Hello, I'm a medical doctor in Asia.
Conducting a prospective cohort analysis, I have two problems.
1. Can I make table 1(baseline characteristics table) after multiple imputation?
Ex.
Below is Table 1 using multiple imputation in the journal 'Circulation'.
The authors wrote "Missing values are imputed with predictive mean matching."
I wonder...
1) Is it possible to do predictive mean matching for categorical values? (Ethnicity, Smoking variables are certainly categorical.)
2) How can I make "Table 1" after multiple imputation?
In my understanding, there will be long form data after multiple imputation, with multiple set of estimated values for the missing values.
In this table, there are correctly 502,635 ethynicity and smoking categorical values.... How can it be?
2. Can I adjust age after doing 'stptime' command?
What I understand is that doing 'stptime', I can get person-years and standardized mortality rate. (For mortality analysis)
There are many explanation for age adjustment for incidence rate.
However, I only found the age adjusted incidence rate for the fixed period.
Like below:
https://www.health.ny.gov/diseases/chronic/ageadj.htm
Can I apply this method for person-years data with time dependent outcome?
Thank you for reading.
Conducting a prospective cohort analysis, I have two problems.
1. Can I make table 1(baseline characteristics table) after multiple imputation?
Ex.
Below is Table 1 using multiple imputation in the journal 'Circulation'.
The authors wrote "Missing values are imputed with predictive mean matching."
I wonder...
1) Is it possible to do predictive mean matching for categorical values? (Ethnicity, Smoking variables are certainly categorical.)
2) How can I make "Table 1" after multiple imputation?
In my understanding, there will be long form data after multiple imputation, with multiple set of estimated values for the missing values.
In this table, there are correctly 502,635 ethynicity and smoking categorical values.... How can it be?
2. Can I adjust age after doing 'stptime' command?
What I understand is that doing 'stptime', I can get person-years and standardized mortality rate. (For mortality analysis)
There are many explanation for age adjustment for incidence rate.
However, I only found the age adjusted incidence rate for the fixed period.
Like below:
https://www.health.ny.gov/diseases/chronic/ageadj.htm
Can I apply this method for person-years data with time dependent outcome?
Thank you for reading.
