NHS PASA CEP
Evaluation details for infusion pumps mentioned in this article were summarised with the consent of NHS Purchasing and Supply Agency Centre for Evidence-Based Purchasing. Reports were originally prepared by the Bath Institute of Medical Engineering and funded by NHS PASA
Intravenous (IV) medications are associated with high risk – a fact that has been recognised for many years. This was recently highlighted in the UK by NHS National Patient Safety Agency alerts regarding injectable medications (see Resources). IV infusion pumps are regularly used for administering total parenteral nutrition (TPN), anaesthetic agents and opiates/patient-controlled analgesia (PCA), because pumps are able to deliver IV fluids in a way that would be impractically expensive or unreliable if fluids were delivered manually. A range of infusion pumps is available on the European healthcare market. These combine different medication administration safety features – for instance, many infusion pumps are controlled by small embedded systems designed in such a way that no single cause of failure can lead to patient harm. Emerging “smart-pump” technologies incorporating dose error reduction systems have been introduced in the hope of preventing most errors.
In summer 2006 an independent analysis of several of the most important infusion pump models on the market was undertaken by the Centre for Evidence-Based Purchasing (CEP), part of the NHS Purchasing and Supply Agency (NHS PASA). CEP aims to provide independent and objective evaluations of medical devices available on the UK market. Specialist centres, mainly in NHS Trusts, undertake evaluations under contract to NHS PASA and the infusion pump evaluations have been carried out by the Bath Institute of Medical Engineering (BIME; see Resources). Testing used the same protocols to allow direct comparison and offer potential purchasers guidance on which product might best meet their needs.
Fresenius Vial Base Primea(1)
CEP tested the Fresenius system regarding its use in target-controlled infusion (TCI) of intravenous anaesthesia delivery. TCI was originally developed to improve convenience and control during intravenous anaesthesia. The anaesthetist sets and adjusts target blood concentration and depth of anaesthesia as required on clinical grounds. Infusion rates are controlled automatically according to a validated pharmacokinetic model.
A usable TCI system requires:
- An acceptable, validated pharmacokinetic model to control drug delivery.
- An accurate pump, with a control system capable of implementing the pharmacokinetic model.
- A safe, usable interface, featuring a clear display of necessary data, a straightforward method for controlling the infusion, and a clear alarm system to indicate problems.
- Regulatory approval for the system to be commercially produced.
CEP tested the Fresenius pump separately for technical performance (focusing on performance in TCI mode) and usability features.
Technical performance testing indicated that performance was accurate enough to be considered clinically acceptable. Drug concentration, calculated from actual fluid delivery, was seen to deviate from target concentration by 10–20% after certain events (syringe change, line occlusion or pump being raised on a drip stand). These errors were reduced to less than 5% within a few minutes in all cases. None of the technical data support making a distinction between this system and other evaluated TCI pumps on the grounds of performance. CEP said the Fresenius system offers greater functionality than Diprifusor pumps in offering alternative pharmacokinetic models and effect site steering as well as plasma steering. Marsh and Schnider models for propofol and the Minto model for remifentanil are offered. Technical performance was equally good in tests on all pharmacokinetic models.
The ergonomic review assessed pump usability by combining three independent assessments of usability problems noted while completing clinically typical tasks (including tasks relating to use of the advanced functionality) and rating these problems based on clinical significance. In total, 126 usability problems were highlighted. Of these problems, 75% were rated as “minor” or “cosmetic”. There were 33 “major” problems but no problems received a “catastrophic” rating. The total number of problems is similar to that found for the Alaris PK system and greater than the numbers found for Diprifusor systems, reflecting at least in part the greater functionality of post-Diprifusor “open” TCI systems. However, the number of major problems is significantly higher than for any previously evaluated TCI system, suggesting the device may be harder to use. This may be explained by the use of a modular system, resulting in the user having to manage information on more than one display. Several of the more serious problems may be overcome through familiarity with the system, so potential purchasers should be aware that appropriate training is essential for novice users before the system is used with patients.
