David Upton, PhD, MRPharmS
Clinical Director, Pharmacy and Medicines Management, Sheffield Children’s NHS Foundation Trust, Sheffield, UK
It is generally accepted that the administration of drugs by the intravenous (IV) route not only conveys significant therapeutic advantages but also carries the penalty of dire consequences when something goes wrong. The continuous infusion of drugs enables the clinician to administer highly potent agents, which often have very narrow therapeutic indices and short half-lives, in a way that permits the titration of dosing against physiological response. However, the very characteristics that provide beneficial therapeutic advantages, such as 100% bioavailability, rapid distribution to the target site of action and avoidance of first-pass hepatic metabolism, mean that the outcome of any error leading to overdose can be swift and dramatic.
Medication errors at the administration stage of the process are a recognised problem, accounting for 34% of adverse events. Although not as high as the rate of error at the prescribing stage (56%), the detection rate prior to administration was alarmingly only 0–2%.1,2 A European study carried out in paediatric care reported that 56% of medication errors were associated with IV medication.
In the past it has been far too easy for healthcare professionals to make a serious error when administering drugs by IV infusion. Whether it be a large volume pump or a syringe pump, the design and capability of these devices has meant that significant overdoses could go undetected – that is, until the patient demonstrates signs of drug toxicity. Pumps have been traditionally ‘dumb’ in that they were unable to recognise that a programmed infusion rate for a given drug was in excess of the recommended safe maximum and could bring harm to the patient. The device would deliver whatever infusion rate was programmed into the pump, no matter how toxic the drug being administered.
As drugs have become ever more potent and infusion pumps have increased in sophistication, pump design has not always worked in the interests of patient safety. In particular, owing to the preponderance of push-button, numeric keypads, it has been as easy to make a mistake in programming the device as to dial a wrong number into a mobile telephone, for example. Incorrect placement of decimal points and zeros has been a particular problem and led to the well-publicised ‘Death by decimal’ cases in the USA in the 1990s. The public profile generated by these fatalities resulted in calls for the infusion devices’ industry to take positive action. The industry’s response was through much-improved design for safety and the development of ‘smart’ pump technology at the end of the decade.
What makes a pump ‘smart’?
The term ‘smart pump’ was originally coined by the Institute for Safe Medical Practices in the USA to describe an infusion pump with an associated drug library that contained correct parameters for all medications to be delivered by the pump. These safety software packages have been available from the major infusion device manufacturers since the early part of this decade. Initial experience in the home market was gained by American manufacturers and the first system was launched in Europe in 2003, with others following quickly. A review published in 2004 compared the features of the systems then available in the UK.
Error-reduction software confers on the infusion device the ability to recognise an attempt to programme an infusion rate outside a pre-determined dosing range. Should this occur, the attempt is blocked and the user alerted. The systems rely on the construction of a ‘library’ of infused drugs within which the limits are identified. Depending on the system or the options preferred, these limits may be soft or hard. A soft limit triggers an alarm that alerts the person programming the pump but also allows that person to override the alert and proceed with the infusion. A hard alert is absolute; the device will not infuse the drug beyond the pre-determined limit. Because lower and upper dosing limits are set, the software also protects against under-infusion as well as against overdose.
When a pump is being programmed, each time an alert is triggered the software will capture a full record of the events leading up to that alert, thereby providing invaluable “near miss” data on errors averted. For the first time a database of potentially harmful adverse events can be constructed without relying on voluntary staff reporting, which, it is generally acknowledged, will only ever capture the tip of the iceberg. Downloading this information can provide valuable evidence to support practice review and training needs assessment.
