Data conversions are not fun, no doubt about that. But if you're lucky, you can usually learn a thing or two from them. This effort was no different. It was not a lot of fun but a lot was learned from it, including some pretty neat PROC SQL and macro tricks. Credit must be given to SAS Tech Support for providing the basis from which this code was developed.

The goal of this conversion was to take safety data from a clinical data management system whose data is stored in Oracle tables residing on a VAX and migrate it to an Adverse Event tracking system built on a SQL/DS database residing on an IBM. Once the data was extracted from the Oracle database and exported to the IBM, SAS was used to perform the data conversion and insert the data into the adverse event (SQL/DS) database. This was accomplished with the use of PROC SQL statements contained in macros. The macros generated and executed program code that contained the SQL code used to insert, update and delete data to/from the database. The SQL code checked status codes and performed rollbacks when errors were encountered during the database update steps.

All of the code was written in SAS 6.07 in the CMS environment and was developed using the Display Manager interface. The PROC SQL Pass-Through facility was used to directly access the SQL/DS database tables.

Paul Doucette is an IS professional with over 10 years experience in systems design and development, project management and technical training and support. Paul has worked with SAS for over 7 years in the insurance and pharmaceutical industries and spent 3 years as a consultant to the pharmaceutical industry. Most recently, Paul joined PAREXEL International as the Manager of the Biostatistics Programming group. The group is responsible for SAS programming for the statistical analysis of clinical data.

Pharmaceutical companies do much Base SAS and Macro Language programming in the process of analyzing their clinical trials data. This is one part in their larger process of collecting, managing, analyzing, and presenting this data. Many companies now provide considerable system support for the early part of this process (data collection and management) and for the later part (document publishing and presentation of results). But less has been done to support the analysis part in between; i.e., the process of developing and executing SAS programs for statistical analysis, and then writing and reviewing an analysis document based upon the results of program execution. This "Program-Analyze-Write-Review" process is a major, time-consuming process in any research organization, and I address automated support for it in this paper. After a top-level discussion of the full Program-Analyze-Write-Review process, I focus in more detail on support for the SAS program development component within this process.

Barry R. Cohen is an information systems consultant and President of Planning Data Systems, Inc., a consulting firm in Ardmore, Pennsylvania. He has 20 years of programming and systems development experience, including 17 years with SAS Software.

Mr. Cohen provides services to a broad range of industries, including a focus on the pharmaceutical industry. His most recent experiences have involved the design and performance testing of SAS-based client/server configurations for analytic processing, general support for SAS processing on the UNIX platform, and design of an Integrated Development Environment for SAS program development.

Mr. Cohen is a co-founder and the President of PhilaSUG, the Philadelphia Area SAS Users Group. He is an accomplished author and invited speaker at SAS User Group and other industry conferences, and also chairs SAS conference sections.

When measuring several response variables, multivariate statistical techniques, such as multivariate analysis of variance, are often more powerful in detecting differences among populations than traditional univariate techniques. The increased power of multivariate techniques is achieved by utilizing the correlation among the various response variables measured on a single experimentation unit. In recent years, a number of graphical techniques and computer software packages have been developed for viewing multivariate data through computer monitors. These techniques use various combinations of color, shape and movement to display data, attempting to describe the multidimensional relationship of the response data in two dimensions. However, these techniques do not always transfer easily to the printed page for use in reports or research documents.

This paper is a review of two traditional graphical techniques, the profile and Andrews plots, which have proven to be very useful for displaying multidimensional data. The pinion plot is introduced as an alternative 2-dimensional graphical technique for displaying multivariate data. The pinion plot is compared to the profile and Andrews plots for describing differences among populations and as a graphical tool for detecting multivariate outliers.

James Schwenke is the Associate Director of the Biostatistics Department for Covance Periapproval Services Inc. located in Radnor, Pennsylvania. He has a Ph.D. in statistics from Kansas State University. His dissertation research involved the development of statistical tests using hold-out data sets for model validation.

In his current position at Covance, Dr. Schwenke is responsible for all statistical activities, including design, analysis, quality assurance and writing reports and manuscripts of Phase IIIb and IV clinical trials. In addition, he consults regularly with data managers, clinicians and business development specialists, and he is responsible for providing statistical training to members of clinical and data management staffs.

Dr. Schwenke has published 31 articles on statistical issues in a wide variety of scholarly journals, and he has presented 24 papers at national conferences. This is his first presentation before a SAS User's group.

Edit check is one of critical steps in clinical trial data processing. It has a major impact on the overall quality of clinical studies. In the past, some SAS based edit check systems only utilize SAS for limited computing service in a central or local environments. With newly introduced SAS/IntrNet, developers are able to add document management functionality within SAS, and deploy comprehensive systems on an intranet setting. This paper illustrates a working prototype of edit check system, which is able to generate standardized edit check queries, track their resolutions and monitor the overall progress of data cleansing process.

The prototype is essentially a multi-level SAS/IntrNet application. Some technique issues are addressed in designing this kind of applications. Furthermore, the author will discuss the likely impact when such a system is built upon a standardized / modularized CRF design.

James Sun is a Sr. System Analyst with Constat Systems Corp. He has worked as a statistician or SAS programmer in pharmaceutical industry. With over 8 years SAS experiences, he has developed several SAS based applications for data processing and clinical trial project management. His current interest is in designing and implementing process control / document management application with SAS/IntrNet for clinical trial data processing. He is an executive committee member of New Jersey SAS User Group. He has presented many papers at SUGI, NESUG, PharmaSUG, and local user groups. Two of his papers received awards at SUGI'98 and PharmaSUG'98.

• Modeling Dynamic Behavior with Object Interaction Diagrams
  Perry Watts, IMS Health
  Il-Yeol Song, Drexel University
• A Color Scheme for SCL Programs
  Rick Aster
  Breakfast Communications
• Using AUTOEXEC.SAS to Customize SAS Sessions
  Jennifer Price