Martin J M, Benamghar L, Martin J
Pathol Biol (Paris). 1979 Mar;27(3):161-7.
To draw the best conclusions from a set of medical data, one must first choose correct, valid, representative and well-expressed information. So it's not sufficient to collect data, but we have to valid them before including them in the records. After having put out the general characteristics of any validation and defined the fields where they can be applied we recommend a method over three steps. First, the physician alone, then the physician with an informatician and again the physician seem concerned by the steps of control. It is particularly important to develop the validation of chronology and generally speaking of evolutive data. Of course, the costs of validation with respect to the large groups of data must be taken into account.
为了从一组医学数据中得出最佳结论，首先必须选择正确、有效、有代表性且表述清晰的信息。因此，仅仅收集数据是不够的，在将数据纳入记录之前，我们必须对其进行验证。在阐述了任何验证的一般特征并定义了可应用验证的领域之后，我们推荐一种分三步的方法。首先，由医生单独进行，然后医生与信息专家一起，再次由医生进行似乎涉及控制步骤的工作。对时间顺序以及一般而言的演变数据进行验证尤为重要。当然，必须考虑针对大量数据进行验证的成本。
