Friday, June 08, 2007

External oversight needed for hospital EMR implementation?

I read a recent AP article here about one hospital organization of appx. 600 beds and associated clinics, Lancaster General Hospital in Lancaster, PA, that plans to spend $60-100 million on an EMR. That figure seems stunningly high.

This reminds me of my earlier post here about another organization in California that spent that level of cash and had little to show for it except perhaps controversy. From that story:

Internal documents show the [UC Davis clinical IT] 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.

Now we have this in Pennsylvania:

... [Lancaster General Hospital] plans to invest $60 million to $100 million in electronic medical records, another step toward transparency. Marion A. McGowan, executive vice president and chief operating officer of Lancaster General, said electronic medical records are a vital step toward giving patients more control over their care. Primary care physicians would be able to access the information during office visits, and the patients themselves can see their information from their home computers.

The other myth I keep seeing is how EMR's and PHR's ("personal health records") will "empower patients" and "give them control over their care." I've seen no evidence of that and do not understand the arguments behind this claim, but such statements do have "punch" and make a great marketing and "selling point" - which brings up another issue:

I am becoming increasingly concerned about whether the push for national-scale EMR's has created a "gold rush" where prospectors of varied honesty and ability have set out to 'strike it rich' -- by sucking every dollar they can out of an already strained healthcare system under pressure to 'revolutionize care' through computerization.

It seems like health IT is entering a a "pre-Flexner report medicine" or perhaps "Roaring 20's" phase. The UK's CfH national EMR initiative seems to be a few years closer to the "Great Depression" than ours; e.g., see "£20bn NHS computer system doomed to fail ". I am concerned we could be headed down the same path.

At the same time, I see schizophrenic articles such as this ("Hospitals oppose plan for uniform reporting")
where the senior vice president for policy and regulatory services at the Hospital and Healthsystem Association of Pennsylvania, an organization representing hospitals, claims that "[computer reporting on hospital infections] is very costly", in fact too costly for small hospitals to implement for participation in a single statewide system.

As a physician from Philadelphia whose father died of a hospital-acquired infection, a medical informatics specialist, and a staunch advocate of clinical data initiatives to improve quality of care, I am highly troubled by such statements.
Such a central state infection rate reporting system could be implemented, for example, by just a few competent IT people and accessed by hospitals via the Web on commodity PC's, transcribed from paper by hospitals if they have to. In fact the hardest part would be development of a standard, agreed-upon state dataset and data definitions, not technology costs.

Such technology is, in fact, dirt cheap in 2007.

I think it more likely that hospitals and their cost-cutting management fear what an integrated, consistent reporting system might make them do - clean up their dirt, literally.

Reducing infection rates is the expensive issue, requiring better facilities management, enforcement of handwashing, etc., but ultimately the benefits outweigh the financial and human costs by several orders of magnitude.

Hence, resistance to such simple technology as a central infection registry at a time when we can send megabytes of information around the world effortlessly in seconds, and when powerful PC's and web technology are commodities, is absurdist at best.

My concern here is that hospitals would rather make such claims to avoid participation rather than perform the truly expensive physical and process cleanups needed to reduce infections.

It costs little to do little. The ultimate economy in healthcare, after all, is death.

Health IT is an industry that has a bit of a split personality (sorry for all the metaphors today). I believe it needs far more transparency that it has, along with other healthcare sectors (nursing homes and pharma come to mind).

As provocative statement of the day, based on an ever growing number of anecdotes and case examples of health IT difficulties, failures, consultant overuse, etc., I am increasingly of the belief that (at least in public hospitals), strong community, external stakeholder, or other oversight is needed for health IT projects.

I increasingly do not believe most hospitals are competent, disciplined and well-managed enough to accomplish enterprise EMR implementation
by themselves without large amounts of overspending, inefficiency, and waste (if not outright failure), and that outside oversight may be essential to ensure these tasks are performed in a manner that makes the best use of limited resources.

