Easy Button: A Process for Generating Standardized Safety
Contents
1. statistical analyses by category for variable that can be classified into continuous or category type All the parameters in this macro are keyword parameters The report created by this macro is aligned with the report layout in Biometrics Standard TL Shells V1 0 published Specifically this macro produces the following reports stipulated in Standard TL Shells V1 0 1 Subject Baseline Characteristics 2 Subject Medical History Conditions Incidence 2 A in One or More Treatment Groups 3 Summary of Subject Disposition 4 Study Drug Exposure 5 Summary of Subject Compliance Input Datasets Macro nonsafety_reporting requires that input data sets comply with CDISC Vertex ADaM Standards Here is the must have ADaM variables in input data sets used in macro nonsafety_reporting call These are the example of parallel design clinical trials e ADSL data set USUBJID TRT01A TRTO1AN SEXN AGE RACEN e ADMH data set USUBJID TRTO01A TRTO1AN MHDECOD MHTERM There may be other data sets and variables required based on the specific report requirement 6 SAS Global Forum 2013 Pharma and Health Care Table 1 t_disposition doc Summary of Subject Disposition Initialization include standard sas let tblhame bquote scan amp PROGNAME 1 Get the data proc datasets library work nolist nodetails copy in ad out work select adsl quit run Process data data adsl set adsl label randfl Ran2. t Subject ith Ad E ts by Syst re cl d Pref er an meee jects wi verse Events eee oe aan and Prefe Automatically read from SAP by headfoot macro Report Name Arm A Arm B AmC Arm D Total Report to be generated N XXX N XXX N XXX N XXX N XXX ae_reporting rptname SOCPT tS ae a s J a a Number of subjects with AEs XXX XX X XXX XX X XX x xxx xx x xxx xx x General disorders and administration site conditions xx XxX X xx xx x xx Treatments likeiii af Rk xx xx x xx Treatment code variable name defined P RRRS ARA aa aa pap A xXx in the study initialization and xx XxX X xx XxX x XX Ili trtcol xx xxx kk xxx xx represented calling macros trtco xx xx x xx xx x xx and irtdef Macro Parameter Label defined in the program as a parameter to the ae ee et ae a O macro xx xx x xx xx x xx Any departure from the original ae_reporting toprow Number of i pa zx ie wag econ cainiion eg aiga a 4 Total is derived in the trtmapping subjects with AEs ve tissue disorders xx XX x xx XX x xx mize Myalgia xx XX X xx XX X XX XX X EE 1 AR Ef xx ER X Arthralgia xx xx x xx xx x xx xx x xx xx x xx xx x Back pain xx XX x xx xXxX x xx Xx x xx xx x xx XX X Muscle spasms xx XxX X xx XX x xx Xx x xx xx x xx xx x Program Name VERTEX Server Final tables t aesocpt sas Creation Date and Time 29
3. up with a modular approach of design architecture The main principle of this design architecture was to follow the CDISC ADaM standard datasets and variables so that we can have minimum number of macro parameters to gain the efficiency and the modular approach is for the better management of the macros In this phase we first identified the number of macros will be needed to cover all the standard TFLs stipulated in our Biometrics Standard TFLs specification document and then developed specification for each macro The macro specification document contains references of macro called programs and department level existing macros to be used name of the CDISC ADaM data set to be used a follow chart and an annotation of TFL shell to describe the design architecture Another principle was to maximize the utilization of our existing department level macros to avoid redundancy in terms of developing SAS Global Forum 2013 Pharma and Health Care Module Structure Approach Department Macro Library Template Macro Assembly e Macro_1 e Macro_2 e Macro_3 T e 7 Macro_1 Macro_n code e Examples Select Population at selpop Map Treatment a trtmapping Derive Statistics cntfrqbyclass_sort etc Write once gt Use many times Below is an example of ae_reporting macro specification References Type Element Location Main Program ae_reporting sas Reporting Event macrolib Program s t aept sa
