tabulating Summary Table

[zimcom]

Dear Community,

I am quite new to SAS programming and has been learning and using SAS to clean and manipulate data for about 2 months (thanks to the great and generous help I got from this forum) and now I want to challenge myself to learn tabulating summary reports (Example 1, 3 and 4 in the attached paper). Below is a simple example I found on internet:

0verall N=70
	------	------------
System Organ Class		Subject
Preferred Term	Event	n(%)
---------------------------------------------------------
Overall	20	14(20)
Cardiac disorders	6	4 (6)
Atrial flutter	1	1 (1)
Cardiac failure	1	1 (1)
Palpitations	3	2 (3)
Tachycardia	1	1 (1)

This example above was written in SAS Macro, which seems very complicated to me for now and I plan to learn later when I have had more experience. For now, I just want to try to use SAS code to generate the summary above. The Marco codes in the attached abstract is too hard to follow now.

For those not familiar with clinical trial data, I have a data set consisting variables like subject id, aeterm, preferred term and system organ class (SOC) of thousand observation, which I would like to turn the data into the summary table like above.  Below is the part of the data

usubjid	aeterm	preferred term	aebodsys
1002	aspiration pneumonitis	Pneumonia aspiration	Respiratory, thoracic and mediastinal disorders
1002	Epistaxis	Epistaxis	Respiratory, thoracic and mediastinal disorders
1003	ischemic stroke	Ischaemic stroke	Nervous system disorders
1007	Multiple Myeloma	Plasma cell myeloma	Neoplasms benign, malignant and unspecified (incl cysts and polyps)
1007	Dehydration	Dehydration	Metabolism and nutrition disorders
1007	Worsening Pleural Effusions	Pleural effusion	Respiratory, thoracic and mediastinal disorders
1011	Orolingual angioedema	Angioedema	Skin and subcutaneous tissue disorders
1011	Worsening of Congestive heart failure	Cardiac failure congestive	Cardiac disorders
1011	Pneumonia	Pneumonia	Infections and infestations
1012	Recurrent Stroke	Cerebrovascular accident	Nervous system disorders
1013	Right femoral pesudoaneurysm	Vascular pseudoaneurysm	Injury, poisoning and procedural complications
1016	Malignant MCA infarction (Index Stroke)	Ischaemic stroke	Nervous system disorders
1018	Progression of Stroke	Stroke in evolution	Nervous system disorders
1021	Acute Coronary Syndrome	Acute coronary syndrome	Cardiac disorders
1026	Progression of Ischemic Stroke	Stroke in evolution	Nervous system disorders

Any help would be greatly appreciated!

Thanks

[zimcom]

Thank you !!

[Reeza]

---

@zimcom wrote:

Dear Community,

I am quite new to SAS programming and has been learning and using SAS to clean and manipulate data for about 2 months (thanks to the great and generous help I got from this forum) and now I want to challenge myself to learn tabulating summary reports (Example 1, 3 and 4 in the attached paper). Below is a simple example I found on internet:

0verall N=70
	------	------------
System Organ Class		Subject
Preferred Term	Event	n(%)
---------------------------------------------------------
Overall	20	14(20)
Cardiac disorders	6	4 (6)
Atrial flutter	1	1 (1)
Cardiac failure	1	1 (1)
Palpitations	3	2 (3)
Tachycardia	1	1 (1)

This example above was written in SAS Macro, which seems very complicated to me for now and I plan to learn later when I have had more experience. For now, I just want to try to use SAS code to generate the summary above. The Marco codes in the attached abstract is too hard to follow now.

For those not familiar with clinical trial data, I have a data set consisting variables like subject id, aeterm, preferred term and system organ class (SOC) of thousand observation, which I would like to turn the data into the summary table like above.  Below is the part of the data

usubjid	aeterm	preferred term	aebodsys
1002	aspiration pneumonitis	Pneumonia aspiration	Respiratory, thoracic and mediastinal disorders
1002	Epistaxis	Epistaxis	Respiratory, thoracic and mediastinal disorders
1003	ischemic stroke	Ischaemic stroke	Nervous system disorders
1007	Multiple Myeloma	Plasma cell myeloma	Neoplasms benign, malignant and unspecified (incl cysts and polyps)
1007	Dehydration	Dehydration	Metabolism and nutrition disorders
1007	Worsening Pleural Effusions	Pleural effusion	Respiratory, thoracic and mediastinal disorders
1011	Orolingual angioedema	Angioedema	Skin and subcutaneous tissue disorders
1011	Worsening of Congestive heart failure	Cardiac failure congestive	Cardiac disorders
1011	Pneumonia	Pneumonia	Infections and infestations
1012	Recurrent Stroke	Cerebrovascular accident	Nervous system disorders
1013	Right femoral pesudoaneurysm	Vascular pseudoaneurysm	Injury, poisoning and procedural complications
1016	Malignant MCA infarction (Index Stroke)	Ischaemic stroke	Nervous system disorders
1018	Progression of Stroke	Stroke in evolution	Nervous system disorders
1021	Acute Coronary Syndrome	Acute coronary syndrome	Cardiac disorders
1026	Progression of Ischemic Stroke	Stroke in evolution	Nervous system disorders

Any help would be greatly appreciated!

Thanks

---

L
Search for the paper, creating complex reports via Cynthia Zender, @Cynthia_sas

You can also look at PROC REPORT. 

To be honest, you can get pretty close using just PROC FREQ. 

[zimcom]

@Reeza

Totally agree! Individually the summary data can all be calculated using Proc Freq, but the challenge is how to put everything nicely in a user-friendly format

[FreelanceReinh]

Hello @zimcom,

Adverse event (AE) summary tables like this are very common in clinical research. That's why many pharmaceutical companies and CROs use macros to create them. So, in practice you would typically have such a macro available and use it without having to know the underlying macro code.

To verify some (or all) of the frequencies and percentages in the table, you can use PROC FREQ and (especially for the subject counts) PROC SQL. For code samples see, e.g., this PharmaSUG paper (p. 5-6), which covers this and other types of AE tables.

Depending on the desired table layout, the PROC REPORT code can be more or less complicated. For example, you may want to indent the preferred terms ("Atrial flutter" etc.) under their SOC name ("Cardiac disorders"), which can be achieved in different ways as is described in this WUSS paper.

[zimcom · Accepted solution]

@FreelanceReinh

Thanks for your reply.

I am working in a smal biotech company, there are no previous SAS code examples to follow. The clinical trial we are currently running is a blinded study and so I will have to change/update the Marco code in order to fit our situation.

Thank you very much again!