HOME QUICK SEARCH AUTHORS KEYWORDS HELP BUGS

PharmaSUG 1997 Papers - Index

52 papers. Source: PharmaSUG (Pharmaceutical Software Users Group) conference proceedings.

Back to Conf
go 1998
# Title
×		 Authors
×		 Abstract
×		 Keywords
×		 Extras Pages
1 A Generalized Macro for Survival Analysis Malini Krishnan,
Hoechst Marion Roussel		 The use of SAS macro facility in writing a user-friendly generalized application to perform survival analysis on clinical data is illustrated here. The example shows a SAS program with the macro invocation and the SAS outputs generated by the macro, for a CHF study of total mortality with active and control treatment groups. The time interval used in the study was month and the event considered was death macro loop, macro definition, macro conditional, proc step, survival analysis, proc gplot, macro facility, macro call 3
2 A Global Utility Menu: MUM Derek Chalmers GlaxoWellcome has developed an in-house database to manage the collection and processing of nearly all clinical trial data. The database, known as GOLD, is an Oracle-based product located on a VAX platform. GOLD is available to the corporate Clinical Data Management groups via three instances: one to serve the US, another in the UK and a third to handle European operating company studies. GOLD has a number of components including among others: Assessment Dictionaries, a Database set-up library, a Processing database and a Reporting database. GOLD is also used in the management of external data processing activities; for example, work carried out by centralised laboratory services or Contract Research Organisations. Although many ancillary utilities are included, GOLD does not incorporate all the tools required within the Medical Data Sciences (MDS) division. For example, SAS is used as the principle reporting software among statisticians, and, it is also one of the major data transfer media used when retrieving externally processed data. However, there is no inherent connectivity between the GOLD system and SAS; leaving a gap between the reporting database and the reporting tools. Because of this, many utilities have evolved around the GOLD database to help link incoming and outgoing SAS data to the Oracle database. In order to better organise and manage this environment the MDS Utility Menu (MUM) was designed. view, proc step, data cleaning, clinical data, dataset, macro variable 3
3 A SAS/AF Class Library for Automated Generation of Clinical Report Programs Carl R. Haske STATPROBE, Inc., has developed a standard SAse programming system that has been in use for a number of years. The pro gramming system is used to produce computer-generated reports for clinical trials. It is set up to maximize flexibility and efficiency during the course of report production for data analysis, whether the report is a basic clinical report, a report for a Data Safety Monitoring Board (DSMB), a New Drug Application (NDA), or Integrated Summaries of Safety or Efficacy (ISSIISE). The programming system consists of a generic macro library and a standard file structure that is created for each project. The standard file structure is organized in a hierarchical manner, al lowing support for multiple protocols and the opportunity to de velop and store reusable program code. The final phase of development in this programming system con sisted of classifying the computer-generated report specifications and creating a class library in SAS/AF corresponding to these classifications adverse events, clinical database, clinical trial, study day 7
4 Accessing Multiple Versions of a Database with One Set of Programs Spring Tseng,
Kirsten MacDonald		 During a clinical trial, there are times when an interim report needs to be written based on the data available. A shapshot of the complete study database is taken and preserved for report writing. A later version of the on-going database containing more data will be used for another report. Accessing and identifying these two versions of a database co-existing on a system can be a challenging task. Changes to the SAS programs are often needed in order to access the correct database and to identify the database version on the computer output. An efficient way to take care of this situation in UNIX can be accomplished by using two SAS macros and a UNIX shell script. Since all the programs are designed to link to these files, no change is needed once a SAS program is written dataset, macro definition, macro variable 3
5 Adverse Events Summary Tables in Clinical Trials: Incorporating P-values Using SAS Macros and SAS/AF Incorporating P-values In this paper, a SAS reporting system is developed to obtain computer-generated summary tables of adverse events (AE). The system consists various customized SAS macros that can be applied by SAS users. One major characteristics of the system is that the adverse events summary tables can contain one of two types of p-values from two common statistical tests. Fisher's exact test is used to calculate the p-values comparing the incidence rates among treatment groups for each of the adverse events reported, while the Wilcoxon rank sum test is applied to compute the p-values for evaluating the difference of severity for an adverse event among treatment groups. Both techniques and guidelines to develop the system are presented in detail in this paper. Also, an ongoing application development of user-friendly interface with SAS/AF is discussed with a focus on module design and screen design adverse events, proc means, fisher exact test, adverse event summary, exact test, macro definition, macro variable, trial elements, sample, preferred term, data step, macro loop 10
6 An Alternative Method of Transposing Data without the TRANSPOSE Procedure Sunil Kumar Gupta While the PROC TRANSPOSE procedure may be flexible enough to generate a variety of report formats, it may not be appropriate if the user requests a custom report. In addition, for some cases, it may not be possible to successfully apply the TRANSPOSE procedure to obtain the desired output. A technique is presented that will provide the user with more control on the number and order of the columns generated, enable new columns to be created, and allow for transposing of several variables from different data files. The features of the two methods are compared and discussed proc transpose, proc step 3