4. 01 doc ID Step User Test Description Expected Results Requirement Reference TST 1 1 N A Capture the current date and time Date captured when execution of the test script Date and time will be referred starts and report it in the Actual to in following steps as the Results section of the Test Script Reference date document TST 2 1 N A Check if required datasets listed in Datasets found and consistent the Input Datasets section of the with document document standard_table reference_docum 12 SAS Global Forum 2013 Pharma and Health Care ents User Guide for ae_reporting doc exists data analysis adsl sas7bdat data analysis adae sas7bdat TST 3 1 USR 1 Generate report t ae summary doc No specific issues detected using the macro from standard_table macrolib Header Titles Footnotes are ae_reporting sas consistent with Shell Use parameter RPTNAME AE in the ae_reporting macro calling to generate t ae summary doc Compare the overall layout against the t_ae_summary doc shell found in standard_table reference_docum ents DRAFT standard TL V9 doc Take an overall visual look at the results layout and values for any issues TST 4 1 USR 1 Generate the report t ae No differences associated to summary_compare doc using the the table results statistics compare macro call qcCompareDocumentBody arg prefix by comparing the report generated w
5. 0CT2012 21 39 Note If subject has multiple events within a system organ class or preferred term the subject is counted once Note Thifs table is sorted in descending frequency ji aa number for each system organ class and preferred term Standard Footnote Report Specific Footnotes Defined in the macro tflsetup Automatically read from SAP by headfoot macro SAS Global Forum 2013 Pharma and Health Care b Development We developed the macros considering the following assumptions which were implemented across all the macros i To identify the population of the TFLs e g Safety Set Full Analysis Set etc the macro selpop will be used ii The column header for treatment column in the report will be generated using trtmapping macro and iii Default Titles and footnotes will come from study specific headfoot macro iv Standard macro parameters name and specification across all the macros Flow of trtmapping macro How trtmapping works Standard Table Reporting Column Statement and Define Statement in PROC REPORT trtdef Global Macro Variables _GRPCDn _GRPLBLn _TRTCDn _TRTLBLn _TRTCNTn fi trtmapping TRTMAPDS Data Set Parameters Specifications The following are the common parameters used across all 5 the macros we developed The specified values of these parameters are not case sensitive POPDS SAS data set Identifies the input population SAS data set The population in this dat
6. 4 1 mean 0 6989700043 1 24 Pop B 24 1 hi 0 6989700043 1 FUW4 Pop B 52 1 lo 0 6989700043 1 FUW4 Pop B 52 1 mean 0 6989700043 1 FUW4 Pop B 52 1hi 0 6989700043 1 For every visit timepoint three records exist the midpoint of the error bar which can either be the mean median etc i e _NAME_ mean one for the lower bar NAME_ lo and the other for the upper bar _NAME_ hi Missing values are allowed Both the numeric as well as character versions of the variable for the time point X axis Also treatment identifier TRTO01A and TRTO1AN are required N variable represents the number of subjects at each visit and is required only if the number of subjects at each visit need to be displayed below each time point The character values of the grouping variable TRT01A and the time point variable AVISIT are used as labels for the legend and X axis respectively Figure 1 f_errorbar1 Error Bar plot for efficacy measure Initialization include standard sas let tblhame bquote scan amp PROGNAME 1 Get the data proc datasets library work nolist nodetails copy in ad out work select adsl adhc quit run Process data data srcdata set adhc if avisitn 950 then avisitn 0 else if avisitn 1001 06 then avisitn 0 1 11 SAS Global Forum 2013 Pharma and Health Care else if avisitn 8888 then avisitn 52 run proc means data srcdata noprint by trt01an trt01a avisitn avisit var ava
