Home » News » Currently Reading:

Monday Morning Update 9/10/07

September 8, 2007 News 9 Comments

From Inside View: “Re: Bronx Lebanon. Eclipsys won. Cerner, who everyone assumed was the winner a couple of months ago, was the surprise loser. But the biggest loser was GE. They should have won, could have won, but didn’t execute and their product is a mess. The rate at which GE is losing LastWord clients is alarming.” Unverified, I should mention. Thanks for the update.

From Abe Froman: “Re: Dubai. In Cerner’s presentation at the conference, they showed how CPOE cuts medication turnaround time from hours to 20 minutes. A hospital director there couldn’t understand how you can put the ordering process in the hands of a physician. What happens if he makes a mistake? In Dubai hospitals, pharmacists are simply glorified salesmen, dispensing whatever the order says. They do not question anything. The nurse provides the review when the order is transcribed, which doesn’t happen with CPOE. The conclusion was that Cerner doesn’t understand that market.”

I mentioned tee shirts that will be worn by former employees of a certain vendor at their “glad I’m out of there” reunion. Folks asked me to run a picture. If you want a look, see here and here.

I got an e-mail from Jason Maude, CEO of Isabel Healthcare, after I mentioned the press its diagnosis product receives. He recommends a book called Super Crunchers: Why Thinking-by-Numbers Is the New Way to Be Smart. It has a healthcare chapter that covers IHI’s 10,000 Lives campaign and devotes five pages to Isabel. A quote: House makes excellent drama, but it’s no way to run a health care system. I’ve suggested to my friend Lisa Sanders, who recommends script ideas for the series, that House should have an episode in which the protagonist vies against data-based diagnostics ala Kasparov vs. the IBM computer. Isabel’s Dr. Joseph Britto doesn’t think it would work. ‘Each episode would be five or seven minutes instead of an hour,’ he explains. ‘I could see Isabel working much better with Grey’s Anatomy or ER where they have to make a lot of decisions under a lot of time pressure.’ Only in fiction does man beat the machine.”

Researchers involved in the development of relational databases now say they’re obsolete and should be considered legacy technology. Reason: row-oriented databases are slower than column-oriented ones, especially for data warehouses. I thought the article might mention Cache’ since it doesn’t have those performance issues (remember they used to call it a “post-relational database”) but the guys quoted have already started their own company, so they’re not likely to bring up competitors.

In the UK, Cerner suspends work on the new Millennium R2 release intended for the London and South trusts.

Check out the new SCI Solutions ad to your left. Kind of a cross between the old Batman and Austin Powers.

Computers stolen from a McKesson office in July contained the personal information of thousands of patients who had signed up for prescription assistance programs from drug companies. The company said it’s “not clear” whether the data was encrypted, which is a nice way of saying it wasn’t (few users encrypt their data).

MPI vendor Initiate Systems signs a deal with Capgemini UK for a children’s services application.

A VIASYS engineer has been charged with hacking the computer system of his employer in 2003. He started by deleting files of his EEG development coworkers, then went after the executives. He was finally caught by an outside firm who located the outdated computer he used along with its hard drive, which he kept locked in a desk drawer.

A UK hospital bans employee access to Facebook after heavy use degrades its network.

The latest issue of The White Stone Group’s newsletter is out (warning: PDF). They run a contest to see which hospital has overturned the largest single denial using their TRACE tracker and workflow software. The winner: a Georgia hospital recovered $452,000 when they showed documentation of precertification notification and response. Even more interesting was the “Taste of Trace” recipe for cheese grits souffle, which of course is the food of the gods for Southerners. If Mrs. HIStalk could cook, I’d dispatch her to the kitchen to whip up a batch, albeit using Velveeta instead of fancy cheddar.

The Northwest Medical Information Symposium will be held on September 13-14 in Spokane, WA. Speakers: Gingrich, Leavitt, and Frisse, among others. Meditech is even on board as an uncommon sponsor since the meeting is put on by its customer, Inland Northwest Health Services.

