Featured Articles

Safety Issues with PCA Part I - How Errors Occur

Patient-controlled analgesia (PCA) has considerable potential to improve pain management for patients, allowing them to self-administer more frequent but smaller doses of analgesia. When used as intended, PCA actually reduces the risk of oversedation, which is an unintended consequence of the more traditional method of nurse-administered analgesia in larger, less frequent doses. In fact, with PCA, patients often develop a synergism with the device and can quickly learn how to manage their pain while avoiding undue mental clouding. However, through the USP-ISMP Medication Errors Reporting Program, the USP MEDMARX program, and a sizeable response we received from readers answering our call for information about PCA problems, it's clear that errors happen frequently, sometimes with tragic consequences.

Just last week, we met with staff at FDA's Center for Devices and Radiological Health (CDRH) to discuss medication errors associated with PCA. At the meeting, we presented in-depth information about the following factors that have frequently contributed to the problem.

PCA by proxy. Several safety features exist with PCA to make sure patients do not receive too much analgesia. These include a lockout interval that specifies the minimum amount of time between each dose, and a maximum allowable amount during 1- or 4-hour intervals. Another "built-in" safety feature that's often overlooked is that the device is intended for patient use. A sedated patient will not press the button to deliver more opiate, thus avoiding toxicity. However, family members and health professionals have administered doses for the patient, by proxy, hoping to keep them comfortable. This well-intentioned effort has resulted in oversedation, respiratory depression, and even death.

Improper patient selection. Since an important safety feature with PCA is that the patient delivers each dose, candidates for PCA should have the mental alertness and cognitive, physical, and psychological ability to manage their own pain. However, the benefits of PCA have led providers to extend its use to less than ideal candidates such as infants, young children, and confused elderly patients. This has facilitated the dangerous practice of PCA by proxy. PCA use in these types of patients has also spurred ethical debates about the potential for undertreatment caused by the poorly coordinated efforts of family members (who are not at the bedside continuously) and clinicians, and the inability of these patients to clearly communicate their pain level. In addition, oversedation has occurred in less than ideal candidates at risk for respiratory depression due to comorbid conditions such as obesity, asthma, or sleep apnea, or use of concurrent drugs that potentiate opiates.

Inadequate monitoring. Even at therapeutic doses, opiates can suppress respiration, heart rate, and blood pressure. Thus, nurses or other caregivers typically monitor patients at frequent intervals while they are using PCA. However, these monitoring activities may not alert caregivers to opiate toxicity. First, patients may not be monitored frequently enough, especially during the first 24 hours and at night when nocturnal hypoxia can occur. But more often, the way that caregivers assess patients may be at the root of the problem. Patients who are experiencing opiate-induced respiratory depression or oversedation can easily be stimulated to a higher level of consciousness and an increased respiratory rate. Thus, if patients are disturbed in order to make the assessment, the level of consciousness and respiratory rate observed is not helpful in detecting toxicity. Once the stimulus is removed, patients quickly fall back into an oversedated state. There's also too much reliance on pulse oximetry readings, which can offer a false sense of security since oxygen saturation is usually maintained even at low respiratory rates, especially if supplemental oxygen is in place.

Inadequate patient education. Most patients who are suitable candidates for PCA can be taught how to use the device successfully. However, patients who have been taught to use the device during the immediate postoperative period have often been too groggy to fully understand its use, and they often report poor pain control during the first 12 hours after surgery. Alert, intelligent patients also have misunderstood the directions for use, most often believing that they must press the button every 6 minutes or so, even when they are sleepy and comfortable. At times, family members have awakened patients so that they can press the button frequently, or they may press the button for the patient if they have not been alerted to avoid PCA by proxy.

Drug product mix-ups. Some of the opiates used for PCA have similar names and packaging, which has led to drug selection errors. Prefilled syringes of meperidine and morphine have been packaged in similar-looking boxes. Morphine is available in prefilled syringes in two concentrations, but the packaging may not help to quickly differentiate the strengths. Differentiation between opiates with and without preservatives is not prominent on labels. All pharmacy-applied labels may look similar on extemporaneously prepared syringes or bags. Name similarities have also led to inadvertent mix-ups between morphine and hydromorphone, or the mistaken belief that hydromorphone is the generic name for morphine. Since opiates are typically in unit stock, these errors are rarely detected and, most often, they have led to significant overdoses; less often, they have led to undertreatment of pain or to an allergic response to the medication.

