Known issues (aka "bugs") for OpenVigil 1 & 2 and OpenVigilFDA

This document lists the known issues/bugs with OpenVigil 1 & 2 and the installations of OpenVigil 1 & 2 at Christian Albrechts University of Kiel. This list was compiled on and is thus up to date to 2016-04-22.

• OpenVigil 1 & and OpenVigilFDA
• Floating point numbers might differ slightly from values calculated in other software. This is due to the concept of the float datatype in the PHP programming language.
• OpenVigil 1
• data loss while importing, affects OpenVigil prior to v1.2.6
  A data field missing in the FDA AERS ASCII data prior 2005Q3 was expected on import, resulting of a loss of all the data in the DEMO table in the affected early FDA quaterly data files.
  Workaround: None. Upgrade to v1.2.6b or later and use the supplied import script to reimport all FDA AERS data.
  
• 0 instead of NULL for empty items, affects all OpenVigil 1 versions
  Calculating of any descriptive statistical measurement is skewed by these incorrectly imported numerical values. They should be set to NULL instead of '0'. A fix for the import script was done for v1.2.7).
  Workaround: None. Reimport your data using the most recent importer (which uses NULL instead of 0) or check your data for plausibility or exclude all zero values.
  Note: The currently imported dataset at Kiel university uses '0' instead of NULL. (2016-04-22)
  
• missing feature: no counting of CASE instead of ISR
  Options to use either all ISR, unique ISR, uniqe CASEs is planned for v1.2.7.
  Workaround: Extract the data with ISR and CASE numbers and count them in an external program.
  
• OpenVigil 2
• Imported datasets with OpenVigil 2.0 and Drugbase 3.0 & drugs@FDA (2013)
  
• Drugs with different USAN/INN/BAN names might be incorrectly imported
  E.g., both "acetylsalicylic acid" (INN) and "aspirin" (USAN) are listed as drugnames in the instance of OpenVigil 2 hosted at Kiel university.
  Workaround: None. Use the ORIGINAL_DRUGNAME field or try using newer drug database data. Validate your dataset using OpenVigil 1 which operates on raw data.
• ATC codes might be mismapped
  E.g., ampicillin was incorrectly assigned ATC-codes of other penicillin antibiotics.
  Workaround: None. Try using a more recent Drugbase dataset.
  
• OpenVigil 2.0
• Buttons for NEXT, 2nd page etc. are not working correctly when using additional search options like "latest report submitted"
  Workaround: use the CSV exporter to extract all results or add manually the required correct parameters to the URL.
  
• Counting and sorting issues in HTML output of "Frequentist methods"
  Very small values are not sorted correctly under some circumstances. Occassionally, a record is split between to pages or not
  shown. Workaround: Use the CSV exporter.
  
• missing feature: no counting of CASE instead of ISR
  Options to use either ISR or CASEs was missing in OpenVigil 2.0
  Workaround: Use OpenVigil 2.1 or extract the data with ISR and CASE numbers and count them in an external program.
• OpenVigilFDA
• 2x2 table calculation broken, fixed in 1.0rc5.
• Results are strongly dependent on the openFDA API. This Elastic/Lucenne-based API appears inconsistent regarding exact-matches. Changes to the syntax might render OpenVigilFDA unusable.
• Feature requests
• Automatic grouping of drugs and/or events
  OpenVigil 2 and OpenVigilFDA allow already using drug classes instead or along of drug names. SMQ and SOC (other MedDRA levels) are required and are currently under development.