It simply works

released 11.12.09

By Barbara Jewett

Common sense and simple, old-fashioned tools saved patients from pain medication related death in one study conducted at an Illinois hospital. Now researchers hope to translate what they learned into an expanded electronic program that can transform hospital practices across the country and significantly reduce medication errors and adverse reactions.

It's estimated that adverse drug events (ADE) cause more than three-quarters of a million injuries or deaths annually in U.S. hospitals and that hospitalized patients are subjected to an average of one medication error per day.

But a team of researchers, including Ian Brooks, head of NCSA's health sciences group, made an intriguing observation: prescriptions are medical algorithms. What would happen, they wondered, if software design principles and debugging methods were applied to improving the quality of prescriptions used in a hospital?

ADEs and medication errors are disproportionately severe when they involve opioids used to treat pain. One in four of all fatal ADE among hospitalized patients are associated with the use of opioids, drugs such as Vicodin, OxyContin, morphine, and fentanyl. So the team—which, in addition to NCSA's Brooks, was composed of nurses, pharmacists, physicians, clinical pharmacologists, quality assurance specialists, and statisticians from the University of Illinois College of Medicine in Peoria, OSF Saint Francis Medical Center in Peoria, and Jesse Brown Veterans Affairs hospital in Chicago—set their sights on reducing the rate of opioid ADEs.

"A prescription is a treatment plan. And that treatment plan is, in its most general sense, an algorithm," says Brooks. "So you can apply the same processes to a prescription as you can to a computer science program."

To test this hypothesis, the team created and debugged a "Patient-Oriented Prescription for Analgesia" (POPA), a pain management protocol that was followed for adult patients at OSF Saint Francis Medical Center in Peoria, Illinois, over a five-year period. The results were astounding: while there were seven severe or fatal patient ADEs in one month during the study's first year, as the POPA protocols were introduced and consistently followed this number steadily declined, reaching zero for the study's final six months. These results were published in Clinical Pharmacology and Therapeutics in 2008.

The initial POPA study concluded in December 2002. In 2003, the POPA protocols were implemented hospital-wide. Amazingly, as of February 2009, the last month for which Brooks has information, the hospital's severe or fatal ADEs related to pain management has remained at zero. That's right, not a single severe or fatal opioid ADE in the hospital in more than six years.

Back to basics

The techniques employed were fairly simple. Once anesthesia and initial surgical pain medicines were out of the patient's system, non-opioid analgesics, such as acetaminophen, ibuprofen, or ketorolac, were administered on a fixed round-the-clock schedule. This cut patients opioid requirements in half. The highly potent opioid fentanyl was then given through a subcutaneous catheter—avoiding the interruptions that can occur with intravenous administration of drugs. The fentanyl was provided via a patient-controlled analgesia pump, thus allowing the patient to have more control over the amount of pain relief.

Patients' pain severity was assessed on a fixed schedule with a visual-analog pain scale—a device that looked like an old-fashioned slide rule. Patients slid the marker to indicate their pain level, from no pain at the left edge to intolerable pain at the right edge. The side of the pain gauge the patients saw was plain, while the underside had numerical markings from zero to 10.

After the patient noted their pain level, the nurse would flip the gauge to read the number, record it on a paper POPA prescription form, and then follow the specified protocol for that pain level. A fingertip pulse oximeter was also used to record the amount of oxygen in the patient's blood. A falling blood oxygen level is an early sign of opioid-associated respiratory depression.

In addition to including explicit instructions for pain medication dosing, pain severity monitoring, and pulse oximetry, POPA includes orders to handle common post-surgical complaints such as nausea and constipation, and most importantly, detailed instructions for recognizing and responding to severe opioid-associated ADEs.

Expanding POPA

The "prescriptions are programs" insight validated by the POPA study, combined with the increasing use of computers on hospital wards and the current push for the expansion of electronic medical records, means the time is right to pursue POPA's potential.

Brooks says the POPA team has regrouped into a new one that includes computer science professors from the University of Illinois, pharmacists, clinicians, Veterans Affairs hospitals, and the Institute for Safe Medical Practices. This team has three objectives, he says.

The first is to make an electronic POPA to interact with hospital medical record systems. Because there are various types of electronic medical records systems, developing an electronic POPA will be challenging, says Brooks. Widespread POPA adoption depends upon creating a tool that can be integrated into all, or nearly all, of the records systems currently on the market.

Another objective is to make the protocols adaptable to other diseases, such as treatments for diabetes or congestive heart failure. Many disease treatments rely on established best practices, making the protocol structure of explicit procedures and subroutines a good choice for standardizing care.

Lastly, the team wants to develop an electronic framework that can support multiple types of treatments. Since many hospital patients have more than one issue, medical staff and the patient would benefit from having all of their treatment plans and medications in the protocol format.

The team is just beginning their work on these new objectives, but they hope it won't be too many years before hospitals and the patients they serve will be benefiting from them.

Why POPA worked

POPA empowered the nurses to rapidly respond. The measurements were easy for the nurses to convert directly into action as the POPA procedures and subroutines were explicitly laid out: First you do a, then b, then c. If you observe x, you do y.

"The ward nurses loved it," says Brooks.

Another facet of POPA's success that was equally simple—yet not widely employed prior to the study—was around-the-clock use of common non-prescription painkillers that have a lower incidence of adverse effects. Maintaining a steady dose of those analgesics aided patients in managing their pain, often reducing a patient's need for stronger painkillers.

POPA is powerful but simple. When a prescription makes sense to a nurse and patient, it is more likely to be properly executed. Which bodes well for the future success of the expansion into electronic POPAs.

This project was funded by the U.S. Food and Drug Administration, Caterpillar, and OSF Saint Francis Medical Center.

Team members
S.M. Belknap
H. Moore
S.A. Lanzotti
P.R. Yarnold
M. Getz
D.L. Deitrick
A. Peterson
J. Akeson
T. Maurer
R.C. Soltysik
G.A. Storm
I. Brooks