The previous articles of the Statistical Round in the Korean Journal of Anesthesiology posed a strong enquiry on the issue of null hypothesis significance testing (NHST). P values lie at the core of NHST and are used to classify all treatments into two groups: "has a significant effect" or "does not have a significant effect." NHST is frequently criticized for its misinterpretation of relationships and limitations in assessing practical importance. It has now provoked criticism for its limited use in merely separating treatments that "have a significant effect" from others that do not. Effect sizes and CIs expand the approach to statistical thinking. These attractive estimates facilitate authors and readers to discriminate between a multitude of treatment effects. Through this article, I have illustrated the concept and estimating principles of effect sizes and CIs.
Anesthesiology ; Confidence Intervals* ; Thinking

Confidence Intervals ; Probability ; Population

The null hypothesis significance test method is popular in biological and medical research. Many researchers have used this method for their research without exact knowledge, though it has both merits and shortcomings. Readers will know its shortcomings, as well as several complementary or alternative methods, as such the estimated effect size and the confidence interval.
Biostatistics ; Confidence Intervals ; Models, Statistical

The estimation of 90% parametric confidence intervals (CIs) of mean AUC and Cmax ratios in bioequivalence (BE) tests are based upon the assumption that formulation effects in log-transformed data are normally distributed. To compare the parametric CIs with those obtained from nonparametric methods we performed repeated estimation of bootstrap-resampled datasets. The AUC and Cmax values from 3 archived datasets were used. BE tests on 1,000 resampled datasets from each archived dataset were performed using SAS (Enterprise Guide Ver.3). Bootstrap nonparametric 90% CIs of formulation effects were then compared with the parametric 90% CIs of the original datasets. The 90% CIs of formulation effects estimated from the 3 archived datasets were slightly different from nonparametric 90% CIs obtained from BE tests on resampled datasets. Histograms and density curves of formulation effects obtained from resampled datasets were similar to those of normal distribution. However, in 2 of 3 resampled log (AUC) datasets, the estimates of formulation effects did not follow the Gaussian distribution. Bias-corrected and accelerated (BCa) CIs, one of the nonparametric CIs of formulation effects, shifted outside the parametric 90% CIs of the archived datasets in these 2 non-normally distributed resampled log (AUC) datasets. Currently, the 80~125% rule based upon the parametric 90% CIs is widely accepted under the assumption of normally distributed formulation effects in log-transformed data. However, nonparametric CIs may be a better choice when data do not follow this assumption.
Area Under Curve ; Confidence Intervals ; Phenothiazines ; Therapeutic Equivalency

OBJECTIVE: To compare different methods for calculating sample size based on confidence interval estimation for a single proportion with different event incidences and precisions. METHODS: We compared 7 methods, namely Wald, AgrestiCoull add z2, Agresti-Coull add 4, Wilson Score, Clopper-Pearson, Mid-p, and Jefferys, for confidence interval estimation for a single proportion. The sample size was calculated using the search method with different parameter settings (proportion of specified events and half width of the confidence interval [ω=0.05, 0.1]). With Monte Carlo simulation, the estimated sample size was used to simulate and compare the width of the confidence interval, the coverage of the confidence interval and the ratio of the noncoverage probability. RESULTS: For a high accuracy requirement (ω =0.05), the Mid-p method and Clopper Pearson method performed better when the incidence of events was low (P < 0.15). In other settings, the performance of the 7 methods did not differ significantly except for a poor symmetry of the Wald method. In the setting of ω=0.1 with a very low p (0.01-0.05), failure of iteration occurred with nearly all the methods except for the Clopper-Pearson method. CONCLUSION Different sample size determination methods based on confidence interval estimation should be selected for single proportions with different parameter settings.
Confidence Intervals ; Sample Size ; Computer Simulation ; Monte Carlo Method ; Probability

BACKGROUND/AIMS: Complete endotracheal tube obstruction is a medical emergency, and partial occlusion causes increased breathing rates and failure to wean off mechanical ventilation. Partial occlusion may be underestimated due to the lack of proper detection methods. We tested whether the sound of an endotracheal tube could be used to detect an endotracheal tube obstruction using an in vitro model. METHODS: An endotracheal tube was connected to a ventilator on one end and a test lung on the other. Sounds were recorded with a microphone located inside the endotracheal tube via a connector. During mechanical ventilation, we changed the endotracheal tube internal diameter from 5.0 to 8.0 mm and different grades of obstruction at different sites were used along the tube. Sound energy was compared among the different conditions. RESULTS: The energy of endotracheal tube sounds was positively correlated with the internal diameter and negatively correlated with the degree of obstruction. The rate of decline in energy differed with obstruction location. When the obstruction was more distal, the rate of decline in endotracheal sound energy was more rapid. CONCLUSIONS: Changes in the sound of an endotracheal tube can be used to detect an obstruction. Further studies are needed for clinical application.
*Airway Obstruction ; Confidence Intervals ; Exhalation ; Humans ; *Intubation, Intratracheal ; Models, Theoretical ; Positive-Pressure Respiration ; Statistics as Topic

OBJECTIVESTo formulate an equation for fine mapping of disease loci under complex conditions and determine the marker-disease distance in a specific case using this equation.

