Medication Errors Involving Patient-controlled Analgesia

Rodney W. Hicks; Vanja Sikirica; Winnie Nelson; Jeff R. Schein; Diane D. Cousins


Am J Health Syst Pharm. 2008;65(5):429-440. 

In This Article


The value of pain control includes improved patient satisfaction, decreased complication rates, shortened stays following surgery, and potentially reduced exposure to litigation.[8,13] Methods for controlling acute pain have advanced greatly over the past three decades, especially since anesthesiologists organized services for the management of acute pain.[12] However, pain control carries its own set of risks.

This study sought to add to knowledge regarding PCA-related medication errors by using secondary data from nearly 500 hospitals and more than 9500 PCA error records from institutions that participated in the Medmarx medication error-reporting program over a five-year span. Slightly more than 60% of the Medmarx subscribers reported at least one PCA error, suggesting that PCA errors are common rather than isolated.

The first priority for patient safety is to minimize harmful errors and the second is to minimize all errors. While PCA events accounted for only 1% of the errors reported to Medmarx, such events were associated with 6.5% of harmful outcomes. Two deaths were attributed to PCA errors. By comparison, during the same period, only 1.5% of all other errors reported to Medmarx led to harm. This represents a fourfold higher relative risk of harm for PCA events. More resources are required to treat patients with harmful errors because affected patients often received additional care, such as closer observation, change in the treatment plan, use of antagonists, interruption of care, diagnostic testing, and life-sustaining interventions.[14] Harmed patients often had extended hospital stays, further stressing already highly strained staffs.[15] From a public health perspective, the finding that a greater percentage of harmful events occurred in older patients is concerning because this age group has greater surgical needs and is growing. The increased risk of harm in elderly surgical patients is likely due to their frailty and comorbidities, making the need to minimize PCA errors ever more imperative.

These findings are supported by recent national initiatives to reduce error, such as the third National Patient Safety Goal (NPSG) established by the Joint Commission.[16] The NPSG deals broadly with the safe use of medications and would be applicable to the processes involved in PCA. As part of the program, institutions are encouraged to review the PCA process and to comply with the goals through strategies such as failure-mode and effects analysis, root-cause analysis, and focus groups working directly with PCA equipment.

The leading type of error identified in this study was an improper dosage or quantity of analgesic medication. The risk of respiratory depression, a well-known adverse effect of PCA,[9] is obviously heightened when excessive dosages are administered. These findings suggest that incorrect programming of the PCA infusion device is often the source of errors; this occurs most often during the administration phase. Operator errors during infusion setup (programming) involved the bolus dose, the lockout interval, and the basal rate. These results are consistent with the literature.[2]

Another reason for excessive dosages was PCA by proxy. In 2004 the Joint Commission issued a sentinel-event alert focusing on PCA by proxy.[17] It recommended educating staff working with PCA devices, as well as teaching family members about the proper use of the PCA pump and the dangers when someone other than the patient activates the controller.

Errors due to an unauthorized or wrong drug were also common. These errors occur for a number of reasons, including confusion between generic names that either look or sound alike. This study found that two drugs (hydromorphone and morphine) were often switched because of name confusion. The manufacturer's use of "tall-man" lettering may help staff visually distinguish among packages, although this product pair was not originally in FDA's name-differentiation project. The packaging may be similar for PCA drug products, which can contribute to the error as well, especially when the products are retrieved from automated dispensing devices that are not directly linked to pharmacy information systems and bar-code medication administration (BCMA) systems. As more institutions adopt BCMA, the risk of administering the wrong drug during PCA may decrease.

Staff are encouraged not to override prompts from automated dispensing devices. Rather, pharmacists should be consulted to assist in resolving any alerts.

The Medmarx program contains more than 60 selection options for the cause of an error. Aligning these options with the NCCMERP taxonomy affords an opportunity to cluster related choices for analysis. Each cluster identifies a different risk for a medication error and requires different interventions to address the cause. Human errors were overwhelmingly present in the sample and included miscalculation, not following policies and procedures, selecting the wrong drug from the dispensing device, and confusion about drug names and concentrations. Miscalculation occurred during the initial programming of the PCA device or during subsequent changes to the infusion variables, leading to multiple errors involving an improper dosage or quantity. Misinterpretation of drug concentrations was also implicated, as was staff confusion between hydromorphone and morphine and between morphine and meperidine, a finding consistent with errors involving an unauthorized or wrong drug. Independent double-checks by two qualified individuals for the right product, the right calculations, and the right pump settings should be performed routinely.

The PCA error records implicated equipment-related issues as the cause of the second largest error cluster. The finding is consistent with the requirement for accurately programming the bolus dose, the basal infusion rate, and the lockout interval. According to the Institute for Safe Medication Practices (ISMP), mis-programming a PCA pump is rather easy, given the sequence of screens pertaining to the drug, the bolus dose, the infusion rate, and the lockout interval.[18] Even with experience and independent double-checking, medication errors can occur during PCA setup. ISMP recommends restricting the number of medications used for PCA to eliminate some of the confusion that can arise when selecting drugs from the PCA screen. Institutions should provide readily available reference material specific to each drug for staff members to consult during the setup process. For institutions that provide medication labels for PCA, ISMP recommends designing labels in which the drug concentration is prominent and clearly legible.

The most commonly identified contributing factor was distraction. Distractions may lead a person to forget to perform a key task and thus may contribute to the chain of events that produces an error.[19] Because of the situational nature of distractions, they are difficult to anticipate and control.

The contributing factor of staff inexperience was noted in one fourth of the PCA error records. This finding suggests unfamiliarity among some with PCA processes. This is an organizational or system failure. A frequent action after such events was to provide training, but it would be far better to ensure that staff are fully trained before errors can occur.

Voluntary reporting systems have been criticized for inherent under-reporting.[12] Nevertheless, widely deployed voluntary systems such as Medmarx afford an economical and practical alternative to real-time data collection and are a suitable source of data for analysis. This study did not determine the rate of PCA errors, although the sample was large and was representative of all U.S. hospitals. Rather, the study encompasses only the rate of reporting of such errors. Hospitals participating in Medmarx have opted to self-report errors and thus endorse an open culture of reporting; this may bias the results. Generalization of these findings to other populations should be made with caution.

A better understanding of the major determinants of medication errors, such as the intersection of human factors with equipment-related factors, could potentially lead to error reduction. Alternative routes of administration in pain management include the intranasal route (for either nonopioid or opioid agents) and the iontophoretic route; these options may address some of the issues of PCA by proxy, but their full effect on patient safety needs further evaluation. Given the findings of this study, additional measures should be taken to address PCA errors to improve overall patient safety and care.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.