Use SPSS to explore and analyze data outliers

When we are conducting clinical analysis of data, sometimes we often encounter when the results of clinical data are different from clinical common sense. For example, according to clinical experience, the index B in the figure below should continue to rise as the index A rises, but when the index A rises between 115 and 126, the index B decreases. Why is this?

We can bring up the data of this index, compare it with other data, find the difference, and analyze it.
Use SPSS to open the data

and click Convert to encode it as a different variable

. Select the A index and select the range option

A index In the section 115-126, convert it to 1, and convert the other indicators to 2

to get the new classification indicator A1.
We can analyze the difference between the two sets of data through one-way analysis of variance and chi-square test. In fact, we use R language The Tableone package is easier to do, but today I’m talking about SPSS, and there are not many variables.
First click on Comparative Analysis-Mean-One-way ANOVA test

A1 as a factor, and other continuous variables are selected into the dependent variable list. The method is to select LSD

and then obtain the results of the outlier group and the non-outlier group in other continuous variables,

and then we perform categorical variables For comparison, use the chi-square test.
Analysis—Descriptive Statistics—Cross



Tabulation. There
are too many chi-square tables to list them all. As far as this data is concerned, the final conclusion is that the preoperative bleeding of the outlier data There is a clear difference between the amount and the non-abnormal part, and any clinical significance needs to be analyzed in detail. The solution is to perform subgroup analysis or increase exclusion. The operation is not difficult, mainly clinical thinking.