Findley T W, Daum M C
Kessler Institute for Rehabilitation, Department of Physical Medicine and Rehabilitation, UMDNJ-New Jersey Medical School, West Orange 07052.
Am J Phys Med Rehabil. 1989 Jun;68(3):150-7.
This article guides the design and execution of a small research project using existing clinical data. The most important point in experimental design is to identify potential difficulties and limitations before you start by 1) review of published studies, 2) use of your clinical experience and 3) review of individual case records. Some of these can be addressed by changes in the study design, but some are inherent in the data and methods we are forced to use. The choice of study design may be based primarily on the quality of the clinical data and available resources for additional data collection. The level of measurement (nominal, ordinal, interval or ratio) of your data must first be determined as it limits the descriptive and statistical techniques you can use. After you decide how many variables to include, a rough guess of sample size will help you select your charts for review. Actual review of three charts will further pinpoint any difficulties and will allow you to revise your study and make an accurate estimate of time to completion. Given the long time span of most projects, accurate record keeping is essential.
本文指导利用现有临床数据进行小型研究项目的设计与实施。实验设计中最重要的一点是在开始前通过以下方式识别潜在困难和局限性：1）查阅已发表的研究；2）运用临床经验；3）查阅个体病例记录。其中一些问题可通过改变研究设计来解决，但有些则是我们被迫使用的数据和方法所固有的。研究设计的选择可能主要基于临床数据的质量以及用于额外数据收集的可用资源。必须首先确定数据的测量水平（名义、有序、区间或比率），因为它限制了你可以使用的描述性和统计技术。在决定纳入多少变量后，对样本量进行大致估算将有助于你选择用于审查的图表。实际审查三张图表将进一步明确任何困难，并使你能够修改研究并准确估计完成时间。鉴于大多数项目的时间跨度较长，准确的记录保存至关重要。