7. SAS Global Forum 2013 Pharma and Health Care Paper 176 2013 Easy Button A process for generating standardized safety and non safety related Clinical Trial Reports Xiangchen Bob Cui Vertex Pharmaceuticals Cambridge MA Mominul Islam Vertex Pharmaceuticals Cambridge MA Jiannan Hu Vertex Pharmaceuticals Cambridge MA Yanwei Han Vertex Pharmaceuticals Cambridge MA Sanjiv Ramalingam Vertex Pharmaceuticals Cambridge MA ABSTRACT Developing a process for standardized reports in terms of Tables Figures and Listings TFLs generations for a clinical study report by developing SAS macro based either reporting tool or template SAS programs is an ongoing focus of the health care and life science industry We have developed SAS macros and template SAS programs macro calls based on our Biometrics Standard TFLs shells for safety analysis like Adverse Events AE LABs ECG Vitals and other non safety analysis for example Baseline Characteristics Disposition Concomitant Medication Drug Exposure Compliance and Figures for Kaplan Meier plot Our newly developed process includes development of reporting macros by utilizing the existing department macros to generate these standard reports The macros have been developed assuming the CDISC ADaM analysis data set standards which enable us to minimize the number of macro parameters for efficient use of the macros by the user This process also shortens the development cycle time and facilita
8. a set determines the population in the report As an option for POPDS after a user can specify the sub setting condition for the population to be selected for the analysis For example POPDS str ADSL SAFFL The default option is FASFL It is a required parameter CLASSDS SAS data set Identifies the input analysis SAS data set It is a required parameter SAS Global Forum 2013 Pharma and Health Care USUBJID Identifies the input subject variable It is a required parameter Default amp _usubjid TRT Identifies treatment to be analyzed It is an optional parameter Default TRTO1AN RPTNAME Identifies the type of analysis and report table layout where DISP Defines that the subject disposition will be summarized across treatments It will generate t disposition report as per standard kkkkkkk Other parameter specification OUTDS SAS data set Specifies the name of an output SAS data which contains the same information as in the report if a report is created If a data set is specified the macro produces the data set only If nothing specified the macro creates a report It is an optional parameter Default null DEBUG Specifies a value of 1 or 0 to determine if debug mode is on 1 or off 0 It is an optional parameter Default 0 In the development phase we developed the following 5 macros to cover our standard safety TFLs Ynonsafety_reporting Macro nonsafety_reporting performs
9. ables needed for summarybyvisit_reporting macro call These are the example of parallel design clinical trials e ADSL data set USUBJID TRT01A TRTO1AN e ADLB data set USUBJID PARAMCD PARAMN AVAL AVISIT AVISITN SAS Global Forum 2013 Pharma and Health Care Table 3 t_eg_summary doc Summary Statistics for Standard Digital ECG by Treatment Group and Visit Initialization include standard sas let tolhame bquote scan amp PROGNAME 1 Get the data proc datasets library work nolist nodetails copy in ad out work select adsl adeg quit run Process data data adeg set adeg where upcase anl02fl Y and aval ne and avisitn ne run Call macro to generate report summarybyvisit_reporting popds sir adsl SAFFL classds adeg rptname eg analvar aval chg shift_reporting Macro shift_reporting performs statistical analyses for Laboratory and ECG baseline vs post baseline variables The report created by this macro is aligned with the report layout in Biometrics Standard TL V1 0 Specifically this macro produces the following reports in Standard TL V1 0 ECG Shift from Baseline including and not including column for Missing Chemistry Shift from Baseline including and not including column for Missing Hematology Shift from Baseline including and not including column for Missing Number and Percentage of Subjects with Hematology Toxicity Grade Shifts N
