Patient Safety Tip of the Week


December 16, 2008

Joint Commission Sentinel Event Alert on Hazards of Healthcare IT         






Joint Commission released is most recent Sentinel Event Alert last week, pointing out the many patient safety dangers actually resulting from technologies introduced to improve patient safety. The report comes just as President-elect Obama is about to make a big push for healthcare IT in his new economic stimulus package and as CMS makes a big push the get all physicians to use e-prescribing.


We, of course, are strong backers of the use of information technology in patient safety. But regular readers of this column have seen us frequently talk about the unintended consequences of many of the information technology strategies and tools.


Lessons from other industries have provided ample warning and examples that introduction of new technologies introduces opportunities for new types of errors and other unintended consequences. Charles Perrow in his classic book “Normal Accidents” (Perrow 1999) talks about how new technologies often simply push the envelope, citing as an example how the introduction of maritime radar simply encouraged boats to travel faster and did little to reduce the occurrence of maritime accidents.


Medicine is no different. Our June 19, 2007 Patient Safety Tip of the Week “Unintended Consequences of Technological Solutions” gave several examples of new healthcare technologies leading to unexpected consequences that harmed patients. We gave examples of a new dual-power source ventilator failing to alert staff that it was using battery power, a telemetry transceiver transposition, a case where a new bed alarm displaced the nurse call button, and cases of unintended consequences of CPOE (including a well-known case where neuromuscular blocking agents ordered remotely by CPOE on a patient not on a ventilator led to a patient’s death).


The latter case led to our first discussion on unintended consequences of healthcare IT - CPOE Unintended Consequences – Are Wrong Patient Errors More Common?  In that discussion we noted multiple factors that may make it easier to order on the wrong patient using CPOE rather than paper. In June 2008 “Technology Workarounds Defeat Safety Intent” we focused on barcoding system workarounds and discussed Ross Koppel’s article identifying the multiple types of workarounds in barcode systems and their underlying causes. They identified 15 types of workarounds and 31 types of causes for the workarounds in barcoding medication administration systems. In August 2008 (Pattern Recognition and CPOE) we talked about how CPOE often removes the pattern recognition that nurses and pharmacists and physicians utilize when they look at a set of orders in toto and how that might lead to new types of unexpected consequences. And in September 2008 (Less is More….and Do You Really Need that Decimal?) we discussed misprogramming infusion pumps.


The new Joint Commission Sentinel Event Alert was written by multiple health information technology experts with diverse clinical backgrounds. The Alert begins with some statistics about adverse events arising from factors related to information technologies, some coming from prior Joint Commission sentinel events and others coming from the MEDMARX® data base. It then discusses some of the factors contributing to adverse outcomes and provides practical guidelines in the planning and implementation of new technologies.


It places a real emphasis on the human-technology interface and clinical workflows. Actual or perceived roles often change substantially after introduction of a technology like CPOE or barcoding bedside medication verification (BMV). Frustration and even resentment at these changes is common. It discusses the tensions that often arise between physicians and nursing staff, the perceived loss of autonomy by physicians, “alert fatigue”, and problems with the technology itself.


It provides a good list of guidelines that organizations should follow in implementing new technology. That includes involving all end-users in the planning process, examining workflows, assessing technology and hardware needs, training/education, extensive testing, continuous monitoring and ongoing maintenance, communication and oversight, judicious use of standardized order sets and alerts/reminders, minimizing distractions during use of the technology, and ongoing surveillance and error detection.


One interesting phenomenon we have seen regarding the frustration around these new technologies is that the frustration is on the “upfront” end of the processes. That is, for both CPOE and BMV there is usually actually an increase in time and work done early in the process. It is counterbalanced by considerable time savings on the “back end”, i.e. fewer phone calls for verification of orders, etc. That is why it is extremely important to capture the true time, effort and workflow involved in a whole “episode” of care. It is the whole episode that usually becomes more efficient and more reliable.


We often find that healthcare organizations fail to include an adequate measurement strategy in their decision-making and planning for new technologies. And even when they do, they tend to focus too heavily on measures of the technology itself rather than on the more important clinical outcomes. And every such measurement strategy must include surveillance for unintended consequences. And, since many or most of those unintended consequences may be unexpected, the only way to measure them is to go out and look for them. That means direct observation of the various processes and interviewing staff using the technologies on patient safety walkrounds. And sometimes you have to resort to low-tech methods for your measurements. For example, determining how long it takes a physician to do order entry may not be accurately captured electronically. You may have to go out and use a stopwatch to capture the true time required for order entry and other processes that are part of the workflow not only for physicians, but also nurses and pharmacists.


A good measurement system looks at the impact of the technology from multiple different perspectives. It does need to include measures of the technology itself (eg. system availability, downtime, mean order entry time, % of alerts approved, etc.). But it also needs measures of end-user satisfaction (physicians, nurses, pharmacists, clerical staff, etc.), measures of hospital and staff efficiency (eg. lab/rad/pharm  turn-around times, # of phone calls for clarification of orders,  etc.), measures of clinical outcomes (eg. CMS core measures, Joint Commission standards, P4P measures, SCIP measures, impact on nosocomial infection rates, etc.), unexpected consequences (see above), and financial impact. All too often the set of measures is determined only after the fact. Any good project develops its measures during the planning/decision phase. After all, the three fundamental questions of performance improvement are (in order): Where do I want to be? How will I know that I’ve gotten there? How will I get there? When making decisions about technology-related projects it’s easy to jump to the last question. However, if you take the time and effort to answer the second question and set up the measures by which you will assess success, you may realize that there are other (low-tech) methods to achieve some of those goals.


The authors of the Joint Commission Sentinel Event Alert are a group that are strong advocates for use of technology to improve patient safety. The main theme here is to remind all that any solution (whether technological or instituting a new approach, etc.) may give rise to unintended consequences. Sometimes they can be anticipated, other times they cannot. Either way, careful vigilance for unintended consequences is necessary. Good planning and thoughtful implementation are important but even the best-laid plans often produce unexpected results.





Update: See also our February 2009 What’s New in the Patient Safety World column “Healthcare IT The Good and The Bad”.








Joint Commission. Sentinel Event Alert. Safely implementing health information and converging technologies. Issue 42, December 11, 2008


Perrow C. Normal Accidents: Living with high-risk technologies. Princeton, New Jersey: Princeton University Press, 1999









Patient Safety Tip of the Week Archive


What’s New in the Patient Safety World Archive