View abstracts below by section:
- Applications Development
- Coders Corner
- Data Management & Validation
- Management and Career Development
- Programming Techniques
- Statistics & Pharmacokinetics
|Paper No.||Author(s)||Paper Title (click for abstract)|
|AD01||Peter Eberhardt||PROC FCMP: I Cannot Function Without It.|
|AD02||Stanley Wei||A Visualization Tool Developed with Visual Basic .Net for Global SAS Macros Management|
|AD03||Xiaofeng Shi||Customize your SAS programming toolbar|
& Jianyong Tong
& Huadan Li
& Xin Deng
|One Click to Generate the PK Analysis Outputs in Clinpharm Study|
|AD05||Zhouming(Victor) Sun||From Coal Mining to Data Mining: Advancing Programming Management for Clinical Projects with Text Analytics|
|Paper No.||Author(s)||Paper Title (click for abstract)|
|CD02||Haishan Kadeerbai||Brief Introduction of Oncology Domains in SDTMIG,Version 3.2|
& Changhong Shi
|Design and Construct Efficacy Analysis Datasets in Later Phase Oncology Study|
|CD04||Kyle Chang||An implementation of XSL-FO techniques to convert define.pdf from define.xml|
|CD05||Vijayalakshmi Gajulapalli||Hoping to do an FDA submission with CDISC compliant? Join me for a Mock Drill !!!|
|CD06||Victor WU||Easy ADaM implementation: Case Study|
|CD07||Junjie Shang||Implementation in the CDSIC Tumor Domains|
|CD08||Walter Hufford||Automating Production of the blankcrf.pdf|
& Changhong Shi
|How to meet one new FDA requirement - OSI (Office of Scientific Investigation)|
|CD10||Yu Zhu||Submission of Pharmacokinetics (PK) Data in a CDISC Compliant Format|
|CD11||Matt Becker||Data Transparency and Sharing: Research Benefits, Risks and the Future|
|Paper No.||Author(s)||Paper Title (click for abstract)|
|CC01||Sanjay Matange||Annotate your SGPLOT Graphs|
& Amarnath Vijayarangan
|Marco free automation to identify and drop variables with all values as missing|
|CC03||Mijun Hu||use self-defined function to indent long AE terms and dynamically repeat SOC in each page|
|CC05||Faruk Basha Mulla||Creating Dynamic Template using GTL|
|CC06||Kriss Harris||I Am Legend|
|CC07||Debpriya Sarkar||Plotting Against Cancer: Creating Oncology Plots Using SAS®|
& Amarnath Vijayarangan
|Automated Log Analyzer Dashboard|
|CC09||Mina Chen||Tips and Tricks about SAS Data Migration among Operating Platforms|
& Shuo Chi
|Using SAS SG Procedures to Create and Enhance Figures in Pharmaceutical Industry|
& Quan Zhou
|Make Sure Your Log Is Clean|
|CC12||Kartik Rajan||PROGRAMMING FIGURES: BEYOND SGPLOT AND GTL|
Data Management & Validation
|Paper No.||Author(s)||Paper Title (click for abstract)|
& Xue Yao
|I Object: SAS® Does Objects with DS2|
|DM02||Matt Becker||Visualizing Clinical Trial Data|
|DM03||Paul Ngai||Disclosure Requirements for Public Registries|
Management and Career Development
|Paper No.||Author(s)||Paper Title (click for abstract)|
|MC02||LIXIANG YAO||My Ideas on Clinical SAS Programmer's Training Program|
|MC03||Yong Zhao||A SAS Programmer's Dream|
& Tracy Turschman
|The Fourth Lie - False Resumes|
|MC05||Jianyong Tong||Make Innovative Progress: Let's Establish Our Own Standard Statistical SAS Macros|
& Jasper Jiang
|Building a Partnership: Constructing a global team between a US Sponsor and a Chinese CRO|
|MC07||Yuqin (Alice) Liao
& Huadan Li
|Follow Me, Then You'll Never Be Freshman Again!|
|Paper No.||Author(s)||Paper Title (click for abstract)|
|PO08||Amarnath Vijayarangan||SAS2VBA2SAS: Automated solution to string truncation in PROC IMPORT excel|
|PO10||Weiquan Xin||To Open SAS Datasets in SAS System Viewer while Compiling the Coding in Enhanced Editor|
|PO12||Yangyang Li||Output flexible and customized Excel spread sheets from SAS data sets using Excel XP Tagset|
& Palanisamy Mohan
& Amarnath Vijayarangan
|Alert when one is not Alert: SAS Call Sound Function|
|Paper No.||Author(s)||Paper Title (click for abstract)|
|PT01||Arthur Li||Essentials of PDV: Directing the Aim to Understanding the DATA Step!|
|PT02||CHAO WANG||The Power of SAS National Language Support - Embrace Multilingual Data|
|PT03||Gaoyamg LI||SAS XML Mapper: A bridge connecting PDF Comments, MS excel files, and SAS datasets|
|PT04||Sanjay Matange||Quick Graphs with ODS Graphics Designer|
|PT05||Kriss Harris||Napoleon Plot|
|PT06||Weston Chen||Hash table and its use in the big data analysis of healthcare claims database|
|PT07||Rajesh Babu Moorakonda
& Mihir Gandhi
|Monitoring Child Growth and Safety Profile using Growth Charts|
|PT09||Lianbo Zhang||A simple way to access the data in EXCEL through SAS v9/ ACCESS® libname and Excel engine|
|PT11||Haiqiang Luo||Collapsing Adverse Event Records|
Statistics & Pharmacokinetics
|Paper No.||Author(s)||Paper Title (click for abstract)|
|SP01||Vicky Pan||Continual Reassessment Method (CRM) in Dose-finding Trials|
& Huadan Li
|Proportions: Stop Worrying and Start Calculating|
|SP03||Ka Chun Chong
& Chung Ying Zee
|Compartmental Models in SAS: Application to Model Epidemics|
& Ju Chen
|Bayesian Augmented Control (BAC) Application by using 'Proc iml'|
& Peng Wan
|Analysis of Multiple Imputation in Conjunction with Robust Regression in Clinical Trials Using SAS|
|SP08||Yan Qiao||Forecast number of events in oncology trials: parametric method and non-parametric method|
& Helen Wu
|Extend SAS by R|
|SP10||Yuan Zhuang||Early Development Stage (EDS) study---A wonderful start for junior SAS programmer|
|SP11||Halimu Haridona||Calculation of Relative Risk and Odds Ratio Using StatXact PROCs|
Applications DevelopmentAD01 : PROC FCMP: I Cannot Function Without It.