10. ata set USUBJID TRT01A TRTO1AN e ADAE data set USUBJID TRT01A TRTO1AN AEDECOD AETERM AEBODSYS AREL ARELN ATOXGR ATOXGRN Table 2 t_ae_soc_pt doc Summary of Adverse Events by System Organ Class and Preferred Term Initialization include standard sas let tblhame bquote scan amp PROGNAME 1 Get the data proc datasets library work nolist nodetails copy in ad out work select adsl adae quit run Process data proc sort data adae by amp _usubjid where aedecod ne and upcase aetvpbfl Y run ae_reporting popds adsl classds adae rptname SOCPT missing Yes sortcols 10 summarybyvisit_reporting Macro summarybyvisit_reporting performs statistical analyses by visit or by category for variable that can be classified into continuous or category type All the parameters in this macro are keyword parameters The report created by this macro is aligned with the report layout in Biometrics Standard TL Shells V1 0 published Specifically this macro produces the following reports stipulated in Standard TL Shells V1 0 Summary of Laboratory Results and Change from Baseline Summary Statistics for Standard Digital ECG by Treatment Group and Visit Summary of Vital Signs Results and Change from Baseline Maximum On Treatment QT QTc Interval Results Maximum On Treatment Increase from Baseline in QT QTc Interval NEON Input Datasets Here is the must have ADaM vari
11. ater details of the 13 SAS Global Forum 2013 Pharma and Health Care macros Due to the modular approach of the macros the managing of the macros become more robust as for any updates we need to make we can update the associated module instead of updating the whole reporting macro These macros can be used for any number of treatment columns and the width of treatment column can be handled using macro parameter For any updates of a table shell the module macro can be re run and automatically update headfoot sas which contains title and footnote information Once the table shell document is setup without the presence of special characters and formats this process requires no human intervention hence facilitates a seamless approach The use of these macro result in significant reduction of programming work load and error prone manual processing to define the treatment columns header and to obtain the titles and footnotes especially following the modifications to table shells It also helps to limit the need to manually enter information for project management activities Another notable benefit comes from its built in capability of producing output reporting data set to help the QC programmer to QC the report in the event of large reporting data set like coming out of LAB or Vital Signs related ADaM data set therefore ensure complete implementation of the table shell requirements The macros are easy to use and improves both work efficiency and qua
12. domized label nomedfl No Study Medication label fasfl Included in Full Analysis Set run proc format value streasn 1 Adverse Event 2 Death 6 Other run Call macro to generate report ERER E classds adsl rptname DISP binaryvar randfl nomeadfl fasfl varord randfl nomedfl fasfl format streasn streasn ae_reporting Macro ae_reporting performs statistical analyses for variable that can be classified into several levels such as PTERM SOCTERM etc All the parameters in this macro are keyword parameters The report created by this macro is aligned with the report layout in Biometrics Standard TL Shells V1 0 published Specifically this macro produces the following reports stipulated in Standard TL Shells V1 0 1 Summary of Adverse Events 2 Summary of Adverse Events by System Organ Class and Preferred Term PT Incidence A in One or More Treatment Groups 3 Summary of Subjects With Adverse Events by Relationship to Study Drug System Organ Class and Preferred Term 4 Summary of Subjects With Adverse Events by Relationship to Study Drug System Organ Class and Preferred Term Relationship presented in a column 5 Summary of Adverse Events by Preferred Term PT Incidence 2 A in One or More Treatment Groups SAS Global Forum 2013 Pharma and Health Care Input Datasets Here is the must have ADaM variables in input data sets used in macro ae_reporting call e ADSL d
13. ithout the ae_reporting macro call with the one generated with ae_reporting macro call Review t ae summary_compare doc for any issues related to the table core results statistics e Release of beta version to be used across the therapeutics After completing the UAT of each of the macro working on projects from different therapeutics we developed the plan for roll of these macros within Biometrics Department The plan consists of following i Developed User Guide for each of the macro containing example of macro calls ii Developed template program for macro calls for each of the standard TFL iii Developed Standard TFL User Manual stipulating functionality of all the