Web-based calculator reduces errors, improves pediatric infusions and chemo
Research conducted at the Johns Hopkins University School of Medicine has found that use of a web-based calculator significantly reduces the total number of errors and eliminates all high-risk errors in the prescribing process for continuous pediatric infusions, while failure modes and effects analysis-guided computerized physician order entry reduces ordering errors in pediatric chemotherapy.
Christoph U. Lehmann, MD, director of clinical information technology at the Johns Hopkins Children's Center and one of the researchers in both studies, says medicating children and infants by continuous intravenous infusion is an error-prone process, with incorrect dosing being the most frequent type of error. "For intravenous forms of drugs," he says in a study report in Pediatric Critical Care Medicine, "the ordering step requires prescribers to have more domain knowledge than for other forms, with higher potential impact of dosing errors because of the rapid bioavailability and toxicity of drugs by that route. Because patients who require continuous infusions are usually located in busy, distracting clinical environments [critical care or emergency departments], they may be at higher risk from ordering errors. For multiple infusions, patients may be at further risk for errors due to the increased complexity of ordering."
Past approaches to error reduction and prevention in pediatric infusions have focused on reducing variation in the medication process through applying technology or system changes:
- At the ordering step, computerized prescription writers and physician order entry guide prescriber input, provide decision support, and generate legible orders, resulting in reduced prescribing errors and decreased need for pharmacist intervention.
- At the administration step, computerized "smart" intravenous pumps automate dosage calculation and documentation tasks and provide rule-based alerts and reminders.
- At the dispensing step, using standard (fixed) concentrations for infusion medications reduces variation by limiting prescriber choices and possibilities for pharmacy errors in their preparation.
Lehmann helped develop an infusion calculator providing computerized decision support in the form of default doses, maximum concentrations, and interactive dosing guidelines and alerts for diluent, drug dosage, and insufficient fluid errors. Knowledge for decision support and alerts was adopted from institutional protocols and reference materials, validated by two pharmacists, and approved by institutional pediatric emergency and critical care divisions.
Before the calculator was deployed, 100% of all infusion orders (129 consecutive orders during February and March 2003) were handwritten. After the calculator was deployed, 88% of consecutive infusion orders were calculator-generated and 12% were handwritten.
Before calculator deployment, 27% of all orders contained at least one error, vs. 13.6% after the calculator was put in use. Of all handwritten orders both before and after calculator deployment, 55% contained at least one error, compared with just 6% of calculator-generated orders.
Following calculator deployment, the number of errors in handwritten orders soared. While 27% of handwritten orders had at least one error before deployment, that rate increased to 70% after the calculator was made available. There had been 45 errors per 100 orders before the calculator and 120 errors per 100 orders afterwards. It would appear that the relatively few people who decided to stay with handwritten orders rather than use the calculator made significantly more errors. In calculator orders, there were six errors per 100 orders.
In handwritten orders, the most frequently observed error types were wrong concentration (10.1%), wrong dose (9.4%), and wrong use of the Rule of Six (7.4%). The most frequent omissions were infusion rates (five per 100 orders), diluent types (five per 100 orders), patient weights (five per 100 orders), and patient identifiers (one per 100 orders), and there were two illegible orders (1.3%). In calculator-generated orders, there was what the researchers termed a "dramatic reduction" of the most frequently observed error types — wrong concentration, 0%; wrong dose, 0%; wrong calculation, 0%, with a corresponding decrease in errors of omission for patient identifier (five per 100 orders), and no illegible orders.
Lehmann says the calculator provided automated calculation and rule-based decision support based on guided user input of body weight and medication. Limiting user choices, he says, it reduced human variation and prevented inappropriate selections such as wrong diluent fluids. Introducing voluntary use of the calculator was associated with a 50% reduction in orders with errors during the study period. And calculator-generated orders, in comparison with handwritten orders, were associated with an 83% reduction in orders with at least one error and elimination of high-risk prescribing errors.
"Our findings reveal that using a web-based calculator makes it less likely to order and give a child the wrong dose or commit other errors, such as omitting patient information, weight parameters, or infusion rates," Lehmann says. "Our calculator stops ordering errors before they can even reach the pharmacy, let alone the patient. Many of these errors may have been caught by pharmacists or nurses before they could have hurt a patient, but we wanted to prevent providers from making errors in the first place."
The calculator computes all doses, advises and warns of drug interactions, and automatically offers default doses and drug dilutions to help doctors avoid over- or underdosing.
"The overwhelming prescriber acceptance for the calculator was reflected by the overwhelming percentage of infusion orders being calculator-generated during the initial voluntary use period," Lehmann says. "The inference is that the user-perceived benefits of using the infusion calculator outweighed the additional time and steps needed to use it."