Friday, 2:45 PM - 3:20 PM, Location: Executive Ballroom B
How many times have you tried to simplify your code with LINK/RETURN statements? How much grief have you put yourself through trying to create macro functions to encapsulate business logic? How many times have you uttered "If only I could call this DATA Step as a function"? If any of these statements describe you, then the new features of PROC FCMP are for you. If none of these statements describe you, then you really need the new features of PROC FCMP. This paper will get you started with everything you need to write, test, and distribute your own "data step" functions with the PROC FCMP. This paper is intended for beginner to intermediate programmers, although anyone wanting to learn about PROC FCMP can benefit.
AD02 : A Visualization Tool Developed with Visual Basic .Net for Global SAS Macros Management
Friday, 1:30 PM - 1:50 PM, Location: Executive Ballroom B
To facilitate programming activities and to achieve higher efficiency as well as quality, plenty of global SAS macros are often developed, which has been one of the must-haves for a mature organization and is also one of the key components for process optimization. However, how to efficiently manage/maintain those hundreds of macros, to help users identify the right macros and use them correctly, is usually not an easy job, especially for new comers. To address this problem, a visualization tool was developed accordingly with visual basic .NET, which has user-friendly interfaces and could be acted as a central platform for global macros management and learning system for SAS programmers. This presentation will demonstrate the typical usages of this tool and how it helps users to make full use of global SAS macros in your organizations.
AD03 : Customize your SAS programming toolbar
Friday, 1:50 PM - 2:10 PM, Location: Executive Ballroom B
In pharmaceutical companies, statistical programmers should make their SAS programs comply with GPP (Good Programming Practices) and other SOPs regarding statistical programming and validation. This paper briefly introduces how to create a SAS toolbar that facilitates program compliance to GPP and SOPs. Using the toolbar buttons, it's easy to create a program template including a standard program header, main sections of a program and other program codes standardized in each company. And then, this paper will extend to discuss how to customize your own toolbar buttons to make the daily work easier.
AD04 : One Click to Generate the PK Analysis Outputs in Clinpharm Study
Friday, 2:10 PM - 2:30 PM, Location: Executive Ballroom B
ABSTRACT In order to improve the delivery quality, increase productivity, and empower better decision making in PK studies we initiate a program called "Standard Statistical SAS Macros for the Analysis of Pharmacokinetic Parameters in EDS PK studies". Through this initiative, we aimed to develop three kinds of tools: 1) Metadata-based centralized macro programs; 2) Programs templates that depend on the study type and use the centralized macros; 3) ADaM templates for generating the ADPP and ADPC. The most important part is centralized macro and it will cover both table and graph. The analysis methods and related output will include 1) any linear fixed/mixed effects model on any log-transformed parameters and its outputs; 2) the standard power model for dose proportionality analysis in a parallel groups (cross-over)design and its outputs 3) the standard non-linear mixed effects model for steady-state analysis in a multiple dose study and its outputs; 4) the standard linear model for steady-state analysis in a multiple dose study and its outputs; 5) computes descriptive statistics on any list of parameter and provides the related summary table; 6)Plot the ratio plot for the ratio of GMR. Key word: Metadata-based centralized macro, PK studies, ADaM
AD05 : From Coal Mining to Data Mining: Advancing Programming Management for Clinical Projects with Text Analytics
Friday, 2:30 PM - 2:45 PM, Location: Executive Ballroom B
ABSTRACT Both coal mining and data mining involve the process of extracting valuable materials from raw resources to form economically useful packages. Due to the theoretical similarities behind these two different types of mining, the process of coal mining, as will be described, serves as inspiration for the development of the Simultaneous Monitoring of Analysis and Reporting Toolkit (SMART), a method for data mining based on the techniques of text analytics using Base SAS@ 9+. Established with a general SAS macro, SMART is a versatile toolkit that allows for the reporting of the real-time status of programming activities. By using the techniques of text analytics to find explicit relationships between documents by classifying documents into predefined or data-driven categories, SMART makes management of clinical project programming more effective and dynamic. This paper will introduce the concepts of SMART, followed by a presentation of its four key processes. It will additionally demonstrate the power of text analytics in extracting useful information while providing a helpful roadmap for project leaders to efficiently manage programming activities independently of a project leader's programming skill or experience.
CDISCCD02 : Brief Introduction of Oncology Domains in SDTMIG,Version 3.2
Friday, 1:50 PM - 2:10 PM, Location: Executive Ballroom B Salon
The final document of CDISC Study Data Tabulation Model Implementation Guide (SDTMIG) v3.2 was released in the end of 2013. All of previous annotations in SDTMIG 3.1.3, originally published in 2012, have been incorporated in SDTMIG v3.2 that includes many enhancements and improvements from earlier versions. This paper presents a brief introduction on the newly added oncology domains (TU, TR and RS) in Findings Observation Class, which are extremely useful in oncology clinical trials, mainly focused on: 1) the purpose and usage of the domains; 2) relationship among the domains; 3) the domains' cross-data (domains) linkability /traceability and relationship with other SDTM domains, such as DD (newly added), DM, SC, and with some ADaM datasets like ADTTE, etc.
CD03 : Design and Construct Efficacy Analysis Datasets in Later Phase Oncology Study
Friday, 2:10 PM - 2:30 PM, Location: Executive Ballroom B Salon
Under the CDISC frames, The Therapeutic Area Standards (TA Standards) have already been the hotspot. The CFAST TA Standards Program was launched to accelerate clinical research and medical product development by facilitating the establishment and maintenance of data standards, tools and methods for conducting research in therapeutic areas important to public health. The Oncology TA is the pioneer of this effort. The SDTM Oncology domain models for Tumor Identification (TU), Tumor Results (TR) and Disease Response (RS) are available in SDTMIG v3.1.3. However, the Oncology TA ADaM has not achieved any standardization. This paper will demonstrate how to design and construct standard efficacy analysis datasets in later phase Oncology studies according to the SDTM Oncology domain models.