7 An Application of the Internet-based Automated Data Management System (IADMS) for a Multi-Site Public Health Project Michele G. Mandel,
Robert E. Schwartz Jr.,
Steven A. Kinchen		 Prevention (CDC) began a multi-center case-control study of breast cancer risk factors, focusing especially on risks associated with use of oral contraceptives and other steroid hormones. Data collection for the study, which is already in progress, will last four years. The study is a collaboration between the National Institutes of Health, National Institute of Child Health and Human Development (NICHD), CDC, five field sites, and a data management contractor. Initially, data processing relied on diskettes and mainframe tapes to transfer data among field sites, the data management contractor, and CDC. Data were loaded onto the CDC mainframe and processed in a batch environment. The growth of the Internet and the availability of the SASiI!> System on a UNIX workstation led to a redesign of the data processing protocol and development of the Internet-based Automated Data Management System (IADMS) data wrangling, libname statement, data governance, dataset 6
8 An Introduction to Data Warehousing, with Emphasis on Its Use With Clinical Data Andrew A. Kramer Data warehousing was originally conceived as a more efficient means of storing, retrieving and analyzing business data. Books on the subject have focused almost exclusively on applications in the financial and manufacturing areas. Thus it is not surprising if clinical data mangers are unfamiliar with data warehousing, or feel it's inapplicable to their work. But some of the concepts of data warehousing are already being utilized by clinical data professionals, although under different terminology than that used by systems developers. In an effort to bridge the gap between clinical data practices and data warehousing, this presentation will explain some of the basic concepts of data warehousing and demonstrate how data warehousing can be beneficial to the analysis of data from clinical trials. clinical data, data warehouse 6
9 Autogenesis: Automatic Generation of a SAS Data Entry Application Derek Chalmers This paper describes a project whose objective was to pilot automatic creation of data entry applications: Autogenesis. Regardless of the software used, manual design, creation and validation of data entry applications can be a resource intensive process. Currently Medical Data Sciences (MDS) at GlaxoWellcome use EntryPoint 90 as the preferred software to build data entry applications. Programming, testing and validation takes typically 4-6 weeks and is carried out by a relatively small pool of available programmers. The company has developed an in-house database to manage the collection and processing of nearly all clinical trial data. The database, known as GOLD, is an Oracle-based product located on a VAX platform. GOLD is available to the corporate Clinical Data Management groups via three instances: one to serve the US, another in the UK and a third to handle European operating company studies. GOLD has a number of components including among others: Assessment Dictionaries, a Database Set-up library, a Processing database and a Reporting database; however, GOLD does not incorporate a data entry subsystem. When a new study protocol is being initiated, MDS will build a relevant database structure by adding to the Assessment Dictionaries and Database Set-up subsystems of GOLD. The challenge of autogenesis is to exploit the dictionary and set-up information; collectively know as Meta-Data, and use it to generate data entry applications automatically. proc step, data cleaning 4
10 Benefits of using a SAS based Electronic Data Collection Application Tracy Millander,
Ronald Fitzmartin		 Electronic data collection (ED C) provides for electronic capture of clinical information at the investigational site and electronic transmission to a central database where data can be distributed rapidly for review and decision analysis. The traditional method of data collection in the pharmaceutical industry has been the paper Case Report Form (CRF). This is a multi-page document designed to capture patient information during the conduct of a clinical trial. Currently, study coordinators enter information onto these forms which are later returned to the Sponsor. Once in-house these forms are reviewed by CRAs and data personnel, data clarification request forms (DCF) are generated, and the data are dual entered into a database. With paper CRFs data entry usually does not begin until the CRF's are in-house. The database can not be locked and released for analysis until all DCF's have been resolved data validation rules, sas base 2
11 Better Titles: Using The #BYVAR and #BYVAL Title Options Arthur L. Carpenter Although it has been possible to place macro variables in titles and footnotes for quite awhile, Releases 6.07 of the SAS System and later also include the ability to place BY variable values and names in the titles and footnotes as well. These options are only documented in SAS Technical Report P-222 and are therefore under utilized. The names of BY variables and the BY variable values are inserted into titles and footnotes using the #BYV AR and #BYV AL options. There are two forms of these options and they will accommodate more than one BY variable. Several examples will be presented using these options by group processing, proc print, proc step 3
