topic Re: multiple imputation code in Statistical Procedures

multiple imputation code

Gopi2 — Fri, 09 Apr 2021 03:55:58 GMT

Hi SAS experts,

I am a beginner, planning to do multiple imputations for a data set.

The data has the following variables ( both continuous and categorical); age, gender, education level, blood pressure, body weight index, cholesterol., mean daily steps, genotype, and outcome of dementia.

All the variables including the outcome, have got missing values.

Please do share what type of multiple imputations I have to go with along with the codes

( the primary aim is to validate a prediction model, i.e. to find how well the variables predict the outcome of dementia, done by finding the ROC curve!)

I have attached the missing pattern table.

Thanks in advance

Re: multiple imputation code

gcjfernandez — Fri, 09 Apr 2021 05:44:20 GMT

Please review the attached PDF file on Missing value imputation.

Re: multiple imputation code

SAS_Rob — Fri, 09 Apr 2021 21:17:55 GMT

The imputation method will depend on the missing data pattern and the variable type. This table from the documentation will be helpful.

https://go.documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=9.4_3.4&docsetId=statug&docsetTarget=statug_mi_details05.htm&locale=en

There is not a way to get a combined ROC curve directly using multiple imputation.

You can however get a combined estimate of the C-statistic as the area under the ROC curve. Below is an example that shows how you might code this.

data roc;
input alb tp totscore popind @@;
totscore = 10 - totscore;
datalines;
3.0 5.8 10 0 3.2 6.3 5 1 3.9 6.8 3 1 2.8 4.8 6 0
3.2 5.8 3 . 0.9 4.0 5 0 2.5 5.7 8 0 1.6 5.6 5 1
3.8 5.7 5 1 3.7 6.7 6 1 3.2 5.4 4 1 3.8 6.6 6 1
4.1 6.6 5 1 3.6 5.7 5 1 4.3 7.0 4 1 3.6 6.7 4 .
2.3 4.4 6 1 4.2 7.6 4 0 4.0 6.6 6 0 3.5 5.8 6 1
3.8 6.8 7 1 3.0 4.7 8 0 4.5 7.4 5 1 3.7 7.4 5 1
3.1 6.6 6 1 4.1 8.2 6 1 4.3 7.0 5 1 4.3 6.5 4 1
3.2 5.1 5 1 2.6 4.7 6 1 3.3 6.8 6 0 1.7 4.0 7 0
3.7 6.1 5 1 3.3 6.3 7 1 4.2 7.7 6 1 3.5 6.2 5 1
2.9 5.7 9 0 2.1 4.8 7 1 2.8 6.2 8 0 4.0 7.0 7 1
3.3 5.7 6 1 3.7 6.9 5 . 3.6 6.6 5 1
;
proc mi data=roc out=mi_roc;
class popind;
var alb tp totscore popind;
monotone logistic;
run;
proc logistic data=mi_roc;
by _imputation_;
ods output ROCAssociation=roc_ds(where=(ROCMODEL='Model'));
model popind(event='0') = alb tp totscore;
roc;
run;
proc mianalyze data=roc_ds;
modeleffects area;
stderr stderr;
run;

Re: multiple imputation code

Gopi2 — Tue, 13 Apr 2021 01:06:23 GMT

thanks will try to do this

Re: multiple imputation code

Gopi2 — Tue, 13 Apr 2021 01:07:03 GMT

thanks will try to do this and reply back if it worked

Re: multiple imputation code

Gopi2 — Tue, 08 Jun 2021 05:16:08 GMT

finally, the code below gave me the required imputation. thanks all!

I used multiple imputation with Fully conditional specification (FCS)

proc mi data= data minimum = 0 0 0 0 MINMAXITER=100 nimpute=10 out= imuputed_data ;
var (variables);
fcs plots=trace(mean std);

fcs discrim( categorical variables) /classeffects=include) nbiter =100 ;
fcs reg (continuous variables) nbiter =100 ;
run;

mini function was used in some to prevent getting some negative minimum values. minmaxter is to re-imupute if the system gets a negative ( or any avoided number in minimum)