Practice-related problems. Misprogramming of the PCA pump is, by far, the most frequently reported practice-related issue. Pump design issues that have led to programming errors are described in the section that follows. Other practice-related issues that have contributed to PCA errors include incorrect transcription of prescriptions into pharmacy computers or medication administration records (often related to look-alike product names), calculation errors when determining the patient's dose or rate of infusion, and IV admixture errors. Unavailability of hydromorphone in prefilled syringes or premixed bags necessitates IV admixture of this opiate. Drug shortages, most notably with fentanyl, also have required the use of less familiar products, such as sufentanil, which has led to serious dosing errors.

Device design flaws. Programming a PCA pump requires multiple steps, but the pump's design is often far from intuitive. In fact, two device models (Abbott Lifecare PCA II and APM Infusers) have been under close scrutiny for years because of the frequency of programming errors, many of which have resulted in patient deaths. Most of these programming errors have resulted because concentration settings for opiates such as morphine default to 0.1 mg/mL or 1 mg/mL, but a higher concentration is available and used. Other design flaws that have facilitated programming errors include pumps that do not require users to review all settings before the infusion starts, and pumps that require users to program the dose in mL, not mg, leading operators to overlook the amount of drug the patient is actually receiving. Siphoning (free flow) also has been reported after entrainment of air into the system due to a fractured glass syringe. It also may happen when a cassette breaks and detaches from a CADD pump that lacks attached anti-siphon tubing. Mechanical problems, such as short circuits, are possible but quite rare.

Other design flaws are related to the patient's use of the pump and are common to many PCA pumps currently on the market. First, the activation button looks just like a nurse call bell, so patients have inadvertently given themselves a dose of analgesic while believing they were pressing the button to call a nurse. Another problem is that, with lockout intervals set, unless the pump provides some visual or auditory feedback, patients cannot tell whether the press of the button has resulted in the actual delivery of medication. As a result, some patients press the button more frequently than they would with proper feedback, and others become frustrated and give up, resulting in undertreatment of pain. While pump manufacturers are required to perform human factors testing, this regulation is loosely enforced and these types of design problems may not be identified up front because practitioners and patients are not closely involved in the testing procedures.

Inadequate staff training. Entering a PCA prescription into a pump requires a number of steps. However, nurses may not always receive adequate training in pump programming, or they may not retain their proficiency once trained if multiple pumps are in use or if PCA is encountered infrequently. Additionally, prescribers may not undergo a credentialing process designed to verify proficiency with this form of pain management. Prescription errors, including improper drugs or doses, have resulted.

Prescription errors. The PCA order itself can be a source of error. Physicians have made mistakes in converting an oral opiate dose to the IV route (most problematic is hydromorphone, which has an oral to IV conversion range of 3:1 to 5:1). They have prescribed a drug to which the patient is allergic, and have selected an opiate that is not appropriate for the patient (meperidine in patients with renal impairment). Occasionally, one opiate has been prescribed but the accompanying dose has been appropriate for a different opiate. Even with correct PCA orders, clinicians have been known to mishear or misread verbal or written orders, sometimes leading to serious errors. Concurrent orders for other opiates (oral or parenteral) while PCA is in use has also resulted in opiate toxicity. As of yet, none of the PCA pumps have the safety features available in new general purpose "smart pumps," which alert when maximum doses or flow rates are exceeded.

Again, we thank our readers for giving us feedback about PCA errors. The information provided allowed us to share insightful examples describing just how PCA errors have happened. Part II in our next newsletter will cover the error-reduction strategies that were shared with FDA, which include a balanced approach of practice-related, system-related, product-related, device-related, and regulatory-related efforts. It's time we work together to reduce the risks associated with this wonderful technology.