12 ClinAccess: An Integrated Client/Server Approach to Clinical Data Management and Regulatory Approval Martin J. Rosenberg Motivated by the desire to better manage an ever increasing volume of information and to reduce the length of time required to introduce new drugs around the world, the pharmaceutical industry and regulatory agencies such as the U.S. Food and Drug Administration have sought to facilitate the drug development and approval processes through innovative uses of computer technology. One such effort that has received great attention is the computer assisted NDA or CANDA and its cousin for biologicals the computer assisted PLA or CAPLA. The first CANDAs were created with much effort, and at great expense, after paper NDAs had been filed. From these early experiments, it became evident that if CANDAs were to ever become routinely used throughout the industry by both large and small companies, their development would need to be simplified and to begin earlier in the drug development process. Medical reviewers at regulatory agencies have frequently commented that for a system to be useful as a CANDA, it should also be beneficial for in-house use. Recent advances in computer technology now permit the construction of sophisticated systems which remain relatively simple to use. However, in order to maintain data integrity, it is essential that data tables do not require transformation or restructuring by the end user. Hence ideally, the development of a CANDA should begin at the time of study definition and data entry. With these goals in mind, MAJARO has explored ways of developing systems which seamlessly integrated data entry and in-house medical review, so as to facilitate CANDA development. This paper discusses ClinAccess a family of integrated clinical trials products developed with SAS software, that combines the data entry and data management capabilities of traditional clinical information systems, with clinical data review features that permit monitors, CRAs, managers, and other members of the clinical staff to monitor the progress and quality of ongoing clinical trials. It achieves its ease of use through a graphical user interface that incorporates pull-down menus, scroll bars, dialog boxes, radio buttons, and a mouse. Initially developed for use on PC compatibles, the code has been written to permit easy porting to other computer platforms and exploitation of client-server technology. clinical trial, clinical data, crf, data review, data governance, clinical research, clinical data management, audit trail, pharmaceutical industry, key, adverse events, data cleaning 9
13 Clinical Software Systems Development in SAS/AF Carl R. Haske Using SAS/AP' as a software development platform permits rapid applications development. SAS supplies several classes of ob jects that the developer can use to prototype and implement sys tems quickly. However, in developing applications for use in the pharmaceutical industry, it is useful to augment these classes with additional classes that provide robust data management and reporting capabilities. Classes that provide generic tools to sub set data, browse data, move data, and report data are useful in the clinical research environment. Managing clinical trials data requires the efforts of a large team of people, including clinical data managers, data entry personnel, programmers, and statisticians. A system that enables concur rent access to data at several stages of the database life cycle enhances the overall work process and accelerates database de velopment clinical data, dataset, clinical database, audit trail, data governance, data cleaning, clinical trial, proc step 11
14 Combining Disparate Outputs from Consistency Checks into a Central Report Greg Henkle A common scenario for many programmers is to face a database full of data that needs to be validated. The programmer, therefore, must find both a means of validating the data and reporting the findings of that validation. In such a scenario, I feel the most efficient approach to SAS validation lies in the use of individual, self-sufficient checks grouped together to produce a common output. While this does increase system usage, the advantages of this approach easily outweigh any loss of time and system efficiency. Using self-sustained programs for consistency checks (CCs) allows the checks to be easily copied from one protocol to another, requiring only minor modifications to the path. This also allows the checks to be run individually (if desired) as well as tested individually. However, if consistency checks are not grouped in some manner during output, the Medical Data Administrator (MDA) can encounter significant headaches. If the checks aren't grouped, preferably by patient, the MDA is left bouncing back and forth between individual reports attempting to determine the number of problems flagged for a particular patient. And, while a patient may only be flagged for one error, there is no way to determine this until all of the CC reports have been examined. It is this dilemma that led me to develop the macros I am discussing today. proc step, proc transpose, data cleaning 3
15 Controlling for Experimentwise Type II Error: Optimum Patient Allocation for a Three Armed, Parallel, Randomized, Controlled Clinical Trial with Binary Outcomes Benjamin Duncan In the drug development process, controlling for type I error (declaring the test drug is efficacious when in reality it is not) typically takes precedence over controlling for type II error (declaring the test drug is not efficacious when it is). Protecting the public from ineffective medicines is naturally of extreme importance. However, mistakenly withholding effective medicines from the public is also important. Either type of error could cause the loss of human life¹ or lead to unnecessary suffering. Multiple outcomes or multiplicity complicates matters further when attempting to control for type I and type II errors. Multiplicity is a concern for each of the following situations: a clinical trial with two or more primary variables of interest, a clinical trial with three or more treatment groups, two or more similar studies testing the same test treatment, and any combination of these. Therefore, how to and when to adjust for type I and type II errors is both nontrivial and controversial. type ii error, sample size, sample, clinical trial, type i error, significance level, alternative hypothesis 13
16 Creating Tables from SAS to Microsoft Word Susan Haviar The following is a macro that can save an enormous amount of time when creating numerous tables that go into summaries, final reports, and submissions in the Microsoft Word environment. The purpose of this macro is to create a comma delimited file of statistical final output that can be imported into the Microsoft Word program to create statistical tables for final and summary reports. %if-%then-%else, macro loop 1