-- SS

6 comments:

Anonymous said...

As a seasoned RN having just spent the last 12 years in health care IT at a large 900 bed hospital working with Clinical and Business Applications and as a Network Director I can tell you that it IS that expensive. I refer to it as the "rape and pillage" of health care by corporate greed.

In the selection process for an EMR implementation, we discovered that a major hospital in Kentucky set aside 75 million for their EMR implementation 3-4 years ago.

When you spend almost 20 million dollars on hardware and software alone--not including a PACS, an OR suite, active/passive tracking, a viable ER product, a viable ambulatory product, cardiology services products, lab products, etc--it is easy to see how the large consulting dollars and implementation dollars feed into the cost. Expand that over a 3-5 year timeframe. They will say it can be done in 18-24 months but that is not true unless you slam dunk in something that is not tailored to the need of the organization and the workflow of the people that work there.

Then add on top of that hardware refresh and upgrades--due to the time lapse.

When a major lab software vendor charges 7-14K for every interface regardless of the fact that they have done the same interface for other facilities multiple times....

When the data is now forever and terrabytes are required and more....the storage industry is having a field day with hospitals.

When you calculate in that the real value of software is not in the software itself but rather the workflow it should support so--the workflow and workflow analysis is key to success. That is huge. Now you have major assessments--much needed--and costs associated with that AND the change management and reassessments required.

It just is incredible. In my mind, the software should be a subscription model so it is operational and the cost should be negligible. The real value is in the workflow and best practice support. We need standardization of data. We need control of costs. We need to unify instead of segment health care....find a way to fight against the self interest of one organization and focus on the interest of the patient, the community, the nation and the world.

Okay...off my soap box.

I am now working in a company that is focused on the patient...meeting that need and doing it cost effectively for the organization. I want to make a difference. The company wants to make a difference. That's the only way I know to impact this tidal wave of greed...one step at a time.

MedInformaticsMD said...

That is a sobering response.

Yes, I imagine we in healthcare have gotten accustomed to rape and pillage by corporate greed.

On top of all these cost issues in health IT is the fact that most of the CIO's and healthcare IT workers I've met are comfortable with this status quo, and at the same time unqualified to be leading EMR projects. I've also worked with some who are, to speak plainly, incompetent, not to mention mean-spirited.

Several of the cases at my site on health IT failure are mine personally.

Anonymous said...

Informatics 101 will tell you that the absolutely cheapest part of any EMR implementation is the computer hardware. It isn't even a significant component of the total cost. It's background noise. Usually the hardware isn't even purchased from the vendor. The most expensive part is personnel.

As to what vendors charge, that's open for discussion. From my perspective, part of the problem is that they often sell vaporware to hospitals and develop the systems during implementation, for whch they charge consulting fees. They are also getting free labor from the hospital staff who often put in a lot of work designing the clinical workflow. These are fruitful areas for exploration.

Our hospital purchased the XYZ System about 5 years ago with a complete order entry system. It turned out that the order entry system was unusable because the pharmacy system hadn't been developed yet. (Minor detail). The implementation is about 4 years behind schedule. I'm sure it's not a unique occurrence.

It's as if you bought a new car and when you picked it up from the dealer, you had a chassis, an engine in a crate and a pile of blueprints and you needed to work with the dealer for a year or two to configure the car before you could drive it. It's that absurd.

MedInformaticsMD said...

Anonymous wrote:

Informatics 101 will tell you that the absolutely cheapest part of any EMR implementation is the computer hardware.

Actually, the informatics courses I teach are at the 600 and 700 level, but let's not go there at this moment.

Some have commented privately that the problems are indeed not hardware costs, but all the other processes, people and brainpower, and that stories of failure are "anecdotal." My response is as follows.

