简单随机化、中心分层区组随机化和最小化法的均衡性比较

目的比较简单随机化、中心分层区组随机化和最小化法的均衡性。方法运用MonteCarlo方法分别进行三种随机化方法的模拟分组,然后比较三种随机化方法在有6个非处理因素时,组间总例数的均衡性及非处理因素组间分布的均衡性。结果中心分层区组随机化和最小化法可保证组间总例数的均衡;在保证非处理因素的组间分布均衡性上,最小化法效果最好,明显优于其他两种方法,中心分层区组随机化与简单随机化在保证非处理因素均衡性上效果接近。结论最小化法既可保证组间总例数的均衡也可保证非处理因素组间分布的均衡,因此在保证均衡性上是首选。     

Bias in recruitment to cluster randomized trials: a review of recent publications

Objectives To assess recruitment bias and the techniques employed to counter this problem in a recent selection of published cluster randomized trials. Design Review of 24 cluster trials published in 2008 in four leading medical journals. Data extraction Studies were assessed by four reviewers to identify if an alternative design could have been employed using individual randomization. Data were also extracted on the randomization procedure and the likelihood of this introducing bias to the selection of participants into the study. Results Of the 24 trials, eight could have used individual randomization as an alternative to cluster allocation. Seven studies could have recruited participants prior to cluster randomization but did not. In eight studies where recruitment bias was possible, more than half (five) demonstrated some evidence of differential recruitment rates. Conclusions Many cluster trials published in leading medical journals are not clear in their justification for the design. We also found significant proportions of cluster trials used suboptimal designs that increase their risk of introducing selection bias. Better design of cluster trials is possible and should be adopted.     

IRINI: random group allocation of multiple prognostic factors

Statistically sound experimental design in pharmacology studies ensures that the known prognostic factors, if any, are equally represented across investigational groups to avoid bias and imbalance which could render the experiment invalid or lead to false conclusions. Complete randomization can be effective to reduce bias in the created groups especially in large sample size situations. However, in small studies which involve only few treatment subjects, as in preclinical trials, there is a high chance of imbalance. The effects of this imbalance may be removed through covariate analysis or prevented with stratified randomization, however small studies limit the number of covariates to be analyzed this way. The problem is accentuated when there are multiple baseline covariates with varying scales and magnitudes to be considered in the randomization, and creating a balanced solution becomes a combinatorial challenge. Our method, IRINI, uses an optimization technique to achieve treatment to subject group allocation across multiple prognostic factors concurrently. It ensures that the created groups are equal in size and statistically comparable in terms of mean and variance. This method is a novel application of genetic algorithms to solve the allocation problem and simultaneously ensure quality, speed of the results and randomness of the process. Results from preclinical trials demonstrate the effectiveness of the method.     

Statistical comparison of random allocation methods in cancer clinical trials

The selection of a trial design is an important issue in the planning of clinical trials. One of the most important considerations in trial design is the method of treatment allocation and appropriate analysis plan corresponding to the design. In this article, we conducted computer simulations using the actual data from 2158 rectal cancer patients enrolled in the surgery-alone group from seven randomized controlled trials in Japan to compare the performance of allocation methods, simple randomization, stratified randomization and minimization in relatively small-scale trials (total number of two groups are 50, 100, 150 or 200 patients). The degree of imbalance in prognostic factors between groups was evaluated by changing the allocation probability of minimization from 1.00 to 0.70 by 0.05. The simulation demonstrated that minimization provides the best performance to ensure balance in the number of patients between groups and prognostic factors. Moreover, to achieve the 1 percentile for the p-value of chi-square test around 0.50 with respect to balance in prognostic factors, the allocation probability of minimization was required to be set to 0.95 for 50, 0.80 for 100, 0.75 for 150 and 0.70 for 200 patients. When the sample size was larger, sufficient balance could be achieved even if reducing allocation probability. The simulation using actual data demonstrated that unadjusted tests for the allocation factors resulted in conservative type I errors when dynamic allocation, such as minimization, was used. In contrast, adjusted tests for allocation factors as covariates improved type I errors closer to the nominal significance level and they provided slightly higher power. In conclusion, both the statistical and clinical validity of minimization was demonstrated in our study.     