17 Design of the Clinical Data Portion of a CANDA to Facilitate Review by CDER Kay Obenshain This paper presents the design of the clinical data portion of a CANDA that was submitted to the Division of Metabolic and Endocrine Drug Products at the FDA's Center for Drug Evaluation and Research (CDER) in 1995. SAS/PH-Clinical software Version 2.0 (PHC) was customized to provide an interface to data presented in the integrated summary of efficacy and safety (ISE/ISS) sections of the New Drug Application (NDA). The focus of this paper is on the technical aspects of designing the PHC portion of the CANDA, referred to here as the current CANDA. dataset, laboratory data, vital signs, proc means, clinical data, clinical trial 5
18 Developing a Suite of Clinical Report Macros for SAS/PH-Clinical Paul M. Gilbert,
Steven Light		 In an effort to streamline the production of clinical reports, many companies are developing SAS report generation tools to produce listings, summary tables and graphics. This presentation discusses critical factors in the design of a suite of SAS-based Clinical Report Macros. Additionally, it describes how these report macros can be used within SASIPH-Clinical as interactive report templates macro call, proc step, proc freq, proc report, clinical data, macro parameter, proc means, sample, adverse events 9
19 Development of an Epidemiology Menu-Driven Query System Kathryn A. Munoz,
Richard M. Radkowski,
Roy A. Hoffman,
Cynthia J. Girman		 Data from govemment agencies, universities and industries on health and health service resource utilization are used in the pharmaceutical industry to answer research questions valuable to epidemiology, health outcomes/services and marketing. Currently no methodology (including data warehousing) exists that will allow easy access to such information. Therefore, we developed a user-friendly, menu-driven, information system that allows the novice and expert user to easily access public and proprietary databases and to obtain statistics, graphics and datasets. Since SAS version 6.11 can manipulate large complex file structures and provide a seamless interface across multiple platforms in a Windows environment, it was used to build this two-component system. The first component, a menu-driven "set-up" program, allows automated addition, deletion, transfer and copy of databases and definitions to Dataland, a data repository, and was developed in SAS/AF utilizing SAS/SCL excel, dataset, data cleaning 6
20 FDA Initiatives: Electronic Data Interchange (EDI) for Drug Safety Monitoring Bill Calvert Past, present, and future Food and Drug Administration (FDA) initiatives in EDI are presented with emphasis on current activities that will significantly impact the way the pharmaceutical industry and The FDA do business in the future. Specific information is provided about ongoing development and testing of ED I and related FDA drug safety monitoring operations. The role of SAS software in these activities is also discussed pharmaceutical industry, adverse events 3
21 Find SAS System Errors and Make Corrections Quicker: An Enhancement of the VMS Eve Editor Lewis Fountain This paper discusses two EVE editor (VMS operating system) commands you can use to easily find SAS errors in LOG files, switch to the line in the SAS file which has the error, and make the correction. The EVE commands can be added to the base EVE editor in less than 1 hour. One command searches LOG files for a variety of strings associated with SAS errors. Upon identifying the line of code that produced the error, you may position the cursor before the text that caused the error and then execute the second command, switching to the corresponding line in the SAS file that produced the error so that it may be corrected. You may then switch back to the LOG file at the point where you exited it and continue looking for more errors, and switch back to the SAS file to make corrections. The two EVE editor commands greatly speed up the time it takes to search for and correct errors error message, error 2
22 How Do I Take the Mean of an Apple and an Orange? Techniques for Integrating Data from Multiple Studies Linda Collins A clinical research program on a drug or device includes a number of studies. Pooling information from many studies is necessary, particularly for safety information. Such pooling is complicated by the fact that the data gathering instruments (case report forms) tend to evolve over time. To create a meaningful integrated analysis, the differences between studies must be rectified by transforming the data into a 'common denominator' format. This format depends on the raw data available, the planned analysis, and, ideally, an institution-wide definition of standards. This paper discusses methods for planning and carrying out this transformation and the analyses that are based on it raw data, macro variable, crf, sample, proc step, adverse events, response variable, dataset, date variable, missing value, proc contents, proc means, macro definition, macro loop 8
23 How To Create a Simple SAS/AF Interface For Existing Batch Applications Bryan K. Beverly Often after a programmer has developed an application for his/her own use, the need arises to make the application accessible to a broader community of interest. This is particularly true for clinical research programmers who develop multifunctional reporting applications. Having a user-friendly interface facilitates the technology transfer process because it makes the application useable by internal and external customers. However, retrofitting an existing application with an interface can be difficult if tasks such as passing source code JCL parameters are involved. This poster presents a simple but effective method of building a SAS/AF interface around a batch application. Using an MVS SASapplication, the reader will learn how to design a screen that will enable a user to specify section criteria for multiple executions of an application without having to reedit the source code batch job, clinical research 4
24 How to Use CD-ROM as a Simplified CANDA Steven Wong An application for approval to market a drug requires extensive regulatory agency review from different scientific disciplines, one of which is statistics. Clinical data and statistical software must therefore be transferred to regulatory agencies. The transfer media could vary from the high-tech sophisticated CANDAs (computer-assisted new drug applications) to simple floppy diskettes, tapes, or CD-ROMs containing data and program files. This paper discusses SAS programming techniques that enhance the transfer, especially if a CD-ROM is the desired media. The strategy developed by Matrix Pharmaceutical, Inc. to resolve these issues and some of the problems encountered along the way wiIl be addressed sample, options statement, macro variable, macro definition 4
