Why we need more data professionals employed with health care providers rather than health care consultancies.
I'm a member of a health care provider industry association that shares information across organizations about our experiences delivery business intelligence and data warehouse solutions. Recently, one of the members of that organization posted a forum question about several packaged industry-specific data warehouse solutions that they'd had pitched to them. Here was my response. I thought it was worth sharing here as well.
My personal experience (as a BI/DW consultant and professional for the past 10 years) is that prebuilt data warehouse solutions for any industry come in three flavors:
- Logical-only models that come with an implementation "project." The physical implementation will be built to suite your organization and usually use language and terminology that your organization is comfortable with. These are full blow projects. Though some ideas come in the box, very little runable code is ever included inside the box. That takes the time and money of professional services.
- Canned data warehouses / data marts that come as a full blown, predefined solution. These might seem like a good idea because so much is already built for you. In my experience, though, they return only a fraction of the business value they seem to promise, though. The model in the canned solution doesn't typically match your own business model closely enough to be as valuable as it appears during the sales meetings. Another challenge with these solutions is the integration of data into them. They're either so high level that they don't have a lot of value; or they're so different from your own organization that the mapping of data into their model is overly complicated.
- Marketing hype - in some cases, the solutions really are just a marketing message. In reality, the vendor is merely saying "We have some people from health care and we have some BI skills. We've put them together on a team for you."
That's not to say vendor solutions aren't a reasonable way to go for you. When you talk with them, be very specific in your questions and very critical. Ask to talk with other client references that are similar in size, systems, and market to you. Do a proof of concept to implement some small portion of the solution free of charge.
I hope I don't sound too harsh on vendor solutions in that response. I don't believe that vendors intentionally oversell their solutions. In some cases, they're ignorant as to what it takes to really build, deploy, (and most often) run, maintain, and enhance a data warehouse for several years after the initial implementation. This was one of my personal realizations as I transitioned from consulting into a corporate job. Other vendors may simply being putting the cart before the horse, and using consulting opportunities to build up their solution offering. No harm in that as long as they're upfront and honest about it.
On December 30, 2009 HHS presented the much anticipated details of the definition of meaningful use of electronic health records. This definition provides more information about the level of functionality and use health care organizations must achieve by certain deadlines in order to receive certain types of government assistance and maintain the highest levels of medicare reimbursement. (I have not read the details.) Soon there after @theEHRguy tweeted a couple of great predictions:
I think it would be better if we paid those of us who actually work for health care organizations $450/hr. Unlikely.
Still, I think he makes a very good point about consulting services, and I'd extend that to certain types of packaged vendor solutions by nature of the services they usually imply. The high demand in the area of health care systems right now is going to be shortly followed by a high demand for BI/DW solutions in the same space. In the rush, I think the industry risks implementing poorly designed packaged vendor solutions will result in a lot of wasted time, money, and effort; and lost opportunities for growth and optimization. Time will tell, but I think it's a long road ahead.