著者
柳川 堯
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.38, no.2, pp.153-161, 2018-03-01 (Released:2018-05-18)
参考文献数
5

Many clinical studies are conducted in Japan with sample sizes that are not deter-mined statistically. Application of Neyman-Pearson type statistical tests to data from such studies is not justifiable and should be stopped. Also 5% significance level that is commonly employed in a clinical study without taking into account disease, drug and other factors is not justifiable. Alternatively, the use of p-value is recommended in this paper as a measure of showing the magnitude of difference of two treatments; it is the role of principal investigator to summarize the study results by considering disease, drug and other factors, sample sizes and p-value.
著者
三中 信宏
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.38, no.2, pp.117-125, 2018-03-01 (Released:2018-05-18)
参考文献数
21

The recent controversy over the use and abuse of p-values in statistical data analysis sheds a light on the epistemological diversity of scientific researches and the nature of science. Since the nineteenth century theoretical statisticians including Karl Pearson, Ronald A. Fisher, Jerzy Neyman, and Egon S.Pearson constructed the mathematical basis of modern statistics, for example, experimental design, sampling distributions, or hypothesis testing, etc. However, statistical reasoning as empirical inference is not necessarily limited to the Neyman-Pearson’s decision-making paradigm. Any kind of non-deductive inference—for example, abduction—also uses statistics as an exploratory tool for relative ranking among alternative hypotheses and models. We must understand not only the proper use of statistical methods and procedures but also the nature of each science to which statistics is applied.
著者
岩崎 学 吉田 清隆
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.26, no.2, pp.53-63, 2005-12-31 (Released:2011-09-30)
参考文献数
16
被引用文献数
3

For the occurrence of a rare event A such as a severe adverse drug reaction, there exists the “Rule of Three” to remind practitioners that “absence of evidence is not evidence of absence.” The Rule of Three actually says that even if the event A was not observed among n patients it would be quite possible to observe three events among other n patients. The present paper examines this useful rule in detail and also extends it to a testing problem for occurrence probability of A.First, the Rule of Three is extended to the case that the number of the event observed among the first n patients is more than zero. We give rules that when k (> 0) events were observed among n patients, nk events would be possibly observed among other n patients. Next, a testing procedure is introduced to examine whether the occurrence probabilities of A for two populations are the same under the condition that k events were observed among n patients for one population. It will be shown that the relevant probability distribution is a negative binomial, and then critical regions for small k's are given. For a possible application of the procedure, we mention the signal detection for spontaneous reporting system of adverse drug reaction.
著者
松山 裕
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.25, no.2, pp.89-116, 2004-12-31 (Released:2012-02-08)
参考文献数
55
被引用文献数
1

Missing data is a prevalent complication in the analysis of data from longitudinal studies, and remains an active area of research for biostatisticians and other quantitative methodologists. This paper reviews several statistical methods that are used to address outcome-related drop-out. We begin with a review of important concepts such as missing data patterns, missing data mechanisms, ignorability and likelihood-based inference, which were originally proposed by Rubin (1976, Biometrika 63, 581-592). Secondly, we review the simple analysis methods for handling drop-outs such as a complete-case analysis, an available data analysis and a last observation carried forward analysis, and their limitations are given. Thirdly, we review the more sophisticated approaches for handling drop-outs, which take account of the missing data mechanisms in the analysis. Inverse probability weighted methods and multiple imputation methods, which represent two distinct paradigms for handling missing data, are reviewed. The analysis methods for non-ignorable drop-outs are also reviewed. Three approaches, selection models, pattern mixture models and latent variable models are presented. We illustrate the analysis techniques using the longitudinal clinical trial of contracepting women reported by Machine et al (1988, Contraception 38, 165-179). We briefly review the analysis methods in the presence of missing covariates. Finally, we give some notice in the analysis of missing data.
著者
矢田 真城 魚住 龍史 田栗 正隆
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.40, no.2, pp.81-116, 2020-06-01 (Released:2020-07-21)
参考文献数
63

