In this paper, we introduce meaning and purpose of confidence interval (CI) in evidence-based medicine, For example, RRR ,ARR ,NNT. It s referance for user and doer of EBM in China.
To show that a new drug is better than, as good as, or no worse than that of a known effective drug. Theoretically, it is necessary to confirm the efficacy of a treatment, but the current practice of clinical trial suggests that there exists many problems in its confirmation including the objectives of clinical investigation vary based on the fact that more and more clinical trials use active controls. Applied statistical methods have to adapt to these changes. In this paper, we illustrated some statistical issues of confirming efficacy in clinical trials, including its conditions, the determination of clinical margin, the forms of the null and alternative hypothesis and confidence intervals, the choice of endpoints and some miscellaneous considerations. We bly suggests that it is necessary to make biostatisticians and clinical trialists understand the importance of using the right statistical methods when investigating clinical trials. We also think these methods introduced in the paper may provide some help in trial design and evaluation.
Missing data represent a general problem in many scientific fields, especially in medical survival analysis. Dealing with censored data, interpolation method is one of important methods. However, most of the interpolation methods replace the censored data with the exact data, which will distort the real distribution of the censored data and reduce the probability of the real data falling into the interpolation data. In order to solve this problem, we in this paper propose a nonparametric method of estimating the survival function of right-censored and interval-censored data and compare its performance to SC (self-consistent) algorithm. Comparing to the average interpolation and the nearest neighbor interpolation method, the proposed method in this paper replaces the right-censored data with the interval-censored data, and greatly improves the probability of the real data falling into imputation interval. Then it bases on the empirical distribution theory to estimate the survival function of right-censored and interval-censored data. The results of numerical examples and a real breast cancer data set demonstrated that the proposed method had higher accuracy and better robustness for the different proportion of the censored data. This paper provides a good method to compare the clinical treatments performance with estimation of the survival data of the patients. This provides some help to the medical survival data analysis.
Predicting the termination of paroxysmal atrial fibrillation (AF) may provide a signal to decide whether there is a need to intervene the AF timely. We proposed a novel RdR RR intervals scatter plot in our study. The abscissa of the RdR scatter plot was set to RR intervals and the ordinate was set as the difference between successive RR intervals. The RdR scatter plot includes information of RR intervals and difference between successive RR intervals, which captures more heart rate variability (HRV) information. By RdR scatter plot analysis of one minute RR intervals for 50 segments with non-terminating AF and immediately terminating AF, it was found that the points in RdR scatter plot of non-terminating AF were more decentralized than the ones of immediately terminating AF. By dividing the RdR scatter plot into uniform grids and counting the number of non-empty grids, non-terminating AF and immediately terminating AF segments were differentiated. By utilizing 49 RR intervals, for 20 segments of learning set, 17 segments were correctly detected, and for 30 segments of test set, 20 segments were detected. While utilizing 66 RR intervals, for 18 segments of learning set, 16 segments were correctly detected, and for 28 segments of test set, 20 segments were detected. The results demonstrated that during the last one minute before the termination of paroxysmal AF, the variance of the RR intervals and the difference of the neighboring two RR intervals became smaller. The termination of paroxysmal AF could be successfully predicted by utilizing the RdR scatter plot, while the predicting accuracy should be further improved.
