Medication Safety: The Right Stuff
January / February 2007
Medication Safety
The Right Stuff
There isn’t any substitute for good clinical judgment.
Generally speaking, the term “decision support” refers to the functionality in computer systems that analyzes data and provides recommendations based on agreed-upon protocols. In medication delivery, clinical decision support refers to the imbedded functionality in all clinical systems that analyzes patient-specific data and provides alerts based on already established and accepted drug and treatment protocols. One would think clinical decision support functionality should be a “no-brainer,” or so it seems to the academics who develop the rules and alerts that govern such functionality in clinical systems. Why wouldn’t clinicians wildly embrace tools that help prevent errors?
In fact, many times during my consulting career, healthcare provider organizations (hospitals and physician clinics) have told me that they need to replace their clinical systems so that they can aggressively move into clinical decision support. I listen with “tongue in cheek” because these desires are being expressed from having heard others talk about how wonderful rules and alerts-based clinical decision support have been to their organizations. What a great opportunity to develop guidelines and suggestions to recommend drug and other medical therapies to standardize our care delivery.
The Problem
There are still many clinicians using departmental applications intended for a single purpose. Examples include the laboratory and pharmacy applications in the hospital, the ambulatory pharmacy system, and the EHR system in the physician’s office. Each of these may have clinical decision support capabilities that work well for the intended purpose in each of these areas, as long as they are kept current. (More on that later!)
How often do you think the outpatient pharmacists reading this column review the most current INR result when they are asked to refill a warfarin prescription? Sure, the original was filled only 10 days ago and says, “May be refilled one time,” and there was nothing in the patient’s medication profile that said this was a problem. Or, how about the orthopedic clinician treating a patient for an ankle sprain who enters an order for ibuprofen 600mg TID into his e-prescribing system? The system agreed that the indication was appropriate, consistent with payer guidelines, the drug is indeed available in a 600mg dose, and TID makes perfect sense. But the system never told the orthopedic physician the patient was prescribed omeprazole at last month’s physical. The clinical decision support functionality in each of the above systems worked well, but something was missing.
Clinical decision support? The functionality works depending on the definition and expectations we apply to the term. Clinical decision support in the ambulatory pharmacy and physician practice world works well; the functionality checks for dose, interactions, indications, availability, and payer guidelines. But our sense of security is false if we don’t have access to all the available data about our patients. There isn’t any substitute for good clinical judgment.
Clinical decision support? Without full integration of all available data? Like playing with a deck of 51 cards, or driving a car with a flat tire…
Are We Over-simplifying?
I have heard hospitals and larger multispecialty clinics say, “We’ll just buy an integrated system and we won’t have this problem.” While that may be true, no one can guarantee your patients will only see your affiliated clinicians, have all their prescriptions filled, and only buy their OTC medications from your affiliated ambulatory pharmacy.
More important to the use of these tools is the governance structure to recommend and implement changes to keep the clinical decision support rules and alerts up to date. Evidence that impacts clinical practice is being published daily. How do we assess these research studies and incorporate them into our rules? How often do we meet to do this? Who do we get involved to do these evaluations? What do we do with exception reports that list all of the rules being ignored and the offending clinicians? How do we get agreement across our entire clinician base?
Where do we start? And we thought integration of all data elements was the only problem?
As a practicing clinician, I can only think about how many times those darn alerts pop up with recommendations about one thing or another. I can also think about how many times I type in “bypass” after reading the alert and quickly deciding it was really intended for the next patient.
And I’m the one who understands this phenomenon and teaches it to other clinicians? I can only think of my associates who don’t like computers to begin with and really become angry with the number of times they have to type in “bypass.”
Managing the Expectation for Success
I will be among the first to admit that the correct use of decision support tools helps enhance quality and reduces the possibility for errors. However, incomplete feeder systems or poor integration of data elements doesn’t help anyone. Let’s not be lulled into thinking we are infallible just because we have decision support capabilities in our clinical systems.
While the implementation of these systems is becoming an expected standard of practice, all provider entities need to understand and accept the requirement to keep these systems working at the optimal level expected by practicing clinicians. At the same time, all clinicians need to recognize the limitations of these systems and work with their respective organizations to truly make decision support functionality a tool that augments the delivery of quality patient care and not just another “nice to have.”
Larry Pawola is associate professor in the College of Applied Health Sciences at the University of Illinois at Chicago. He also is president of his own healthcare information technology and clinical services consulting firm, Lincolnshire Consulting Associates LLC. As a respected industry consultant for more than 25 years, Pawola has worked with a variety of ambulatory clinics, community hospitals, and academic medical centers. His work has focused on assessing clinical systems needs, operational improvements, strategic planning, and education. He has also consulted with healthcare companies for the strategic positioning of their technology products. Pawola is a member of the Editorial Advisory Board for Patient Safety and Quality Healthcare and may be contacted at lpawola@uic.edu.