D. John Doyle MD PhD

December 2004

Patient-controlled analgesia (PCA) is a computer-based medical technology now used extensively to treat postoperative pain via the self-administration of analgesic agents [1]. Potential benefits include superior pain control, automatic electronic documentation and improved utilization of nursing resources. Unfortunately, however, analgesics are also a frequent cause of adverse drug events [2, 3].

A typical PCA machine contains an embedded computer programmed to give, for instance, 1 mg of morphine intravenously every time the patient pushes a button on the end of a cable. To help prevent excessive drug administration, the onboard computer ignores further patient demands until a lockout period (usually set for 5–10 minutes) has passed. In addition, 1-hour or 4-hour cumulative limits also are available in some models.
Despite such safety features, numerous reports of respiratory depression and death associated with PCA pumps have appeared [4-9].

One particularly notorious unit is the  Abbott Lifecare 4100 PCA Plus II machine. This machine is easily misprogrammed by caregivers, who must manually enter the PCA parameters, and is in need of a more sensible and forgiving user interface [10]. A number of patients have received opiate overdoses as a result of PCA errors when using this product: insertion of a 5 mg/mL morphine cartridge when the machine is expecting a 1 mg/mL concentration, or acceptance of the default (initial) drug concentration when the correct action is to scroll up to the correct value, among other errors [11-13].

In 1997, ECRI documented 3 deaths that occurred while patients were connected to the Lifecare 4100 [11]. In at least 2 of the cases, the alleged reasons for the deaths were the same. In the mode of operation in use, when nurses program the drug concentration the Lifecare 4100 display shows a particular concentration (e.g., 0.1 mg/mL). Nurses can either accept this initially displayed value or modify it using the arrow controls. The critical flaw in the design is that in this situation the Lifecare 4100 offers the minimal drug concentration as the initial choice. If nurses mistakenly accept the initially displayed minimal value (e.g., 0.1 mg/mL) instead of changing it to the correct (and higher) value (e.g., 2.0 mg/mL), the machine will "think" that the drug is less concentrated than it really is. As a result, it will pump more liquid, and thus more narcotic, into the patient than is desired.

It should be emphasized, however, that other mechanisms of PCA overdoses may occur. Doyle [14] has proposed a possible classification for such overdoses based on a review of the literature. 

[1] Use of wrong drug or wrong cartridge (e.g., insertion of a 5 mg/mL morphine cartridge when a 1 mg/mL cartridge is required).

[2] Accidental misprogramming, sometimes as a consequence of a hostile user interface.

[3] False triggering, for example, due to a short circuit in the PCA button or for other reasons.
[4] PCA triggering by proxy (e.g., relatives pushing the PCA button because “Granny” is too sleepy to do it herself).

[5] Drug accumulation in IV deadspace.  This may occur with large IV deadspaces under low flow conditions.

[6] Runaway fluid column due to "siphoning".  (Should a crack occur in the PCA drug cartridge, entrainment of air into the system may lead to a free-flow of drug into the patient. Some manufacturers incorporate anti-siphon valves into their designs to prevent this).

[7] PCA machine malfunction due to hardware failure.

[8] PCA machine malfunction related to software design error. 

[9] Retrograde flow of PCA analgesic drug into a secondary IV set (e.g., for administering antibiotics) due to a temporarily blocked IV catheter.  When the IV catheter is subsequently unblocked, the PCA drug that has accumulated in the secondary IV bag is then suddenly released into the patient. (The use of a one-way valve on the secondary IV set will prevent this).

[10] Bad medical judgment in formulating PCA prescription, or opiate orders from other physicians unaware that PCA orders have been written.

[11] Anaphylaxis (either de novo or despite knowledge of risk of reaction).

[12] Extraordinary sensitivity to opiates resulting in unexpected respiratory depression.

In addition to these situations, there is one safety situation that, while theoretically possible, is unlikely to be encountered in real life.

[13] Reprogramming with criminal intent.

