Image  Click on image to enlarge. 
Title  Cumulative Incidence for Alanine Aminotransferase (ALT): Elevations > 3 Times the Upper Limit of Normal 
Graph_Subgroup  Labs Liver 
Labs_Liver_Clinical_Question 
Baseline: 2. Is there a temporal relationship between treatment and lab value? 
Description 
This is a cumulative incidence plot. The cumulative incidence of those subjects with > 3 times the upper limit of normal for the liver function test alanine aminotransferase (ALT) is plotted for treatment versus placebo over time. Standard error bars are added to the graph and are calculated using the following formula: Lower Bound = P – 1.96*SQRT(P(1P)/n) where P is the proportion of subjects with an elevation (i.e. number o f subjects with an elevation > 3xULN divided by the total number at risk) and n is the total number at risk. Upper Bound = P + 1.96*SQRT(P(1P)/n) where P is the proportion of subjects with an elevation (i.e. number o f subjects with an elevation > 3xULN divided by the total number at risk) and n is the total number at risk. Another formula that gives similar results that is used in the literature for the crude standard error of cumulative incidence is as follows: Lower Bound = P – 1.96*P/SQRT(n) where P is the proportion of subjects with an elevation (i.e. number o f subjects with an elevation > 3xULN divided by the total number at risk) and n is the total number at risk. Upper Bound = P + 1.96*P/SQRT(n) where P is the proportion of subjects with an elevation (i.e. number o f subjects with an elevation > 3xULN divided by the total number at risk) and n is the total number at risk. The standard error calculation can be modified as necessary within the code. Upper limit of normal is calculated by dividing the standard result ( SDTM data set variable lbstresn) in numeric units by the standard high range limit (SDTM dataset variable lbstnrhi). Original units can be substituted if needed as the upper limit of normal is normalized (has no units). Below the graphic are both a legend indicating the treatment groups depicted on the graph and a legend indicating the at risk subject population by treatment group at each specific visit. Note that due to the uniform axis separation, and the visit schedule of the study, there may be blanks in the at risk population legend if there is no data for a specific visit. Note that the placement of the treatment group legend can be adjusted within the code based upon end user preferences. It is important to consider that if a color graphic is not presented, the line styles will allow for distinction between treatment groups. In addition, the amount of white space (yaxis range) can be adjusted within the programming. The current graph depicts the yaxis range from 0 to 0.6 and the line separation between treatment groups is adequate so that the viewer can easily discern the two groups. However, if the lines were closer it may be advantageous to lessen the yaxis range and therefore broaden the separation between close lines. There is also the possibility of adding another analysis group (i.e. >5 times the upper limit of normal for example). In this instance, it is possible that there would be two lines for each treatment group and the programmer would need to choose line styles/colors appropriately. *Note: The actual data had very few measurements that were >3 xULN. Therefore, for the graph to offer a beneficial representation, the ULN values were quadrupled (multiplied by 4). This allows for a good separation and representation of the effectiveness of the graph in such situations. This multiplication factor is NOT included in the attached code. It is mentioned here purely in an effort to explain why running the attached code produces almost a flat line graphic with no separation and not the one depicted on this page. 
Contributor/Email  Robert Gordon (rgordon2@its.jnj.com) 
Additional_Contributor_Info 

Background  This graphic was inspired from a GSK cumulative incidence graphic supplied by Susan Duke. 
Keywords  Cumulative, Incidence, Step 
References 

Datasets 

CDISC_Data 

Data_Format  Dataset(s) necessary for the graphic are created within the code. 
CLASSIFICATIONS 

Graph_Type 

Variable_Relationship  Continuous versus Continuous 
Data_Types  Continuous 
Special_Considerations  Depending on the lab parameter and the level of concern, clinically meaningful values (e.g. > 3 times upper limit of normal) and interpretation will differ. It is important to specify these values and properly incorporate them into the code before producing a graphic for interpretation. 
Code_Available  Yes 
Software Program  SAS 
Software  SAS v9.2 
RCode  Attachment 

RCode 

SASCode  Attachment  Gordon  Cumulative Incidence  SAS Doc 
SASCode  v9.2  Attachment 

StataCode  Attachment 

StataCode 

Other Code  Attachment 

Other Code 

Disclaimer  The views expressed within CTSpedia are those of the author and must not be taken to represent policy or guidance on the behalf of any organization or institution with which the author is affiliated. 
Permission 
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF ERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT OLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. 
Click here to add comments 

Reference Image 