CD04 : An implementation of XSL-FO techniques to convert define.pdf from define.xml
Friday, 2:30 PM - 2:45 PM, Location: Executive Ballroom B Salon
It's getting more common and important for the electronic submissions of clinical data to include define.xml, which specifies the CDISC standard for providing case report tabulation (CRT) data definitions in XML format. Although define.xml has proven to be a useful mechanism can easily help regulatory reviewers to navigate transmission of CRT metadata, it is not able to provide original features (e.g., hyperlinks, bookmarks&etc.) while they're printed out. One solution is to generate a printable define.pdf document with the same content as the define.xml. The printable define.pdf file accommodates bookmarks and hyperlinks functionality for online review, and it can also be printed out for hardcopy review. Providing define.pdf documents is not only to have an easy way to review clinical data submission package, but also to fulfill submission requirements for sponsor and regulatory authority. This paper provides an approach that using Extensible Stylesheet Language Transformations (XSLT) template converting "define.xml" to formatting objects for easy transformation to "define.pdf".
CD05 : Hoping to do an FDA submission with CDISC compliant? Join me for a Mock Drill !!!
Friday, 2:45 PM - 3:20 PM, Location: Executive Ballroom B Salon
In today's modest environment the reduction of the time taken to reach the market is critical to a drug and hence the company's success. The proper approach of its Regulatory Affairs activities is therefore of considerable importance for the company. Inadequate reporting of data or in a form that the regulators can not easily review may prevent a timely positive evaluation of a marketing application. A new drug may have cost many millions of dollars, pounds or Euros to develop and even a two-month delay in bringing it to the market has considerable financial considerations.
CD06 : Easy ADaM implementation: Case Study
Friday, 3:40 PM - 4:10 PM, Location: Executive Ballroom B Salon
With the popularization of CDISC, more and more interests are shown on how to implement CDISC standards to create CDISC-Conformed datasets. Here I would like to show how to create ADaM datasets step by step. Proposed solutions to most frequently encountered issues will be discussed, such as assigning attributes to dataset/variables, selecting baseline, and deriving new variables vs deriving new records; and several related macros will be introduced to demonstrate how to implement easily.
CD07 : Implementation in the CDSIC Tumor Domains
Friday, 4:10 PM - 4:30 PM, Location: Executive Ballroom B Salon
Assessment of the change in tumor burden is an important feature of the clinical evaluation of cancer therapeutics: both objective response and disease progression are useful endpoints in cancer clinical trials and in most studies the tumor burden is evaluated using the RECIST criteria. This paper will introduce how RECIST 1.1 data are streamlined in CDSIC since there comes up with 3 new SDTM tumor domains - TU (Tumor Identification), TR (Tumor Response) and RS (Disease Response). All tumor domains are finding structure and have a distinct purpose respectively. Moreover, they can be linked using additional linking variables and the connection can be established through RELREC domain. This paper will introduce how to implement these tumor domains and their relationship.
CD08 : Automating Production of the blankcrf.pdf
Friday, 4:30 PM - 4:50 PM, Location: Executive Ballroom B Salon
The annotated blank Case Report Form (blankcrf.pdf) is a critical component of the NDA submission. Per FDA guidance, source data domain, variable name and controlled terminology for each case report form (CRF) item included in the tabulation datasets submitted should be displayed on the blankcrf.pdf. Production of the blankcrf.pdf is a tedious, non-programming task that is increasingly becoming the responsibility of the statistical programmer. This paper describes an easy to use, automated method of annotating the CRF.
CD09 : How to meet one new FDA requirement - OSI (Office of Scientific Investigation)
Friday, 4:50 PM - 5:10 PM, Location: Executive Ballroom B Salon
Previously known as Division of Scientific Investigation (DSI), the Office of Scientific Investigation (OSI) is now the new FDA requirement for the NDAs consisted of the completed Phase II and Phase III clinical trials. It is administrated under the Office of Compliance, in the Center for Drug Evaluation and Research (CDER). The OSI verifies the integrity of efficacy and safety data and assures that the rights and welfare of clinical participants are protected by conducting risk based site inspections, and also facilitates the FDA's decision on the timely selection of the appropriate sites. Currently, the OSI request only affects submissions to CDER with respect to clinical data. In this paper, the 3-part components of OSI will be introduced and an implementation example will be discussed per a Merck project.
CD10 : Submission of Pharmacokinetics (PK) Data in a CDISC Compliant Format
Friday, 1:30 PM - 1:50 PM, Location: Executive Ballroom B Salon
The legacy Pharmacokinetic (PK) data are usually produced from different sources with different data format: sample collection in CRF, concentration result from Lab findings, and parameters from pharmacokineticist. This brings a lot challenge when preparing SDTM and ADaM datasets for submission. This paper introduces some of the challenges and solutions associated with mapping the concentration data and the calculation of PK parameters into SDTM (e.g. PC and PP) and subsequent creation of ADaM.
CD11 : Data Transparency and Sharing: Research Benefits, Risks and the Future
Saturday, 1:30 PM - 2:00 PM, Location: Executive Ballroom B
Whether called data transparency or data sharing, there is a movement to give more researchers greater access to patient-level clinical trial data. The goal is to create an environment for innovation in clinical research. Join this presentation to discuss what is being done, including exploring the value to the overall health care system of creating a multi-sponsor environment that gives researchers access to larger pools of data.
Coders CornerCC01 : Annotate your SGPLOT Graphs
Saturday, 8:30 AM - 9:00 AM, Location: Executive Ballroom B
The SG procedures provide you multiple plot statements to create many different kind of graphs. These plot statements can be used together in creative ways to build your graph. However, even with this ability to customize, there are times when you need more than what you can get using just the plot statements. You need a way to add custom information anywhere on the graph. With SAS 9.3, the SG procedures support the ability to annotate the graph using data set based information. This annotation functionality is designed in a way similar to the annotate facility available with the SAS/GRAPH procedures. There are a few differences and enhancements. If you already know annotation from SAS/GRAPH, or if you are new to it, this paper will show you how to add custom annotations to your graphs.