ObjectiveTo set reference intervals of the levels of thyroid hormones among normal pregnant women without presence of thyroid antibodies during three trimesters of pregnancy in Quanzhou city, Fujian province. MethodsA total of 490 pregnant women during 4-39 week pregnancy without presence of thyroid antibodies were enrolled in Quanzhou city, Fujian province. Levels of thyroid stimulating hormone (TSH), free triiodothyronine (FT3), free thyroxine (FT4) and thyroid peroxidase antibodies (TPO-Ab) were detected through the electrochemistry immunoassay (ECL) method. In addition, a total of 51 healthy women without pregnancy were enrolled to set the reference intervals of levels of thyroid hormones among normal pregnant women without presence of thyroid antibodies. ResultsThe median levels of TPO-Ab were in the reference intervals provided by the pharmaceutical factory. Levels of FT3 and FT4 gradually decreased from the first to the third trimester (P < 0.01), levels of serum TSH gradually increasd from the first to the third trimester (P < 0.01). Compared with those of pregnant women, levels of thyroid hormone in normal non-pregnant women were higher in the first trimester, lower in the second and the third trimesters (P < 0.01). During three trimesters, the reference intervals of FT3 in the three trimesters were (first: 3.75 to 7.23; second 3.31 to 4.9; and third: 3.16 to 4.48 pmol/L); the reference intervals of FT4 were (first: 12.85 to 25.3; second: 12.03 to 20.14; and third: 11.02 to 19.43 pmol/L); and the reference intervals of TSH were (first: 0.01 to 3.79; second: 1.09 to 4.19; and third: 1.08 to 5.95 mIU/L), respectively. ConclusionThrough this detection, we set the levels of thyroid hormones among normal pregnant women without presence of thyroid antibodies during three trimesters of pregnancy in Quanzhou city.
Objective To investigate the effectiveness and long-term stability of small fenestration vertebral bone grafting and transpedicular bone grafting in the treatment of Denis types A and B thoracolumbar burst fractures. Methods Between January 2012 and February 2014, 50 patients with Denis type A or B thoracolumbar burst fractures were treated with vertebroplasty and pedicle screw rod fixation system, and the clinical data were retrospectively analyzed. Small fenestration vertebral bone grafting by trans-interlaminar approach was used in 30 cases (group A), and bone grafting by unilateral transpedicular approach was used in 20 cases (group B). X-ray and CT examinations of the thoracolumbar vertebrae were performed routinely before and after operation. There was no significant difference in sex, age, cause of injury, time from injury to operation, fracture type, injury segment, and preoperative Frankel classification, the percentage of the anterior body height of the injured vertebra, and visual analogue scale (VAS) score between two groups (P>0.05). There was significant difference in preoperative Cobb angle of kyphosis between two groups (P<0.05). The Cobb angle of kyphosis, the percentage of the anterior body height of the injured vertebra, and the recovery of neurological function were recorded and compared between two groups. Results The patients were followed up for 16-31 months (mean, 19.1 months) in group A and for 17-25 months (mean, 20.2 months) in group B. Primary healing of incisions was obtained in the two groups; no nerve injury and other operative complications occurred. The neurological function was improved in varying degrees in the other patients with neurological impairment before operation except patients at grade A of Frankel classification. The lumbar back pain was relieved in two groups. There was significant difference in VAS score between before operation and at 3 months after operation or last follow-up in two groups (P<0.05), but no significant difference was found between at 3 months and last follow-up in two groups and between two groups at each time point after operation (P>0.05). X-ray examination showed that there was no breakage of nail and bar, or dislocation and loosening of internal fixation during follow-up period. The bone grafts filled well and fused in the fractured vertebra. The vertebral height recovered well after operation. The percentage of the anterior body height of the in-jured vertebra and Cobb angle of kyphosis at 1 week, 3 months, and last follow-up were significantly better than preope-rative ones in two groups (P<0.05), but there was no significant difference between different time points after operation (P>0.05), and between two groups at each time point after operation (P>0.05). Conclusion For Denis types A and B thoracolumbar burst fractures, vertebral bone grafting and pedicle screw internal fixation through interlaminal small fene-stration or transpedicular approach can restore the vertebral height, correct kyphosis, and maintain the vertebral stability, which reduce the risk of complications of loosening and breakage of internal fixators. The appropriate bone grafting approach can be chosen based on the degree of spinal canal space occupying, collapse of vertebral and spinal cord injury.
Etiological and prognostic studies always directly reported effect size with its 95% confidence interval, hence, data transformation was needed when performing meta-analysis based on these studies. Using the data of risk ratio, hazard ratio, odds ratio and 95% confidence interval as an example, this paper introduces the process of using RevMan 5.3 software to convert data and perform meta-analysis.
