Grunberg S M, Groshen S, Steingass S, Zaretsky S, Meyerowitz B
Division of Medical Oncology, University of Southern California Comprehensive Cancer Center, Los Angeles, USA.
Qual Life Res. 1996 Feb;5(1):65-72. doi: 10.1007/BF00435970.
Two common formats for grading quality of life parameters are descriptive choices (mild, moderate, severe) and visual analogue scales. However the quantitative relationship between descriptive terminology and visual analogue scale scores has not been determined. A content neutral questionnaire was administered to 213 evaluable subjects who were asked to place the descriptors 'mildly', 'moderately', and 'severely' (presented in random order) on 100mm visual analogue scales. Visual analogue scales were presented without and then with hashmarks at 25mm, 50mm, and 75mm. Median visual analogue scale values for the descriptive terms differed significantly without hashmarks ('mildly' = 24mm', moderately' = 43mm, 'severely' = 84mm; p < 0.001) and also with hashmarks ('mildly' = 31mm, 'moderately' = 49mm, 'severely' = 85mm; p < 0.001). Comparison of interquartile range values (25th-75th percentile) revealed a distinct meaning for 'severely' (68-93mm) but marked overlap between 'mildly' (10-45mm) and 'moderately' (22-53mm). Errors of order (order other than 'mildly' < 'moderately' < 'severely') were made by 91 subjects. The discrepancy 'moderately' < 'mildly' accounted for most of these errors (72 subjects). Median values for 'mildly', 'moderately', and 'severely' are distinct and approximately linear on a visual analogue scale for large populations. However there is significant confusion between the terms 'mildly' and 'moderately' for individual subjects. Visual analogue scales can reveal finer quantitative differences than descriptive terms but require a significant time commitment for instruction and administration. Descriptive terms on a word-graphic scale or descriptive terms with numerical values to reenforce order of severity (0 = absent, 1 = 'mildly', 2 = 'moderately', 3 = 'severely') may be reasonable alternatives.
生活质量参数评分的两种常见形式是描述性选择（轻度、中度、重度）和视觉模拟量表。然而，描述性术语与视觉模拟量表评分之间的定量关系尚未确定。对213名可评估的受试者进行了一份内容中立的问卷调查，要求他们将描述词“轻度”“中度”和“重度”（以随机顺序呈现）放在100毫米的视觉模拟量表上。视觉模拟量表开始时没有刻度标记，然后有25毫米、50毫米和75毫米的刻度标记。描述性术语的视觉模拟量表中位数在没有刻度标记时差异显著（“轻度”=24毫米，“中度”=43毫米，“重度”=84毫米；p<0.001），有刻度标记时也差异显著（“轻度”=31毫米，“中度”=49毫米，“重度”=85毫米；p<0.001）。四分位间距值（第25百分位数至第75百分位数）的比较显示“重度”有明显的含义（68 - 93毫米），但“轻度”（10 - 45毫米）和“中度”（22 - 53毫米）之间有明显重叠。91名受试者出现了顺序错误（顺序不是“轻度”<“中度”<“重度”）。这些错误大多是由“中度”<“轻度”的差异造成的（72名受试者）。对于大量人群，“轻度”“中度”和“重度”的中位数在视觉模拟量表上是不同的且大致呈线性。然而，对于个体受试者，“轻度”和“中度”这两个术语之间存在显著混淆。视觉模拟量表比描述性术语能揭示更细微的定量差异，但需要大量时间用于指导和实施。词图量表上的描述性术语或带有数值以强化严重程度顺序的描述性术语（0 = 无，1 = “轻度”，2 = “中度”，3 = “重度”）可能是合理的替代方法。