CC02 : Marco free automation to identify and drop variables with all values as missing
Saturday, 9:00 AM - 9:15 AM, Location: Executive Ballroom B
Missing values do have a vital role to play in handling data either numeric or character. Especially when a variable is completely missing it taxes the execution speed and also the space that the dataset occupies. However, if we are able to identify these missing variables even before the analysis or study and drop them it would significantly increase the execution speed and would also save a good impactful amount of space in the server. There are several solutions to this problem, but this paper introduces a macro free automation to identify and then drop the completely missing variables, as any use of macros or even a simple proc freq would run into a risk of either a complete failure or slowdown of the process due to insufficient macro length or numerous levels in the dataset. Various SAS procedures can be used to identify the variables with completely missing values and then an implementation of a data step or proc step will drop these missing variables. But this is a onetime solution for smaller datasets. In case of a repetitive task or larger datasets it is always efficient to execute it through an automated process. Here we propose an automated solution and to make it even more efficient and simple we do not use any macros in it thus making it completely a macro free automated process to identify and drop the variables with completely missing values
CC03 : use self-defined function to indent long AE terms and dynamically repeat SOC in each page
Saturday, 9:15 AM - 9:30 AM, Location: Executive Ballroom B
we have two common cosmetic problems in AE table to overcome, one is that preferred terms is too long to be present in one line so that it has to be wrapped to the next line, nevertheless the text-indent in the next line cannot be maintained. The other is that preferred terms pertaining to the same SOC display across multiple pages, which entails the need of repeating of SOC on top of each page. For the first problem, we already have a bunch of paper introducing macros to automatically indent wrapped AE terms and this time I want to accomplish this purpose by using a self-defined function because of several advantages. And using this function along with the aid of compute block in proc report, we can dynamically repeat SOC on top of each page regardless how many rows available to filled in each page.
CC05 : Creating Dynamic Template using GTL
Faruk Basha Mulla
Saturday, 9:45 AM - 10:00 AM, Location: Executive Ballroom B
GTL is a comprehensive language developed for capturing the definition of potentially very complex graphs, unlike traditional graphics. It helps in creating graphs with simple and concise syntax and also provides the means to modify the default Graphical templates provided by SAS for creating sophisticated and/or highly customized graphs. This presentation talks about the dynamic, special dynamic and macro variables. Dynamic variables give the flexibility to produce the similar graphs with different variables without redefining the template. Special dynamic variables are useful to produce the user defined graphs without hardcoding the variable names in GTL. The creation, declaration and use of macro variables in GTL to produce multiple graphs from a single program, and finally how to draw an inset as a table of text, positioning and creating an inset values with the Template. The dynamic template brings in the benefit of reusability that saves the programming time and increases program efficiency.
CC06 : I Am Legend
Saturday, 10:00 AM - 10:15 AM, Location: Executive Ballroom B
Have you ever produced a legend on a plot that was taking up too much space, hence making the actual graph too small? Have you ever removed a legend because it was taking up too much space? Have you ever wanted to just produce a legend? Have you ever wondered that there must be a more efficient way of producing a legend then using the exact same legend on every BY variable of your output? This paper will demonstrate solutions to the above problems using Graph Template Language (GTL) in SAS® 9.2, in particularly using the SERIES, VECTOR and SCATTERPLOT statements.
CC07 : Plotting Against Cancer: Creating Oncology Plots Using SAS®
Saturday, 10:30 AM - 11:00 AM, Location: Executive Ballroom B
Graphs in oncology studies are essential for getting more insight about the clinical data. This presentation demonstrates how ODS Graphics can be effectively and easily used to create graphs used in oncology studies. We discuss some examples and illustrate how to create plots like drug concentration versus time plots, waterfall charts, comparative survival plots, and other graphs using Graph Template Language and ODS Graphics procedures. These can be easily incorporated into a clinical report.
CC08 : Automated Log Analyzer Dashboard
Saturday, 11:00 AM - 11:15 AM, Location: Executive Ballroom B
The importance of validating a SAS program through the generated log file is inevitable. A successful execution would require an ERROR, WARNING and other system message free log. Though, the severity of NOTE or WARNING might not be very high, but there are chances for multiple NOTES or WARNINGS together in a program can cause severe problems or incorrect results equal to an error message. The SAS products in various domains generate very large and N number of log files when executed. For instance, Clinical Research domain demands 100 + SAS programs to be either executed in a batch mode or interactive mode for a final delivery, which is validated manually. This means a programmer needs to review 1000 + lines of code in the multiple logs manually where certain seemingly unimportant messages might be overlooked and also a manual review is really time consuming process. The proposed automated log analyzer will help in a 3600 degree review of the generated SAS Logs; a program is developed to ensure that no system generated message is overlooked. The automated log analyzer scans each and every single line of every log file in the directory for any system generated messages and provides a visual report: a dashboard using SAS. The dashboard provides an overall summary on various system generated messages by SAS logs or across SAS logs through quality charts which will allow the programmer to quicken the validation process. The dashboard generates three reports 1) System messages captured from each log 2) Frequency messages 3) a visual presentation. This ensures not only accuracy in the validation process but also a significant reduction in the time taken to validate the programs.
CC09 : Tips and Tricks about SAS Data Migration among Operating Platforms
Saturday, 11:15 AM - 11:30 AM, Location: Executive Ballroom B
In general, SAS data (datasets and catalogs) created on an operating system is not necessarily accessible after being transferred to a different operating system. SAS provides multiple approaches in helping to migrate and access SAS data among different operating environments. There are constraints in some of these methods. For example, Cross-Environment Data Access (CEDA) does not support SAS catalogs and indexes other than datasets. The purpose of this paper is to provide an overview of basic techniques to move SAS datasets and formats catalogs between operating environments so data remain accessible afterwards. The focus will be on situations when moving data with customized formats from Windows SAS9.2 to Unix SAS9.2. The concept introduced is also applicable in different scenarios. The pros and cons will be elaborated.
CC10 : Using SAS SG Procedures to Create and Enhance Figures in Pharmaceutical Industry
Saturday, 11:30 AM - 11:45 AM, Location: Executive Ballroom B
SAS/GRAPH statistical graphics (SG) procedures provide a simple syntax for creating graphics commonly used in exploratory data analysis and for creating customized statistical displays. This paper will focus on 1) how to use SAS SG Procedures to create pharmaceutical figures, 2) Introduce the advantages and enhancements of SAS SG Procedures, 3) New features in SAS SG 9.3 version to enhance both graphic visualization and printer-friendly result, 4) How to prepare statistics for better and flexible graphic visualization using SAS SG.
CC11 : Make Sure Your Log Is Clean
Saturday, 11:45 AM - 12:00 PM, Location: Executive Ballroom B
SAS log file provides lots of information about the execution of SAS programs. In clinical trial practice, it is usually unacceptable to leave error, warning, or certain special notes in the log. Programmers spend lots of time investigate the log file and fix the issues they might find. However, the messages in SAS log are not always straightforward for programmers to understand and fix. This paper would like to summarize the common issues in SAS log we face in our daily work, and propose solutions on how to fix them. The summary and solutions will be presented in the following order: issues that might be generated from SAS DATA steps; from SAS macros; and then from some SAS procedures. Examples will be provided.