Sequential balancing: a simple method for treatment allocation in clinical trials

Although minimisation methods have frequently been advocated for treatment allocation in clinical trials, they are not widely used. As this may partly be due to the complexity of the methods, we devised a new and simple minimisation method to balance for prognostic factors, called sequential balancing. Each factor is dealt with sequentially and when a new subject enters the trial, he or she is allocated the treatment that leads to improved balance of the first factor over the treatments. If the balance of the first factor was already satisfactory, then the treatment is allocated that leads to improved balance of the second factor and so on. The algorithm requires no calculations. We simulated a realistic trial and compared the performance of this method to the performance of alternative allocation strategies: the variance minimisation method, simple randomisation and stratification. The sequential balancing method led to better balance than randomisation and stratification. In the case of four factors or less, the performance of the sequential balancing method and the variance minimisation method were comparable and the sequence of the factors was not very relevant. When more factors were introduced, the balance of the sequential method remained comparable with the balance achieved with the variance minimisation method for the first four factors, but it started to decrease from the fifth factor onwards. We conclude that the ease and simplicity of the new method make it an attractive option when balance is required for four factors or less. If there are more than four factors, the sequential balancing method may still be an acceptable option, but the advantage of simplicity has to be weighed against the loss of performance compared to other minimisation methods.     

Minimization in crossover trials with non-prognostic strata: theory and practical application

To obtain meaningful results in any clinical trial, patients need to be allocated to treatments in such a way that valid analysis can be carried out. Balancing treatment groups before analysis is carried out is more desirable than trying to compensate for incomparability at a later date. Therefore, the development of allocation procedures to produce comparable groups in which prognostic factors are equally represented is important. Minimization, a deterministic allocation method, aims to ensure balance on such factors, particularly in small trials when traditional randomization methods are likely to fail. However, views on the use of conventional analysis following minimization are divided. The use of minimization in two randomised crossover trials is described where, in addition to the comparisons between randomised treatments, it was desired to have balance between groups based on differential trial procedures. Theoretical concerns about the use of minimization are not applicable in this setting, and therefore minimization is shown to be a useful technique for obtaining balance.     

Teaching evidence-based complementary and alternative medicine: 4. Appraising the evidence for papers on therapy

Practicing evidence-based complementary and alternative medicine (CAM) requires practitioners to develop an ability to appraise the quality of published studies addressing questions related to their clinical practice. This paper describes a process by which CAM practitioners can determine the validity of studies evaluating therapeutic interventions. The process requires asking two broad questions: (1). Do the treatment and control group begin with the same prognosis? and (2). Do the treatment and control group remain the same with respect to important prognostic factors? Answering these questions requires determining whether studies used effective randomization, preserved randomization through intention-to-treat analyses, used blinding, and had adequate follow-up of trial participants.     

Accumulated fluid balance in patients admitted to the ICU: is it really reliable?

In most of the textbooks, it is considered that the balance calculated after admission and the losses measured and/or estimated is an inexact way of establishing the real balance. Thus daily monitoring of the weight variations is recommended as a single possible alternative. On the other hand, there are few studies that have strictly studied the reliability of the fluid balance calculated. We also have not found any study in middle-long stay critical patients. These circumstances have led us to design an observational prospective study that will allow us to know if the accumulated balance calculated after admission and loses adequately reflect the weight changes in middle-long stay patients. We include 20 patients who were weighed every 48 hours (at least 3 times each one) and we compare the weight changes with the balances calculated. We find that, above all after the 6th day, the accumulated balance calculated adequately reflected the weight changes (mean error/day < 250 ml), regardless of the presence or not of fever, sweat, oral diet, feces or mechanical ventilation. When weight on admission to the ICU was less than 75 kg, the changes in the balance calculated adjusted even more to the weight change, the contrary occurring when the weight was greater than 75 kg. These findings suggest that the accumulated balance calculated represents a valid alternative to daily weighing of the patients and that factors such as body mass and/or surface should be taken into account to reach more exact estimations.     

The advantages of community-randomized trials for evaluating lifestyle modification

Observational studies may provide suggestive evidence for the results of behavior change and lifestyle modification, but they do not replace randomized trials for comparing interventions. To obtain a valid comparison of competing intervention strategies, randomized trials of adequate size are the recommended approach. Randomization avoids bias, achieves balance (on average) of both known and unknown predictive factors between intervention and comparison groups, and provides the basis of statistical tests. The value of randomization is as relevant when investigating community interventions as it is for studies that are directed at individuals. Randomization by group is less efficient statistically than randomization by individual, but there are reasons why randomization by group (such as community) may be chosen, including feasibility of delivery of the intervention, political and administrative considerations, avoiding contamination between individuals allocated to competing interventions, and the very nature of the intervention. One example is the Community Intervention Trial for Smoking Cessation (COMMIT), which involved 11 matched pairs of communities and randomized within these pairs to active community-level intervention versus comparison. For analysis of results, community-level permutation tests (and corresponding test-based confidence intervals) can be designed based on the randomization distribution. The advantages of this approach are that it is robust, and the unit of randomization is the unit of analysis, yet it can incorporate individual-level covariates. Such covariates can play a role in imputation for missing values, adjustment for imbalances, and separate analyses in demographic subsets (with appropriate tests for interaction). A communityrandomized trial can investigate a multichannel community-based approach to lifestyle modification, thus providing generalizability coupled with a rigorous evaluation of the intervention.     