Using the web-based calculator is now mandatory at Lehmann's facility along with standardized dosing. They also have added smart pump technology.
Lehmann and his colleagues say they believe in data sharing and have made the calculator available on-line (e-mail Lehmann at the address, to arrange access to the calculator) so others can apply it to their own situations and develop their own technology.
In his second study, reported in the May Archives of Pediatric and Adolescent Medicine, he says the complexity of pediatric chemotherapy makes it vulnerable to errors because it combines the risks for errors in pediatric drug dosage calculation (which is dependent on age, weight, and body surface area) and the risks associated with the narrow therapeutic profile and high potential for acute and cumulative toxicities of antineoplastic agents.
"These risks combined with potential interruptions in the execution of detailed protocols (due to the labile physiological conditions of patients), create a higher probability and impact of errors and worse outcomes in an already vulnerable population," Lehmann explains. "Error management in this domain requires global and specific consideration of the medication process and the information that controls it and coordination and communication among physicians, nurses, pharmacists, and allied health professionals. In addition, there must be ongoing surveillance and data collection and analysis to evaluate the effects of interventions on the medication process and to identify the introduction of new types of errors."
One approach to reducing and preventing errors in medicating patients is to model the process as a series of steps performed by individuals, such as physicians, nurses, clerks, and pharmacists, who pass information from one to another.
Lehmann says according to literature and experience of those working in his medical center, most medication errors in general, and pediatric chemotherapy errors in particular, occur at the prescribing/ordering step. The research team's recommendations for interventions at this step included automatic calculations via embedded calculators; reducing free-text and handwritten entry through drop-down medication lists; reducing prescriber memory load through autocompletion of specific data fields; and enforcing prescriber completion of essential data fields via pop-up alerts.
While incorporation of computerized order entry has been advocated as an effective way to reduce medication errors, Lehmann says it must be implemented and deployed in a way that is tailored to the needs of specific clinical processes. In pediatrics and neonatology, computerized order entry has been demonstrated to mitigate errors in specialized medication processes such as total parenteral nutrition and continuous intravenous infusions, and those deployments have succeeded through guidance by domain and information technology experts familiar with the processes.
Lehmann says it's important that error management include continued surveillance for persistence of old errors and anticipation of the introduction of new ones. He notes that in the first iteration of their computerized order entry process, there was an unanticipated but statistically significant decrease in the matching of orders (chemotherapy drug doses) to specific protocols. A data review suggested several reasons including the lack of an automated drug-protocol linkage, user interface problems, and human transcription failure. Those issues were the basis for the upcoming next iteration of the system.
"One of the important lessons from this study is that machines do not prevent errors and, in fact, may initially introduce new potential hazards," said co-investigator Robert Arceci, MD, director of pediatric oncology for the Johns Hopkins Kimmel Cancer Center in Baltimore, where more than 200 children receive cancer treatment each year. "It takes a whole culture of safety and vigilance to prevent errors."
Lehmann tells Drug Formulary Review that this was not research he would have started on his own, but it came to his attention through a safety concern following a sentinel event in chemotherapy administration.
System is ahead of commercial models
Because there is a lot of complicated math involved in determining the appropriate dosage, they wanted to automate that effort as much as possible. While it took Lehmann and his team about two weeks to develop the infusion calculator, they worked with a vendor for more than a year to build a satisfactory computerized order entry system. He says commercial off-the-shelf systems don't have the detailed features needed for pediatric oncology, and those available on the market today still don't have some of the features built into the Johns Hopkins system two years ago.
"We designed the system knowing what the errors would be," he says. "Safety is incredibly important to us. We have a pediatrics chief who wants to be able to sleep at night. Because humans make unpredictable errors, we'll use machines to standardize processes any time we can."
Based on the findings, researchers recommend adopting digitized systems for intravenous drug infusions, as well as chemotherapy orders, in high-risk clinical areas such as pediatric oncology centers, pediatric intensive care units, and emergency rooms.
"What we found extends well beyond pediatric oncology and has potential application in all areas of medicine," says George Kim, MD, lead author of the chemotherapy study. "This system is part of a whole new culture of safety that requires not only change in the system, but also change in the mindset."
[Editor's note: Contact Dr. Lehmann at (410) 955-4582 or e-mail [email protected].]
Research conducted at the Johns Hopkins University School of Medicine has found that use of a web-based calculator significantly reduces the total number of errors and eliminates all high-risk errors in the prescribing process for continuous pediatric infusions, while failure modes and effects analysis-guided computerized physician order entry reduces ordering errors in pediatric chemotherapy.You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content