However, in spite of foreseen difficulties in adjusting to the modular system, a survey of anaesthetists using the Fresenius system found initial setup and inputting of patient data to be relatively easy, and target control very easy. Overall, the product was found to have a usability rating of 3.9 (on a 1–5 scale), with the main complaints relating to poor-quality instruction manuals and counterintuitive functionality. This may relate to inexperience with modular systems, although the survey also found customer support and training provided by the manufacturer to be excellent.
Overall, combining technical performance and usability, the Base Primea system was found to be an excellent and highly flexible system which was strongly recommended by both CEP and the survey panel. It was noted that using it necessitates a steep learning curve, so it was not recommended for clinicians lacking experience in regular TCI procedures. However, the pump is “failsafe” in that one cannot proceed if the startup sequence is incorrect.
Mckinley T34 syringe pump(2)
The Mickinley T34 syringe pump (commonly referred to as a syringe driver due to its small size) aims at a particular clinical niche: ambulatory small-volume infusion. It was included in the CEP report since similar products in this area universally fail several of the fundamental safety requirements delineated by IEC 60601-2-24 (relating to the provision of mechanical interlocks within the syringe and provision of alarms during incorrect operation). The T34 is the first ambulatory syringe pump which addresses all the safety features necessary for compliance. This compliance is a strong, competitive selling point for this simple pump. Multidisciplinary palliative-care teams with palliative-care pharmacists may particularly benefit. The pump retains the small size of similar devices but is heavier and has a large protrusion on the reverse surface to accommodate the mechanism for detection of syringe size. This means the pump can be programmed in ml/h rather than mm/h or mm/24h, reducing opportunities for user error and the associated possibilities of serious under- or overdose. Other ambulatory pumps provide no means of detecting syringe size, so the avoidance of dosing errors depends on the user to use only one syringe size. Unusually for a pump of this size, the device includes history logging – a record of button presses maintained within the pump which can be reviewed later.
Technical assessment of the T34’s long-term accuracy examined a full range of flow rates from 1 to 1,000 ml/h. A range of syringe brands and sizes were used in testing. The T34 gave a delivery rate well within expected tolerances in all tests.
The short-term accuracy of syringe pumps is expressed in terms of constancy index, which is measured at 1ml/h and indicates the minimum period of time over which the flow rate remains within 10% of the mean flow rate. The constancy index for the T34 is 1.5 minutes, which is deemed excellent for an ambulatory pump of this type. Previously evaluated small-format syringe drivers have all had a longer constancy index (ie, less smooth flow patterns).
Usability of the T34 was determined via assessment on four user sites – an oncology ward, a gynaecological oncology ward, a hospice and a health centre. In terms of setup, automatic calculation of the rate by the pump was found to be much easier than setting the rate with other previously used syringe drivers, particularly in terms of being able to lock the syringe in place, and the automatic detection of different syringe brands and sizes. Alarms and indicators were generally found to be very good, with the volume-to-be-infused (VTBI) light particularly useful. However, one user found the backlighting did not stay on long enough and had to relight it using the info button. Other niggles noted by users included the unit’s weight and the small size of function buttons. In general, users commented that the system was reliable and easy to use and said they hoped to eventually be able to use it permanently. CEP specifically made note of the fact that all issues regarding improvements in technical performance and usability suggested by CEP and at the four review sites were proactively responded to by McKinley, with a revised version of the pump being supplied to the users.
Codan Argus 708(3)
The Codan Argus 708 is a volumetric infusion pump offering standard features, as well as some specialised therapies such as blood transfusions and parenteral nutrition. A central docking station can be supplied for use with multiple pumps, offering use of a single power cable and connection to patient data management systems. P-versions of the docking stations are available without the data transfer capability. The pump can be programmed for rate, volume to be infused (VTBI) and volume over time. Both preset and manual bolus deliveries are offered. Comprehensive alarms are offered and a large dual red light on the pump provides a distinctive alert. Unfortunately, the limited seven-segment display restricts the usefulness of alarm messages, making understanding and rectifying alarms problematic at times. Occlusion alarm settings include special low-pressure settings designed for neonatal applications. However, CEP noted that usability problems encountered with the pump could preclude use of the pump for critical care applications. Accessories for the pump include a drop detector, barcode reader and associated software, and a bottle holder.