CC12 : PROGRAMMING FIGURES: BEYOND SGPLOT AND GTL
Saturday, 12:00 PM - 12:15 PM, Location: Executive Ballroom B
Setting up programs for figures for CSRs or otherwise comes with its own set of unique challenges, for in addition to accurately presenting data, figures also need to be clean and understandable (more so than tables and listing). This means a programmer often may need to spend a significant amount of time tweaking code and experimenting with the display of a figure in order to get an output that's "just right". This need for customization comes at a cost of increased programming complexity (lots of macro parameters if the figure code is in a macro) or sometimes SAS doesn't directly support what you need to produce. In such cases one needs to go beyond what SGPLOT and GTL offer, for instance by re-building plots using what is on offer, or "tricking" SAS into doing what you need.
Data Management & ValidationDM01 : I Object: SAS® Does Objects with DS2
Saturday, 1:30 PM - 2:05 PM, Location: Executive Ballroom A
The DATA step has served SAS® programmers well over the years, and although it is powerful, it has not fundamentally changed. With DS2, SAS has introduced a significant alternative to the DATA step by introducing an object-oriented programming environment. In this paper, we share our experiences with getting started with DS2 and learning to use it to access, manage, and share data in a scalable, threaded, and standards-based way.
DM02 : Visualizing Clinical Trial Data
Saturday, 2:40 PM - 3:15 PM, Location: Executive Ballroom A
Today, all employees at health and life science corporations may need access to view operational data. There may be visualization needs for a business analyst to compare clinical trial spend versus similar past trials; for a clinical research associate to easily see what research sites may need more monitoring; for a data scientist to quickly and easily explore adverse event data for outliers; for a medical writer to have graphical patient profiles; for a CEO to immediately have high-level dashboards of operational performance. In this presentation we will discuss SAS Visual Analytics and demonstrate the benefits of visualization for health and life science organizations.
DM03 : Disclosure Requirements for Public Registries
Saturday, 2:05 PM - 2:40 PM, Location: Executive Ballroom A
Management and Career DevelopmentMC02 : My Ideas on Clinical SAS Programmer's Training Program
Friday, 4:10 PM - 4:40 PM, Location: Executive Ballroom B
Abstract: This paper is to introduce the training programs proved to the junior clinical SAS programmer in the first few months. Compared with pure SAS programmers in other industries, clinical SAS programmer should know a lot of additional knowledge rather than SAS to be a skilled one. And to provide clinical SAS programmers the trainings logically is quite important to let a junior one become a middle one in a short time. And we will introduce what skills can be provided to clinical SAS programmer and in what way we transfer the knowledge to the clinical programmer. It is assumed that the programmer to take this program has SAS knowledge.
MC03 : A SAS Programmer's Dream
Friday, 3:40 PM - 4:10 PM, Location: Executive Ballroom B
With the introduction of the latest Chinese government, a concept of China dream has been given a lot of publicity. It may sound very vague and remote, but I am sure every SAS programmer has his/her own dream in career development. Being a SAS programmer is a unique experience we all enjoy. We have our own culture and characteristics because we're both programmers and analysts. We also work in a very special pharmaceutical environment which has a lot of regulations and requires a great deal of quality work. In order to be able to climb in this kind of corporate ladder, it requires a lot of efforts from each individual. The focus of the presentation is to have a positive discussion on how to fulfill our dreams. It will include topics such as how to set up your goals, how to improve your technical and communication skills and how to get ready for your next position. I am going to use psychological theories and my personal experience as manager with a large group of direct reports to launch an open discussion on the topic we all care deeply about, how to fullfill our dreams. I'd propose some suggestions on areas we may be able to improve and to invite people to share their ideas.
MC04 : The Fourth Lie - False Resumes
Friday, 4:40 PM - 5:00 PM, Location: Executive Ballroom B
The fourth lie - false resumes. Twelve years ago we were surprised to see an increase in the number of resumes which exaggerated the sender's breadth of experience and work history. And now, twelve years later, we are still receiving resumes which are suspect. Unfortunately, not all false resumes are exposed and individuals are getting jobs that may endanger a clinical trial. Over time we've discovered that false resumes can be vetted before they hit the hiring managers desk. In this presentation, Tracy Turschman speaking for Ernest Pineda, President of the Gerard Group will discuss measures his company has taken to validate the accuracy of resumes using pre-planned questions, inexpensive background checks, and industry knowledge. He will share experiences, observations, policies and tools that have helped his firm expose under qualified, and falsely represented candidates.
MC05 : Make Innovative Progress: Let's Establish Our Own Standard Statistical SAS Macros
Friday, 5:00 PM - 5:20 PM, Location: Executive Ballroom B
The common model between the programming team of Global Pharmaceutical Company and their regional division is that the global standard macro team provides the standard macro and we use the macro. In order to make our own contribution to the company, to change the passive working model of the AP programming team and improve the efficiency of our work on PK studies in EDS (Early Development Statistics), we decided to initiate a project named "Standard Statistical SAS Macros for the Analysis of Pharmacokinetic Parameters in EDS PK studies", the team members of which include IT, EDS statistics and programmer. To make this project run smoothly, we took several necessary actions, such as specific task allocation, information and resource sharing, biweekly project meeting, consultation to the Global (programming wise and statistical wise), set up accurate and reasonable timeline, etc.. Up until now, we have accomplished ADaM templates for generating the ADPP and ADPC and 6 Metadata-based centralized macro programs and more will come in the future. Through this project we learned that it's important to have the spirit of innovation, it matters when you start to do something and you never know what you are capable of if you don't try.
MC06 : Building a Partnership: Constructing a global team between a US Sponsor and a Chinese CRO
Saturday, 11:30 AM - 12:00 PM, Location: Executive Ballroom A
There may be many reasons for a sponsor to decide to work with a CRO but no matter the reason both teams always want to ensure a successful collaboration. Careful planning prior to the study work beginning is an important step towards this goal in any situation, no matter where the companies are located, but when they are in different countries planning becomes imperative. Whose SOPs will be followed? How will issues be communicated? What are the responsibilities of the leads within each company? By discussing, and coming to agreement, on these types of questions before study work starts, the teams avoid slowdowns due to confusion, miscommunications and under realized expectations. To avoid these issues, the Santen and Rundo teams have worked together to construct a solid plan, prior to study work getting underway, to clearly set expectations, anticipate challenges and develop processes to mitigate identified risks. This paper is intended to describe the steps taken and the working model for the partnership that resulted from these discussions.
