Wednesday, July 17, 2002

Notes from Ojai, second topic - CPOE Implementation Case Studies. Big panel. 1st speaker is Doug Jones (ex-Kaiser), now CIO at Cedars Sinai. Have built a CPOE system called Patient Care Expert. PCX is a pure thin client app, web-based UI. PCX challenges: dev/test of content; way too many requests for enhancements; training 7000 people is a huge challenge; conversion of good medical advice to rule sets is difficult; change management (PCX is only way to place orders). Hospital made PCX competency a term of employment. Doug mentioned that a strong business case led CS to make the investment in PCX. I'd love to get my hands on that business case. PCX does Patient Mgmt, Orders Mgmt, MR Coding, Patient Accting. Runs on AIX. Have built own CDR. Has dozens of interfaces. Wonder who Doug used to build PCX - IBM proservices? Next speaker - Northwestern Memorial Hospital. Ambulatory EMR strategy. Chose Cerner for their CPOE project. This speaker had a lot of good lessons learned, but he had to blow through them very fast. Next speaker - the new VP/CIO of Stanford Medical Center. Lots going on here: have de-installed Oacis, have turned over CIO a couple of times; have installed IDX Lastword. What went well?  (1) pharmacy integration and (2) portable wireless devices. What would they do differently? (1) mandatory MD training, (2) offer advanced training for shortcuts and power users, (3) pay more attention to use of rule sets and workflow, (4) put a process in place to deal with verbal orders, (5) would have done nursing documentation before CPOE, (6) paid more attention to organizational readiness and change management - had LOTS of churn with people and systems in the post-Oacis stage. This guy's slides are a good template for why you need great proservices help in any CPOE project. Just ran into Orlando Portale. Looks like he's now Sun's Healthcare Industry Manager. Cool...good for him. Seems he's moved back to San Diego. Missed another speaker from Loma Linda. Advocated Big Bang approach for CPOE - good list of why. User training keeps being mentioned as the long pole in the tent for CPOE. Once again, it's not the tech, it's the people/process that is important. Last speaker - John Fleming, Cerner. Presented Cerner's "Transformation Model". Looks like good content, but he's run out of time. <The moderator of this panel should be punished.> Spent time talking about a strategic planning roadmap, the big picture for rollout. Point is the #1 reason for problems is poor change management, not the app. Last last speaker - an MD from Ohio State, director of ICU. Avowed complete non-digital guy. Interesting strategy - introducing himself via comprehensive denigration. Don't think I'll use this... This is really pathetic - the guy has been talking extemporaneously for 10 minutes, while 2 people feverishly worked to get his PC  booted and his slides projected, with no luck. Doesn't anyone prepare!?!? OK, finally, an insight. The OSU guy notes that focusing on workflow is of limited value, because in his world care is a set of responses to random events. Good point. I guess the trick is focusing on the pre-programmed responses, and explicitly assuming that a response (workflow) can/will happen at any instant. comment []3:33:14 PM

Notes from PCC, Ojai. Feels good to be back at Ojai - it's been 10 years. Wasn't here 10 minutes and ran into some old friends and familiar people - Bill Spooner, Lee Ann Slinkard, Ed Kopetsky, Rich Rydell. Running into Lee Ann was particularly nice (other than introducing her as Lee Marley...ackkk). It helps to see some familiar faces. First speaker talked about system selection - at least ostensibly. He cited Tufte's "Presentation of Visual Information" www.edwardtufte.com/1715886774/tufte/, plus a book called "The Logic of Errors". He used space shuttle O ring data to illustrate his point about presenting information properly. He spent some time describing why Google is great. The only link I could find to system selection is the notion that you need a good data model and flexible, specialty-specific presentation capability. Duh...He also cited the Gartner 8-part EMR model. VP/CIO of Loma Linda: "...when you get right down to it, any vendor's product can do the job...". Implication is that selection must be made on some basis other than the product's function. Another speaker - good insight - multi-tier architectures make system performance complex, even problematic. Pay attention to performance guarantees. Spooner: advocates scenario-based selection process, because all vendors can/do lie in RFP responses. LL CIO advocates detailed Q&A being attached as part of the contract. Usabililty, workflow, integration are cited as things you can get from scenario-based selection vs. RFP-based. 1st speaker keeps citing Google as the model. Bad example, I think. Google can withstand casual use and even errors - it's not mission critical. EMR systems cannot afford to present a single error - very different problems. LL CIO cites 5 year strategic plan as his saving grace when other executive staff turned over. CGEY person made the point that if you're not ready to really, really transform the way you do business, don't bother acquiring an EMR or major app. LL CIO is doing Cerner rollout right now. Good system selection sources: (1) Product specific message groups; (2) Gartner and Dorenfest; (3) one of the speaker's own web pubs (KLAS?). Interesting question - how decoupled is the data store from logic and presentation in the big apps (Cerner, Siemens, IDX)? If the data store is decoupled, you can do a best of breed strategy. Otherwise, it's big bang. comment []2:03:50 PM

© Copyright 2002 Jeff Nichols.