An investment analyst surveys US users of its Emdeon system and “… believes that concerns about falling demand from medical practices and possible glitches in the software are overdone. However, it said that it had uncovered evidence that the group, whose core business is accounting software for small firms, had made only mediocre efforts to market its medical software.”

Throughput software vendor Premise Corporation announces events from the first half of the year: client based doubled, 100% “would buy again” results from KLAS, 100% employee growth, new financing, and new operational and board leadership.

A mother sues the makers of Tylenol #3 when her newborn baby dies, apparently poisoned by the her codeine-containing breast milk. If your hospital flags breastfeeding moms and successfully issues CPOE or pharmacy lactation precautions for specific drugs, it would be the first I’ve seen. Tylenol #3 probably wouldn’t be on the list anyway since it’s not a common problem. FDA issued a warning last month.

I’m moving on to Brev+IT from here, so if you pounced on this e-mail notification, you might have time to sign up before I send it out.

E-mail me. How hard can it be?


View/Print Text Only View/Print Text Only


HIStalk Featured Sponsors

     

Currently there are "9 comments" on this Article:

  1. Apparently GE lost a EMR deal at Fletcher-Allen by not bidding. Epic picked up the deal. Got it from a recently departed senior manager.

  2. Re: Column-centric databases – sure, that works for population-centric OLAP models. The bottom-line is that architects and developers need to specify and use the right tool for the right job. Vendors and developers typically try and jam everything into one model when it doesn’t work. Would you use a claw hammer to break down a wall, you could, but using a sledge hammer would be better.

    Regardless, I find the article a little misleading. It doesn’t give any workable examples. When you look at the author’s website (hey, no possible conflict of interest there – is there?), they state that this is the end of the one-size-fits-all database. Uhhhhh….. maybe that is why each of the “relational” databases has extensions for other types of data, for analytics, and optimizations for transaction read-write. Different schemas are implemented for different requirements.

    It gets down to (as already commented upon) using the right tool for the job – AND – defining a supportable (by people and within cost requirements) infrastructure. Adding more complexity to the technology environment for a given application will add costs. Adding more vendors to the mix may add costs. Does the benefit outweigh the performance inconvenience?

    IMHO, the best database for healthcare transactions and look-up is the E-A-V schema (http://en.wikipedia.org/wiki/Entity-Attribute-Value_model). Shahid Shah has commented about these topics before. It’d be interesting to hear him chime-in and some CTOs from the major vendors (GE, Eclipsys, Cerner). The column-centric schema works great for population-centric queries WITH DEFINED QUERY possibilities. If you don’t know what your clients are going to query, the columnar model can still be a little challenged (but not so bad that it isn’t a compromise to make). It’d be hard to justify maintaining three schemas in my mind (E-A-V, columnar, 3NF).

  3. Re: New Ads – Mr. HISTalk, what is new about SCI Solutions? It still seems as gaudy and cartoonish as before. Healthcare starts with SCI? I don’t think so.

  4. Art,

    They are talking about how the data is stored internally, not just at the schema level. I think they are pretty clear that column oriented DBs beat the snot out of row oriented RDBs for things like data warehousing that store complex or loosely structured data that is mostly read. Row oriented will be faster for data is that has more writes, like OLTP systems. Which leads to their point, RDB’s are not appropriate for everything. Sure you can add an analytics package, but it is all still be stored in a RDB at the end of the day. So, you basically put the proverbial lipstick on a pig.

  5. Re: relational databases. I read the article and I have a few comments:

    1. The ComputerWorld article is quoting a guy who is, guess what, selling a “column-based” DBMS. So the original source is not exactly an impartial observer (not that I’m disagreeing with him, just keep it in mind). OOPS. rereading the comments, Art has already mentioned this.

