如何考虑三臂非劣效性检验的置信区间问题—— interpret_noninferiority {confinterpret}

Posted 2023-02-21 统计学小王子

tags:

篇首语：本文由小常识网(cha138.com)小编为大家整理，主要介绍了如何考虑三臂非劣效性检验的置信区间问题—— interpret_noninferiority {confinterpret}相关的知识，希望对你有一定的参考价值。

0引言

近年来,随着科技和医疗卫生事业的发展,国内外创新药物临床试验逐年增多,临床上某些领域虽然已存在一些有效药物,但是这些药物存在一定的缺陷,如副作用多、价格昂贵、使用繁琐、患者依从性低等。而一些新研发的药物虽然疗效上未必能够提高,但因其具有安全性高、耐药性小、价格低廉、使用方便、患者依从性高等优点仍然值得推广使用。如何有效筛选和评价这些药物,借助常规无效假设的优效性临床试验已难以胜任,非劣效试验¹设计应运而生并越来越受到大家的青睐。

1. 初识R包

1.1 安装与载入

> install.packages("confinterpret")
试开URL’https://mirror-hk.koddos.net/CRAN/bin/windows/contrib/4.1/confinterpret_1.0.0.zip'
Content type 'application/zip' length 118786 bytes (116 KB)
downloaded 116 KB

package ‘confinterpret’ successfully unpacked and MD5 sums checked

The downloaded binary packages are in
        C:\\Windows\\Temp\\RtmpcdwjA8\\downloaded_packages
> library(confinterpret)
Warning message:
程辑包‘confinterpret’是用R版本4.1.3 来建造的

1.2 主要函数

> ls("package:confinterpret")
 [1] "confinterpret"                   "interpret_equivalence"          
 [3] "interpret_noninferiority"        "interpret_superiority"          
 [5] "interpretation_result"           "interpretation_set"             
 [7] "interpretations_equivalence"     "interpretations_noninferiority" 
 [9] "interpretations_superiority"     "plot_interpretation_result_list"
[11] "validate_interpretation_result"  "validate_interpretation_set"

2. 函数interpret_noninferiority

由于篇幅原因，本文只介绍interpret_noninferiority函数，其他函数读者感兴趣的自行探索。
函数体

> interpret_noninferiority
function (ci, actual_null = 0, ni_margin = 0.1, groups = c("Control intervention", 
    "Test intervention"), beneficial_outcome = TRUE)

例子

> # Establish a test confidence interval
> ci_test <- matrix(c(-0.05, 0.05),
+                   nrow = 1, dimnames = list("estimate",
+                                             c("2.5 %", "97.5 %")))
> interpret_noninferiority(ci_test, 0, 0.1, c("Treatment as usual",
+                                             "New treatment"))
Object of class 'interpretation_result', with interpretation values:
  $interpretation_short: Non-inferior
  $interpretation:       New treatment non-inferior to Treatment as usual but not shown to be superior
  $interpretation_md:    New treatment **non-inferior** to Treatment as usual but not shown to be superior

And parameters:
  $parameters$ci:
         2.5 % 97.5 %
estimate -0.05   0.05

  $parameters$interpretation_set:
interpretation_set object. For details use 'print(x$parameters$interpretation_set)'.

  $parameters$interpretation_set_name:  interpretations_noninferiority

  $parameters$boundaries:
[1] -0.1  0.0

  $parameters$comparison_labels:
comparison_intervention     tested_intervention 
   "Treatment as usual"         "New treatment" 

  $parameters$low_to_high:   TRUE>