I did not write "In an era when computer hardware has become "dirt cheap," why does it cost such an incredible amount of money to implement a hospital electronic medical record?" Instead, I wrote this:

At the same time, I see schizophrenic articles such as this ("Hospitals oppose plan for uniform reporting") where the senior vice president for policy and regulatory services at the Hospital and Healthsystem Association of Pennsylvania, an organization representing hospitals, claims that "[computer reporting on hospital infections] is very costly", in fact too costly for small hospitals to implement for participation in a single statewide system.

As a physician from Philadelphia whose father died of a hospital-acquired infection, a medical informatics specialist, and a staunch advocate of clinical data initiatives to improve quality of care, I am highly troubled by such statements. Such a central state infection rate reporting system could be implemented, for example, by just a few competent IT people and accessed by hospitals via the Web on commodity PC's, transcribed from paper by hospitals if they have to. In fact the hardest part would be development of a standard, agreed-upon state dataset and data definitions, not technology costs.


Building a focused system in one domain is not extremely difficult in an era when hardware is indeed dirt cheap. It is a weak excuse at best to not develop a statewide registry of hospital infections.

OTOH, in addressing the issue of "In an era when computer hardware has become "dirt cheap," why does it cost such an incredible amount of money to implement a hospital electronic medical record?", some comparisons for context are relevant:

$100 million begins to approach the entire annual budget for the pharma division I worked in, Research Information Systems in a multinational pharma's Research Labs, supplying research computing and biomedical information management of proprietary research information and clinical trials data to 10,000 R&D employees (among others). As a group director I had an annual capital and operating budget of appx. $13 million, which was one of the higher budgets in the division due to the expense of scientific information assets. We did not receive academic discounts from the aggregators and vendors, and the salaries of 50+ information and IT professionals were also supported via this budget.

From another angle, $100 million could pay for the contruction of a full service 100-bed hospital (~ $1 million/bed), or a larger addition to an existing facility.

I simply cannot easily accept that information systems on which the vendors have been working for years should rise to that level of expenditure, and that there is little critique of such costs from the medical informatics community.

Regarding "successes", in my own experience as CMIO implementing enterprise EMR at a 1,100-bed two-hospital system, I saw implementation errors being made or almost being made that ran - or could have run - costs up without causing the system to "fail", often times due to workarounds, and my own efforts to steer the implementation team around these issues. I was, however, hired in part if not primarily to resuscitate one subspecialty system that had failed, and failed badly.

I believe one major problem is that there are no well-accepted metrics for what at EMR should cost, how efficient and cost-effective an implementation has been, and what constitutes "success" among other factors. In other industries I believe such a situation would be deemed unacceptable.

I believe health IT vendors have indeed been put in a very favorable position, with multiple stakeholder groups demanding EHR's and little true competition in the cost sense. I believe they could likley be taking advantage of this in terms of pricing.

Where's the pressure for critical evaluation of the industry? People complain about costs of healthcare, and demand clinician reimbursement and pharmaceutical costs be constrained. I see no such clamoring in the health IT industry. Why? At some organizations, there is beginning to be backlash (eg UC Davis). However, the health IT industry seems to largely have a captive audience, where every wasted or padded dollar spent on IT takes away care from some child or poor person or raises the costs for everyone else.

This issue has real social repurcussions. I am on the dissertation committee of a PhD candidate in my college whose thesis revolves around gauging pre-implementation EHR attitudes from a social perspective, to identify opportunities for proactive intervention. It was to be deployed at her large medical system in Philadelphia, serving many underserved populations, a system that was readying rollout of EHR.

Her project had to be put on hold because the healthcare system decided it simply could not afford an EHR due to the costs. Several other area hospitals are similarly affected. What is evolving is a have/have not two-tiered scenario that feeds into healthcare disparities. I see it in my own city.

As for "anecdotal evidence" [of clinical IT implementation problems, not just of outright system implosion], those anecdotes are so plentiful from so many sources that a number of us in the CIS-WG are authoring a book aimed at CIO's and other IT personnel that shares lessons-learned from these anecdotes. The stories come from our health informatics students taking our courses who are working in the trenches, from practitioners, from our own observations, and from the press amongother sources. One can ignore a cornucopia of anecdotes at peril to one's real-world effectiveness.

