Clinical Standards Toolkit 1.5 ± How Do I Know My...   PhUSE 2012

PhUSE 2012
Paper TS01
Clinical Standards Toolkit 1.5 ± How Do I Know My Metadata is Right?
Gene Lightfoot, SAS Inc., Cary, NC, USA
ABSTRACT
®
One of the key features in SAS Clinical Standards Toolkit 1.5 (Toolkit) is the ability to validate the metadata used by the
Toolkit for each supported standard. Currently when installing Toolkit, an installation qualification/output qualification (IQ/OQ)
document is provided to ensure the user the product was installed properly. But as users modify files associated with each
standard, add references to study-level files, and define and add their own custom standards, how can they be confident that
all the component metadata is in sync for optimal Toolkit performance? This paper focuses on the two primary types of Toolkit
metadata, Global Library metadata defining each standard and study-level metadata representing each study of interest, while
also covering changes to the Toolkit in the form of revised metadata, new metadata, the addition of internal validation checks,
and the process of running these checks.
INTRODUCTION
Prior to the release of 1.5, Toolkit relied on many assumptions that the users have expected metadata in a certain form and
that metadata is consistent across multiple metadata files. However, with few exceptions, there was no systematic
assessment of the validity of Toolkit metadata. The previous assessment was generally limited to:
-­‐
running the cstutil_checkds framework macro to assess the structure and content of the SASReferences data set
(performed as a setup step running Toolkit sample drivers).
-­‐
requiring the specification (and existence) of various metadata files as parameters when calling selected Toolkit
macros (e.g. cst_registerStandard requires the specification of valid standards and standardsasreferences data sets).
To further enhance Toolkit it needed to provide validation of the Toolkit metadata as part of the definition, registration, and
maintenance of each standard and optionally as a setup step for each submitted Toolkit process. In addition, this internal
validation uses the existing Toolkit validation framework used for specific standards such as CDISC-SDTM.
TOOLKIT INTERNAL VALIDATION OBJECTIVES AND USAGE SCENARIOS
The two primary objectives of the Toolkit Internal Validation functionality are:
-­‐
-­‐
To make sure SAS-supplied Toolkit metadata files are consistent and correct
To make the functionality available to users to help them validate custom standards
In developing Toolkit 1.5, the following usage scenarios were identified to describe the range of functionality desired by Toolkit
Internal Validation:
-­‐
-­‐
-­‐
-­‐
-­‐
-­‐
-­‐
-­‐
-­‐
-­‐
Support IQOQ assessment and reporting
Support registration of a new standard and/or updates to an existing standard
Assess metadata consistency across files
Assess structural validity of a metadata file
Assess content of a metadata file
Validate a SASReferences data set
Evaluate validation check metadata
Perform a process pre-check
Perform run-time process validation
Evaluate code module header compliance
The expectation is that a mature Toolkit Internal Validation suite should be able to support each of the usage scenarios
described above, unless information is inserted rejecting the scenario.
SUPPORT IQOQ ASSESSMENT AND REPORTING
Through the 1.4 release, the Toolkit IQOQ runs a representative set of standard-specific driver modules to illustrate that
Toolkit is installed and configured satisfactorily. Toolkit Internal Validation is designed to make sure proper metadata is where
PhUSE 2012
it belongs and contains what it is supposed to contain. Toolkit uses a subset of the Internal Validation tools and processes as
the first step of the product IQOQ. This involves calling a series of Internal Validation tools to validate specific standards or
files, and/or running a validation process against a subset of Internal Validation checks (e.g. validation_IQOQ.sas7bdat).
SUPPORT REGISTRATION OF A NEW STANDARD AND/OR UPDATES TO AN EXISTING STANDARD
When a new standard is registered, offer the option to perform Internal Validation on the metadata for that standard. Likewise,
when the metadata for a registered standard is updated, optionally perform Internal Validation on the metadata for that
standard. Both of these scenarios require that tools and checks be callable at the standard level.
ASSESS STRUCTURAL VALIDITY OF A METADATA FILE
Verify the structural validity of any given metadata file against a fixed reference, such as a template.
ASSESS METADATA CONSISTENCY ACROSS FILES
The interrelationship among Toolkit metadata assumes that information in certain files is consistent with information in other
files. Three examples illustrate this:
-­‐
It is expected that the table.column pairs specified in the reference_columns data set have a matching record for
each table in the reference_tables data set
-­‐
The metadata content in source_tables and source_columns is expected to be consistent with (though not
necessarily match) the data sets found in the source data library. Toolkit would not expect, for example, a column
documented in the source_columns data set that does not exist in the corresponding data set in the source library.
