Case 1: Two patients were to receive intrathecal methotrexate with fluoroscopic guidance, Patient A in the morning and Patient B in the afternoon. The pharmacist dispensed the two methotrexate doses and the syringe for Patient A was delivered to his ward that morning. However, the syringe was not delivered before Patient A was transported to radiology. When Patient A arrived in the radiology suite, the technician called over to pharmacy looking for Patient A’s medication. In the pharmacy, the tech saw a single syringe labeled for intrathecal use and delivered it to radiology. This error was discovered when the pharmacist made afternoon deliveries and saw that Patient A’s syringe was still on the ward even though Patient A had already returned from the procedure. Patient A had received Patient B’s medication.
Case 2: Chemotherapy agents were given via the intrathecal route by the radiologist without verifying the order. Cytarabine was given intrathecally, but ordered subcutaneously.
Case 3: Methotrexate intrathecal was sent to radiology for administration instead of intrathecal cytarabine. The error was detected before starting the procedure.
Fifty years after the first death from accidental injection of the wrong chemotherapy agent intrathecally, we continue to struggle with ensuring that these accidents do not recur. Wrong medication errors for intrathecal chemotherapy fortunately are rare, but, when they occur, can have severe consequences. None of the above cases resulted in significant harm. They serve as warning signals that our defenses against such events need bolstering.
It is particularly important to pay attention to near misses and minor events in settings where the potential for harm is rare and extreme. “As systems become safer, it becomes progressively more difficult to detect the remaining vulnerabilities.” (Franklin et al)
Discussions with CHPSO members and review of the CHPSO event database identified the following challenges with intrathecal chemotherapy:
It is difficult for the person performing the lumbar puncture to scan the syringe barcode just before administration, as his/her hands are engaged in performing the procedure.
Bar code medication administration may not be implemented yet in the radiology suite.
Missing orders can result in time pressure for the pharmacist, something that should be avoided for any chemotherapy order.
Patients receiving chemotherapy often receive both intravenous and intrathecal chemotherapy the same day.
A long free-text chemotherapy order may be truncated by some CPOE systems.
Defense in depth is required for protection against intrathecal chemotherapy errors. The reference by Franklin et al describes the typical measures taken in UK hospitals and is worth reviewing. In addition to those measures, designing processes that make bar code scanning occur reliably during intrathecal procedures is important. However, while bar code medication administration does help with wrong patient or wrong time errors, it is not a strong protection for wrong-route errors. A correct scan is not assurance that the drug should be given intrathecally.
Ultimately, the new NRFit connector, designed so that intravenous syringes are not interconnectable with intrathecal or other neuraxial devices, will provide an additional layer of protection. Supply may not be sufficient yet for change-over, and hospitals should check with their suppliers before starting the transition. The main barrier to the change currently appears to be supply of infusion sets needed for continuous labor epidurals. Infusion sets should come into the market late this year.