25 Integrating SAS Programs with a Multi-Functional Computer Assisted NDA (Canda) Kathryn B. Eisen Providing a user-friendly multi-capability environment that will speed the FDA review of Drug and Biological Applications has become a challenge for many Pharmaceutical and Biotech companies today. One of Covance's current CANDAICAPLA models is a non SAS-based application which provides robust text, image and data capabilities. This paper takes an in-depth look at the way in which Covance integrates SAS programs with free form queries to create a user-friendly system which can facilitate review of massive amounts of clinical data. A quick demo of the system shows that even Medical Reviewers, with no knowledge of SAS programming techniques, can use prewritten SAS programs to perform summary statistics on subsets of data they define. The hardware, software and the process of converting data to PC format is discussed. By modifying the SAS mainframe programs which create NDA reports to work with subsets of data on the PC, the system quickly and efficiently produces meaningful SAS output data cleaning, validation set 4
26 Integration of Oracle Clinical databases with SAS/PH-Clinical Daniel Gronell This paper addresses the process of making Oracle Clinical data available to SASIPH Clinical. As well, much of this paper will also be useful to sites that store their raw data in Oracle format but are not using Oracle Clinical, as many of the tasks discussed (e.g., use of SAS/ Access to create SAS datasets from Oracle data) apply in both cases. This paper does not assume knowledge of Oracle Clinical, but does assume familiarity with data dictionary concepts of SASIPH -Clinical dataset, proc sql, clinical data, data dictionary, view, clinical trial, sample, proc contents, proc step, select 10
27 Keeping Track of Electronically Transmitted Data M. Rita Thissen This article discusses the monitoring and control of electronically transmitted data. A SAS language application is used to report the arrival of data, perform identification checks and test for duplicate data as part of the monitoring step. Observations that pass the validity test are moved to a separate storage area, and those that fail are noted in an exception report and held for evaluation. Data arrival events are recorded in a control system database, and periodic reports aid in tracking the progress of data collection. Together, these operations allow the identification, monitoring and processing of electronically transmitted data. Shifting Technology Advances in the transfer of data offer attractive alternatives to traditional manual methods. While motives for adopting new technologies vary from one organization to another, some frequently cited reasons include improvement in timeliness, increased quality, enhanced performance or service dataset, data cleaning 6
28 MATCH-MERGING: Techniques, Tricks, and Traps Malachy J. Foley Match-merging, or BY merging, is the most common merging technique used in SAS . Yet, how it works is not always obvious. This presentation covers all the basics of how match merging works, shows many of the technique's nuances and subtleties, and demostrates typical traps and how to avoid them by group processing, merge statement 8
29 Migration of SAS Environment from IBM Mainframe to UNIX CLIENT/SERVER Setting Dorothy Hu Huang,
Hoechst Marion Roussel		 To cut expenses and consolidate environments, a decision was made to migrate production SAS jobs from an IBM mainframe to a UNIX client server setting. After ?? years of using the mainframe, we, the Scientific Programming Group, were given three months to accomplish the switch; time was tight. The larger UNIX server would be accessed from both workstations and PC's. Complicating matters, a SAS version switch was also needed to migrate platforms; the mainframe ran SAS 5 while the UNIX machines ran SAS 6. We defined the tasks and objectives to be accomplished. They included the following: (1) Design file structure and naming conventions for SAS programs and datasets. (2) Investigate and use UNIX system security. (3) Migrate all in-house SAS application systems, e.g. TABLETS (table generation), GRAPHIX (graph generation), CLAS (clinical laboratory data), PAS (pharmacokinetics analysis), AE (adverse events), from IBM to UNIX. (4) Define a method to access historical IBM programs and datasets. (5) Design a PC connection: NFS (Network File System for sharing files from a PC or workstation). Reflection-X (graphical protocol to log into UNIX from a PC). (6) Configure laser printers. (7) Investigate full screen editors available on UNIX. (8) Test migration on a pilot project. We met weekly to discuss progress and identify problems. We divided up the tasks and addressed each problem and finally came up with a workable solution. decimal, adverse events 3
30 More About Missing Character Data Hany Aboutaleb Many data managers like to use some of the high ASCII characters to enter more information about missing values in their data, i.e., using' i. , for 'Cannot Enter', using 'Yz ' for 'Unknown' or 'j , for 'Illegible.' If you are tired of seeing these values ('0', 'i.,' , 'Yz', 'y..' , 'j' , '": , '"', or any other values used for missing data) in your databases, there is a solution. This paper describes a macro that will search all of your databases (library) for you and replace any high ASCII character value using your format values missing value, proc step, %if-%then-%else 2
31 New Architecture for Linkage of SAS/PH-Clinical Software with Electronic Document Management Systems Peter Villiers The purpose of this paper is to provide the reader with an overview of the architecture that will enable SAS PharmaTechnology Center products to interact with document management systems. As a whole, this is known within the PharmaTechnology Center as the PH-Document Linker Interface or PH DLI. The paper will cover the following major areas: Business case for the interface Design goals which are to be achieved Host computing environment impact Architecture components involved * The overall PharmaTechnology Center Process The paper will not cover lower level details of the code that has been written clinical trial, clinical data 7