MC07 : Follow Me, Then You'll Never Be Freshman Again!
Yuqin (Alice) Liao
Friday, 5:20 PM - 5:30 PM, Location: Executive Ballroom B
Abstract For a new SAS programmer in pharmaceutical industry, how to train himself or herself to gain essential knowledge and to be an expert in this field efficiently in a short time? In this paper, I will lead you to catch a big picture of how to improve yourself to get rid from being a freshman in clinical trial and related SAS skills by sharing with my two years experiences, useful guidance and resources. Follow me and I'll help you to figure out a clear and effective clue about how to move forward to next level from a freshman stage. Key Words: Pharmaceutical industry, Clinical trial, SAS
PostersPO08 : SAS2VBA2SAS: Automated solution to string truncation in PROC IMPORT excel
SAS Proc import is one of the greatest approaches for converting other software's data files into SAS datasets. While importing the files string truncation of a character variable is a common pitfall to every SAS users. Most of the times, values of a character variables are having dissimilar length such as product names, descriptions & retailer address. In this case, Proc import may not be able to retain the values as it is since it fixes the length attribute using GuessingRows as 20. Alternatively data step with infile statement can be used which requires significant effort even for a smaller file and is error prone. Our proposed approach is making use of SAS & VBA to resolve the length issue of a character variable. Fortunately, Proc import generates the data step program in the log. VBA reads the log for character variables and then reads in excel file to get maximum length for each variable from excel by scanning length of each values for every variable. The data step program will be modified with this maximum length for the character variables using VBA. This new modified data step program then will be executed to import the excel file without any truncation of character variable. This approach also does the importing of several files at a time.
PO10 : To Open SAS Datasets in SAS System Viewer while Compiling the Coding in Enhanced Editor
Oftentimes, SAS programmers open the newly created SAS datasets for a check purpose while they are writing codes in SAS enhanced editor. To do this, they activate the SAS explorer, open the SAS library, locate the SAS dataset, and double click the dataset to open it. Then, VIEWTABLE is used to open the dataset by default. These steps are rather time-consuming and VIEWTABLE is not so user-friendly. Therefore, in the article, we discuss two alternative methods. One convenient method is presented to relieve this issue. While compiling the coding in the enhanced editor, you only need to select the dataset name, and press two shortcut keys, and then the dataset will be opened in SAS system viewer. The other one, to open SAS dataset in the SAS system viewer from SAS explorer, is also briefly discussed.
PO12 : Output flexible and customized Excel spread sheets from SAS data sets using Excel XP Tagset
In HE&OR studies, customers often not only want to view the report, but also be able to 'manipulate' the data, for example, to sort, filter, add up, and etc. And Excel is a popular tool that most customers can easily handle with. However, while many programmers are familiar with ODS RTF, PDF, HTML, they may find dealing with Excel format is not that an easy stuff. As Excel tool applies unique language rules, the interactive mechanism between SAS and Excel appears more complicated than others. This presentation demonstrates how to generate flexible and customized Excel spread sheets from SAS data sets using Excel XP Tagset. Using standard HE&OR deliverables as output examples, we will discuss how Excel XP Tagset works on customized headers, titles and footnotes, background, colors, font styles, frames, rules, cell borders, alternating rows and columns, and etc.
PO13 : Alert when one is not Alert: SAS Call Sound Function
Multitasking is the order of the day and many SAS programmers work on multiple tasks simultaneously. For long running programs, the programmer assumes that this particular program might take certain amount of time, but in reality the program might get completed earlier or later than the expected time based on the load on server. The only way a programmer can check the completion of the program is by checking the log intermittently or by setting up an email alert, which can also be missed when the person is working offline. A prompt notification after the execution of the program can help and save lot of time for the programmer. SAS can alert the programmer through CALL SOUND function which can play music after a DATA step or a PROC step or program completion. The advantage of an alert function is that 1. Programmer is notified even when he/she is not glued with SAS environment and is held up with other tasks. 2. Sometimes when there are multiple users working on a project, the programmer sitting adjacent to the current programmer can also hear the alert and can act upon it and submit next task assigned to him when the actual programmer is not around 3. When combined with Scheduler this can also serve the purpose of an Alarm The purpose of this paper is to leverage the functionality of Call Sound function which would help in effective time and resource management
Programming TechniquesPT01 : Essentials of PDV: Directing the Aim to Understanding the DATA Step!
Friday, 1:30 PM - 2:30 PM, Location: Executive Ballroom A
Beginning programmers often tend to focus on learning syntax without understanding how SAS® processes data during the compilation and execution phases. SAS creates a new data set, one observation at a time, from the program data vector (PDV). Understanding how and why each of the automatic or user-defined variables is initialized and retained in the PDV is essential for writing an accurate program. Among these variables, the following variables deserve special attention, including variables that are created in the DATA step, by using the RETAIN or the SUM statement, and via by-group processing (FIRST.VARIABLE and LAST.VARIABLE). In this paper, you will be exposed to what happens in the PDV and how these variables are retained from various applications.
PT02 : The Power of SAS National Language Support - Embrace Multilingual Data
Friday, 2:30 PM - 3:00 PM, Location: Executive Ballroom A
As clinical trials take place around the globe, data in different languages may need to be handled. The usual solution for this issue is to translate the data to English, after some manipulations, results then will be translated back to the original language. Since version 9.1.2, SAS software can provide a specific function called National Language Support, this issue would be easy to handle. In this paper, an introduction will be made on: 1) how to build the ready SAS session to import and display the mixed data with NLS system option 2) how to use the edge tool - SAS K-Function (including K-macro Function) to do data manipulations and 3) how to display any character in reports with the Unicode support. Here would use a multilingual (Chinese and English) AE dataset as an example on SAS 9.3 PC version. The related applications on the UNIX server and SAS enterprise guide would also be discussed.