When a causal effect between treatment and outcome variables is observed, effects on the outcome are of interest to investigate the mechanisms among the outcome and treatment. Indirect effect is defined as the causal effect of the treatment on the outcome via the mediator. Direct effect is defined as the causal effect of the treatment on the outcome that is not through the mediator. In this paper, we discuss the estimation of direct and indirect effects based on the framework of potential response models focusing on the 4-way decomposition. Direct and indirect effect estimations are illustrated with two examples where the outcome, mediator, covariate variables are continuous and categorical data. Moreover, we discuss the estimation of clausal effects and the effect decomposition in the settings that include confounder of mediator and outcome affected by treatment, multiple mediators, or time-varying treatment in the presence of time-dependent confounder.
著者
黒木 学 小林 史明
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.32, no.2, pp.119-144, 2012-03-31 (Released:2012-06-08)
参考文献数
89

This paper reviews basic ideas of Structural Causal Models (SCMs) proposed by Judea Pearl (1995, 2009a). SCMs are nonparametric structual equation models which express cause-effect relationship between variables, and justify matematical principles of both the potential outcome approach and the graphical model approach for statistical causal inference. In this paper, considering the difference/connection between SCMs and Rubin's Causal Models (RCMs) (Rubin, 1974, 1978, 2006), we state that (1) the expressive power of the potential outcome approach is higher than that of the graphical model approach, but (2) the graphical model approach. From these consderations, we conclude that we should discuss statistical causal inference based on both approaches.
著者
佐藤 俊哉
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.38, no.2, pp.109-115, 2018-03-01 (Released:2018-05-18)
参考文献数
21

On March 7th, 2016, the American Statistical Association released its “ASA Statement on Statistical Significance and P-values,” which provided 6 principles to improve the conduct or interpretation of quantitative research. Misunderstanding and misuse of statistical tests and P-values were discussed many times in the epidemiologic field. In this paper, I gave a summary of the ASA statement and its translation process into Japanese. Then, I discussed how to avoid misunderstanding or misuse of statistical tests or P-values in epidemiologic observational studies.
著者
手良向 聡
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.41, no.1, pp.37-54, 2020 (Released:2020-12-04)
参考文献数
43

Fisher’s randomization rule has been widely viewed as a revolutionary invention in experimental design. The three rationales of randomization in clinical trials are (i) randomization ensures that known and unknown confounders are asymptotically controlled, (ii) the use of randomization itself provides the basis of statistical inference, supposing patients in a clinical trial are a non-random sample of a population, and (iii) the act of randomization mitigates selection bias by providing unpredictability in treatment allocation. Randomized controlled trials have been the gold standard for more than five decades, while such trials may be costly, inconvenient and ethically challenging. Some Fisherian statisticians have emphasized the importance of design-based inference based on randomization test, however some statisticians does not agree with them. From the Bayesian point of view, the randomization sequence is ancillary for a parameter of interest, and randomization itself is not absolutely essential although it may sometimes be helpful. In this review, I provide an overview of the rationales of randomization and the related topics, and discuss the significance and limitations of randomization in clinical trials.
著者
鵜飼 保雄
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.32, no.Special_Issue, pp.S1-S17, 2011-05-31 (Released:2011-09-05)
参考文献数
15
著者
武田 健太朗 大庭 真梨 柿爪 智行 坂巻 顕太郎 田栗 正隆 森田 智視
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.36, no.1, pp.25-50, 2015-07-20 (Released:2015-09-08)
参考文献数
47
被引用文献数
1 3

It is expected to develop new drug more efficiently by incorporating historical data into the current study data. Borrowing historical data which is sufficiently similar to the current data allows increasing power and improving the accuracy of the estimated treatment effect. On the other hand, if the historical data is not similar to the current data, there is a potential for bias and inflated type I error rate. Power prior and hierarchical model are widely known as the Bayesian approaches with borrowing strength from historical information. They have the advantage of deciding the amount of historical information continuously depending on the similarity between historical data and current data. Our goal is to introduce power prior and hierarchical model while showing some examples, and provide a review of points to keep in mind when these approaches are used in the clinical trials.
著者
浜田 知久馬 中西 豊支 松岡 伸篤
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.27, no.2, pp.139-157, 2006-12-01 (Released:2011-09-25)
参考文献数
50
被引用文献数
4 3

