Continue in 2 seconds

Harvard School of Public Health Tracks Research Projects with CrossGraphs

Published
  • February 01 1999, 1:00am EST

BACKGROUND: As statistical consultants in the Harvard Medical area, we help medical researchers plan, carry out and analyze clinical and observational trials to test the effectiveness of new medical treatments. These studies often collect many hundreds of variables of interest on a set of patients who may be studied repeatedly over time.

PLATFORMS: The CrossGraphs data visualization and graphical reporting system runs on the following workstations: Windows 3.1, Windows 95, Windows NT, PowerPC Macintosh, Sun Solaris and HP-UX. CrossGraphs interfaces to a variety of data sources including Oracle, ASCII data files, SAS, dBase and ODBC-compliant sources.

PROBLEM SOLVED: Researchers at the Harvard School of Public Health needed easy-to-use data visualization and graphical reporting tools to track the progress of their research projects, to help with data cleansing and checking, to explore subtle relationships in their data and to generate figures for publication which illustrate their results.

PRODUCT FUNCTIONALITY: CrossGraphs is a powerful yet easy-to-use tool with several useful features: 1) It facilitates the generation of arrays of graphs, one graph per individual or subgroup, to assist data cleansing, checking for outliers and exploring the data on a patient-by-patient basis. For example, in a study of premature infants, researchers may wish to see a summary of each patient's vital signs, such as blood pressure, temperature and respiratory status, collected each hour of the day. 2) It creates displays of multivariate data to help us visualize higher dimensional relationships in data sets and assist the monitoring, analysis and presentation of research data. For example, in a study of premature infants, it is easy to create side-by-side "plots" of each patient group receiving a different treatment, while simultaneously stratifying by other variables, such as the sex of the child, the quantile of birthweight and the medical status of the mother. These plots allow the researcher to consider relationships among many variables. For example, arrays of scatterplots let researchers view relationships among six variables at a time.

STRENGTHS: CrossGraphs' strength lies in its ability to handle a variety of 15 plotting types. These allow display of traditional graphs such as X-Y scatterplots, boxplots and bar graphs, as well as advanced graph types such as time line summary graphs, delta graphs and Kaplan-Meier plots. Within each plot, it is straightforward to drill down on individual plots or individual points to display additional information. Because CrossGraphs reads data directly from other packages such as SAS and Oracle, it is easy to integrate CrossGraphs with an existing data management system. It allows researchers to monitor the status of the study data, search for outlying observations that merit further investigation, explore relationships in the data and create publication quality figures.

WEAKNESSES: Creating arrays of plots is easy, but comprehending and interpreting these graphs takes practice. While the manual offers examples of the different plotting types, the more sophisticated graphs can be tricky to set up. The newest release provides many options for configuring the graphs which has made the configuration dialogs busier.

SELECTION CRITERIA: CrossGraphs was selected for use because it quickly produces many graphs without programming, so we can review, compare and report complex, multivariate data more effectively. CrossGraphs supports a wide range of traditional and advanced graph types, so we can select the most effective data presentation technique. The figures created by CrossGraphs are suitable for publication. The software presents data graphically, using an elegant and spare design, without distracting visual elements.

DELIVERABLES: We use CrossGraphs for exploratory analysis, data review and final presentation of results in formal publications. It is straightforward to generate useful plots in CrossGraphs within a few hours of using the product. Integrating the package into an existing investigation takes thought and planning, but it is well worth the investment.

VENDOR SUPPORT: Support is excellent. Turnaround time for e-mailed questions and problems was less than an hour, and all of the questions were answered immediately and professionally.

DOCUMENTATION: The extensive on-line documentation complements the well-organized hard-copy manual. Both were easy to navigate.


Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access