PT03 : SAS XML Mapper: A bridge connecting PDF Comments, MS excel files, and SAS datasets
Friday, 4:55 PM - 5:10 PM, Location: Executive Ballroom A
For almost every Clinical SAS programmers, Adobe PDF documents, MS excel files and SAS dataset are the top 3 documents types when conducting data analysis. First, we need review CRF or aCRF, which typically is of PDF documents. Sometimes, they are of RTF or WORD documents. Then, several working documents would be created in MS excel documents, such as domain summary, programming specifications, and external data not from database. And Third step also the most import step is analysis data in SAS datasets. These 3 types' documents are interactive documents, and usually most people deal with them separately or manually. SAS XML Mapper provides a bridge to connect them even if they are not using a standard XML formats. This make a automatic or programmed way happen when. In this paper, XML file, PDF comments as XML Forms Data Formatted (XFDF) file, SAS XML Mapper are esplained. A real example was presented on how to import and parse annotations contained in a blank aCRF to create domain summary and fields details by SAS program with the help of XML Mapper. This example would inspire readers to create other documents in the same way.
PT04 : Quick Graphs with ODS Graphics Designer
Friday, 3:40 PM - 4:10 PM, Location: Executive Ballroom A
You just got the results of the study and you want to get some quick graphical views of the data before you begin the analysis. Do you need a crash course in the SG procedures just to get a simple histogram? What to do? The ODS Graphics Designer is the answer. With this application you can create many graphs including histograms, scatter plots, scatter plot matrices, classification panels and more using an interactive 'drag-and-drop' process. You can render your graph in batch with new data and output the results to any open destination. You can view the generated GTL code as a leg up to GTL programming. You can do all this without cracking the book or breaking a sweat.
PT05 : Napoleon Plot
Friday, 4:10 PM - 4:40 PM, Location: Executive Ballroom A
Do you want to produce a very useful plot? Okay, do you want to produce a plot that for each subject shows the number of treatment cycles, the number of days on treatment, the doses that were received, whether the subject has discontinued treatment, and the cohort the subject is in? This paper will demonstrate how to do the above in SAS® 9.4.
PT06 : Hash table and its use in the big data analysis of healthcare claims database
Friday, 4:40 PM - 4:55 PM, Location: Executive Ballroom A
With the increasing needs on the analysis of big data-enormous amounts of information collected in the healthcare information system, how to develop more efficient codes has become one of the key skills SAS users have to master. SAS itself provides multiple ways for efficiency considerations. Hash table is one of them. This paper will show you how to create a hash object in data steps and some useful techniques to address big data match-merge and/or join operations. In addition, some performance comparisons will be made between typical data lookup techniques and hash table with the use of real-world big healthcare data to show the impressive efficiency improvement by hash object. Some additional practical tips about creating Hash objects for efficient programming would be introduced as well.
PT07 : Monitoring Child Growth and Safety Profile using Growth Charts
Rajesh Babu Moorakonda
Friday, 5:10 PM - 5:25 PM, Location: Executive Ballroom A
In child nutritional trials, graphical presentation of subject-level anthropometric parameters over time is useful to visualize child growth trajectory according to age. The trajectories of selected anthropometric parameters are usually evaluated with reference to a growth chart developed using growth standards, such as World Health Organization (WHO) child growth standards. Growth charts provide expected percentile values of the anthropometric parameters based on children of same gender and age-group from the general population. It is usually important in child nutritional trials to monitor the child growth for potential effect of study product feeding, adverse events and concomitant medications. This paper describes step-by-step procedure for plotting subject-level trajectory of an anthropometric parameter during the trial period on the growth chart developed using WHO child growth standards. The chart also displays details of onset of adverse events as well as intake of concomitant medications. It uses PROC GPLOT and SAS Annotate facilities.
PT09 : A simple way to access the data in EXCEL through SAS v9/ ACCESS® libname and Excel engine
Friday, 5:25 PM - 5:40 PM, Location: Executive Ballroom A
Aiming to summarize libname access and Excel engines which is useful to read excel file into SAS dataset but not familiar to many people who prefer to 'PROC IMPORT' or 'infile input'. This paper will discuss the principle and useful techniques about data transform between SAS and Excel, including named range, spreadsheet, libname, Excel engines. And also solving some problems we may meet. We use SAS® 9.3 64-bit and Excel 2013 64-bit in win7 x64. SAS/ACCESS® software for PC files must be available when we use the Excel engine.
PT11 : Collapsing Adverse Event Records
Friday, 3:00 PM - 3:20 PM, Location: Executive Ballroom A
In clinical trials, the way of collecting adverse events (AEs) is diverse and collapsing AEs sometimes becomes necessary. Categorizing AEs into collapsible and non-collapsible, and collapsing the collapsible AEs are challenging tasks for SAS programmers in the clinical trials. This paper will introduce how to categorize AEs into collapsible and non-collapsible. Three types of collapsible AEs will be discussed. 1) Multiple AEs with same onset date; 2) Multiple AEs with a time contiguous sequence; 3) Overlapping AEs. This paper will introduce the implementation of collapsing and sample SAS codes for each type of collapsible AE records.
Statistics & PharmacokineticsSP01 : Continual Reassessment Method (CRM) in Dose-finding Trials
Saturday, 8:30 AM - 9:00 AM, Location: Executive Ballroom A
Traditionally, a dose-finding trial is based solely on toxicity categorized as binary outcome (i.e., YES or NO), that found wide application in clinical trials including the 3+3 design. However it is neither the most effective nor the most efficient statistical tool that can be utilized in such trials. Continual reassessment method (CRM) is based on a parametric model on dose-toxicity relationship, and uses accumulating data to continuously update the estimate of the dose-toxicity relationship so that we can make the best decision of dose assignment for the next patient. This reflects what clinicians actually do in practice and has been shown to work well even in trials with relatively small total sample size. This poster provides a high level summary of various CRM techniques including demonstration by simulation.
SP02 : Proportions: Stop Worrying and Start Calculating
Saturday, 9:00 AM - 9:15 AM, Location: Executive Ballroom A
In clinical trial, it is often necessary to obtain statistical results by using confidence intervals for unknown proportion based on binomial sampling. The calculation of confidence intervals can be straightforward using normal approximation based on central limit theorem. However, the usual approximation is known to be poor for small sample size, or when the true proportion is close to Zero or One. This paper will summarize some usual methods to calculate confidence intervals for single proportion and the difference of two proportions. It also demonstrates various calculation methods through SAS PROCs or other DATA steps for usual or limited situations.In clinical trial, it is often necessary to obtain statistical results by using confidence intervals for unknown proportion based on binomial sampling. The calculation of confidence intervals can be straightforward using normal approximation based on central limit theorem. However, the usual approximation is known to be poor for small sample size, or when the true proportion is close to Zero or One. This paper will summarize some usual methods to calculate confidence intervals for single proportion and the difference of two proportions. It also demonstrates various calculation methods through SAS PROCs or other DATA steps for usual or limited situations.
