More than 10,000 children in the United States are diagnosed with cancer annually, according to data from the American Cancer Society.1 However, rates of survival in this younger patient population have risen considerably across decades. The 5-year survival rates for childhood cancer were approximately 58% in the 1970s; today, they exceed 80%.1 These advances in treatment and survival are largely due to complex chemotherapy regimens, which have rendered once-deadly childhood cancers like leukemia and lymphoma treatable and even curable.2
Errors in delivery of childhood chemotherapeutic regimens can lead to serious adverse events, and even death in certain cases.3 Yet research has shown that due to the complex nature of pediatric regimens, the risk for error is high.4 Up to 7% of pediatric cancer patients treated in the outpatient medical setting experience errors in chemotherapy delivery, according to research published in Journal of Clinical Oncology.3 Researchers and treating physicians have noted the need for quality improvement in care delivery for this patient population.
Beginning in 2010, researchers from Cincinnati Children’s Hospital (Cincinnati, OH) have been studying the application of improvement science to reduce chemotherapy errors for childhood cancer patients. The researchers implemented specific measures related to chemotherapy delivery, including a chemotherapy near-miss reporting system; interventional models, including daily chemotherapy huddling and the use of headphones to reduce interruptions during chemotherapy ordering; and hospital-wide standards for chemotherapy ordering. In a paper published in Journal of Oncology Practice,5 the researchers reported a reduction in chemotherapy errors over the course of the program’s first 22 months, with sustained shifts from the baseline rate maintained for more than 4 years. Additionally, they observed a significant and sustained reduction in errors after 16 months of improvement implementation (P < .001).
Journal of Clinical Pathways spoke with Brian D Weiss, MD, director of the neuroblastoma program, associate director of the oncology research program, and associate director of safety and compliance at Cincinnati Children’s Hospital, regarding these advances, and the continued potential for improvement science’s utility across the field of childhood cancer treatment.
What type of errors are seen in childhood cancer treatment, and how common are they?
Our institution has a pretty strict definition of chemotherapy error. By that, I mean that there are some medications that could be given either orally or intravenously, or that could be given over the course of 15 minutes or 60 minutes. In cases like that, both are right. So we called an “error” any deviation from intention, even if that deviation was still technically correct. Another part of this is that what we are trying to avoid is the kind of error that occurred at Dana-Farber Cancer Institute in the mid-1990s, when the health journalist Betsy Lehman died after being given 5 days’ worth of chemotherapy in a total daily dose.6 That was what we would consider a major error, the kind of thing everyone tries to avoid.
The difference between a major error getting through and a minor error getting through is often just dumb luck. We started by considering prescription errors. I assumed that most errors we would see would be errors in the writing of prescriptions. But when we cast a broad net across all the things that can occur that deviate from intention, prescription errors were really a minority. None of the errors that we saw at our hospital rose above a Level C or Level D error, the latter of which is not life-threatening at all. But as I said, some of this is just dumb luck. If a Level C error can get through, something even more serious can likely get through.
So what were your main findings?
We found that prescription errors do not occur all that commonly. That is certainly helped by computerized protocols, which are built so that once the correct protocol is loaded, it does all the math, and keeps all the items—fluids, antiemetics, chemotherapy, labs—together in one packet. With that in place, the opportunity for prescription error really decreases. The kind of error that we are left open to is an error where some chemotherapy is dosed by a kidney function test, and if the glomerular filtration rate changes from what it was 2 weeks ago, it is possible to base the chemotherapy order off of the wrong test. These are the kinds of errors we are seeing with the new, computer-based protocols.
The other thing we found was that errors in administration and errors in dispensary were more common not because physicians do not make mistakes, but because there are now fewer steps between nurses and patients than there are between orders and patients. I may write that something is to be given over 15 minutes rather than 20 minutes, but that is likely to be caught by the checks that occur. But if a nurse programs the pump wrong, there are no steps left between her and the patient. We have considered adding a second nurse to act as a check, but that is not the same thing.
How would you define improvement science?
There is a definition, which is that improvement science uses statistics and scientific methods to ensure that changes seen in patient situations are not happening by chance. Improvement science allows doctors and researchers to determine whether changes in error rates are happening just by chance, or whether actions taken to improve these areas are having the desired effect.
Do you think this kind of research will continue in future studies?
I hope so. We have found that we not only precipitated a significant decrease in error rates, but that those decreases have held steady over time. We are working on other changes we can make to keep these going. Some of these are technical—for instance, Epic can now communicate with certain pumps. One consistent error we still see is when, for example, a nurse programs a chemo pump for 6 mL an hour when the order is supposed to be 60 mL an hour. Even if no harms occurs because of this, we would still consider it a chemotherapy error. But there are new pumps that will read the order directly from Epic, with no transcription involved, so that such an error can be prevented before it would even start. Things like that can close the gaps that are made through human error.
Is there anything you would say to other hospitals considering the implementation of this kind of system?
None of this could be done without a buy-in from nursing staff, pharmacy, and all of the other team members who contribute to cancer care in this patient population. Our nurses, in particular, have really driven a lot of this change. I cannot say enough about what a collaborative effort is involved in building the kind of program we have created. The effort has been well worth it.
References
1. Key statistics for childhood cancer. American Cancer Society website. https://www.cancer.org/cancer/cancer-in-children/key-statistics.html. Accessed June 1, 2017.
2. Treating children with cancer. American Cancer Society website. https://www.cancer.org/cancer/cancer-in-children/how-are-childhood-cancers-treated.html. Accessed June 1, 2017.
3. Walsh KE, Dodd KS, Seetharaman K, et al. Medication errors among adults and children with cancer in the outpatient setting. J Clin Oncol. 2009;27(6):891-896.
4. Gandhi TK, Bartel SB, Shulman LN, et al. Medication safety in the ambulatory chemotherapy setting. Cancer. 2005;104(1):2477-2483.
5. Weiss BD, Scott M, Demmel K, Kotagal UR, Perentesis JP, Walsh KE. Significant and sustained reduction in chemotherapy errors through improvement science. J Oncol Pract. 2017;13(4):e329-e336.
6. Altman LA. Big doses of chemotherapy drug kill patient, hurt 2. New York Times. March 24, 1995. https://www.nytimes.com/1995/03/24/us/big-doses-of-chemotherapy-drug-killed-patient-hurt-2d.html?pagewanted=all. Accessed June 1, 2017.
