From the January 12, 2006 issue
Problem: In our November 3, 2005 newsletter, we reported an error that resulted from pressing a number key once on an Alaris SE pump (formerly Signature Edition GOLD Infusion System) and getting an unintended repeat of that same number (i.e., the pump recorded 366 mL per hour, not 36 as intended). The reporting facility had uncovered 11 similar double key bounce events with high-alert medications within the past 2 years-all involving different nurses working in various patient care units who felt certain that they had pressed the number key just once. This problem differs from an accidental double keying error in which a number or letter key is accidentally pressed twice, perhaps due at times to an unsteady finger. Such errors have been reported many times in the past regardless of the type of pump in use.
To further investigate double key bounces, we asked our newsletter readers to let us know if they had encountered similar problems with this or any other infusion pump. One nurse noted a similar problem with an Alaris SE pump. She thought she had programmed a propofol infusion to run at 25 mL per hour. Fortunately, right before she left the patient's room, she noticed that the pump was actually delivering 225 mL per hour. At first, she thought she had pressed the number 2 key too long, causing a double key bounce in the same manner that a depressed computer key might run a string of numbers. But later, when she tested the pump, she could not recreate the error.
Her inability to recreate the error in this case may be explained by the hospital that first reported the problem to ISMP. During an investigation, this hospital found that double key bounces appear to be directly related to the pressure used to activate the key as well as the duration of holding the key down. They were able to duplicate the double key bounce error only by pressing softly, or partially, on the key for a time, as one might do if programming the pump from an odd angle rather than standing directly in front of it. They found that firmly pressing a fully engaged key for several seconds, as the nurse in the other hospital had done to try to duplicate the error, did not cause a double key bounce.
Another reporter told us he was able to replicate double key bounce errors on both the SIGMA 8000 and 6000+ infusion pumps (from Sigma International) about 10-20% of the time by partially depressing the number keys to a shallow depth. He also reported that he was able to replicate the number entry up to three times during the double key bounce. ISMP has not been able to confirm these observations.
According to our recent responses, there are some additional conditions under which double key bounces or double keying errors may not be detected-high levels of background noise when programming the pump, and confirmation bias when viewing pump screens. Pumps will give an audible beep for each digit registered, so it will beep twice if either type of error occurs, not once (as long as the device's tones have not been turned off). However, some staff involved in these errors felt they had never really heard the two beeps because of distracting background noise on the unit, especially in critical care units and the emergency department. Thus, this auditory cue was not always helpful in detecting errors, nor was the following visual cue. A display of the entered value appears on infusion pump screens, thus there is visual evidence if a double key bounce or a double keying error has occurred. In some of the cases investigated, involved staff felt they had overlooked the extra digit on the screen because they saw only what they expected to see-the correct infusion rate.
Other hospitals may be experiencing double key bouncing without recognizing it as a hazard. Clinicians may notice the discrepancy (either through manual review or smart pump technology) and simply correct it without reporting the event, often believing they caused the error by inadvertently pressing the key twice. If they recognize that the pump, itself, contributed to the mistake, clinicians may still just "accept" the phenomenon as typical for high-tech medical devices. Clinicians also may overlook a double key bounce or double keying error, especially if the higher infusion rate did not result in patient harm.
Safe Practice Recommendations: ECRI, an impartial health services research company, is also interested in double key bounce errors. They have begun testing various brands of infusion pumps to better understand the nature and scope of the problem. Meanwhile, until we know more, please alert staff to the potential for double key bounces when programming infusion pumps. ISMP also suggests implementation of the following recommendations to help prevent or quickly detect double key bounces as well as double keying errors with any infusion pump.
Verify screen displays. When programming pumps or changing settings, always compare the patient's prescribed therapy on the medication administration record, original order, or bar code device, to the displayed pump settings for verification before starting or restarting an infusion.
Independent double check. Require an independent double check of pump settings by another practitioner before starting or changing infusions with hospital-selected high-alert medications.
Listen then look. Keep pump tones functional on all pumps and focus on listening to the number of beeps while programming IV pumps; each beep should correspond to a single digit entry. Before leaving the patient's room, actually look at the IV tubing drip chamber to see if the observed rate of infusion looks faster or slower than expected. (A rate of 36 mL per hour is visually much slower than an erroneous rate of 366 or 336 mL per hour.)
Dose alerts. Use smart infusion pumps with activated dosage error reduction software that will alert when safe doses and infusion rates have been exceeded when programming the pump. This will help detect most double key bounces and double keying errors before the infusion begins. Keep in mind, some hospitals may not have added alerts to the pump library for IV solutions that are dosed by infusion rate alone (e.g., TPN), or nurses may not always use the dose alert features. It's also important to review overridden dose alerts and adjust dosing guidelines as appropriate to reduce the frequency of clinically insignificant alerts, thus minimizing the chance of disregarding an alert from a double key bounce or double keying error.
Medical device vendors may have to begin exploring standards for the design and function of keyboards, keypads, and switches used in healthcare. This is a difficult task given the wide variation of designs currently in use. As noted above, holding down a key on a computer keyboard for more than about one-half second will result in repetition of the letter until the key is no longer depressed. Yet holding down the key on many telephone keypads will result in registering just the single number. Some fax machines, calculators, and microwaves allow double key bouncing when a depressed key is held, but others do not. Even the arrangement of number keys on telephones differs from those on computer keyboards or calculators. Thus, keyboard/keypad standards for medical devices could be helpful to reduce the risk of misprogramming.
We thank all who have already submitted comments and reports about double key bounces. Please continue to bring any events to the attention of ISMP (email@example.com). If known, include the pump manufacturer, model, and the specific conditions under which the double key bounce occurred.