Meta-analysis is defined to be ‘the statistical analysis of a large collection of analysis results from individual studies for the purpose of integrating the findings'. Since the 1980s there has been an upsurge in the application of meta-analysis to medical research. The rapid increase in the number of meta-analysis being conducted during the last decade is mainly due to a greater emphasis on evidence based medicine and the need for reliable summaries of the vast and expanding volume of clinical studies. Over the same period there have been great developments and refinements of the associated methodology of meta-analysis. When judging the reliability of the results of a meta-analysis, attention should be focused on ‘publication bias’. Publication bias is the term for what occurs whenever the research that appears in the published literature is systematically unrepresentative of the population of completed studies. This bias can provide a flaw of the result of meta-analysis. In this article, the causes and origins of publication bias are reviewed, and then the history and some findings of publication bias in medical research are presented. Several statistical methods that have been developed to detect, quantify and assess the impact of publication bias in meta-analysis are demonstrated.
著者
篠崎 智大 横田 勲 大庭 幸治 上妻 佳代子 坂巻 顕太郎
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.41, no.1, pp.1-35, 2020 (Released:2020-12-04)
参考文献数
65

Prediction models are usually developed through model-construction and validation. Especially for binary or time-to-event outcomes, the risk prediction models should be evaluated through several aspects of the accuracy of prediction. With unified algebraic notation, we present such evaluation measures for model validation from five statistical viewpoints that are frequently reported in medical literature: 1) Brier score for prediction error; 2) sensitivity, specificity, and C-index for discrimination; 3) calibration-in-the-large, calibration slope, and Hosmer-Lemeshow statistic for calibration; 4) net reclassification and integrated discrimination improvement indexes for reclassification; and 5) net benefit for clinical usefulness. Graphical representation such as a receiver operating characteristic curve, a calibration plot, or a decision curve helps researchers interpret these evaluation measures. The interrelationship between them is discussed, and their definitions and estimators are extended to time-to-event data suffering from outcome-censoring. We illustrate their calculation through example datasets with the SAS codes provided in the web appendix.
著者
田中 司朗
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.40, no.1, pp.35-62, 2019-08-01 (Released:2019-09-18)
参考文献数
37

A central problem in medical research is how to make inferences about the causal effects of treatments or exposures. In this article, we review fundamental concepts for making such inferences in randomized clinical trials or observational studies. The statistical framework consists of potential outcomes, an assignment mechanism, and probability distributions. Randomization-based and model-based methods of statistical inference are illustrated with a series of extracorporeal membrane oxygenation (ECMO) clinical trials, which are thought-provoking in that each trial used different assignment mechanisms.
著者
三輪 哲久
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.38, no.2, pp.163-170, 2018-03-01 (Released:2018-05-18)
参考文献数
12

In 2016 the American Statistical Association published “ASA Statement on Statistical Significance and P-Values.” In this statement it seems that the use of statistical tests or p-values is discouraged because they are misused and misinterpreted. I doubt whether a statistical procedure such as test of significance should be rejected because it is misused and misinterpreted. I pose some questions about the ASA statement.
著者
佐藤 恵子 岩崎 学 菅波 秀規 佐藤 俊哉 椿 広計
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.35, no.1, pp.37-53, 2014-08-31 (Released:2014-10-15)
参考文献数
18
被引用文献数
1

All statisticians are expected to produce statistical outcomes of high quality and reliability. To ensure reliability in statistical performance and outcomes and to meet societal expectations, certain standards of conduct (SOC) must be established such that individual statisticians embrace their own principles and so that the community of statisticians as a whole functions with more self-control.In 2008, the Biometric Society of Japan began revision of the code of conduct, and the working group drafted an SOC. This particular draft re.ected the opinions of statisticians and the basic concepts which aligned well with ethical guidelines of the American Statistical Association and the International Statistical Institute. As forced guidelines rarely result in full compliance and increased ethical conduct, the SOC offers a framework to encourage individual biostatisticians to establish and hold their own principles and to act responsibly with integrity.The SOC comprises a preamble, mission statement, values, ten principles and background information. The draft SOC was approved by the Council of the Biometric Society of Japan in November 2013.The SOC will help statisticians improve their capacity to perform sound statistical practices, improve the working environment, cultivate the next generation of statisticians with professionalism, and acquire societal trust.
著者
坂巻 顕太郎 兼清 道雄 大和田 章一 松浦 健太郎 柿爪 智行 高橋 文博 高沢 翔 萩原 駿祐 森田 智視
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.41, no.1, pp.55-91, 2020 (Released:2020-12-04)
参考文献数
44