Stratification in clinical trials of febrile neutropenia

Patients with febrile neutropenia are characterized by well described prognostic factors leading to heterogeneity of risk of serious clinical outcomes. These prognostic factors complicate the design and interpretation of clinical trials of antimicrobial therapy in this population. Stratification is a method by which comparability of groups is ensured. This method can be employed prior to randomization or during the analysis. Factors that should be considered for stratification prior to randomization include the underlying neoplasm (leukemia vs solid tumor vs BMT), use of granulocyte growth factors, comorbid conditions and study site (in multicenter trials). In stratified analyses, trend in neutrophil count, site of infection, organism and susceptibility should be considered.     

The method of minimization for allocation to clinical trials. a review

Minimization is a largely nonrandom method of treatment allocation for clinical trials. We conducted a systematic literature search to determine its advantages and disadvantages compared with other allocation methods. Minimization was originally proposed by Taves and by Pocock and Simon. The latter paper introduces a family of allocation methods of which Taves' method is the simplest example. Minimization aims to ensure treatment arms are balanced with respect to predefined patient factors as well as for the number of patients in each group. Further extensions of the method have also been proposed by other authors. Simulation studies show that minimization provides better balanced treatment groups when compared with restricted or unrestricted randomization and that it can incorporate more prognostic factors than stratified randomization methods such as permuted blocks within strata. Some more computationally complex methods may give an even better performance. Concerns over the use of minimization have centered on the fact that treatment assignments may be predicted with certainty in some situations and on the implications for the analysis methods used. It has been suggested that adjustment should always be made for minimization factors when analyzing trials where minimization is the allocation method used. The use of minimization may sometimes result in added organizational complexity compared with other methods. Minimization has been recommended by many commentators for use in clinical trials. Despite this it is still rarely used in practice. From the evidence presented in this review, we believe minimization to be a highly effective allocation method and recommend its wider adoption in the conduct of randomized controlled trials.     

Adaptive allocation in randomized controlled trials

Adaptive allocation has been proposed as a procedure to reduce the risk of chance imbalance of important prognostic factors in randomized controlled trials when the number of prognostic factors is large. In this article, minimization, a type of adaptive allocation, is compared to simple randomization and stratified allocation in a series of Monte Carlo simulations. Three outcomes are studied: estimated treatment effect, size of the rejection region, and power. Minimization produced an unbiased estimate of treatment effect and increased power when compared to simple randomization. Student's t test was conservative for both minimization and stratified allocation. Minimization and stratification produced similar improvements in power but there was some evidence that minimization might produce higher power than stratification when some prognostic variables cannot be included in the stratified allocation scheme.     

An application of a modified constrained randomization process to a practice-based cluster randomized trial to improve colorectal cancer screening

BackgroundWhen designing cluster randomized trials, it is important for researchers to be familiar with strategies to achieve valid study designs given limited resources. Constrained randomization is a technique to help ensure balance on pre-specified baseline covariates.MethodsThe goal was to develop a randomization scheme that balanced 16 intervention and 16 control practices with respect to 7 factors that may influence improvement in study outcomes during a 4-year cluster randomized trial to improve colorectal cancer screening within a primary care practice-based research network. We used a novel approach that included simulating 30,000 randomization schemes, removing duplicates, identifying which schemes were sufficiently balanced, and randomly selecting one scheme for use in the trial. For a given factor, balance was considered achieved when the frequency of each factor's sub-classifications differed by no more than 1 between intervention and control groups. The population being studied includes approximately 32 primary care practices located in 19 states within the U.S. that care for approximately 56,000 patients at least 50 years old.ResultsOf 29,782 unique simulated randomization schemes, 116 were determined to be balanced according to pre-specified criteria for all 7 baseline covariates. The final randomization scheme was randomly selected from these 116 acceptable schemes.ConclusionsUsing this technique, we were successfully able to find a randomization scheme that allocated 32 primary care practices into intervention and control groups in a way that preserved balance across 7 baseline covariates. This process may be a useful tool for ensuring covariate balance within moderately large cluster randomized trials.     

Zelen randomization: attitudes of parents participating in a neonatal clinical trial

Recruitment to randomized controlled trials can be difficult for all parties involved. An alternative to the standard process has been suggested for trials in which the control group receives standard treatment or nontreatment. In this approach (the Zelen design), randomization precedes consent, which is only sought from those allocated to the experimental arm of a trial. The control group is thus unaware that randomization has taken place. As a controversial method, this approach has been often suggested but rarely used. Here we describe how 44 parents recruited to a difficult neonatal trial that used conventional randomization reacted to the idea of Zelen randomization. The arguments they gave for and against the method pertain to four areas: the giving or withholding of information, the effect on decision making, the use of data without parental knowledge, and the long-term impact for parents. The parents were evenly divided in accepting or rejecting the method. Further analysis showed that those rejecting Zelen randomization were more likely to be parents of infants allocated to the control group. This suggests that those from whom consent would not be sought, the group that this approach is primarily meant to protect, are most likely to find it unacceptable.     

