The Art of DR Auditing By Acmeware, Inc. Edward Chisam – Senior Consultant
Course Overview Our Experience with Auditing Writing your own Audits DR Auditor Task Management
Lots of Examples Examples are derived from real problems experienced by our clients Some examples have been simplified for purposes of this discussion We will provide tools demonstrated in this session to all those who are interested in them via
Acmeware’s experience with Auditing Glen was the manager at MEDITECH responsible for development of the DR product Intimately involved in original design decisions Has written audits, reports, and applications against DR while at Acmeware Ed was a senior developer in DR at MEDITECH Lead developer for the DR Manager product Has run numerous audits using DR Auditor
The MEDITECH Auditor MEDITECH is working on its own Auditor of DR. Drawbacks to it at the current time include: Slow performance Limited choices of sample rates Uses same routines as initial load to populate sample tables By definition there is a lack of independent objectivity
Developing your own Audits Possible approaches that might be tried: NPR Report Writer Use Report Writer to save DPM subscripts to a file Import the file into a SQL table using Data Transformation Services Write a computer program that compares your table against the SQL table from MEDITECH Open Database Access Use 3 rd party ODBC or.NET product to extract DPM subscripts, and save these to a SQL Table Write a computer program that compares your table against the SQL table from MEDITECH
Developing your own Audits Drawbacks to these methods: Programming background required (C#,ODBC,.NET,SQL, etc.) Your program will break with the next MEDITECH schema change Cost of acquiring 3 rd party tools Not optimized for speed
DR Auditor Overview DR Auditor from Blue Elm Relatively inexpensive Has a long history of uncovering problems leading to MEDITECH correction. Audits may be scheduled Easy to use – no programming background needed
Using the DR Auditor After a MEDITECH update, it is critical to run a Schema Upload before running further audits This automates the process of accounting for MEDITECH changes to table structures Only takes about 45 minutes for this routine to execute
Using the DR Auditor Tables with <= 100,000 rows should be audited at a 1:1 sample rate Tables > 100,000 rows may be audited at a rate such that 100,000 rows will sampled. ( 1 million row would be audited at a 1:10 sample rate ) However, I have run audits on 12 million row tables at a 1:1 sample rate when it is absolutely required to guarantee table accuracy.
Using the DR Auditor The DR Auditor has the great ability to group multiple related audits from the same DPM together The method of sampling all the data works equally as fast if the DR Auditor is sampling ten tables at time, or if it is only sampling one Take advantage of this ability when sampling MEDITECH DPMS containing large amounts of data, such as BAR.PAT or LAB.L.SPEC.
Using the DR Auditor Re-analyzing vs. Re-sampling Before submitting errors to MEDITECH, use the Re-analysis routine to run a second comparison of the sample versus the MEDITCH table. This prevents errors due to MEDITECH transmission lag Audits more than 3 days old should be re- sampled instead, as significant edits may invalidate the sample The re-sample takes longer to run than the re- analysis
Using the DR Auditor Only Audit DR tables you are either reporting from or are planning on using MEDITECH has limited resources for Data Repository support, you only want them to work on the issues that can most help you Audits do use up some system resources
Using the DR Auditor When reporting issues, always include examples of missing data Obtain these by running a Report on an audit, then choosing the Export to a text file routine Copy the contents of the text file and paste it into the MEDITECH task, so they have the example available to work with
Using the DR Auditor A new ability involves using the DR auditor to re-queue missing rows This process has been automated for Saint John’s Hopkins -- Bayview After Audits are run, missing rows are saved to a file, and then are automatically re-queued to transmit to the Data Repository When the number of missing rows is not too large, this helps preserve data integrity