-­‐
Each validation check is expected to have a corresponding description in the messages data set.
Toolkit Internal Validation supports comparison of metadata between two files. A simple example is comparing
dataset1.column1 against dataset2.column2. Toolkit Internal Validation also includes support of a full file comparison, such
that one file may be deemed to be structurally equivalent with (or different from) another. There is no expectation that the
content of one file be deemed to be equivalent to (i.e. a copy of) the content of another file.
ASSESS CONTENT OF A METADATA FILE
Toolkit checks that the content of any given metadata file meets expectations in terms of having the correct number of
columns, expected number of rows, and the expected column values. This runs and reports on all metadata files, or by
standard, or for any specified file.
VALIDATE A SASREFERENCES DATA SET
Toolkit has enabled this both as a runtime option (current CST 1.4 behavior using cstutil_checkds) and as a standalone option.
In developing this process the following aspects were taken into consideration:
-­‐
-­‐
-­‐
-­‐
-­‐
-­‐
-­‐
Is the data set structurally correct? It may have optional, user-extended columns.
Can the referenced input and output files and folders be reached?
Does any required look-through to Global Library defaults work?
Are all global macro variable¶s references valid and are values expected based on the assessment context?
Is any derived format search path valid?
Is any derived autocall path valid?
Are all called macros reachable?
EVALUATE VALIDATION CHECK METADATA
The process of defining the metadata to support the running of a specific validation check in Toolkit is essential to proper
functionality. The following Toolkit validation metadata issues must be correct for the check to run successfully:
-­‐
-­‐
-­‐
-­‐
-­‐
-­‐
-­‐
-­‐
Is the checkid unique?
Is the standard and standardversion information consistent with available, registered standards?
Does the combination of fields that identify this as a unique check actually uniquely identify the check?
Is the codesource (check macro) available?
Are all utility macros and required macro variables found?
If lookup information is provided, can the referenced data set or format catalog be found?
Does the check status suggest the check is ready for use?
Is any specified or referenced codelogic consistent with the requirements of the referenced check macro in
codesource?
PhUSE 2012
It is not expected that this assessment should be made at run time. However, it may be necessary that the calling validation
SURFHVV HVWDEOLVKHV WKH H[SHFWHG VHVVLRQ ³HQYLURQPHQW´ OLEUDULHV DOORFDWHG IRUPDW VHDUFK SDWh enabled, etc.) and that this
evaluation occur for the purposes of validating each standard-specific validation_master data set before that master set of
checks is used to validate study data.
PERFORM A PROCESS PRE-CHECK
This DGGUHVVHV WKH TXHVWLRQ ³$UH WKHUHDQ\ SUREOHPV ZLWKP\ Toolkit SURFHVV DVFXUUHQWO\ GHILQHG"´ 7KDW LV ZLOOP\GULYHU
module run successfully? Theoretically, this can be part of an IQOQ pre-step before submission of any driver modules used
as part of the IQOQ. This utility is most useful as an aid to customers building their own CST processes.
PERFORM RUN-TIME PROCESS VALIDATION
1RPDWWHUKRZPXFK³SUH´YDOLGDWLRQKDVEHHQ performed, submission of any Toolkit process may include run-time changes in
the FRGHPDFURFDOOVWKH³JOREDOPDFURHQYLURQPHQW´FRQWHQWRIWKHSURFHVV6$65HIHUences data set, etc. Through Toolkit
WKLV LV WKH RQO\ ³LQWHUQDO YDOLGDWLRQ´ SHUIRUPHG ± the specific call to cstutil_checkds which occurs as part of the call to
cstutil_processsetup(). Run-time process validation will typically occur prior to and/or as a part of the call to the primary task
macro (e.g. sdtm_validate). If problems are detected, default behavior is to abort the process, reporting the problems. Runtime validation can optionally be performed on any metadata files expected/required by that process.
EVALUATE CODE MODULE HEADER COMPLIANCE
Each SAS-supplied macro, tool, and driver module now conforms to an expected naming convention and commenting style in
the header that conforms to a SASDOC standard so that the code interface documentation can be automated for Toolkit user
documentation.