32 Obtaining Group Sequential Test Boundaries with SAS William J. Stager,
Hoechst Marion Roussel,
Jessie J. Yun,
Hoechst Marion Roussel		 The following article describes a procedure for determining boundaries applicable to group sequential tests (GST) for normally distributed data. Methods are given for both the boundary specification and Type I and II error spending function approaches. These include standard formulations, symmetric and asymmetric, for one-parameter boundaries described by Pocock (1977), O'Brien and Fleming (1979) and Wang and Tsiatis (1987) and for spending functions given by Lan and DeMets (1983) and Hwang, et. al. (1990). Extensions allowing early rejection of the alternative hypothesis were developed by Emerson and Fleming (1989) and Pampallona and Tsiatis (1994). A numerical integration procedure based on Romberg quadrature incorporating PROBNORM and PROBBNRM is used to generate the probabilities associated with a set of boundaries. This procedure appears to be very efficient for the convolution of conditional normal densities used in this setting. Screens for simplifying implementation were developed using SAS/SCL. sample size, sample, null hypothesis, alternative hypothesis, clinical trial, macro definition 4
33 Playing with Macros: Take the Work Out of Learning to Do Macros Arthur L. Carpenter The macro language can add an additional level of complexity to programs written in SAS~. Macro statements are often complicated and may use logic that appears to defy the rules used by all proper SAS programmers. Regardless of the complexities, learning to use the fundamentals of the macro language need not be difficult This workshop will take users familiar with the basics of the SAS System and introduce them not only to the basics of the macro language, but also to a process for writing their own macros. The course will start with macro variables and will examine their differences and similarities to data set variables. Existing code will then be converted to macros and named and positional parameters will be introduced. Examples will include discussion on the resolution of macro variables and the use of the double ampersand (&&). Several macro programming statements, such as, %IF will be introduced along with selected macro functions macro variable, macro facility, macro loop, %if-%then-%else, proc step, macro call, if-then-else, macro definition 6
34 Printing with SAS 6.10 and 6.11 Hany Aboutaleb SAS has been an excellent product for analyzing and displaying data. However, the output controls are all based on the character position within the line. This will work fine if the font is fixed-spaced (where the inter-character spacing is the same whatever the character), but fails when using a proportional font (where different characters have different widths). After all, If you have large saved output files and would like to print them, you will need to load them to the SAS output window first. This paper describes a macro to help you print all of your files by using the printer command macro variable, data step, sas output, proc step, macro definition, %if-%then-%else 4
35 PROC TRANSPOSE as a Useful Intermediary Step in Generating Tables and Listings Tony T. Salehi proc step, proc transpose, sample, clinical data, proc freq, laboratory data 11
36 SAS Macro Plus Proc Report: A Code-free Approach to Table-driven Data Reporting Jun Zhao This paper demonstrates a way to integrate SAS macro and PROC REPORT and make it a table driven data reporting tool proc report, proc step, dataset, macro loop, macro definition 5
37 SAS Output To Final Polished Tables And Listings: An Automated Procedure Using Microsoft Wordbasic And Microsoft Visual Basic R. J. Wilson Technological advancements in computing software have resulted in greater expectations for the visual quality of statistical tables and listings. Historical methods for producing final tables have resulted in clinical writers recreating tables from the output provided. Careful quality control is needed to ensure this process is reliable. A method currently being developed ensures a validated step from SAS dataset output files to customised tables and listings. SAS program code is used to produce a dataset which contains information to be presented in the tables. Microsoft Word Basic macros are then called from Microsoft Visual Basic. It is unnecessary for the user to open Word, meaning no changes are made manually. Alternatively, Word Basic programming can be used directly in Visual Basic i.e. the whole system can be written in Visual Basic, offering flexibility in formatting the final output proc means, dataset, sas output, confidence interval, sas log, proc step 5
38 SAS Software Distribution Issues Gerardette M. Furlow The SAS software consultant for a university or large organization is often faced with the tremendous task of distributing the SAS System to a large user base of customers. These customers (i.e. faculty, staff, and students, etc.) generally request the SAS System for a variety of operating systems. As the SAS System continues to grow, the SAS software consultant must decide which software distribution methods will work the best in a heterogeneous environment This paper will review the methods used by the North Carolina State University (NCSU) Computing Center (CC) to distribute the SAS System to campus users grid computing, order by 3
39 Software Validation for the Rest of Us Michael C. Harris Statisticians and programmers alike are coming to realize that the issue of software validation in the clinical development area of our industry is here to stay. We have quickly discovered that the Software Development Life Cycle (SDLC) approach that works well for automated manufacturing and process control or for developing large scale interactive systems doesn't easily map to what we do. While certain types of analyses, such as analysis of baseline patient characteristics, can easily be standardized and therefore made part of a relatively static reporting system, others, such as endpoint-specific or exploratory analyses can't. Another distinguishing characteristic of our work is that the programmer is frequently the sole user of the program. In a team environment, programmers and statisticians receive requests from other team members, and rapid turnaround is essential data cleaning, crf 9