Objectives To detect expressions of heat shock protein 70 (HSP70) and glial fibrillary acidic protein (GFAP) , and to estimate the post-injury interval after concussion of brain via the ratios of percentage of HSP70/GFAP-positive cells. Methods We established a brain concussion model of rat. Tissue levels of HSP70 and GFAP were determined by immunohistochemical staining at different time points after injury. Finally, the relationship between the ratio of percentage of HSP70/GFAP-positive cells and the post-injury interval was measured. Results The ratio of percentage of positive cells (increased from 7.15 to 11.73) and the percentage of HSP70-positive cells (P<0.05, compared with control group) increased, and the percentage of GFAP-positive cells did not change remarkably (P<0.05, compared with control group); the post-injury interval was between 0.5 hour and 3 hours. High ratio (>6.66) and high percentage of HSP70 and GFAP-positive cells (P<0.05, compared with control group) indicated the post-injury interval was between 3 and 12 hours. A low ratio (<6.66) and high percentage of HSP70 and GFAP-positive cells (P<0.05, compared with control group) suggested that the post-injury interval was later than 12 hours. Conclusion By analyzing the variation rule of the ratio of percentage positive cells after brain concussion, the post-injury interval after concussion of brain could be estimated.
Most statistical data in observational studies is expressed as the effect value and its 95% confidence interval (95% CI), which do not correspond to the data format used for traditional meta-analyses, so special data conversion is to be needed when Review Manager software is applied to do a meta-analysis for this type of data, which will make the operation complicated and cumbersome. In addition, Stata software is such a powerful statistical software that can be used directly to conduct a meta-analysis with the effect value and its 95% CI. Therefore, it is an indispensable statistical tool for meta-analysis in observational studies. And this study will give a brief introduction how to use Stata software to conduct a meta-analysis with effect value and its 95% CI based on the published meta-analysis data.
ObjectiveTo investigate the effect of the interval between neoadjuvant chemoradiotherapy (nCRT) and surgery on the clinical outcome of esophageal cancer.MethodsPubMed and EMbase databases from inception to March 2018 were retrieved by computer. A random-effect model was used for all meta-analyses irrespective of heterogeneity. The meta-analysis was performed by RevMan5.3 software. The primary outcomes were operative mortality, incidence of anastomotic leakage, and overall survival; secondary outcomes were pathologic complete remission rate, R0 resection rate, and positive resection margin rate.ResultsA total of 17 studies with 18 173 patients were included. Among them, 13 were original studies with 2 950 patients, and 4 were database-based studies with a total of 15 223 patients. The results showed a significant positive correlation between the interval and operative mortality (Spearman coefficient=0.360, P=0.027). Dose-response meta-analysis revealed that there was a relatively better time window for surgery after nCRT. Further analysis for primary outcomes at different time cut-offs found the following results: (1) when the time cut-off point within 30-70 days, the shorter interval was associated with a reduced operative mortality (7-8 weeks: RR=0.67, 95% CI 0.55-0.81, P<0.05; 30-46 days: RR=0.63, 95%CI 0.47-0.85, P<0.05; 60-70 days: RR=0.64, 95%CI 0.48-0.85, P<0.05); (2) when the time cut-off point within 30-46 days, the shorter interval correlated with a reduced incidence of anastomotic leakage (RR=0.39, 95%CI 0.21-0.72, P<0.05); when the time cut-off point within 7-8 weeks, the shorter interval could achieve a critical-level effect of reducing the incidence of anastomotic leakage (RR=0.73, 95%CI 0.52-1.03, P>0.05); (3) when the time cut-off point within 7-8 weeks, increased interval significantly was associated with the poor overall survival (HR=1.17, 95% CI 1.00-1.36, P<0.05). Secondary outcomes found that the shorter interval could significantly reduce the positive resection margin rate (RR=0.53, 95% CI 0.38-0.75, P<0.05) when time cut-off point within 56-60 days.ConclusionShortening the interval between nCRT and surgery can reduce the operative mortality, the incidence of anastomotic leakage, long-term mortality risk, and positive resection margin rate. It is recommended that surgery should be performed as soon as possible after the patient's physical recovery, preferably no more than 7-8 weeks, which supports the current study recommendation (within 3-8 weeks after nCRT).