ADDITIONAL METADATA FOR TOOLKIT 1.5
To support some of the assessments listed above, several modifications were needed to existing Toolkit metadata to allow for
internal validation. These modifications consist of additional required metadata. Including, but not limited to, additional
information about a standard, whether a data set was considered input or output, location of key files, and Boolean style
information (Y or N). The following modifications were incorporated into Toolkit to perform internal validation. Examples shown
here access the default installation file paths, customer Toolkit installation paths may be different.
C:\cstGlobalLibrary\standards\cst-framework-1.5\control
STANDARDLOOKUP.SAS7BDAT
± Addition of 2 new columns template and templatetype. Template is used to identify an entity that can be used to
populate the metadata for a file or data set. For example, this would name a blank data set (e.g. sasreferences)
provided by Toolkit that would be used to compare the metadata against the populated Toolkit version of sasreferences
to verify the metadata is correct.
Templatetype defines what kind of template this is. Current valid values are DATASET, FOLDER, CATALOG, and
FILE.
STANDARDS.SAS7BDAT
± Addition of 3 new columns studylibraryrootpath, controlsubfolder, and templatesubfolder. Studylibraryrootpath
identifies the root path for the sample study. Controlsubfolder identifies the folder relative to the standard rootpath that
contains Toolkit control files (ie standardlookup, standards, etc.). Templatesubfolder identifies the template folder
relative to the standard rootpath. These metadata are used to verify directories and files exist.
STANDARDMACROVARIABLES.SAS7BDAT
± New data set containing all of the global macro variables accessed by Toolkit and used to verify the existence of
Toolkit global macro variables.
STANDARDMACROVARIABLESDETAILS.SAS7BDAT
PhUSE 2012
± New data set containing all of the valid values for the global macro variables accessed by Toolkit and used to verify
Toolkit global macro values are correctly initiated.
STANDARDSASREFERENCES.SAS7BDAT
± Addition of 3 new columns IOtype, filetype, and allowoverwrite. IOtype defines whether or not the entity is input,
output, or both. Filetype describes what the entity is and can be folder, dataset, catalog, or file. Allowoverwite informs
Toolkit if the entity can be written to (Y) or not (N).
C:\cstGlobalLibrary\standards\cst-framework-1.5\messages
MESSAGES.SAS7BDAT
± Addition of new messages for handling Toolkit internal validation. These messages are prefaced with CSTV (eg
cstv001).
C:\cstGlobalLibrary\standards\cst-framework-1.5\programs
VALIDATION.PROPERTIES
± New file containing global macro variable settings for internal validation.
C:\cstGlobalLibrary\standards\cst-framework-1.5\templates
MESSAGES.SAS7BDAT
± New metadata template data set for messages data set. Contains structural content of the messages data set such
as columns, label, lengths, etc., but contains no data for the messages data set (0 obs). Is used to verify structure.
PhUSE 2012
STANDARDMACROVARIABLEDETAILS.SAS7BDAT
± New metadata template data set for global macro details data set. Contains structural content of the
standardardmacrovariabledetails data set such as columns, label, lengths, etc., but contains no data (0 obs). Is used to
verify structure.
STANDARDMACROVARIABLES.SAS7BDAT
± New metadata template data set for global macro variables data set. Contains structural content of the
standardardmacrovariables data set such as columns, label, lengths, etc., but contains no data (0 obs). Is used to verify
structure.
VALIDATION_STDREF.SAS7BDAT
± New metadata template data set for validation standard references data set. Contains structural content of the
validation_stdref data set such as columns, label, lengths, etc., but contains no data (0 obs). Is used to verify structure.
C:\cstGlobalLibrary\standards\cst-framework-1.5\validation\control
New folder containing Toolkit internal validation validation_master data set.
VALIDATION_MASTER.SAS7BDAT
± New data set containing all of the internal validation checks for Toolkit.
ADDITIONAL TOOLKIT TOOLS
In addition to new metadata, new tools were also identified and developed to help with Toolkit internal validation. These tools
are all written as macros and therefore, in keeping with the core requirement of Toolkit, are open source to all customers. The
following table itemizes a sampling of the new tools that generally can be run stand-alone with a single macro invocation.
These tools can be called by each other or as a part of a Toolkit process that performs some larger function (e.g. validation,
IQOQ). This paper was authored prior to final approval and development of the tools listed below, as such, some tools may
not appear in the final product or new unlisted tools may be developed in the interim.
Tool / Description
Structural file comparison
Find a referenced file
Macro confirms presence and validity of referenced external file
(supported file types: FOLDER, DATASET, CATALOG, FILE)
Can a file be created
Validate SASReferences
This is a replacement for current cstutil_checkds macro.