Regarding long-term technical accuracy, the Codan Argus 708 pump was tested over the full range of flow rates offered by the device, using the dedicated Codan IV set, IVIP 708 B86-Y. The other dedicated sets offered for this pump (IVIP 708 B88 for blood infusions and IVIP 708 V86) were not tested, as all these sets are made from the same materials (PVC) and should perform similarly in flow accuracy tests. The manufacturer recommends changing the administration set every 24 hours or after 2.5 litres of fluid have been delivered, whichever occurs first. The results of flow-rate accuracy testing showed a slight tendency for the test pump to overdeliver. The greatest overdelivery recorded was +5.2% at the maximum flow rate of 999.9 ml/h.
With respect to short-term accuracy the Argus 708 showed constancy index of 4.5 minutes. Based on this, drugs with a half-life shorter than 4.5 minutes would not be recommended for delivery via this pump at low flow rates (1ml/h and lower). Startup time – the time taken to start delivering steadily – for this pump is very rapid (0.5 minutes), as is typical for volumetric pumps.
Laboratory-based user assessment of the Codan Argus 708’s usability was conducted using a task list comprising common required procedures for this type of pump. The task list was constructed using a combination of interviews with appropriate users and inspection of the history logs of recently clinically used pumps of this type. Overall, the system was found to offer a wide variety of configurable features, but was limited by its somewhat rudimentary display. The two findings judged to be in the most severe band were (1) use of the alarm silence key simultaneously triggers a confusing change in the displayed information, and (2) the power-off button was configured on the test pump to be instantaneously effective and also cleared previous displayed settings, leaving the user vulnerable to accidental and/or unnoticed switch-off and loss of parameters without evident means of reversing the loss. This problem would not arise if the pump was configured differently. The highly flexible configurations offered by the Argus 708 are intended to make the pumps more marketable to a wider range of customers.
CEP concluded that a more limited, safety-oriented configuration would have scored better for good usability features. This finding complies with NPSA and MHRA advice advocating standardisation of infusion pumps in Trust settings – subtle changes in pump functionality that are not obvious to the user, but are configurable through the technician menu, can lead to hazardous confusion on the part of users. Codan responded that configurations of pumps for clinical use would be discussed fully with Trusts before being set up, alleviating problems noted here where configuration was a contributory factor.
Overall the Codan Argus 708 is a pump with a basic user interface that has some usability problems, some of which would decrease with familiarity. Some of the difficulties arise due to features being unconventionally configured. Furthermore, limitations of the text message display can leave the user unsure how to recover errors or proceed with required procedures.
It should be noted that the usability evaluation approach used is a negative process, primarily designed to identify features which are suboptimal. Numerous good features were also identified; examples of these were the pump’s attractive design and compact size. The pump is available at comparatively low cost. For familiar users with only routine requirements the pump would probably perform well and safely.
B. Braun Infusomat Space volumetric pump(3)
The Braun Infusomat Space volumetric pump offers extensive features to meet the needs of both general ward use and critical care applications. It is designed to be used on its own or in combination with its sister pump, the B. Braun Perfusor Space. Both pumps can be fitted interchangeably in the B. Braun Space Station to facilitate control and monitoring of multiple infusions, while also providing power to multiple pumps using a single cable. The Space Station also features advanced networking facilities (including drivers to link to hospital information systems, and the ability to connect to a wireless LAN) and an enhanced alerting system utilising large warning lights. Multilevel security settings and dose-limiting software reduce the likelihood of tampering and use error. As with any new device, the pumps will appear unfamiliar to new users, but the overall Space concept has clearly been developed with a user-centred approach in mind, and the clear messages and logical programming steps promote safe use.