It is common to use hypothesis testing to decide whether an investigational drug is ineffective and to determine sample size. However, it may not be good practice that only hypothesis testing is used for sample size determination, go/no-go decision making, and drug development decisions, especially in exploratory clinical trials. That is because important factors for decision making, such as treatment effects, drug development costs, and gains after launch, are not considered in hypothesis testing. The Bayesian decision theory is one of the approaches to consider such factors for decision making. The utility, which is defined by using important information such as cost, benefit, and disease severity, is used for decision making in the decision theory. In consideration of uncertainties of data and parameters, the expected value of the utility is used for decision making in the Bayesian decision theory. In this article, we explain basic concepts of the Bayesian decision theory, backward induction for calculation of expected value of utility in sequential decision-making, and introduce some approaches using the Bayesian decision theory in clinical trials. We summarize actions, utilities and sample size determination for applications of Bayesian decision theory in future clinical trials.
著者
平川 晃弘 浅野 淳一 佐藤 宏征 手良向 聡
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.39, no.2, pp.85-101, 2019-01-31 (Released:2019-05-11)
参考文献数
62
被引用文献数
1 1

In oncology, next generation sequencing and comprehensive genomic profiling have enabled detailed classification of tumors using molecular biology. It, however, may be unrealistic to conduct phase I-III trials according to each subpopulation based on the molecular subtypes. Common protocols that assess the combination of several molecular markers and their targeted therapies by means of multiple sub-trials are required. These protocols are called “master protocols,” and are drawing attention as a next-generation clinical trial design. In this review, we provide an overview of clinical trials based on master protocol including basket, umbrella, and platform trials along with their recent examples. We also discuss the statistical challenges encountered in their application.
著者
手良向 聡
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.29, no.2, pp.111-124, 2008-12-01 (Released:2011-09-15)
参考文献数
32

The aim of single-arm clinical trials of a new drug is to determine whether it has sufficient promising activity to warrant its further development. For the last several years Bayesian statistical methods have been proposed and used. Bayesian approaches are ideal for earlier phase exploratory trials or proof-of-concept studies as they take into account information that accrues during a trial. Posterior and predictive probabilities are then updated and so become more accurate as the trial progresses. If the relevant external information is available, the decision will be made with a smaller sample size. The goal of this paper is to provide a review for statisticians who use Bayesian methods for the first time or investigators who have some statistical background. In addition, a clinical trial is presented as a real example to illustrate how to conduct a Bayesian approach for single-arm clinical trials with binary endpoints.
著者
寒水 孝司 杉本 知之 濱崎 俊光
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.34, no.1, pp.35-52, 2013-08-31 (Released:2013-09-20)
参考文献数
61
被引用文献数
1

Clinical trials often employ two or more primary endpoints because a single endpoint may not provide a comprehensive picture of the intervention’s effects. In such clinical trials, a decision is generally made as to whether it is desirable to evaluate the joint effects on all endpoints (i.e., co.primary endpoints) or at least one of the endpoints. This decision defines the alternative hypothesis to be tested and provides a framework for approaching trial design. In this article, we discuss recent statistical issues in clinical trials with multiple primary endpoints. Especially, we introduce the methods for power and sample size determinations in clinical trials with co-primary endpoints, considering the correlations among the endpoints into the calculations. We also discuss the methods to alleviate conservativeness of testing co-primary endpoints.
著者
川口 淳
出版者
日本計量生物学会
雑誌
計量生物学 (ISSN:09184430)
巻号頁・発行日
vol.33, no.2, pp.145-174, 2013-02-28 (Released:2013-03-07)
参考文献数
116

Imaging techniques have been used for effectively studying the brain in a non-invasive manner in several fields, for example, psychiatry and psychology. In this review, we focus on two imaging techniques that provide different views of brain structure and function. Structural magnetic resonance imaging (sMRI) provides information about various tissue types in the brain, for example, gray matter, white matter, and cerebrospinal fluid. Functional MRI (fMRI) measures brain activity by detecting changes in cerebral blood flow. These techniques enable high-quality visualization of brain activity or the location of atrophies; moreover, these techniques facilitate the study of disease mechanisms in the healthy brain and might lead to the development of effective therapies or drugs against such diseases. However, raw MRI data must be statistically analyzed to obtain objective answers to clinical questions. Therefore, statistical methods play a very important role in brain research. Here, we briefly review the most commonly used statistical analyses, namely, data pre-processing, general linear model, random field theory, mixed effect model, independent component analysis, network analysis, and discriminant analysis. Further, we provide information about brain imaging data structure and introduce useful software to implement these methods.