Infusion rate calculation
Reflecting its American clinical practice origins, the library of drugs uploaded by the user into the ‘smart’ pump is based on one or more standardised concentrations for each drug. Therefore, rather than infusing an individually prepared weight-based concentration at a fixed rate, the pump automatically calculates the volumetric infusion rate for a standardised solution required to deliver the prescribed dose per kilogram. In many respects this function is as valuable an aid to medication safety as the error-reduction software itself. It removes the need for complex calculations, discourages prescribing in millilitres rather than drug mass units and significantly increases the feasibility of drug infusions being prepared in ready-to-administer forms, either commercially or by local pharmacy production units.
Few infusions ‘smart’ protected
Despite recommendations that hospitals should use double-checking systems, such as an independent check by another practitioner and dose-checking software in ‘smart’ infusion pumps and syringe drivers, uptake of this technology in Europe remains low compared with that in the USA. Even allowing for the fact that the technology originated in the USA, the difference in utilisation rates is still profound nearly a decade after the initial availability in Europe. Manufacturers vary in their estimates. One reports single-figure percentages of pumps being sold with error-reduction software in Europe, versus approximately 80% in the USA. Another quotes 65–75% of large volume pumps and 50–60% of syringe pumps in the USA as being ‘smart’, versus a “low and in pockets” uptake in Europe. The waters can be muddied by the fact that pumps are often sold with the technology installed, but it is then up to the user whether or not to install the drug library in order to activate the full functionality. In the UK, it is estimated by one manufacturer that, although the number of ‘smart’ pumps sold is rising constantly, the proportion currently providing full dosage error software protection remains under 10%.
‘Smart’ resistance in Europe
There are six key reasons why ‘smart’ pump technology is not used more widely in Europe.
Lack of standardisation of equipment
In the past, individual wards in many hospitals were allowed to select and purchase their own equipment and this inevitably led to a variety of devices in use. It is not feasible to implement ‘smart’ pump technology unless one manufacturer’s system can be adopted; however, this can lead to concerns about being totally dependent on one supplier. Some organisations regard installing one brand of volumetric device and another brand of syringe pump as standardisation, but even this can cause confusion when two different versions of error-reduction software are in use in the same hospital. Significant progress has been made in this area in recent years, however, and hospital-wide equipment policies are becoming the norm.
Low investment or poorly targeted investment in new equipment
Healthcare has generally lagged behind the commercial sector in terms of investment in technological solutions, although less so in countries such as Germany than in the UK, for example. In addition, new equipment is generally first introduced into intensive care units. There is undoubtedly a level of IV drug administration error in the ITU, as demonstrated in a French study by Tissot and colleagues, but it could be argued that staff in this arena are experts in the field and technological aids to support patient safety are needed more elsewhere. A review of IV therapy-related litigation claims against the UK National Health Service has shown that critical care areas generate fewer claims than general wards.
Resistance to change (particularly adopting standardised concentrations)
‘Smart’ pump technology has its origins in the USA, so it is designed around US models of practice. Initially, for example, error-reduction software was only available for volumetric infusion devices because the administration of drugs via syringes is much less common in the USA than in Europe.
US practice also involves more extensive use of standardised concentrations, whereas the traditional European model is based on practices inherited from an era when pumps were not as reliable, accurate, or sophisticated as today. Then, it was safer to infuse at a fixed rate of, for example, 1mL per hour and vary the dose of drug administered to the patient by varying the concentration of drug solution. The adoption of standardised concentrations seems to be a major impasse because for many it presents a large change in practice. For some, this is a step too far and in certain areas, particularly paediatrics, the concept can be extremely difficult to implement.
Resistance to adopting new technology has manifested itself as deliberate avoidance of the error-reduction software capability. For rapid access in medical emergencies, all pumps have the facility to shortcut the safety check function and users have a tendency to follow this path inappropriately if it is made too easily available during pump programming.
