Thursday, October 12, 2006

$70 million for an Electronic Medical Records system?

Seen on HISTalk News (a popular blog in the healthcare information technology community) on 10/11/2006:

Geisinger Health System (PA) contracts with IBM to build a big clinical data warehouse. They called it "partnering," although I'm guessing that only Geisinger will be writing checks. Nothing interesting in the announcement, other than the little pearl that Geisinger's Epic project has cost $70 million so far and has reached "a quasi endpoint in how far we will get in simple decision support functions." Since they're calling it done, how about sharing the ROI and outcomes data?

Epic is an developer and vendor of healthcare information systems. This story is stunning if the facts are correct - or even near correct. A $70 million figure for a "quasi endpoint", and even more at another healthcare system (UC Davis, link), should scare the daylights out of any healthcare CEO. Just how much does it cost to build an entire hospital these days?

Internal documents show the [UC Davis] project -- with the final bill estimated to be anywhere from $75 million to $100 million -- is two years behind schedule and up to a fifth of the budget went to an outside consulting firm whose expense reports are now the subject of an internal UC audit.

The $70 to $100 million figures also seem concerning on their face. For example, these figures are far in excess of what it costs a trans-national company such as a pharma to implement major research information systems. It represents nearly 10 times the annual budget of my former department in one such pharma, that served the scientific information needs of 6,000 scientists and tens of thousands of others worldwide within the company.

It’s clear that metrics are needed such that a health system’s implementation can be benchmarked to others. What did that $70 million buy? Were the internal resources, expenditures, project management, use of consultants, etc. in line with other organizations? Were the cost/bed, cost/clinician, and other figures in line with what other medical centers experience? Was internal expertise adequate to properly manage the ubiquitous sociotechnical issues, i.e., the change-resistance and friction among all stakeholders including IT, administration and clinicians?

(I’m sorry, hospital CIO’s; your skills in management information systems don’t seem to work well when dealing with those issues in the “hiding in plain sight” invisible complexity of healthcare delivery environments.)

Finally, were vendor or software-created problems a cause of cost overruns, project rework, "peek-a-boo template delays" (a term used by a frustrated healthcare executive some years ago), false go-lives, end user disenfranchisement and other cost-increasing problems?

Without sharing of implementation difficulties and metrics, how can we ever really know if a healthcare IT project was implemented in a reasonable, cost-effective manner, even assuming it functions “successfully?”

What other industry is immune from accountability? Even clinicians are coming under greater scrutiny; a capability facilitated by clinical IT. Why is the health IT industry effectively excused from such scrutiny itself?

Of course, another issue raised in HISTalk that would prevent information sharing about inferior offering and best implementation practices -- that health IT vendors may threaten customers to stay quiet about problems-- is certainly borne out by my observations of almost complete lack of cross-institutional information sharing and publication about health IT problems.

Such a vendor stance makes perfect sense. If institutions don’t or can’t share, it means more business for the vendors – as well as the “remediation consultants” called in to clean up the various messes.

Unfortunately, healthcare doesn’t have the capital for clinical IT misadventures, and I believe when the issues become more public in this industry sector and information flows about mismanagement and abuses (as is happening in the UK ’s Connecting for Health project), the fallout won’t be pretty.

Addendum: on expensive "consultants", a common feature of the health IT landscape: the hospital at which I led the implementation of EMR and development of specialty clinical IT, Christiana Care in Delaware, spent some $500,000 back in the late 1990's for a similar Clinical Data Warehouse engagement with another of the large management consultant organizations.

This project was initiated by senior management without consulting with me, the Director of Clinical Informatics, or my staff. We had experience in this area ourselves.

For $500,000 I could have hired a team of people with specific expertise and built a functional, useful clinical data warehouse.

Additionally, if this had been done, at least there would have been an actual working deliverable, rather than just a pretty spiral-bound consultant's report which, at $500K, works out to about $250,000-per-pound. I'm not sure even weapons-grade plutonium is that expensive.

When healthcare organization leaders start to recognize that it's not OK for others to know their business better than themselves, and that it's not OK to believe it acceptable for others to know their business better than themselves, they will stop the massive waste on dubious external consulting engagements. (Assuming, of course, that there are not other motivators for such engagements, which as readers of HCRenewal know is not always the case).

2 comments:

EMRNurse said...

Good piece on Geisinger. 70 Million for an EMR. Reminds me of my experience with UC Davis Medical Center who also is implementing Epic. When the project started in 2002 they projected costs of 50-60 million. By mid 2004 it was 70 million. This spring when responding to an article on the project delays and bonuses to the CIO, Dr Tinstman, by the Sacramento Business Journal, said somewhere between 75 and 100 million. UCD is about 1/2 of the way through the project with order entry and the bulk of documentation not yet to be impemented in the hospital. The origional timeline called for all of the above to be completed by late 2005. UCD spent 17.2 million on Deloitte for consulting services from 2002-2005 that produced very little if any measurable deliverables and now UCD is trying to rebuild its team after a big turn over and is investing heavily in "remediation consultants". The EMR is under CIS and there is another IT department under different leadership that is even bigger if you can believe it. But who cares, its only public money and in academia timelines are of little importance.

MedInformaticsMD said...

Odd that Tinstman is in academia; when he was an executive at Cerner in the mid 1990's and I a faculty member of the Yale Center for Medical Informatics, I was unable to even secure an interview with Tinstman or Cerner.

I guess I didn't have enough experience.