COX regression plays an important role in statistics. Most of it is used in the content of survival analysis such as tumors and blood diseases. In the previous content, we taught you how to use the R language to establish COX regression and draw a nomogram (Nomogram). COX regression is an overview, so I won’t talk about it here. In the previous section, we already talked about how to use SPSS to perform multiple imputation on missing data. Today we talk about how to use SPSS to perform COX regression analysis on imputed data. First open the data in our previous chapter, 1 set of original data and 5 sets of imputed data, a total of 6 sets of data,
and then click Analysis-Survival Analysis-COX regression
to select time as time, status as status, and other indicators Select the covariate, and Imputation select the stratification (this indicator is generated by SPSS interpolation, if you do ordinary COX regression, you don’t do this stratification, the others are the same),
click the classification, put sex, phecog Include categorical covariates, and change the reference category to the first
graph option: Survival analysis function graph
Save selected: Survival analysis function
Select among options: Exp(B) 95% confidence interval
defines the status of the status (that is, the state of death), I here is 2 means death. The
final point is determined, and the following figure is obtained. This is the HR and credibility interval of each data, and finally the credible interval after the merger, and the last article reports the credible interval after the merger
The 6 lines of the survival function in the figure below are almost coincident, indicating that the imputed data is basically the same as the original data, and there is no obvious change, indicating that we have imputed very well and did not cause significant bias in the results.
If you want to see the difference in the survival of men and women, you can use gender stratification
OK. We have already presented all the data needed in SCI papers. Is it easy to
pay attention to it? More exciting articles are available at zero Basic research