Code**
cstutilcomparestructure(
_cstReturn=_cst_rc,
_cstReturnMsg=_cst_rcmsg,
_cstBaseDSName=,
_cstCompDSName=,
_cstResultsDS=work._cstCompareStructure)
cstutilfindvalidfile(
_cstfiletype=,
_cstfilepath=,
_cstfileref=)
cstutilcheckwriteaccess(
_cstfiletype=,
_cstfilepath=,
_cstfileref=)
cstutilvalidatesasreferences (
_cstDSName=&_cstSASRefs,
_cstStandard=CST-FRAMEWORK,
_cstStandardversion=1.2,
_cstSASRefsGoldStd=,
_cstallowoverride=,
_cstResultsType=LOG,
_cstPreAllocated=N,
_cstVerbose=N)
PhUSE 2012
Tool / Description
Evaluate global macro variables
Code**
cstutilcheckglobalmvars(
filepath=,
fileref=,
macrovar=)
Assess metadata file content
cstutilevaluatefilecontent(
Metadata file has the correct number of columns, expected number of
filepath=,
rows, and the expected column values.
fileref=,
targetds=,
referenceds=)
Evaluate check metadata
cstutilevaluatecheckmetadata(
Evaluate a single check metadata data set (e.g. validation_master)
_cststd=CDISC-SDTM,
record
_cststdver=3.1.2,
_cstcheckds=,
_cstchkid=)
Is xml file valid
cstutilcheckxmlfile(
Input of this macro would be an XML file and a schema that the XML
xmlfilepath=,
file should validate against.
xmlfileref=,
schemapath=,
schemaref=)
Parse SAS Log for problems
cstutilparselog(
logtype=Current|File,
filepath=,
fileref=)
Summarize a CST results data set
cstutilsummarizeresults(
No errors or warnings noted in the results data set.
dstype=Current|File,
resultsds=)
Process pre-check
cstutilchecksubmittedprocess(
³$UH WKHUH DQ\ SUREOHPV ZLWK P\ &67 SURFHVV DV FXUUHQWO\ GHILQHG"´
driverpath=,
That is, will my driver module run successfully?
driveref=,
checktype=Log|Results|Both)
Report installation status
cstutilreportinstallationstatus(
Generate a report that itemizes installation status, by standard, and
_cststd=CDISC-SDTM,
provides an overall product status
_cststdver=3.1.2)
Evaluate code module compliance
cstutilparsecodeheader(
filepath=,
fileref=)
This macro is called in the internal validation driver programs and builds cstutilbuildstdvalidationcode(
a job stream for all registered standards passed to
_cstStdDS=work._cstStandardsforIV,
cstutilbuildstdvalidationcode() in the _cstStdDS parameter data set.
_cstSampleRootPath=_DEFAULT_,
_cstSampleSASRefDSPath=_DEFAULT_,
_cstSampleSASRefDSName=_DEFAULT_)
Builds the framework reference_tables and reference_columns data
cstutilbuildmetadatafromsasrefs(
sets from a SASReferences data set .
cstSRefsDS=work.stdvalidation_sasrefs,
cstSrcTabDS=work.source_tables,
cstSrcColDS=work.source_columns)
**macro names and signatures may change
THE TOOLKIT ³,NTERNAL VALIDATION´6TANDARD
The Toolkit framework for 1.5 is handled as a standard that supports validation. It is registered and contains all of the
metadata one would associate with any Toolkit standard undergoing validation and includes a sample library providing users
with examples of data and sample programs. Notably absent In the section above describing additions and modifications to
the Toolkit global standard library, is the metadata folder that stores the reference_tables and reference_columns data sets.
Since internal validation is dynamic in the sense it can be run or modified at any time, and as new standards, studies, etc. are
added by the customer, most of the metadata are generated into temporary work files when needed. The generated
temporary tables include reference_tables, source_tables, reference_columns, and source_columns. Below is an example of a
generated reference_tables data set.
PhUSE 2012
In addition to being registered as a Toolkit standard, internal validation also has its own Toolkit Sample library. In these
folders, users will find the validation_control data sets. Multiple validation control data sets have been created to anticipate
different types of internal validations ± Framework IQOQ, Standard IQOQ, and Standard metadata. Because different
metadata exists in different areas within the Toolkit, internal validation specific sasreferences data sets are also included. New
to Toolkit is the introduction of SAS views accessing the validation_control data set to subset out the different internal
validation checks based on type.