METHODSLewontin's linkage disequilibrium (LD) measure D' was used to formulate an equation for mapping disease genes in the presence of phenocopies, locus heterogeneity, gene-gene and gene-environment interactions, incomplete penetrance, uncertain liability and threshold, incomplete initial LD, natural selection, recurrent mutation, high disease allele frequency and unknown mode of inheritance. This equation was then used to determine the distance between a marker ( epsilon 4 within the apolipoprotein E gene, APOE) and Alzheimer's disease (AD) loci using published data.

RESULTSAn equation was formulated for mapping disease genes under the above conditions.If these conditions are present but ignored, then recombination fraction theta between marker and disease loci will be either overestimated or estimated with little bias. Therefore, an upper limit of theta can be obtained. AD has been found to be associated with the marker allele epsilon 4 in Africans, Asians, and Caucasians. This suggests that the AD- epsilon 4 allelic LD predates the divergence of peoples occurring 100 000 years ago. With the age of AD- epsilon 4 allelic LD so estimated, the maximal distance was calculated to be 23.2 kb (mean 5.8 kb).

CONCLUSIONS(1) A method is developed for LD mapping of susceptibility genes. (2) A mutation within the APOE gene itself, among others, is responsible for the susceptibility to AD, which is supported by recent evidence from studies using transgenic mice.

Number needed to treat (NNT) is a simple and effective index for clinical therapeutic effect assessment worldwide accepted in recent years. By calculation of absolute risk reduction (ARR) of classified variables, it made the effect estimate reflect objectively the therapeutic effect of an intervention. However, clinical application of this index was introduced rarely in Chinese literature. With the examples from some published clinical reports, the calculation of NNT and its 95% confidence interval were demonstrated in this paper, and its application was illustrated by some relevant terms explanation and Meta-analysis methods introduction.
Confidence Intervals ; Humans ; Meta-Analysis as Topic ; Numbers Needed To Treat ; Outcome Assessment (Health Care) ; methods ; Treatment Outcome

Comparative pharmacokinetic (PK) analysis is often carried out throughout the entire period of drug development, the common approach for the assessment of pharmacokinetics between different treatments requires that the individual PK parameters, which employs estimation of 90% confidence intervals for the ratio of average parameters, such as AUC and Cmax, these 90% confidence intervals then need to be compared with the pre-specified equivalent interval, and last we determine whether the two treatments are equivalent. Unfortunately in many clinical circumstances, some or even all of the individuals can only be sparsely sampled, making the individual evaluation difficult by the conventional non-compartmental analysis. In such cases, nonlinear mixed effect model (NONMEM) could be applied to analyze the sparse data. In this article, we simulated a sparsely sampling design trial based on the dense sampling data from a truly comparative PK study. The sparse data were analyzed with NONMEM method, and the original dense data were analyzed with non-compartment analysis. Although the trial design and analysis methods are different, the 90% confidence intervals for the ratio of PK parameters based on 1000 Bootstrap are very similar, indicated that the analysis based on NONMEM is a reliable method to treat with the sparse data in the comparative pharmacokinetic study.
Area Under Curve ; Confidence Intervals ; Humans ; Nonlinear Dynamics ; Pharmacokinetics ; Sampling Studies

Scientific journals are important scholarly forums for sharing research findings. Editors have important roles in safeguarding standards of scientific publication and should be familiar with correct presentation of results, among other core competencies. Editors do not have access to the raw data and should thus rely on clues in the submitted manuscripts. To identify probable errors, they should look for inconsistencies in presented results. Common statistical problems that can be picked up by a knowledgeable manuscript editor are discussed in this article. Manuscripts should contain a detailed section on statistical analyses of the data. Numbers should be reported with appropriate precisions. Standard error of the mean (SEM) should not be reported as an index of data dispersion. Mean (standard deviation [SD]) and median (interquartile range [IQR]) should be used for description of normally and non-normally distributed data, respectively. If possible, it is better to report 95% confidence interval (CI) for statistics, at least for main outcome variables. And, P values should be presented, and interpreted with caution, if there is a hypothesis. To advance knowledge and skills of their members, associations of journal editors are better to develop training courses on basic statistics and research methodology for non-experts. This would in turn improve research reporting and safeguard the body of scientific evidence.
Confidence Intervals ; Editorial Policies ; Journalism ; Normal Distribution ; Peer Review ; Publications ; Research Design ; Research Report