During technical assessment, the Infusomat Space was tested over the full range of available flow rates, from 0.1 ml/h to 1,200 ml/h. Standard 250 cm Infusomat Space Lines were used for all tests. The alternative sets offered by Braun feature the same pumping segment as the standard set, so similar fluid delivery performance would be expected. Excellent delivery accuracy was seen in all tests at flow rates greater than 1 ml/h. A relatively large delivery error was measured at 1 ml/h and an unacceptably large error was seen at 0.1 ml/h. These errors are worse than have been seen on most comparable pumps that have been evaluated. Low flow rates should be avoided when using volumetric pumps since accurate delivery cannot be assumed. Efforts to avoid low flow rates should be made for fast acting drugs, and the use of syringe pumps should be considered when low flow rates are clinically indicated since they can offer more consistent fluid delivery, provided the required volume does not involve a syringe change.For short-term accuracy the constancy index for the Infusomat Space was measured at 16 minutes. As with all volumetric pumps this is a long time compared with modern syringe pumps. A long constancy index indicates the pump is unsuitable for use with short half-life drugs, particularly at low flow rates. Constancy is relatively poor compared with other volumetric pumps, and as expected is worse than for syringe pumps.
User assessment of the pump was requested from 28 NHS trusts and private clinics recommended by Braun. Questionnaires were then sent to users who offered to provide opinions on their experiences with the pump. Twenty responses were received, from five NHS trusts. Most respondents were nurses, with between two weeks’ and 10 months’ experience with the pump, spanning a range of clinical areas (in ITU/high-dependency units, and haematology and rheumatology departments). Two responses were received from medical equipment engineers. The results of the survey showed, among other points, that an excellent training package was supplied along with a clear and concise manual. It took some time for users to become accustomed to the pump’s relatively small size, but the stackable configuration was rated very positively. The device was found to be both robust and easily configurable (if a little fiddly owing to small size), with clear displays and excellent alarms.
The laboratory-based usability test noted that Braun has designed much of the functionality of the user interface on these pumps around CD players, on the grounds that users will be universally familiar with the symbols. This is a laudable incentive towards implementing standards, but users did find themselves unfamiliar with some of the symbols in the context of a medical device. It remains to be seen whether drawing on standards used in other fields will ultimately reduce or increase user confusion. The device’s physical layout means the set loads from right to left, which is conventional for Continental Europe, but not in the UK, and this unfamiliarity might lead to errors. Acclimatisation to the system would reduce the likelihood of error here.
The Braun Infusomat Space volumetric pump has many features and generally clear messages and alarms. Increased versatility would normally result in a larger number of usability problems than with a simpler pump. In view of this, the Space volumetric pump scored particularly well in this usability assessment, as further evidenced by the large number of positive evaluator comments. However, it may take users some time to become familiar with the interface’s unfamiliar format and layout. The most significant problem identified is the removal from view of the interface and the guiding messages while the user completes core procedures. This is an inevitable consequence of locating the display on the door. There is an increased risk of user error during periods of familiarisation, where old procedures are forgotten and replaced by new ones. Purchasers should ensure all users are adequately trained before using the system on patients.
Each of the systems tested by CEP displayed different features and levels of functionality. Technical performance testing demonstrated that the reviewed systems are suitable for clinical use but manufacturers’ default configurations and system-specific features suggest that some pumps are more suited to certain clinical applications than others. Built-in dose error reduction software has the potential to reduce many associated problems with IV medication infusion, but this does not reduce the need for user-friendly interfaces, clear displays, comprehensive instruction manuals and manufacturer after-sales support. It must be noted that following CEP’s evaluation of these pumps, each of the manufacturers addressed the negative issues found during technical assessment and each has retrospectively implemented system improvements. With a predicted move towards the integration of electronic patient records and clinical information systems, future infusion pump purchasing decisions should include consideration of the pumps’ (and/or docking stations’) ability to interface with other systems, including USB, infrared and wireless networking facilities.
1. Centre for Evidence-Based Purchasing. Report 06001: target controlled infusion (TCI) systems. Part three: Fresenius Vial Base Primea. London: NHS PASA; 2006.
2. Centre for Evidence-Based Purchasing. Report 05076: McKinley T34 syringe driver. London: NHS PASA; 2006.
3. Centre for Evidence-Based Purchasing. Report 06022: B. Braun Space volumetric and syringe pumps and Codan/Argus 708 volumetric pump. London: NHS PASA; 2006.
Tel: 020-7972 6080
Bath Institute of Medical Engineering
UK National Patient Safety Agency