Skip to main content

TreeExtraction Documentation

    Basic Details
    Date Posted
    Status
    In progress
    Description

    TreeScan™ is a free data mining software that allows users to analyze large data sets using different versions of the tree-based scan statistic. TreeExtraction is a reusable cohort extraction program designed to create an analytic dataset, which can be analyzed with TreeScan™ software.

    Included on this page:

    1. TreeExtraction Documentation and a SAS Package Toolkit, version 1.2: Select features of this version include the ability to collect data for multiple episodes of exposure, and an exposure exclusion feature.
    2. TreeExtraction Documentation and a SAS Package Toolkit, version 1.3: Select features of this version include the ability to use ICD-9-CM only or ICD-9-CM and ICD-10-CM together, and ability to create a washout period for exclusions.
    3. TreeExtraction Documentation and a SAS Package Toolkit, version 1.4: Select new features of this version include allowance for differential follow-up time, the ability to make the searching risk window parameters more flexible, and the additional of a primary diagnosis option to define outcomes and exclusions.
    4. Supporting Tree and Mapping Files: FDA center-specific tree and mapping files include curated trees of ICD-9-CM and ICD-10-CM codes utilized for specific Sentinel analyses. Each tree file is formatted into a comma-separated, two column ASCII file that can be read into TreeScan. The first column is the child node and the second column is the parent node. Each tree file was used in prior TreeScan studies, and was curated by the respective FDA center listed in the naming scheme. The naming scheme indicates which trees were used in ICD-9-CM only studies and which were used in mapped ICD-9-CM to ICD-10-CM studies. Full trees with all codes included in the Clinical Classification Software (CCS) can be downloaded through the Agency for Healthcare Research and Quality (AHRQ) Healthcare Cost and Utilization Project, (H-Cup). In mapped ICD-9-CM to ICD-10-CM studies the relevant mapping file for CDER or CBER must be used in conjunction with the relevant trees files in TreeExtraction.  
    Workgroup Member(s)

    David Cole, BM; Judith C. Maro, PhD; Inna Dashevsky, MS; Department of Population Medicine, Harvard Pilgrim Health Care Institute and Harvard Medical School, Boston, MA

    Martin Kulldorff, PhD; Division of Pharmacoepidemiology and Pharmacoeconomics, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA