## 2011年10月28日 星期五

### Comments in the writing course (10/27)

1. meaningful chance表示臨床意義上之改變，並不等於統計上的顯著改變。
2. 名詞太長，動詞放得太後面，此寫法不易看得懂，例如： The minimum and maximum values...were analysed。
3. 在研究中，我是取最小值和最大值之90% CI of RCIp，並非分析最小和最大值，使用analyza此字不適當。
4. 句子過長，不易理論，如下斜體字。
5. practical和useful為類似描述，使用一字即可。
6. 在前言中，概念性和技術性理念需要說明清楚。例如：90% CI is examined for detecting of significant change, representing that 5% ...需要在前言說明，不是放在方法。

Introduction
A range of 90% confidence interval (CI) of a reliable change index modified for practice (RCIp) estimates random measurement error and provides cut-score values for determining whether there is a meaningful change, after considering the practice effect. The score change in pre-and post-test is beyond the range of cut scores would be taken account of a significant improvement or deterioration. The magnitude of 90% CI of RCIp is practical and useful for clinicians and researchers in interpreting the test results.
Data analysis

The minimum and maximum values of 90% CI of RCIp across three comparisons at 4 time points (i.e. Time 0-1, Time 1-2, Time 2-3) were analyzed for providing the conservative criteria to interpret score changes in patients with stroke. The formula of 90% CI of RCIp was calculated:
90% CI of RCIp = mean practice effect ± 1.645 × SEdiff
SEdiff = √2(SEM)2
SEM = √1-rxx
where mean practice effect is the mean of difference between two test scores; SEdiff is the standard error of the differences; SEM is the standard error of measurement; and rxx is the ICC value. 90% CI is examined for detecting of significant change, representing that 5% of patients may be anticipated to fall above and 5% of patients may be expected to fall below the cutoff, because of chance rather than true change beyond the practice effect.

Introduction
The reliable change index (RCI) estimates random measurement error and calculates whether a change in an individual’s score is statistically significant. A reliable change index modified for practice (RCIp) which takes into account practice effect is the RCI corrected for practice. 90% confidence interval (CI) of RCIp is examined for detecting of significant change, representing that 5% of patients may be anticipated to fall above and 5% of patients may be expected to fall below the cutoff. A range of 90% CI of RCIp is used for providing cut-score values. The score change in pre-and post-test is beyond the range of cut scores would be taken account of a significant improvement or deterioration. The magnitude of 90% CI RCIp is useful for clinicians and researchers in interpreting the test results.
Data analysis

We chose the minimum and maximum values of 90% CI of RCIp across three comparisons at 4 time points (i.e. Time 0-1, Time 1-2, Time 2-3) for providing the conservative criteria to interpret score changes in patients with stroke. The formula of 90% CI RCIp was calculated:
90% CI RCIp = mean practice effect ± 1.645 × SEdiff
SEdiff = √2(SEM)2
SEM = √1-rxx
where mean practice effect is the mean of difference between two test scores; SEdiff is the standard error of the differences; SEM is the standard error of measurement; and rxx is the ICC value.