40 Some Less Known but Powerful Proc SQL Features Jun Zhao This paper will introduce four less known but useful PROC SQL features. They are the MIX join, the outer join with where clause, the multi level CASE expression, and the statistics through Boolean expression. This paper will also give sample code to demonstrate the usage of these features proc sql, inner join, proc step, where statement, if-then-else 4
41 Statistical Programming Workflow Process Improvement at Fujisawa USA Jay Erdman,
Paul M. Gilbert		 In an effort to reduce time to production of an NDA, an evaluation of quality and efficiency of statistical programming at Fujisawa USA was undertaken by the company with the assistance ofDataCeutics, Inc. This paper describes the evaluation process, the opportunities for improvement which were identified, and implementation of changes. A combination of procedural and technical changes led directly to substantially increased quality, decreased production time and reduced staffing needs. History Fujisawa USA, a US subsidiary ofa large Japanese company, was formed in the late 1980's. The company was tasked with the clinical development of drugs to facilitate human organ transplantation. Clinical trials involving transplant patients tend to be rather complex due to a very sick patient population with disease complications. The studies were noteworthy in the following ways: Studies were long, up to three years in duration clinical programming, adam dataset, clinical data management, sas programming, clinical data, data governance, data cleaning, proc sql 4
42 SuperSASaPharmalisticTelecommutosis Eric C. Brinsfield Eric C. Brinsfield, Meridian Software, Inc., Raleigh, NC Can a SAS programmer in the pharmaceutical industry telecommute? This paper attempts to address this question by surveying the technical, political, and sociological issues that arise when you start to consider telecommuting. What types of connections and software can a SAS programmer use? Is telecommuting more efficient or less efficient than being there? Is security a problem? Who are good candidates for working remotely? Why isn't everyone doing it? sas programmer, pharmaceutical industry 6
43 System Development Life Cycle for SAS Applications Development Incorporating a Procedure for As-Built Validation of In-Use Products Wayne Lanier SAS is increasingly being used as an applications development language. Originally an analysis lan guage, SAS tradition has only recently begun to in clude a body of formal System Development Life Cycle (SDLC) procedures. Of concern is validation for SAS-developed products used in data manage ment or data analysis during clinical studies in appli cation for FDA drug licenses. IBAH Resource Biometrics Clinical Software (RBI) is a Contract Research Organization (CRO) developing SAS-based software for Data Management and Data Analysis in the biotechnology and pharmaceutical industries. We are now publishing these software products for sale. It is necessary to validate these systems as part of current Good Clinical Practice (cGCP.) Initially we defined a simple RBI System Develop ment Life Cycle for new-product development which sets out the tasks for development, validation, re lease, and maintenance of software products data cleaning, clinical trial, validation documentation 7
44 The Evil Twins of Real Numbers That May Cause Unexpected Results in SAS Applications Aileen L. Yam The real number system is infinite. Because of computer hardware limitations, real numbers are being converted internally and are represented with an equivalent finite set of numbers. Some real numbers cannot be represented exactly and can vary very slightly according to different platforms on which the SAS System runs. Thus, numeric representation has the potential for imprecision. This paper is a non-technical guide to understand numeric representation, with examples to illustrate when numeric representation can cause unexpected results and how to minimize representation problem &variable, do loop 3
45 The Jonckheere-Terpstra Nonparametric Ordered Alternatives Test Nestor Rohowsky The Jonckheere-Terpstra test is a nonparametric method for evaluating k>2 comparison (e.g. treatment) groups under the null hypothesis of equal outcomes among the groups against the one-sided alternative that outcomes follow a definite ordering, e.g. T1 =< T2 =< ... =< Tk, with at least one strict inequality. This method may therefore be regarded as an analysis of trend or dose response. If the treatments are a priori appropriately ordered and the outcomes are assumed to follow this ordering, then this test is more powerful than the nonparametric Kruskal-Wallis test or a standard one-way Analysis of Variance, for which the alternative hypothesis does not take treatment ordering into account. null hypothesis, proc step, proc freq 10
46 The SAS/IntrNet Application Dispatcher Edmund Burnette The Application Dispatcher component of SAS/INTRNET Software allows a user to run a specified SAS Application program on demand and return the generated Internet Content to a Web Browser. The Application Dispatcher is one of the compute services of SAS/INTRNET software and is composed of the following components: 1. The Application Broker CGI The link between the Web Browser and the compute services. Resides on the Web Server and forwards requests for processing to the Application Server(s). 2. The Application Server An SCL program that listens for requests from the Dispatcher on a defined TCP/IP socket. The Application Server runs the specified SAS Application. 3. SAS Applications SAS programs that perform specified processing. They generate the results (Internet Content) which are sent back to the Web Browser by the Application Server via the Dispatcher. SAS/INTRNET software includes several Dispatcher Applications. The majority are customer supplied macro variable, macro definition 9