Balancing multiple baseline characteristics in randomized clinical trials

It is of vital importance to the success of a randomized clinical trial to maintain the balance of baseline characteristics (covariates) that could potentially confound the outcomes of the trial. Various randomization methods have been proposed to increase the likelihood of having balanced covariates at the end of a trial, most of which only apply to categorical covariates. An optimization approach to maintaining the balance of multiple covariates in randomized clinical trials is proposed, which is applicable to both continuous and categorical covariates and allows the covariates to be ranked according to their clinical importance as perceived by the clinical trial practitioners. Simulation results show that the proposed algorithm significantly outperforms the standard randomization approach via flipping an unbiased coin. The proposed randomization method can be easily implemented and generalized for cases where there are multiple treatment arms with equal or unequal randomization probabilities.     

A dynamic block-randomization algorithm for group-randomized clinical trials when the composition of blocking factors is not known in advance

We present an algorithm for randomizing units in blocks for controlled trials when the composition of blocking factors is not known in advance. For example, suppose the desired goal of an intervention study is to randomize units to one of two interventions while blocking on a dichotomous factor (e.g., gender), but the total number of units, and therefore number or composition, of males and females among those units assembled for randomization cannot be determined in advance. This situation arises in randomized trials when subjects are scheduled to be randomized as a group, but not all of the subjects show up for the visit. Since investigators do not know which of the scheduled subjects will or will not show up, a dynamic randomization scheme is required to accommodate the unknown composition of the blocking factor once a group of subjects (units) is assembled for randomization. These settings are further complicated when there is more than one blocking factor. In this paper, we present an algorithm that ensures the integrity of the randomization process in these settings.     

Statistical process control in nursing research

In intervention studies in which randomization to groups is not possible, researchers typically use quasi-experimental designs. Time series designs are strong quasi-experimental designs but are seldom used, perhaps because of technical and analytic hurdles. Statistical process control (SPC) is an alternative analytic approach to testing hypotheses about intervention effects using data collected over time. SPC, like traditional statistical methods, is a tool for understanding variation and involves the construction of control charts that distinguish between normal, random fluctuations (common cause variation), and statistically significant special cause variation that can result from an innovation. The purpose of this article is to provide an overview of SPC and to illustrate its use in a study of a nursing practice improvement intervention.     

Properties of the urn randomization in clinical trials

In this article we review the important statistical properties of the urn randomization (design) for assigning patients to treatment groups in a clinical trial. The urn design is the most widely studied member of the family of adaptive biased-coin designs. Such designs are a compromise between designs that yield perfect balance in treatment assignments and complete randomization which eliminates experimental bias. The urn design forces a small-sized trial to be balanced but approaches complete randomization as the size of the trial (n) increases. Thus, the urn design is not as vulnerable to experimental bias as are other restricted randomization procedures. In a clinical trial it may be difficult to postulate that the study subjects constitute a random sample from a well-defined homogeneous population. In this case, a randomization model provides a preferred basis for statistical inference. We describe the large-sample permutational null distributions of linear rank statistics for testing the equality of treatment groups based on the urn design. In general, these permutation tests may be different from those based on the population model, which is equivalent to assuming complete randomization. Poststratified subgroup analyses can also be performed on the basis of the urn design permutational distribution. This provides a basis for analyzing the subset of patients with observed responses when some patients' responses can be assumed to be missing-at-random. For multiple mutually exclusive strata, these tests are correlated. For this case, a combined covariate-adjusted test of treatment effect is described. Finally, we show how to generalize the urn design to a prospectively stratified trial with a fairly large number of strata.     

Analysis of randomized and nonrandomized patients in clinical trials using the comprehensive cohort follow-up study design.

In clinical research, randomized trials are widely accepted as the definitive method of evaluating the efficacy of therapies. The random assignment of patients to their treatment ensures the internal validity of the comparison of new treatments with controls. An assessment of the external validity of trial results can best be achieved by comparing the study population to the population of patients who met the eligibility criteria but did not consent to randomization. A part of the data of the Coronary Artery Surgery Study (CASS), in which coronary artery bypass surgery is compared to conventional medical therapy in patients with coronary artery disease, is used to illustrate a strategy of multivariate analysis of randomized and nonrandomized patients which allows an investigation of both internal and external validity. The method used Cox's proportional hazards regression model with inclusion of covariates for randomization status and corresponding interactions in addition to the usual covariates for treatment and the important prognostic factors.