macros along with example macro calls in different TFLs iv Developed training slides and provide live demonstration of macro calls and functionality within Biometrics Department in a macro training session v Identified resources to address questions or concerns if any brought by the users of the macros CONCLUSION In summary the SAS macros and template programs we developed can be used to automate the standard TFLs generation aligned with standard TFL shells document The macros have been developed considering ADaM data sets which are based on CDISC and Vertex standards and conventions which helped us to restrict the number of macro parameters at a minimum to enable the user to use the macro efficiently without know the gre
14. l output out wgt4 drop _type__freq_ n n mean mean stderr stderr Iclm lo uclm hi run proc transpose data srcdata out final prefix bar by trt01an trt01a avisitn avisit mean stderr n var lo mean hi run merrorbar inds final err_mean mean yvalue bar1 xlabel onrstr Study Week ylabel str Mean and 95 Cl of Log HCV RNA leg_pos inside left top ticks C countvar y Listings For listings we have developed template programs instead of macros Most of the listings were done using CDISC SDTM data sets The following two module macros used in the listing programs checkvar sas automatically detect if information for optional column exist if yes the column will be created Y varlen sas set variable length in report c User Acceptance Testing UAT We first developed Test Specification and Test Script for each of the macro The UAT of the macros were done while working on the real life project We developed QC programs for standard TFLs without using any macros and then compare the production reports with that of QC of reports With this principle we have been able to test the functionality of the macros as well as have been able to deliver the project deliverable with quality soeed and values Example of Test Specification for ae_reporting macro Test Specifications Implementation of the tests described below is done using the document standard_table reference_documents qc ae_reporting TestScript0
15. lity ACKNOWLEDGEMENTS Appreciation goes to Qunming Dong PhD and Leif Bengtsson for their review and comments SAS and all other SAS Institute Inc product or service names are registered trademarks or trademarks of SAS Institute Inc in the USA and other countries indicates USA registration Other brand and product names are registered trademarks or trademarks of their respective companies CONTACT INFORMATION Your comments and questions are valued and welcome Contact the authors at Xinagchen Bob Cui Ph D Vertex Pharmaceuticals Incorporated 130 Waverly Street Cambridge MA 02139 4942 617 341 6069 Mominul Islam Vertex Pharmaceuticals Incorporated 130 Waverly Street Cambridge MA 02139 4942 617 961 0913 Jiannan Hu Vertex Pharmaceuticals Incorporated 130 Waverly Street Cambridge MA 02139 4942 617 961 7524 Sanjiv Ramalingam Vertex Pharmaceuticals Incorporated 130 Waverly Street Cambridge MA 02139 4942 617 961 0831 Yanwei Han Vertex Pharmaceuticals Incorporated 130 Waverly Street Cambridge MA 02139 4942 617 341 6736 14
16. s Reporting Event tables Called t aesocpt sas t aesummary sas Macro s selpop trtmapping entfrqbyvar Reporting Event macrolib Data Sources Data Set Description Required Variables Required Formats Adsl Population amp _usubjid amp _trtcd amp _trtfmt amp _trtcd Adae Adverse Events amp _usubjid aedecod aebodsys Note s 1 Variables usubjid and trt01an are the default variable names associated to the macro variables amp _usubjid and amp _trtcd The macro variables are created and assigned in the study initialization file Reporting Event standard sas 2 Eventdata sdtm ADaM datasets located in directory Reporting Eventdata analysis and SDTM datasets located in directory Reporting SAS Global Forum 2013 Pharma and Health Care Process Flow Chart Initialization Select Population selpop Map Treatment strtmapping Derive Statistics ssummaryStatsInRows ecntfrqbyclass scntfrqbyclass_sort sae scentfrqbyvar Generate Report smacro dorpt tflsetup Get Report sheadfoot stflpre strtcol strtdef stflpost smend dorpt Annotation on Standard Output Standard Output Standard Headers Vertex Pharmaceuticals Incorporated BLINDED RESULTS Labels defined in the study initialization file standard sas gt Page 1 of 19 Protocol VXXX XXX XXX A Phase II Study of VX XXX Table 14 3 1 2a Titles Numb d P