TOOLKIT INTERNAL VALIDATION VALIDATION
During development, internal validation of Toolkit was broken into 3 areas:
x
Global Library and Sample Library
x
Standard reference metadata
x
Standard source metadata
Global Library checks attempt to assure that the definition of standards within the CST Global Library is consistent. Two
points of emphasis:
1)
2)
That the structure and content of the Global Library metadata folder standards, standardsasreferences, and
standardlookup data sets is correct
That the metadata for each standard within the Global Library is complete and consistent with expectations.
Three values were added to the checktype variable in the validation_master data set. These allow the user to subset the
validation_master based on the type of internal validation required. The three values are:
PhUSE 2012
GLMETA ± pertains to validation checks that are specific in nature to the Toolkit global library. Considered an IQOQ check.
STDIQOQ ± Standard level IQOQ checks
STD ± Standard level checks that are not considered part of an IQOQ.
Internal validation runs like any other Toolkit validation process. It uses the driver programs supplied with the product, but it
also demonstrates the flexibility of Toolkit. This flexibility is seen in the creation of multiple driver programs to handle different
types of validation (ie Global Library versus a Standard), multiple validation control views and/or data sets representing various
subsets of the validation_master data set. Several internal validation driver programs are listed below.
VALIDATION DRIVER MODULES
In Toolkit, all driver modules are example SAS programs provided to illustrate Toolkit functionality. For Internal Validation,
these driver modules are located in the <cstSampleLibrary>/cst-framework-1.5/programs folder. The following examples show
actual code snippets and explain the workflow that is used by each driver.
VALIDATE_STANDARD.SAS
The objective of this driver module is to perform all standard-specific validation checks defined in validation_master where
FKHFNW\SH1(µ*/0(7$¶
The current driver workflow is described in the following steps:
a)
Choose the standard(s) of interest and populate work._cstStandardsforIV with these standard(s):
*******************************************************************;; * User defines standard(s) of interest in the following data step *;; * Note exclusion of CST-­FRAMEWORK standard. *;; *******************************************************************;; %cst_getRegisteredStandards(_cstOutputDS=work._cstAllStandards);; data work._cstStandardsforIV;; set work._cstAllStandards (where=( (upcase(standard) = 'CDISC-­ADAM' and standardversion='2.1') or (upcase(standard) = 'CDISC-­CRTDDS' and standardversion='1.0') /* or (upcase(standard) = 'CDISC-­ODM' and standardversion='1.3.0') or (upcase(standard) = 'CDISC-­ODM' and standardversion='1.3.1') or (upcase(standard) = 'CDISC-­SDTM' and standardversion='3.1.1') or (upcase(standard) = 'CDISC-­SDTM' and standardversion='3.1.2') or (upcase(standard) = 'CDISC-­SEND' and standardversion='3.0') or (upcase(standard) = 'CDISC-­TERMINOLOGY' and standardversion='NCI_THESAURUS') */ ));; run;; b)
Modify the standard validation SASReferences to point to the validation_control view of interest. SAS views are being
used to make definition of the various check subsets more dynamic. If problems arise using SAS SQL views within or
external to Toolkit (e.g., within CDI or SDD), Toolkit can revert to the use of physical SAS data sets.
******************************************************************************;; * Modify the sample SASReferences data set to point to the run-­time *;; * validation_control data set identifying the validation checks of interest. *;; ******************************************************************************;; * libname below accesses c:/cstSampleLibrary/cst-­framework-­1.5/control *;; libname _cstTemp "&studyrootpath/control";; data work.stdvalidation_sasrefs;; set _cstTemp.stdvalidation_sasrefs;; if type='control' and subtype='validation' then do;; filetype='view';; memname='validation_control_std.sas7bvew';; end;; run;; PhUSE 2012
The view validation_control_std is one of several provided with Toolkit 1.5. The development code used to create these
views is listed below:
*******************************************************************************;; * Create the various internal validation validation_control data sets/views *;; *******************************************************************************;; proc sql;; create table cstcntl.validation_control as select * from refcntl.validation_master;; create view cstcntl.validation_control_fw as select * from cstrcntl.validation_master as a where upcase(a.checktype)="GLMETA";; create view cstcntl.validation_control_std as select * from cstrcntl.validation_master as a where upcase(a.checktype) in ("STD","STDIQOQ");; create view cstcntl.validation_control_stdiqoq as select * from cstrcntl.validation_master as a where upcase(a.checktype) in ("STDIQOQ");; quit;; c)
Call the process setup macro to perform all file and library allocations. The returned &_cstSASRefs data set contains fully
resolved paths and memnames.
