative and an absolute change, with different cut-offs with regard to the drug, the route of administration and the OA localization. Objective: To propose a simplified set of responder criteria with a similar cut-off, whatever the drug, the route or the OA localization. Methods: Data driven approach: (1) Two databases were considered The 'elaboration' database with which the formal OARSI sets of responder criteria were elaborated and The 'revisit' database. (2) Six different scenarios were evaluated: The two formal OARSI sets of criteria Four proposed scenarios of simplified sets of criteria Data from clinical randomized blinded placebo controlled trials were used to evaluate the performances of the two formal scenarios with two different databases ('elaboration' versus 'revisit') and those of the four proposed simplified scenarios within the 'revisit' database. The placebo effect, active effect, treatment effect, and the required sample arm size to obtain the placebo effect and the active treatment effect observed were the performances evaluated for each of the six scenarios. Experts' opinion approach: Results were discussed among the participants of the OMERACT VI meeting, who voted to select the definite OMERACT-OARSI set of criteria (one of the six evaluated scenarios). Results: Data driven approach: Fourteen trials totaling 1886 CA patients and fifteen studies involving 8164 CA patients were evaluated in the 'elaboration' and the 'revisit' databases respectively. The variability of the performances observed in the 'revisit' database when using the different simplified scenarios was similar to that observed between the two databases ('elaboration' versus 'revisit') when using the formal scenarios. The treatment effect and the required sample arm size were similar for each set of criteria. Experts' opinion approach: According to the experts, these two previous performances were the most important of an optimal set of responder criteria. They chose the set of criteria considering both pain and function as evaluation domain and requiring an absolute change and a relative change from baseline to define a response, with similar cut-offs whatever the drug, the route of administration or the CA localization. Conclusion: This data driven and experts' opinion approach is the basis for proposing an optimal simplified set of responder criteria for CA clinical trials. Other studies, using other sets of CA patients, are required in order to further validate this proposed OMERACT - OARSI set of criteria. (C) 2004 OsteoArthritis Research Society International. Published by Elsevier Ltd. All rights reserved.

Systematic evaluation of patient-reported outcome (PRO) protocol content and reporting in UK cancer clinical trials: the EPiC study protocol

Introduction Emerging evidence suggests that patient-reported outcome (PRO)-specific information may be omitted in trial protocols and that PRO results are poorly reported, limiting the use of PRO data to inform cancer care. This study aims to evaluate the standards of PRO-specific content in UK cancer trial protocols and their arising publications and to highlight examples of best-practice PRO protocol content and reporting where they occur. The objective of this study is to determine if these early findings are generalisable to UK cancer trials, and if so, how best we can bring about future improvements in clinical trials methodology to enhance the way PROs are assessed, managed and reported. Hypothesis: Trials in which the primary end point is based on a PRO will have more complete PRO protocol and publication components than trials in which PROs are secondary end points.

Methods and analysis Completed National Institute for Health Research (NIHR) Portfolio Cancer clinical trials (all cancer specialities/age-groups) will be included if they contain a primary/secondary PRO end point. The NIHR portfolio includes cancer trials, supported by a range of funders, adjudged as high-quality clinical research studies. The sample will be drawn from studies completed between 31 December 2000 and 1 March 2014 (n=1141) to allow sufficient time for completion of the final trial report and publication. Two reviewers will then review the protocols and arising publications of included trials to: (1) determine the completeness of their PRO-specific protocol content; (2) determine the proportion and completeness of PRO reporting in UK Cancer trials and (3) model factors associated with PRO protocol and reporting completeness and with PRO reporting proportion.

Ethics and dissemination The study was approved by the ethics committee at University of Birmingham (ERN_15-0311). Trial findings will be disseminated via presentations at local, national and international conferences, peer-reviewed journals and social media including the CPROR twitter account and UOB departmental website (http://www.birmingham.ac.uk/cpro0r).

Calculation of sample size for stroke trials assessing functional outcome: comparison of binary and ordinal approaches

Background Many acute stroke trials have given neutral results. Sub-optimal statistical analyses may be failing to detect efficacy. Methods which take account of the ordinal nature of functional outcome data are more efficient. We compare sample size calculations for dichotomous and ordinal outcomes for use in stroke trials. Methods Data from stroke trials studying the effects of interventions known to positively or negatively alter functional outcome – Rankin Scale and Barthel Index – were assessed. Sample size was calculated using comparisons of proportions, means, medians (according to Payne), and ordinal data (according to Whitehead). The sample sizes gained from each method were compared using Friedman 2 way ANOVA. Results Fifty-five comparisons (54 173 patients) of active vs. control treatment were assessed. Estimated sample sizes differed significantly depending on the method of calculation (Po00001). The ordering of the methods showed that the ordinal method of Whitehead and comparison of means produced significantly lower sample sizes than the other methods. The ordinal data method on average reduced sample size by 28% (inter-quartile range 14–53%) compared with the comparison of proportions; however, a 22% increase in sample size was seen with the ordinal method for trials assessing thrombolysis. The comparison of medians method of Payne gave the largest sample sizes. Conclusions Choosing an ordinal rather than binary method of analysis allows most trials to be, on average, smaller by approximately 28% for a given statistical power. Smaller trial sample sizes may help by reducing time to completion, complexity, and financial expense. However, ordinal methods may not be optimal for interventions which both improve functional outcome