Agenda and Details

9:00‑9:10

User-to User Questions and Answers

Moderators - DCSUG officers

9:15‑10:20

Andrew Karp, Sierra Information Services

PROC MEANS: The Basics and Beyond

Abstract: Attending this “super” tutorial gives you in-depth insights in to the wide range of data analysis and management tools available in this extremely powerful and versatile “workhorse” BASE SAS procedure. Starting with the basics, this presentation introduces core PROC MEANS capabilities to calculate and report the values of descriptive statistics and to “group” or “classify” those statistics “by” the values of other variables. We will also look at how to select (from a total of 32 available) statistics we want calculated and how to control their display in our output. Then we will move to an in-depth look at how to use PROC MEANS syntax to create data sets “holding” the statistics and summaries it computes, and how to customize these data sets using the AUTONAME and AUTOLABEL options. Next, we’ll explore the often little-known methods available in the PROC to generate multiple output analytic data sets in a single “pass” through a large data set, and how to speed the generation of these output data sets using the CHARTYPE option and multiple OUTPUT Statements. We’ll then go even further to reduce system resources needed to complete these tasking by learning how to take advantage of the TYPES Statement, and to include extreme values in our output data sets using the IDMIN and IDMAX options. Last, we’ll see how the PRELOADFMT and COMPLETETYPES options, as well as the multilabel format facility, provide even more power to get our work done in less time with PROC MEANS.

Andrew Karp is well known to DCSUG’s members for his many presentations to our group over the years. He is a 30-year SAS Software user who for the past 17 years has had his own SAS Software consulting and training firm, Sierra Information Services, located in the California wine country. Andrew has spoken at numerous events for SAS users in a total of eleven countries and has been an invited speaker at 17 consecutive SAS Global Forum events. He earned his undergraduate and graduate degrees from The George Washington University in Washington, DC.

10:20‑10:35

Ryan A. Leonard, NY City Health and Hospitals Corporation

Determining Number and Lengths of Inpatient Hospital Admissions Using PROC MEANS and the SAS Array Statement

Abstract: This presentation demonstrates how Kings County Hospital Center, one of the largest public hospitals in the United States, recently used a combination of PROC MEANS and SAS programming language tools to analyze our inpatient utilization files. The program we will share can be applied to any SAS data set with up to 100 of unique pairs of admission and discharge dates per patient and requires minimal user intervention to run in as the number of rows in the initial analysis file changes. It rapidly converts a "tall/skinny" file with one row per unique patient admission/discharge "pair." to a "short/wide" file with one record per patient medical record number, This "short/wide" file contains all the admit/discharge "pairs" per medical record number as well as the length of say for each pair and the total number of inpatient days recorded for that patient, making easy to calculate the number of admissions, length of stay for each admission and the number of inpatient days per patient.

Ryan Leonard is Assistant Director of Hospitals at the NY City Health and Hospitals Corporation, where he is assigned to the Office of the CEO at Kings County Hospital Center, the largest public hospital in Brooklyn, NY. He uses SAS regularly to manage and analyze the hospital's clinical, financial, and other operational data for health care utilization and delivery redesign projects. Ryan graduated with honors from the University of Pennsylvania in 2010.

11:00‑11:50

Andrew Karp, Sierra Information Services

Creating the Perfect Plot with PROCs EXPAND and SGPLOT: The Dynamic Duo of Time Series Display

Abstract: SAS users often face critical challenges when tasked to prepare and display a series of values ordered in the time domain. For example, values may be missing from the series, or the observed data need to be either aggregated (e.g., “rolled up” from daily observations to quarterly values) or interpolated (weekly estimates derived from monthly observations). Or, a “moving time window statistic” requires computation and display. Attending this presentation will equip you an understanding of how to apply PROC EXPAND’s wide range of time series preparation capabilities and then how to use powerful PROC SGPLOT resources to create the graphics images you need to “tell the story” contained in the data values to others.

Directions for those taking Metro:

• Take the Red Line to Union Station (map, courtesy of StationMasters).
• Exit the station via the Amtrak exit, but do not go up to the train level.
• Instead, after going up to and out through the farecard gates, turn to the left where there is an exit to the street (First Street NE).
• Directly across from the exit are the employee and visitors' entrances to the building. Satellite view (note that the green arrow is accurately located, but the associated address does not actually exist)

Drivers: get map and directions from MapQuest.

After checking in with security, take the main elevators down to the G level. BRING A PICTURE ID!

If you are new to DCSUG and are planning to attend this meeting, please E-MAIL Beth Schreiber by December 2^nd and leave your name. If you personally and directly received a DCSUG announcement e-mail about this meeting, that means you are on the list, and no action is necessary. Security at the Bureau of Labor Statistics is very tight and having your name on the list of possible attendees will make your entrance to the building easier.

DCSUG Home