I've spelled out my views in depth at http://www.ischool.drexel.edu/faculty/ssilverstein/medinfo.htm , my site on health IT difficulties.

I wish someone could convince me that the problems are anecdotal only. Yet, since the site's inception in 1999 and its existence since then, and many thousands of hits annually from many different countries, I get frequent feedback that I am correct in my assessments. I almost never get any feedback saying otherwise.

A recent example:

Dr Silverstein,

On your webpage, you ask for those who visit to let you know how they
found you.

John Haffty, President of Navin, Haffty & Associates, a healthcare IT
consulting firm, mentioned your page in his monthly newsletter that goes
out to several thousand MEDITECH users around the world. He praised
highly your willingness to investigate and publicize these failures,
given that very few of the industry publications would do so.

Reading through the site, there is obviously a great deal to be learned
from the mistakes others have made. I'm pleased to be able to say that
my facility has managed to avoid most of them. A common theme I saw was
an unwillingness by IS staff to really listen to what their end-users
really needed. I'm currently working on a presentation entitled "No Man
Is An Island: Setting Aside Traditional Divisions of Labor to Improve
Workflow In an Integrated Information System", to be given at the
Western US Regional MUSE Conference ( www.museweb.org ). I'm a firm
believer that communication, in all directions is an invaluable tool,
and your case studies seem to bear that out.

Thank you very much for your work in the healthcare informatics field.
We may be new, but we need all of the genuine research we can get to
make sure we're all headed in the right direction.

Beth

[name witheld]
Information Services Project Manager/Team Leader for MEDITECH & Clinical
Applications
[name] Regional Health
[city], KS


On another angle about anecdotal evidence, I once was a Medical Programs Manager for a regional transit authority. There were anecdotal stories that subway operators were using drugs and racing the cars underground on long straight stretches of track for a "thrill." We in Medical were not permitted to act on "anecdotal evidence" mainly due to labor union opposition to any form of drug testing, rising to the level of personal threats (I myself was threatened to have my head blown off by an irate operator subject to drug testing.)

In 1990 or so, a huge underground accident due to racing of a train on a long underground stretch led to hundreds of injuries and several deaths. The toxicologist at Smithkline told me the cocaine metabolite level of the operator "blew the lid off the GC/MS." The operator involved had been reinstated after rehab as part of a labor agreement depsite a history of substance abuse because evidence of accidents uafter doing so was, at that time, anecdotal. We had no EMR or computer system to prove otherwise

The agreement called for the employee NOT to be tested on return to work but to be subject to random testing. I put in writing my "against my best medical judgment" opposition. Further, the medical dept was not permitted to use its own computer to select testing dates because there were anecdotal stories of medical dept. personnel "sabotaging" safety-sensitive employee's jobs through bias in date selection (yes, I did not understand that claim either). So, the date selection process was left to MIS.

MIS simply forgot to enter this employee into the random testing program after reinstatement, and the accident occurred only five or six months later. The employee had not been tested in the interim.

There was a claim of mechanical factors contributing to the accident, however, there was no doubt the operator was racing the train and that he was likely impaired. Innocent people died because all the evidence up to that point about hazards in this situation were "anecdotal."

I do not lightly dismiss cornucopias of anecdotes, especially where IT is involved, any longer.

Finally, regarding health IT failure, some "anecdotally" believe £20bn national health IT efforts of entire countries may be doomed to fail.

I rest my case.

Anonymous said...

Lancaster General is actually an impressive healthcare organization - look in US News and World Report from 2005 - they were listed for something like six diseases.

MedInformaticsMD said...

Anonymous said...Lancaster General is actually an impressive healthcare organization

I don't doubt it. However, I don't think Information Technology is the core competency of any hospital.