A recent vendor’s assessment report and advisory for a local hospital that is midway through an EMR implementation illustrates some reasons why EMR projects fail, and how common the most common mistakes are.

1. Lack of coordination. Wait, isn’t that what project management is all about? Coordination? The vendor’s report reads, “It appears that there is little oversight or accountability in the review of builds against the approved design decisions.” I’d say that’s a pretty serious lack of coordination.

2. Inadequate testing. Testing is pretty important when you are developing a tool that if poorly designed can cost a patient their life.  Typically in EMR work, testing is done at three or four distinct points in development. If those tests aren’t done or are poorly done, the finished product is defective, and patients welfare is at risk. And as this vendor pointed out to its client, it also results in costly rework and user (doctor and nurse) dissatisfaction.

3. Ignoring important information. Vendors typically send out informational bulletins about their products, as well as software updates, which the hospital or other provider’s technical staff needs to look at, analyze, and act on. In this case, critical bulletins were either ignored or not properly acted on. Software updates were late or improperly applied, resulting–again–in less than optimal functionality of the system, and–again–dampened satisfaction among the clinical users. Again, delaying the updates also can result in costly rework later on.

4. Making the product more complex than necessary. Technical people are prone to use a gadget or a software configuration option just because it’s there. Using a particular piece of functionality may not make things easier for the end user. It may actually make things more difficult, and may scuttle a workflow that was working fine before the new technology was introduced. My favorite Einstein quote is “Make everything as simple as possible, but not simpler”. That little piece of wisdom should be tattooed on the arm of every CIO and IT analyst in the nation.

5. The physicians (and other clinicians) do not have enough input in the process. This particular vendor’s report calls this point repeatedly. “We recommend these decisions be reconsidered with input from end users, including physicians.” This recommendation several times. As well as the physicians, all the end users–nurses, pharmacists, dietitians, and therapists–need to be consulted, and the work that they do has to be thoroughly understood by the people that are configuring the software. Whatever it takes to facilitate this communication has to happen before the installation manual is even opened.

6. Best practices are rarely followed consistently. Everyone has days when they can’t follow the simplest instructions. Bad hair days, fuzzy Fridays, call them what you will. Your implementation plan, at a high level, has to account for the human factor in everything you are trying to accomplish in your EMR implementation. As part of your design, you have to design reliable feedback and quality control mechanisms that will get you back on track if you should divert for one of these human reasons.

We’ve been doing EMR implementations long enough now. there are volumes full of best practices. Why do we still have little hospitals trying to reinvent the wheel? And it’s not just little hospitals. Big famous ones, like Cedars-Sinai in Hollywood, routinely miss human error in the planning stage, and it haunts those projects and those humans all through implementation.

Too-tight timelines, poor communications between departments and work groups all can spell disaster at any time, and safeguards aren’t designed in, time and money is lost, and–back to item 2.–patient’s lives can be endangered.