180 likes | 271 Views
Secrets of a Successful Software Switch (& Database Conversion). Jeannette Jackson-Thompson, MSPH, PhD Operations Director, Missouri Cancer Registry Research Associate Professor, Health Management & Informatics, University of Missouri-Columbia jacksonthompsonj@health.missouri.edu.
E N D
Secrets of a Successful Software Switch (& Database Conversion) Jeannette Jackson-Thompson, MSPH, PhDOperations Director, Missouri Cancer RegistryResearch Associate Professor, Health Management & Informatics, University of Missouri-Columbia jacksonthompsonj@health.missouri.edu This project was supported in part by a cooperative agreement between the Centers for Disease Control and Prevention (CDC) and the Missouri Department of Health and Senior Services (DHSS) (#U58/DP000820-01) and a Surveillance Contract between DHSS and the University of Missouri.
Acknowledgments • Missouri Cancer Registry (MCR) staff • Sue Vest, Nancy Cole, Mary Jane King, Iris Zachary, Saba Yemane – input for presentation • All MCR staff who willingly worked long hours to make this latest software & database conversion a success • Registry Plus development & support team
Objectives • Provide an overview of MCR’s central cancer registry (CCR) software and database conversion experiences • Identify key considerations in making a software switch • Highlight components for success
Background on MCR • Historical: 1972 – voluntary agreement • State mandate – 1984, implemented ‘85 • Federal mandate (Cancer Registries Amendment Act of 1992) – MO funded 1995 • NPCR reference year = 1996 • Expanded reporting – 1999 (192.650 – 192.657 RSMo)
Background – MCR Data • 45,000 – 50,000 records/year • c. 28,000 unduplicated MO cases/year • c. 400,000 unduplicated cases in database • primarily1996 and later • Historical database – > 325,000 records • 1995 and earlier
Background – Software #1 • ≤ 1996: State-developed system • Non-standard layout • Non-standard data elements • Leased hardware (Nixdorf mini – mainframe connection for data storage) • CTRs had no direct access to data
Background – Software #2 • 1996: NPCR funds used for CCR software • Vendor type: Not-for-profit, multiple CCR clients • Standard data elements, NAACCR layout • PC-based system • New Database Manager • CTRs have direct access to data
Background – Software #3 • 1997: NPCR unobligated funds used for CCR software • Vendor type: University, MCR sole client • Standard data elements, NAACCR layout • MO-customized software on super-PC (DEC Alpha, Oracle tables • Stable system (“down” > 24 hrs 1 time in 7 yrs) • Database integrity well-protected • Historic database still separate (match, not update)
Background – Software #4 • 2004: NPCR funds for CCR software available in contract but DHSS they must make purchase • Vendor type: For profit, small number of CCRs use • Standard data elements, NAACCR layout • PC-based, vendor controls access to “back end” • Did not “go live” until 10/05 (not 2/05 as needed) • Lack of shared vocabulary (conversion v. migration) • > 325,000 “dirty” historic records imported(pending) • System not stable, database integrity in question
May 2007 – MCR Quandary • DHSS contract w/ vendor – 3rd year 9/07 • System not stable • Database integrity • MCR staff frustration • Productivity and morale impacted
MCR Solution • Visit another CCR (PA) • See another CCR software system in use • Obtain hands-on experience with CRS Plus • Meet with current vendor • Enter into dialogue with CDC/NPCR project officer and Registry Plus development/support team • MCR already using Abstract Plus, Web Plus
Key Considerations • If we don’t switch software now, what happens? • Productivity & morale decline further • Database integrity may be impaired beyond repair • Contract deliverables, NPCR requirements in question • Is a viable, affordable alternative available? • How many existing problems will be solved? • Will new challenges be created? • What do key staff want to do? • Is a switch feasible?
Considerations for Success • Knowledge of software, hardware & network • Well-trained, competent, motivated staff with needed expertise • Two-sided communication with vendor • Both sides willing to listen, learn, cooperate • CCR & vendor have common interest • Clean, high-quality data (not “bottom line”) • Ongoing, mutually-beneficial relationship
More Considerations • Make sure CCR & vendor using common vocabulary • Data dictionary • Manuals, on-line help • Mutual respect for knowledge, expertise • Willingness to learn, adapt to reach desired outcome
Software #5 • Discussions 6/07 – 7/07 • Decision 8/07 • Notify DHSS – they notify vendor • Shut down #4 9/07 • Most 2005 cases edited, consolidated • Many 2006 cases not entered • Conference calls, Web demos 8/07-10/07
Software #5 (cont’d) • Establish: • Task list for CCR & vendor • Identify responsible parties • Time frame for tasks/actions • Monitor tasks/actions/steps/stage in process • Modify tasks/responsibilities/steps if needed • Create extract files • Test new system • Go live!
Success!!! • Target date – mid-Oct; Actual ‘go live – early Nov. 07 • Active and historical databases combined • Few problems experienced • NAACCR submission – 10-day extension; uploaded day 6 • GOLD certification • NPCR data submission – on schedule • > 95% at 24 months, > 90% at 12 months (1st time)
Satisfied QA Staff Comments - Deb Smith, BA, CTR Questions? Thanks to MCR & Registry Plus staff