SP03 : Compartmental Models in SAS: Application to Model Epidemics
Ka Chun Chong
Chung Ying Zee
Saturday, 9:15 AM - 9:30 AM, Location: Executive Ballroom A
Compartment modeling is useful to quantify the spread of elements in a dynamic system. Apart from the Pharmacokinetics-Pharmacodynamics analysis, epidemic modeling is another broad application of compartmental models. In this paper, we demonstrate how to use PROC MODEL and arrays in DATA steps to generate and fit the epidemic models such as the Kermack and McKendrick model and SEIR model. Practical application is demonstrated for the 2009 pandemic A/H1N1.
SP06 : Bayesian Augmented Control (BAC) Application by using 'Proc iml'
Saturday, 10:30 AM - 11:00 AM, Location: Executive Ballroom A
Bayesian statistical methods are being used increasingly in clinical research because the Bayesian approach is ideally suited to adapting to information that accrues during a trial, potentially allowing for patients receive better treatment and for statisticians plan optimal clinical trial design. Due to the limitation of budget control, patients' demographics or some other unforeseen reasons, sometimes a study cannot recruit sufficient patients, reach the minimum sample size requirement and detect the effect between control group and treatment group. In this situation, it becomes a task for statisticians to design an efficient experiment to make full use of the patients. Bayesian augmented control, known as BAC, has provided a way for statisticians to use the limited patients' number to get an enough power. By borrowing historical information from previous studies and publications, the sample size of control group will be increased, which result in the fact that more patients can be allocated into treatment group and an adequate power can be calculated. By applying 'proc iml' in SAS, BAC method can be implemented in SAS and pharmaceutical companies can reach the goal of saving patients' numbers as well as controlling budget.
SP07 : Analysis of Multiple Imputation in Conjunction with Robust Regression in Clinical Trials Using SAS
Saturday, 11:00 AM - 11:15 AM, Location: Executive Ballroom A
In clinical trial, statistician usually design the statistical method based on normal distribution assumption. But sometimes the actual data may turns out departure from normality. If we continue to use the original method, the result would be misleading and inappropriate In this paper, we would introduce the SAS code for an alternate method that could be used to supplement this issue: multiple imputation for handling missing data and repeated sampling, robust regression method for analysis and get the final result using Rubin's formaula .
SP08 : Forecast number of events in oncology trials: parametric method and non-parametric method
Saturday, 9:30 AM - 9:45 AM, Location: Executive Ballroom A
In this paper, we will discuss 2 methods to calculate the number of events at a future timepoint in oncology clinical trials: parametric method and non-parametric method. Parametric method assumes gamma distribution for event rate and drop-out rate, and the parameter of the distribution can be either based on the assumption from the planning phase, or updated with avaibalbe data when the trial is ongoing. Randon samples are drawn from the gamma distribution. For each subject, a probability of event at time t is calculated. (If an event already occured to this subject, the probability of event is 1; if this subject already dropped out, the probability of event is 0.) The the sum of event probabilities is calsulated for each sample. The median and CI can be calculated across all random samples. Non-parametric method generates bootstrap samples from the original dataset. For each subject without event or drop-out in the bootstrap samples, the subject's event/drop-out time can be drawn from KM distribution based on the data from subjects with longer event/drop-out times. For example, if subject i is in the bootstramp sample and has no event or drop-out. Select all the subjects in the bootstramp sample with longer observed times than the observed time of subject i. Draw KM curve of the event/drop-out time from all the selected subjects. If this KM curve dosen't reach 0, an exponential model needs to be fitted to the tail, to get a complete KM curve. Randomly select event/drop-out time based on the probability in this newly created KM curve and assign them to subject i. Determine the event status for subject i at t based on the randomly assigned event/drop-out time. Repeat this for all the subjects without event or drop-out in this bootstrap sample and for all bootstrap samples. Then the sum of events can be calculated for each bootstrap sample and the median and CI can be calculated across all random samples.
SP09 : Extend SAS by R
Saturday, 9:45 AM - 10:00 AM, Location: Executive Ballroom A
Figures, simulations, and new statistical methodologies have become increasingly important and needed in pharmaceutical industry. Even though SAS is a powerful leading statistical analysis software in pharmaceutical industry, it is still not realistic to depend only on SAS to deal with all figures, simulations and new methodologies. Therefor extend SAS by some other tools are needed and there are several possible ways to extend SAS. One attractive option is to extend SAS by R, a free statistical environment which offers a wide of variety of statistical and computing techniques. This paper will share the knowledge about how to extend SAS by R and illustrate corresponding application using graph programming.
SP10 : Early Development Stage (EDS) study---A wonderful start for junior SAS programmer
Saturday, 10:00 AM - 10:15 AM, Location: Executive Ballroom A
As junior SAS programmers, given the fact of lacking experiences, a relatively less complicated and direct study design would be a better choice to obtain hands on experiences for projects. Fortunately Early Development Stage (EDS) study provides us such an opportunity to begin with. In this paper, basic knowledge and concepts of EDS study will be introduced. Meanwhile, commonly used variables (AUC0_last, Cmax, T_half, Tmax), main content in tables and graphs for EDS study (Geometric Mean, Geometric Mean Ratio and Root Mean Square Error) and related calculation methods (Proc Mixed procedure) will be elaborated in detail. Furthermore, some commonly used programming skills for both table and graph generation will be presented as well. Keyword: EDS, Geometric Mean, Geometric Mean Ratio, Proc Mixed
SP11 : Calculation of Relative Risk and Odds Ratio Using StatXact PROCs
Saturday, 11:15 AM - 11:30 AM, Location: Executive Ballroom A
In clinical trials, statistical inference using asymptotic and approximate statistical methods is not applicable for rare events, e.g., calculating the relative risk/odds ratio for low incidence binomial endpoints. This paper introduces how to calculate the relative risk and odds ratio using StatXact PROCs, which is a more reliable method. StatXact PROCs is a statistical package for Exact Nonparametric Inference, developed for SAS users. It can be used to make reliable inferences by exact and Monte Carlo methods when the data are sparse, heavily tied, skewed, or the accuracy of the corresponding large sample theory is in doubt.