17. t medications by Treatment Group ATC1 ATC2 PT Input Datasets Here is the list of must have ADaM variables needed for cm_reporting macro call These are the example of parallel design clinical trials e ADSL data set USUBJID TRT01A TRTO1AN e ADCM data set USUBJID ATC1TERM ATC2TERM CMDECOD CMPRIOR Table 5 t_conmed doc Number and percentage of Subjects with concomitant medications by Treatment Group ATC1 ATC2 PT Initialization include standard sas let tolhame bquote scan amp PROGNAME 1 Get the data proc datasets library work nolist nodetails copy in ad out work select adsl adlb quit run Call macro to generate report cm_reporting trivar trt01an popfl saffl phase cmtvpbil 10 SAS Global Forum 2013 Pharma and Health Care sortcols 9 10 options atcpt Figures Merrorbar Error Bar Plot Macro Merrorbar produces a Kaplan Meier plot for the input analysis dataset Input Dataset Macro Merrorbar requires that the input dataset be of the structure as below Po siT ROMA SCS asn N NAME YVALUE TRTOIAN BL Pop B 0 14 lo 6 3375857982 1 BL Pop B 0 14 mean 6 6833152447 1 BL Pop B 0 14 hi 7 0290446911 1 4 Pop B 4 14 lo 0 8175023054 1 4 Pop B 4 14 mean 1 1017145048 1 4 Pop B 4 14 hi 1 3859267042 1 12 Pop B 12 9 lo 0 5053253246 1 12 Pop B 12 9 mean 0 8472426434 1 12 Pop B 12 9 hi 1 1891599621 1 24 Pop B 24 1 lo 0 6989700043 1 24 Pop B 2
18. tes the adoption from SAS programmers to clinical reporting During the development of the reporting macros the macros have been tested using the ongoing Clinical trials of different phases We also developed User Manual and standard Template Programs for easy use for our programmers The process stipulated in this paper will reduce the report generation time significantly and achieve the quality by design principle Keywords CDISC TFLs ADaM AE LAB ECG VITALS INTRODUCTION This paper introduces a process which includes several SAS macros and template SAS programs to automate the safety related TFLs generation on the basis of our Biometrics Standard TFLs which was developed by a working group consists of Biostatisticians Statistical Programming Medical Monitor and Medical Writing who worked together to develop this document After approving and releasing this Biometrics standard TFLs within Biometrics the Statistical Programming group initiated a project to develop macros to generate those TFLs stipulated in the Biometrics Standard TFLs document Under this project a process has been developed which consists of the following 4 steps Planning Development User Acceptance Testing UAT Release of beta version to be used across the therapeutics aoop a Planning In the planning phase we developed a working group to work on developing the macros aligned with Biometrics Standard TFLs specification The working group came
19. umber and Percentage of Subjects with Chemistry Toxicity Grade Shifts IEUS Input Datasets Here is the list of ADaM variables needed for shift_reporting macro call These are the example of parallel design clinical trials e ADSL data set USUBJID TRT01A TRTO1AN ADLB data set USUBJID TRT01A TRTO1AN AVISITN PARAM LBRPTLBL ANRIND BTOXGRN BTOXGRHN Table 4 t_lb_shift_grade doc Chemistry Toxicity Grade Shift from Baseline Initialization include standard sas let tblhame bquote scan amp PROGNAME 1 SAS Global Forum 2013 Pharma and Health Care Get the data proc datasets library work nolist nodetails copy in ad out work select adsl adlb quit run Process data proc sort data adlb out adlb nodupkey where abifl Y and parcat1 CHEMISTRY and TABLESFL Y by usubjid Ibrptib BTOXGRN MXGR_T run Call macro to generate report shift_reporting popds adsl classds adlb rptname CHEMT OX base BTOXGRN shift MXGR_T thr hcv debug 1 cm_reporting Macro cm_reporting performs statistical analyses for variable that can be classified into several levels such as ATC1TERM ATC2TERM CMDECOD All the parameters in this macro are keyword parameters The report created by this macro is aligned with the report layout in Biometrics Standard TL V1 0 Specifically this macro produces the following reports in Standard TL V1 0 1 Number and percentage of Subjects with concomitan
Download Pdf Manuals
Related Search