CCMDB Data cleaner.mdb general information

From CCMDB Wiki
Revision as of 18:18, 8 December 2009 by TOstryzniuk (talk | contribs) (→L_Aux)
Jump to navigation Jump to search

The data cleaning tool is used by the Data Processor to flag and fix internal inconsistencies in data already submitted by data collectors.

It is an Access program which is also known as the CCMDB Cleaner.mdb program.

The goal is to pull as many of these checks ahead to be done in CCMDB.mdb, but some checks (e.g. those that required TISS data to be available) need to be included in this in office data cleaner.mdb program.

Progress

The Data Cleaning process is a work in progress.

  • some checks Julie usually runs have not yet been implemented
  • some checks need to be added to the CCMDB.mdb instead of cleaner (do it sooner)

Discussion

Template:Discussion

  • Trish/Pagasa, which check would you like to work on next? Ttenbergen 12:28, 8 December 2009 (CST)

File Location

File name is CCMDB Cleaner.mdb

A master copy of the file is on X:\Data_cleaning\CCMDB Cleaner.mdb

For use, Pagasa copies the file to her C:\ drive. The file can grow to >1GB in size during use so working with it on the network would be too slow.

L_Aux

The table L_Aux contains some calculated values that are exported by TMSX. They are:

  • Apache - not sure where you are getting your data from for this. The registry screen maybe? Ed scores APACHE in the APACHE II file for each patient and also records the total Apache score into the Registry file for each patient.
  • Inits - Collector initials - what do you mean collector initial are calculated?--TOstryzniuk 15:39, 8 December 2009 (CST)
  • LOS -length of stay. admit date/time & discharge date time.
  • Enc - Encounter - currently not used. A sorting utility program that use to tally the number of encounter to the database. Sorting was done using, FN, LN, DOB and PHIN? We had suspending using it when we found many data errors in DOB or name spelling, PHIN. SInce we have worked on fixing this problem we have not continue to use it.
  • Index - not calculated. This was a old file system number we used.
  • Serial - serial numbers are collection site organization number used during collection. These are not calculated. No sure what you mean?
  • HMORT -this is not calculation just "S"urvive or "E"xpired if we know it. Dan wanted us to track hospital outcome but we were not able to get this data easily.
  • apspts - APACHE - Acute Physiological points SeeAPACHE Scoring table. (For added detail Ed's program showed APS score using APACHE Scoring table items 1-11 and GCS score by itself. APS = APS + GSW (which includes all physiological variables).
  • totApache (how is this different from "Apache" above?)- see story under Apache above.
    • TotApache = APS + GCS + chempts + agepts.

Discussion

Template:Discussion

  • How are these calculated, or where are they derived from? Ttenbergen 13:54, 8 December 2009 (CST)
    • the wiki is annoyingly slow.........I am planning to abandon this for today.--TOstryzniuk 17:18, 8 December 2009 (CST)

History

  • Original SAS data checking/cleaning queries developed by Julie in SAS who obtained input from Trish, data collectors or others where she needed.
  • Tina was developing these checks in Access so that others who are not familiar with SAS programming can run or add additional queries into it when needed.
  • Pagasa was running the queries and distributing problem ID, by email, fax or phone to data collectors.
  • data cleaner.mdb was outputting many false +ve's, and collectors were complaining about the number of items they were being required to check that were not errors.
  • The data cleaner.mdb was reviewed by Trish with Pagasa. The problems found and also the original items in the data cleaning.mdb list from Julie is posted below. The list of checks for this program that were are X drive were also transcribed to this list.

Related articles

Retrieved from "https://ccmdb.kuality.ca/index.php?title=CCMDB_Data_cleaner.mdb_general_information&oldid=16161"