    2. In your summary you mention “relational databases are legacy technology”, but the article is actually more focused on “the current crop of relational DBMS engines”. If you’ve ever visited http://www.dbdebunk.com/ (a fun read–and angrier than anything you’ll read, anywhere else) – if you’ve visited Fabian’s dbdebunk site, you’ll find out that the “relational model” and “the current crop of relational DBMS’s” are two different things.

    3. EAV model: gross. I’m shocked most of the time at how much of each hospital system is re-implemented every time. I wish we could just have a “base model hospital” template we could adapt so we didn’t have to store 90% of our data in EAV tables.

    I think everyone should be appaled at how much customization everyone does on their own systems.

  6. Couple of quick comments:
    1) Is that investment analyst that did the survey of Emdeon customers by any chance a sell-side analyst? If yes, well there is a conflict of issue situation here iif there ever was one.

    2) Jumping into the database discussion (disclaimer, I’m no database jockey), there is a new company over in the ERP market started by the founder of PeopleSoft, Dave Dufffield called Workday. They are claiming to also do away with the relational database model, via adoption of what they refer to as OMS, (Object Management Server). This is made possible via the hosted model for their solution, i.e., Software as a Service (SaaS). Though I have not personally seen a demo, a couple of analyst friends have and they say it is quite impressive. More found at: http://www.workday.com/what_we_offer/innovation/technology/index.php

  7. Costanza – Thanks for the clarification – in re-reading, you’re right. The only reason for the lipstick on the pig is to minimize the variation in the technology environment – that is, not having 20 vendors playing in one space. I am not saying it is right – just saying I think that is a major driver. Vendors don’t want to manage and support their integration with 20 sub-contractors either.

    p_anon – I believe there is differentiation in the data model (performance, flexibility, cost-effectiveness). E-A-V may be gross but (IMHO) with the environment changing (new tests added, new fields added to tests), some form of it is required so that every single aspect of care isn’t discretely modeled. The user experience and analytics are other differentiators as well.

    p_anon – you want a base model hospital template – develop it under open source creative commons and make the world a better place.

  8. re: base model hospital template: the open source/creative commons approach is exactly what I want. I think this could be a great way for the large players to distinguish themselves from the competition–“hey look, we’ve got a thriving user community and lots of examples of how to make our products work.” The point is, they can provide a base template (hopefully) without assuming the implied liability if/when errors are found in the base template.

    So I’m basically agreeing, but throwing the onus back on the vendors themselves. They’ve got loads of money, and they’re already doing this on a small scale (see MEDITECH’s NPR report repository). If they provide some full-time staff to work on populating (and marketing) their ‘community templates’, this could REALLY provide some value.

    What I’m getting at is that it’s in the company’s best interest to do this anyway, even if it costs them.

  9. Re: EAV, I just plain don’t like it. My irrational fears developed after working with it on a small scale 🙂 Call it a “battle scar”







Subscribe to Updates

Search


Loading

Text Ads


Report News and Rumors

No title

Anonymous online form
E-mail
Rumor line: 801.HIT.NEWS

Tweets

Archives

Founding Sponsors


 

Platinum Sponsors


 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Gold Sponsors


 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Reader Comments

  • Lets Be Real: Canula -- That is a lot of Monday Morning Quarterbacking.....all the language tied to HIT100 was not about Twitter netwo...
  • HeDothProtestTooMuch: You didn't need to use so many words to try and explain HIT. And then you also didn't need to conflate HIT with social ...
  • Canula: There's nothing wrong with the HIT100 as a marketing-style measure of twitter reach and nothing more. We have plenty of ...
  • Emperor Disrobed: Thanks for confirming that the HIT mutual masturbation choir is really just preaching to other since "there are very few...
  • BlockedByGurus (Steve S): Where's the methodology that @HeDothProtestTooMuch references and the 'scoring tool' that @HIT Realist notes? I'm not s...

Sponsor Quick Links