47 The Standard Programs Management System (SPMS): A Utility to Manage Report Generation within a SAS/AF Application Fern Grillo A SAS/AF/SCL application was developed to process bioavailability (Phase I) studies. This application, known as the Biostatistical Analysis Reporting Tool, (BART), handles the analysis, tabulation, listing and graphing of demographic, safety and blood data. Later, the BART system was modified to produce standardized tables and listings for demographic and safety data from clinical studies. The most recent addition to BART is the Standard Programs Management System. This component documents and categorizes standard program entries, and provides an audit trail of program modifications. Reports generated by the management system assist statisticians in the determination of programming needs for study report generation. A tracking system is provided to verify that study report data listings contain all pertinent case report form data. The system specifications and requirements are discussed, as well as the system's importance as a planning and management tool clinical trial, adverse events 2
48 The Version 7 Output Delivery System SAS Institute (a company) This paper discusses advances made in the output generated by all SAS Software procedures. New features provided by the Output Delivery System [ODS] include HTML generation, output datasets from any printed output and a persistable output document that can be re-organized and re-played long after the PROC that created it has completed. ODS also produces all legacy line-mode printing. The talk establishes the base line ODS features to be delivered in Version 7, as well as a roadmap to future enhancements. History SAS Output has been criticized for its Green-Bar characteristics. [Green-Bar being a generic term for "traditional" computer paper mounted in a chain printer - perforations along the sides, 66 lines deep, 132 characters wide, horizontal green stripes to aid readability ... The stuff legends are made of. ] Until Version 7, SAS procedures that produce printed output (i.e, destined for the SAS Listing File) have had this "page of paper" as their model when generating output output delivery system 10
49 Using SAS/AF as a Front End for Developing Report Program Skeletons Jeff Hamilton,
Jennifer Lester		 With the growing demands being placed on information system departments, the opportunity to offload "simple" programming tasks to non-SAS programmers adds efficiencies into the pro gram development life cycle. Using a SAS/AF application as a front-end entry tool can build such efficiencies into program de velopment. Standardized reporting tools developed at STAT PROBE, Inc., have been designed to allow the programmer to make table shell modifications very easily. These modifications have been automated via SAS/AF allowing a non-SAS program mer to create table shells with minimal involvement of the pro gramming staff. Standard report program "skeletons' are created by the user through AF Frames. Essential program modules are then combined to create usable SAS code. The SAS program created using SAS/AF produces a table shell which can later be modified by a SAS programmer to populate the table shell sas programmer, study day, macro definition 4
50 Visual Component Analysis: Interactive Graphical Modeling in SAS and ViSta/XLispStat Richard A. Faldowski As desktop computing machinery has become faster, more powerful, and easier to use, traditional limitations on desktop sta tistical visualization system capacities are being lifted. Recent innovations provide users with capacities for not only interactively viewing high-dimensional data and models, but also with capabili ties for dynamically modifying a statistical model in conjunction with what is observed. Visual component analysis, as presented by Faldowski (1995), is a user-guided graphical modeling system that supports progressive refinement of principal component mod els. Written within the ViStaJXLispStat programming environment (Young & Lubinsky, 1995), features of the visual component analy sis system will be compared to features provided by the most closely analogous SAS products, SAS/INSIGHT and JMP. Potential directions for the development of future graphically-ori ented SAS products will be offered proc princomp, estimate, proc sgplot, data science, correlation 13
51 Year 2000 Hex Storage Kirk Paul Lafler,
Sunil Kumar Gupta,
Charles E. Shipp		 Legacy systems that store, display or print dates in two-character representation for years ('97, '98, '99, ... ) will have an ambiguity problem when any of the year values go beyond the year 1999. Serious time-span calculations and other problems will then occur. The SAS option YEARCUTOFF provides a user-selected 100-year window and is good for many situations. Another solution is to read and write the storage using hex to retain the old record layout. Two hexadecimal bytes can represent year values from 0 to over 65,000, removing the 100-year limit. This method has the advantage of not needing to redesign old file structures, and the added advantage that other languages can also read the files. Several other options are presented flat file, dataset 3
52 Year 2000 Solutions Kirk Paul Lafler,
Sunil Kumar Gupta,
Charles E. Shipp		 SAS Institute has provided an excellent system/program option, YEARCUTOFF=19yy, but there are several other options for SAS programmers available also. This poster discusses several of these options. Companies will need to examine the role of their major systems that have dates and select the solution appropriate for each SAS program/system. Steps for a planned transition to the year 2000 are also presented 4

Back to Conf go 1998

This site serves as a convenient hosting location and does not work as a publisher. All copyrights belong to their respective owners.

The information on this site is provided for general reference purposes only. Whilst every effort is made to ensure that the site is up to date and accurate, no warrant, nor any responsibility or liability for, the accuracy or completeness of the content or for any loss which may arise from reliance on information contained in this site is provided. Links to other sites are provided for your convenience but no responsibility or liability for the content of those sites or of any external site which links to this site is accepted. References to any specific commercial product, process, service, or corporation are also for information only, and do not constitute endorsement, recommendation or favoring.

SAS is a registered trademark of SAS Institute Inc. SAS, Statistical Analysis System, 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.