Under ideal circumstances, the continuous use of a pulse oximeter and/or other forms of respiratory monitoring would prevent many of these tragic deaths. In the real world, such technologies are often too expensive and / or plagued with false alarm problems for use on ordinary nursing floors. There is evidence, however, that this situation is improving. Pulse oximeters can now be purchased on eBay for under $300. And new signal processing technologies are making pulse oximeter technology much more reliable, as evidenced by new designs from Masimo and other technology leaders.

In addition to improved respiratory monitoring technology, attention to improving the PCA user interface is desperately required, such as has been carried out at the University of Toronto [10, 15]. When PCA pumps are eventually redesigned to make them more intuitive and easier to program (using human factors engineering techniques) another important safety layer will be in place.

Other suggestions to improve safety include [13]:

When an adverse clinical event possibly related to PCA usage occurs, the room and equipment should be immediately secured and an investigation started. (It is very difficult to determine what has happened when the PCA drug cartridge is discarded following a patient death from respiratory depression). 

Caregivers should report any problems or complications associated with PCA pumps (as well as all other medical devices) to the FDA, and hospital authorities. (To file an online FDA report visit

When a PCA pump is programmed, an independent check by a second caregiver should be conducted.

Only a single concentration of morphine cartridge (e.g., 1 mg / mL) should ordinarily be stocked to avoid confusion. Where this is not practical (as with some oncology wards), special precautions should be instituted when higher concentration morphine cartridges (e.g., 5 mg / mL) are also available.

Hospitals should avoid purchasing PCA pumps known to have design flaws (e.g., Abbott Lifecare 4100 PCA Plus II). In addition, there is a critical ongoing need for experts to formally and informally review various medical products in a “Consumer Reports” fashion. (See for an excellent example).


1.Ferrante FM, Ostheimer GW, Covino BG. Patient-Controlled Analgesia. Boston: Blackwell Scientific; 1990.

2.Classen DC, Pesotnik SL, Evans RS, Burke JP. Computerized surveillance of adverse drug events in hospital patients. JAMA 1991; 266: 2847–51. 

3.Bates DW, Cullen DJ, Laird N, et al. Incidence of adverse drug events and potential adverse drug events. Implication for prevention. JAMA 1995: 274: 29–34. 

4.Grey TC, Sweeney ES. Patient-controlled analgesia. JAMA 1988; 259: 2240.

5.Grover ER, Heath ML. Patient-controlled analgesia. A serious incident. Anaesthesia 1992; 47: 402–4.

6.Geller RJ. Meperidine in patient-controlled analgesia: a near-fatal mishap. Anesth Analg 1993; 76: 655–7.

7.Etches RC. Respiratory depression associated patient-controlled analgesia: a review of eight cases. Can J Anaesth 1994; 41: 125–32.

8.Baxter AD. Respiratory depression with patient-controlled analgesia (Editorial). Can J Anaesth 1994; 41: 87–90. 

9.Kwan A. Overdose of morphine during PCA (Letter). Anaesthesia 1995; 50: 919.

10.Lin L, Isla R, Doniz K, Harkness H, Vicente KJ, Doyle DJ. Applying human factors to the design of medical equipment: patient-controlled analgesia. J Clin Monit Comput 1998;14:253-63. 

11.Anonymous. ECRI. Abbott PCA Plus II patient-controlled analgesia pumps prone to misprogramming resulting in narcotic overinfusions. Health Devices 1997; 26:389-91

12.Anonymous. Design flaw predisposes Abbott Lifecare PCA Plus II pump to dangerous medication errors. ISMP Medication Safety Alert 2000;5:2. Huntingdon Valley (PA): Institute for Safe Medication Practices; 2000.

13.Vicente KJ, Kada-Bekhaled K, Hillel G, Cassano A, Orser BA. Programming errors contribute to death from patient-controlled analgesia: case report and estimate of probability. Can J Anesth 2003; 50: 328–32

14.Doyle DJ. Programming errors from patient-controlled analgesia.  Can J Anesth. 2003; 50:855-6

15.Lin L, Vicente KJ, Doyle DJ. Patient safety, potential adverse drug events, and medical device design: a human factors engineering approach. J Biomed Inform 2001; 34: 274–84