%cstutil_processsetup( _cstSASReferencesLocation=&workpath, _cstSASReferencesName=stdvalidation_sasrefs );; d)
Re-FUHDWHZRUNVWGYDOLGDWLRQBVDVUHIVVHWWLQJBVUFILOH ¶):9$/¶
data work.stdvalidation_sasrefs;; set &_cstSASRefs;; attrib _srcfile format=$8. label='File source for record';; *********************************************************************;; * Framework validation sasreferences: cstcntl.stdvalidation_sasrefs *;; *********************************************************************;; _srcfile='FWVAL';; run;; e)
Call the code-generator macro to build the job stream for each standard:
filename incCode CATALOG "work._cstCode.stds.source" LRECL=255;; %cstutilbuildstdvalidationcode( _cstStdDS=work._cstStandardsforIV, _cstSampleRootPath=_DEFAULT_, _cstSampleSASRefDSPath=_DEFAULT_, _cstSampleSASRefDSName=_DEFAULT_ );; SAS Code is generated by the macro above for each standard that is specified in the _cstStdDS parameter. In this
example the 2 standards are ADaM and CRT-DDS. Once this code is created it is then %included and the following
workflow is persisted.
1.
Initialize work._cstTempSASRefDS to accumulate SASReferences records from all standards of interest for later
use by cstvalidate().
PhUSE 2012
2.
Look for standard-specific standardSASReferences data set from the Global Library. If found, run
cstutil_processsetup() using this file.
3. Append the fully-resolved work._cstSASRefs to the work._cstTempSASRefDS created in 1. above, setting
BVUFILOH ¶67'¶
4. Look for standard-specific sdtvalidation_sasrefs data set from the Sample Library. If found, run
cstutil_processsetup() using this file.
5. Append the fully-resolved work._cstSASRefs to the work._cstTempSASRefDS created in 1. above, setting
BVUFILOH ¶678'<¶
6. Remove any duplicate records from work._cstTempSASRefDS (keys=standard standardversion type subtype).
7. Run %cstutilbuildmetadatafromsasrefs(cstSRefsDS=work._cstTempSASRefDS,
cstSrcTabDS=work.source_tables, cstSrcColDS=work.source_columns). This macro dynamically builds
reference_tables and reference_columns data sets from a SASReferences data set.
8. Set _cstSASRefs=work._cstTempSASRefDS ± the cumulative ready-to-go SASReferences data set
9. Call cstvalidate(), which uses the validation_control_std view from b) above
10. Remove standard-specific records from work._cstTempSASRefDS, in anticipation of appending new records for
the next standard to the framework records that remain.
f)
Repeat steps 1-10 as described above for *each* standard selected in a) above, in this case ADaM and CRT-DDS.
g)
Results are collated in cstrslt.validation_results. An excerpt of process results is provided in the following figure:
VALIDATE_IQOQ.SAS
This driver is designed to provide a workflow that enables the user to run all internal validation checks identified as supporting
Toolkit IQOQ.
The current driver workflow is described in the following steps:
PhUSE 2012
Choose the standard(s) of interest and populate work._cstStandardsforIV with these standard(s).
validate_standard.sas example listed above for a code excerpt.
a)
See a) in the
Modify the standard validation SASReferences to point to the validation_control view of interest.
****************************************************************************************;; * The validation_control_fw view of the validation_master data set includes just those *;; * core framework checks that look only within the <cstGlobalLibrary>/metadata folder. *;; ****************************************************************************************;; libname _cstTemp "&studyrootpath/control";; data work.stdvalidation_sasrefs;; set _cstTemp.stdvalidation_sasrefs;; if type='control' and subtype='validation' then do;; filetype='view';; memname='validation_control_fw.sas7bvew';; end;; run;; The view validation_control_fw is one of several provided with Toolkit 1.5. The development code used to create this view
is listed below.
proc sql;; create view cstcntl.validation_control_fw as select * from cstrcntl.validation_master as a where upcase(a.checktype)="GLMETA";; quit;; b)
Call the process setup macro to perform all file and library allocations. The returned &_cstSASRefs data set contains
fully-resolved paths and memnames.
