Abstract 590

When large data sets are collected from patients over an extended time, front-end clinicians often need to access and analyze this data in a fast, efficient and cognitively meaningful format. As machine generated clinical databases proliferate, supposedly to guide nurse/physician's decisions at medicine's front-end, bedside care providers often suffer cognitive overload. Immediate bedside data reporting in a meaningful graphic format, easily reconfigured for a wide variety of different users is essentially unavailable. Proprietary graphic systems for clinical application are usually hard-coded, do not have rapidly scalable axes and cannot accept links to a wide range of different clinical databases. We have developed "MultiPlot", a multifunctional graphing utility to assist bedside caregivers in rapid data analysis. MultiPlot was originally designed to extract blood gas data from a relational database and to present many different graphical representations for quick trend and comparative analysis.

MultiPlot is written in Visual Basic 5.0 and runs on any 32-bit PC with a Microsoft Windows 95 platform. Data can be extracted from any relational database that supports Standard Query Language (SQL). MultiPlot's primary functional value is the ability to rapidly plot and replot several (1 to 21+) parameters simultaneously with individual graphs either stacked along adjacent axis, or overlapping along a shared axis. MultiPlot graphs data along two-dimensional axes that are updateable in "real time". The axes are fully scalable and provide rapid capacity (less than one second) to zoom into a user-defined time period with the click of a mouse. The maximum and minimum values are immediately adjusted to the current data set, allowing the program to provide clear images of widely variant data ranges. MultiPlot is simple enough to accommodate most user requirements, it has several advanced features for fine-tuning. MultiPlot can plot points, line graphs and/or bar graphs, report graph coordinates by pointing the mouse, toggle between a single graph and multiple graphs and allow the user to define color schemes.

MultiPlot eliminates tedious review of extensive clinical data by allowing clinicians to analyze graphically several hours or days worth of data, facilitating quick identification of acute deterioration episodes. Blood gas parameters were used as a model during program development. MultiPlot's zoom function can show very clearly what happened between "recorded" vital signs for any patient who is attached to machines or monitors that automatically upload data to an SQL compliant database (seconds to minutes). MultiPlot was built with medical domain transferability in mind, and can be useful for analysis of laboratory results, vital statistics and population studies. We think the functionality of this graphic utility exceeds the performance of any graphing modality currently available for medical uses.

Supported by Grant #5T15LM07124 from the National Library of Medicine