Lack of robust evidence for effectiveness
Device manufacturers have not invested adequate resources in support of high quality research. Does ‘smart’ pump technology really make a significant difference to patient safety? Does the risk avoidance afforded by this system justify the level of investment required? These key questions have yet to be answered as, to date, there is insufficient evidence in the literature to demonstrate benefit unequivocally. There are reports showing achievement of a reduction in IV errors but, at the same time, some published studies from respected medication safety researchers have failed to demonstrate significant benefit.11,12 The design of some studies has not helped the cause, in that nurses were presented with the option of bypassing the error-reduction software at the point of pump programming.
Under-promotion by manufacturers
Given the claims made by the manufacturers, it was surprising that ‘smart’ pump technology arrived in Europe more with a whimper than a bang. There were no major launch events or conferences. Even now, conversation with professionals responsible for delivering infused therapy often gives the impression that there is a widespread lack of awareness of ‘smart’ pump technology, particularly so among hospital pharmacists. The companies offering these products have arguably not invested adequately in ongoing promotion. Could they have focused more on promoting the benefits to hospital pharmacists? Possibly, but, given the comments below, they could be forgiven for not regarding this as a rewarding investment.
Hospital pharmacists: the sting in the tail
Last but not least, and probably most crucially, there has been a lack of interest from hospital pharmacists. For whatever reason, many (probably the majority) of pharmacists do not seem to regard an involvement in the safe administration of intravenous medication as a key component of medicines management.
A position statement from 2004 advocated that: “Pharmacists’ involvement in the implementation and management
of technology for medication delivery is vital to providing pharmaceutical care. For pharmacists, understanding new technologies is as essential as understanding a new class of drugs… In the past, pharmacists usually did not work with infusion devices. Today, participation in the correct delivery of intravenous medications is as much a part of pharmacy responsibilities as ensuring correct drug compounding.”
Nevertheless, an informal poll carried out at a UK Clinical Pharmacy Association workshop in 2009 indicated that the majority of clinical pharmacists in the UK do not routinely check that IV therapy is being administered correctly; checking infusion pump programming against the prescription is not routine practice. Also, at a strategic level, pharmacists are not usually involved in the selection of new equipment or the training on its safe use.
The major obstacle to the implementation of error-reduction software seems to be the availability of pharmacist time to build the drug library. Pharmacists also do not seem to be championing the benefits of standardising drug solution concentrations. Considerable work is involved in constructing the drug library for upload onto the infusion devices; it is a multidisciplinary process but the involvement of the pharmacist is vital. This effort is surely a wise investment given that it is addressing the highest risk therapy area. Hospital pharmacists should be taking the lead on this process and managers should make staff time available as a priority.
The industry is trying to do more to ease the process through the provision of template drug libraries, but it should be considered that there is also a strong value attached to individual hospitals going through the process themselves. Rather like the discussions that take place during the construction of a drug formulary, the process can be as valuable as the product itself, and it is often found that practice is critically reviewed and amended even before the library goes live on the ‘smart’ pump.
Arguments have been put forward about a lack of personnel, particularly from countries where hospitals are staffed by a small number of pharmacists. It could be argued strongly that this technology is in fact ideally suited to these situations. Two weeks’ intensive effort in drug library construction can result in every infusion rate set in the hospital,
24 hours a day, being given a pharmacist check ‘by proxy’ through comparison against the safety parameters that have been established.
‘Smart’ pumps have been around for a long time but uptake of this safety technology across Europe has been low. Although they do not provide the full solution to eradicating IV infusion errors, they are an invaluable component of a package of safety strategies that also includes standardisation of practice, bar coding and pre-manufacture of ready-to-administer intravenous medication.
For example, combining smart pump technology with standard drug concentrations and user-friendly medication labelling brought about a 73% reduction in reported errors relating to continuous infusion of medication. Rather than continuing to do what we have always done, in a time of increasingly restricted resources pharmacists need to base prioritisation of their efforts much more on evidence of where the major risks to patient safety lie in their organisations. We should be placing much more emphasis on the safe delivery of intravenous therapy and becoming much more involved, not only in the process itself, but also in utilising the technology that already exists to bring this to fruition.