%cstutil_processsetup(_cstSASReferencesLocation=&workpath, _cstSASReferencesName=stdvalidation_sasrefs);; c)
Re-FUHDWHZRUNVWGYDOLGDWLRQBVDVUHIVVHWWLQJBVUFILOH ¶):9$/¶
data work.stdvalidation_sasrefs;; set &_cstSASRefs;; attrib _srcfile format=$8. label='File source for record';; *********************************************************************;; * Framework validation sasreferences: cstcntl.stdvalidation_sasrefs *;; *********************************************************************;; _srcfile='FWVAL';; run;; d)
(New step) Dynamically build reference_tables and reference_columns data sets from a SASReferences data set and call
cstvalidate(), which uses the validation_control_fw view.
%cstutilbuildmetadatafromsasrefs( cstSRefsDS=work.stdvalidation_sasrefs, cstSrcTabDS=work.source_tables, cstSrcColDS=work.source_columns );; %cstvalidate;; e)
(New step) Modify the standard validation SASReferences to point to the validation_control_stdiqoq view for any standard
specific processing requested in step a).
The view validation_control_stdiqoq is one of several provided with Toolkit 1.5. The code used to create this view is listed
below:
PhUSE 2012
proc sql;; create view cstcntl.validation_control_stdiqoq as select * from cstrcntl.validation_master as a where upcase(a.checktype)=" STDIQOQ";; quit;; f)
Call the code-generator macro to build the job stream for each standard:
filename incCode CATALOG "work._cstCode.stds.source" LRECL=255;; %cstutilbuildstdvalidationcode( _cstStdDS=work._cstStandardsforIV, _cstSampleRootPath=_DEFAULT_, _cstSampleSASRefDSPath=_DEFAULT_, _cstSampleSASRefDSName=_DEFAULT_ );; SAS Code is generated by the macro above for each standard that is specified in the _cstStdDS parameter. In this
example the 2 standards are ADaM and CRT-DDS. Once this code is created it is then %included and the following
workflow is persisted.
1.
Initialize work._cstTempSASRefDS to accumulate SASReferences records from all standards of interest for later
use by cstvalidate().
2. Look for standard-specific standardSASReferences data set from the Global Library. If found, run
cstutil_processsetup() using this file.
3. Append the fully-resolved work._cstSASRefs to the work._cstTempSASRefDS created in 1. above, setting
BVUFILOH ¶67'¶
4. Look for standard-specific sdtvalidation_sasrefs data set from the Sample Library. If found, run
cstutil_processsetup() using this file.
5. Append the fully-resolved work._cstSASRefs to the work._cstTempSASRefDS created in 1. above, setting
BVUFILOH ¶678'<¶
6. Remove any duplicate records from work._cstTempSASRefDS (keys=standard standardversion type subtype).
7. Run %cstutilbuildmetadatafromsasrefs(cstSRefsDS=work._cstTempSASRefDS,
cstSrcTabDS=work.source_tables, cstSrcColDS=work.source_columns). This macro dynamically builds
reference_tables and reference_columns data sets from a SASReferences data set.
8. Set _cstSASRefs=work._cstTempSASRefDS ± the cumulative ready-to-go SASReferences data set.
9. Call cstvalidate(), which uses the validation_control_stdiqoq view from e) above.
10. Remove standard-specific records from work._cstTempSASRefDS, in anticipation of appending new records for
the next standard to the framework records that remain.
g)
Repeat steps 1-10 as described above for *each* standard selected in a) above.
h)
Results are collated in cstrslt.validation_results.
VALIDATE_FRAMEWORK.SAS
This driver is designed to provide a workflow that enables the user to run all internal validation checks pertaining to the Toolkit
Global Library metadata folder (<cstGlobalLibrary>/metadata).
The current driver workflow is described in the following steps:
a)
Choose the standard(s) of interest and populate work._cstStandardsforIV with these standard(s).
validate_standard.sas example listed above for a code excerpt.
b)
Modify the standard validation SASReferences to point to the validation_control view of interest.
See a) in the
****************************************************************************************;; * The validation_control_fw view of the validation_master data set includes just those *;; * core framework checks that look only within the <cstGlobalLibrary>/metadata folder. *;; ****************************************************************************************;; libname _cstTemp "&studyrootpath/control";; data work.stdvalidation_sasrefs;; PhUSE 2012
set _cstTemp.stdvalidation_sasrefs;; if type='control' and subtype='validation' then do;; filetype='view';; memname='validation_control_fw.sas7bvew';; end;; run;; The view validation_control_fw is one of several provided with Toolkit 1.5. The code used to create this view is listed
below:
proc sql;; create view cstcntl.validation_control_fw as select * from cstrcntl.validation_master as a where upcase(a.checktype)="GLMETA";; quit;; c)
Call the process setup macro to perform all file and library allocations. The returned &_cstSASRefs data set contains fully
resolved paths and memnames.
