How to improve computerized physician order entry

Computerized physician order entry (CPOE) is being rolled out across our nations’ hospitals. The old days of written, and often illegible, orders from doctors are fast becoming a thing of the past. The potential for this measure to improve patient safety and transform medical practice is unquestionable.

As a physician who has worked in several different hospitals since finishing my residency, my time in practice has coincided with the new age of technology in medicine (the iPhone was released as I was finishing my residency). Yet one common theme has been present in all of the hospitals I’ve practiced in. Despite all the promising technology, the computerized order entry systems that have been introduced have largely been slow, tedious, and difficult to work with.

Having been intimately involved in CPOE implementation myself, I cannot help but feel slightly disappointed that it hasn’t lived up to expectations. The process can only be as good as the final infrastructure allows it, and unfortunately the implementation is frequently happening on suboptimal software platforms. It would be akin to planning to a great traffic system over roads that are completely broken and don’t allow the cars to go at the desired speed. The wrong way round to do things! There’s a large dichotomy between the idea and the technology infrastructure. And to be fair to health care organizations, policies such as meaningful use have meant that hospitals have been forced into a corner of needing to roll everything out quickly in order to receive significant and much-needed federal incentives.

From my perspective as a hospital medicine physician, let’s go through a typical everyday scenario. This has happened in every hospital I’ve worked in, and is not unique to any one institution. The nurse will request an order from the doctor, who may need to drop whatever they are doing and find a computer to log into (typically during the day, a free computer can be hard to find).

Logging in will take several seconds, and booting up the system several more. Then we have to deal with a less than optimal interface, multiple clicking will be involved, typed data entry using a keyboard, then a password. Following this, a few seconds to make sure the order has gone through. This applies to something as simple as ordering a Tylenol. A complex scan or test that requires some text entry to make the order clear, can take minutes.

Why is order entry so cumbersome? It’s often quicker to order an item online than it is to enter a simple order into a medical system! And whenever we do this it takes time away from our patients in an arena where time is already scarce. A recent study published in the Journal of General Internal Medicine revealed the disappointing statistic that medical interns now spend only 12% of their time in direct patient care, and up to 40% in front of computers. Slow order entry undoubtedly contributes to this.

And it’s not just doctors who are affected. The situation applies to everyone involved in the order entry and retrieval process. Everyday, I also see nurses frantically wheeling around their portable computers for most of their shift, reviewing and confirming orders, frequently spending more time with their screens than the patients who really need them.

In our new technological age, here’s what we need:

  • touch screens (undoubtedly the future)
  • quick password entry and rapid screen loading (we cannot afford situations where the software takes up to 30 seconds to load up)
  • minimal clicking or scrolling to get to where we want within the program
  • user-friendly interface.
  • ease of updating and modifying the system if problems are discovered after implementation (this is often very difficult to do, but a vital need for any health care institution)

Many of our current systems utilize older platforms, require constant use of the mouse, and have a pretty terrible user interface. Cars, planes, our home devices — the mouse is on the way out, so why do we use them so much in health care? Other technologies have progressed rapidly, yet the health care industry finds itself behind the curve. Our computerized systems may be complicated and have a stringent need for security, but they are still too slow compared to where they should be. I once spoke to a physician administrator who was involved in implementing a hospital IT order entry system, who told me that it wouldn’t necessarily be designed to save the doctor any time. This seemed to be taken as a given. It shouldn’t be.

Computers and technology need to be optimized with the workflow of frontline physicians. Let’s view the clinical interaction with the computer as a triangle: the doctor, the patient, and the technology. Too often, it is a straight line with the computer coming in between the doctor and patient. This also means that the patient will not feel the full benefit of the new technology.

I’ve witnessed many IT project teams that will contain one physician who hasn’t practiced medicine in several years, with very limited knowledge of IT. I’m sure the idea of involving these physicians is well-intentioned, but how can someone who doesn’t see patients possibly be best placed to design a system for practicing doctors? Whether it’s order entry or data entry, we need to have frontline physicians with knowledge of IT at the forefront of software development, completely involved in every stage of design and implementation.

Well before rolling out any new system, give other hospital clinicians a chance to test it out and provide feedback. To use Apple as an example, the initial designers were all users of the product they were creating, who understood the need for providing a good end-user experience. Successful products like the iPhone were the result of their endeavors. We need to adopt the same philosophy for health care technology.

Physicians should not be required to stop their daily workflow for anything longer than a couple of seconds to request orders. The final aim should be for the process to take not much longer than simply speaking. This applies equally to when the physician is entering any other orders, such as admission or medication instructions.

Hopefully in a few years, this debate would have moved on when we’ve finally got some better systems in place. It’s probable that some of them are already under development — but we needed them yesterday, not in ten years’ time. A handful of hospitals are already getting there, and enable order entry with a few touches of a smart screen device. That’s the way forward. Medical professionals must not be hindered by the computers which are supposed to help them. The successful health care IT of the future will be the software that enables the doctor or nurse to spend maximum time with their patients. Collectively we must do better, and doctors, IT experts and hospital administrators need to all work together to achieve this goal.

Suneel Dhand is an internal medicine physician and author of Thomas Jefferson: Lessons from a Secret Buddha and High Percentage Wellness Steps: Natural, Proven, Everyday Steps to Improve Your Health & Well-being.  He blogs at his self-titled site, Suneel Dhand.

Comments are moderated before they are published. Please read the comment policy.

  • azmd

    These are all excellent points, but you are missing the biggest point of all: as long as clinicians (doctors and nurses alike) volunteer their own time to accommodate for the inefficiencies of a slow/clumsy/non-intuitive CPOE, administrators have ZERO incentive to spend the huge amounts of money it would take to properly purchase and implement a user-friendly CPOE system. As long as hospital administrators are not demanding a user-friendly CPOE platform, the private sector will not develop one.
    There is NO MARKET for such a system, because clinical workers are not part of the market, and we won’t be, until we go on strike, or organize ourselves into unions, or take some action that has the potential to present more of a cost/inconvenience factor to the system than the opportunity cost of bringing in a new CPOE.