%cstutil_processsetup( _cstSASReferencesLocation=&workpath, _cstSASReferencesName=stdvalidation_sasrefs );;
d)
Re-FUHDWHZRUNVWGYDOLGDWLRQBVDVUHIVVHWWLQJBVUFILOH ¶):9$/¶
data work.stdvalidation_sasrefs;; set &_cstSASRefs;; attrib _srcfile format=$8. label='File source for record';; **********************************************************************;; * Framework validation sasreferences: cstcntl.stdvalidation_sasrefs *;; **********************************************************************;; _srcfile='FWVAL';; run;; e)
(New step) Dynamically build reference_tables and reference_columns data sets from a SASReferences data set and call
cstvalidate(), which uses the validation_control_fw view.
%cstutilbuildmetadatafromsasrefs( cstSRefsDS=work.stdvalidation_sasrefs, cstSrcTabDS=work.source_tables, cstSrcColDS=work.source_columns );; %cstvalidate;; f)
(New step) Modify the standard validation SASReferences to point to the validation_control_std for any standard specific
processing requested in step a).
g)
Call the code-generator macro to build the job stream for each standard:
filename incCode CATALOG "work._cstCode.stds.source" LRECL=255;; %cstutilbuildstdvalidationcode( _cstStdDS=work._cstStandardsforIV, _cstSampleRootPath=_DEFAULT_, _cstSampleSASRefDSPath=_DEFAULT_, _cstSampleSASRefDSName=_DEFAULT_ );; SAS Code is generated by the macro above for each standard that is specified in the _cstStdDS parameter. In this
example the 2 standards are ADaM and CRT-DDS. Once this code is created it is then %included and the following
workflow is persisted.
PhUSE 2012
1.
Initialize work._cstTempSASRefDS to accumulate SASReferences records from all standards of interest for later
use by cstvalidate().
2. Look for standard-specific standardSASReferences data set from the Global Library. If found, run
cstutil_processsetup() using this file.
3. Append the fully-resolved work._cstSASRefs to the work._cstTempSASRefDS created in 1. above, setting
BVUFILOH ¶67'¶
4. Look for standard-specific sdtvalidation_sasrefs data set from the Sample Library. If found, run
cstutil_processsetup() using this file.
5. Append the fully-resolved work._cstSASRefs to the work._cstTempSASRefDS created in 1. above, setting
BVUFILOH ¶678'<¶
6. Remove any duplicate records from work._cstTempSASRefDS (keys=standard standardversion type subtype).
7. Run %cstutilbuildmetadatafromsasrefs(cstSRefsDS=work._cstTempSASRefDS,
cstSrcTabDS=work.source_tables, cstSrcColDS=work.source_columns). This macro dynamically builds
reference_tables and reference_columns data sets from a SASReferences data set.
8. Set _cstSASRefs=work._cstTempSASRefDS ± the cumulative ready-to-go SASReferences data set.
9. Call cstvalidate(), which uses the validation_control_fw view from b) above
10. Remove standard-specific records from work._cstTempSASRefDS, in anticipation of appending new records for
the next standard to the framework records that remain.
h)
Repeat steps 1-10 as described above for *each* standard selected in a) above.
Results are collated in cstrslt.validation_results.
CONCLUSION
In the end, when a user is running any validation software, it is important that the user has a high level of confidence that any
supplied standards metadata is not only installed correctly but contains accurate information. This paper described the
process designed by SAS for its Clinical Standards Toolkit product to ensure the product is installed properly and all required
metadata is where it should be and accurate. In addition to verifying standards shipped with the product, the user can also
verify their customized standards.
MORE INFORMATION
SAS Clinical Standards Toolkit information can be found at: http://support.sas.com/rnd/base/cdisc/cst/index.html CONTACT INFORMATION
Your comments and questions are valued and encouraged. Please contact the author at:
Gene Lightfoot
SAS Institute Inc.
SAS Campus Drive R4318
Cary, North Carolina 27513 USA
+1 (919) 677-8000
+1 (919) 677-4444 (Fax)
[email protected]
www.sas.com
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.

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project