Safe IV therapy – innovative solutions

Most IV incidents are preventable but often there is a failure to implement safer systems, said Laurence Goldberg (Independent Pharmacy Consultant, Lancashire, UK) introducing the symposium. Common examples include purchasing intelligent pumps but turning off the safety features because of the work involved in setting them up and not purchasing ready-to-administer (RTA) injections because they are perceived to be expensive, although proper pharmacoeconomic evaluations are rarely performed, he added.

Centrally prepared, standardised, injectable medicines could help to reduce errors with parenteral therapy, according to Claudia Langebrake (Clinical Pharmacist, Haematology and Oncology, University Hospital, Eppendorf, Hamburg, Germany). IV therapy is a complex healthcare technology in which the preparation and administration of injections is often undertaken by nurses on wards. The process is open to errors such as injection by the wrong route, dose calculation errors, selection of the wrong drug or even administering the dose to the wrong patient, she continued.

A UK study published in 2003 had shown that at least one error occurred in almost 50% of IV doses. The majority of errors were classified as moderate but about 1% were severe. A multicentre study involving 113 intensive care units in 27 countries examined errors in the administration of parenteral drugs. The results showed that there were 74.5 events per 100 patient-days. The most common types of error were administration at the wrong time and missed doses, but 12 patients experienced permanent harm or fatal injuries. The contributory factors included complexity of care, a high workload causing stress and fatigue and poor communications. The authors of the study had concluded that parenteral medication errors at the administration stage were common and were a serious safety problem in intensive care units.

A more recent study evaluated the preparation and administration of 568 IV doses by 107 nurses on six wards. There was at least one clinical error in 70% of the doses and 25.5% of errors were categorised as ‘serious’. Three important findings from the study were that infusion pumps were rarely used, the use of a pump was not associated with a reduction in errors and bolus administration was associated with a threefold increase in errors. In addition, patient identification was only checked on about 50% of occasions but when it was performed it reduced errors by more than 50%. The root causes of errors were skills and knowledge deficits, in some cases, and violations of routine procedures likely to have been learned in the workplace.

Turning to microbial contamination of IV doses, Dr Langebrake said that a systematic review and meta-analysis of 19 studies involving 7,277 doses had shown that when prepared in clinical areas there were errors in 5% of individually prepared doses and in 2% of doses prepared as batches. For doses prepared in pharmaceutical environments, there were errors in 1.9% of individually prepared doses and none in those prepared as batches.

A study conducted in one German hospital examined the accuracy of preparation by comparing the concentrations of drugs in injections prepared in the pharmacy with those prepared on wards. The results showed that the actual concentrations of drugs varied from the labelled concentration by more than 5% in 53% of manually-made injections and in 16% of machine-made (in the pharmacy) injections. Moreover, 15% of the manually-made injections deviated by more than 15% from the nominal concentration. Hydrocortisone and amiodarone injections were examples of preparations where actual concentrations deviated widely. The preparation of hydrocortisone injection is a two-step process and this may have contributed to the variability in final concentrations, noted Dr Langebrake.

The error-prone steps relating to injectable treatment include communication and the preparation process itself, especially if this is a multistep procedure. Additional contributory factors include fatigue, stress, lack of awareness and absence of routine checks, she said. Measures to make infusion therapy safer would include the provision of ready-to-use (RTU) or RTA products and the implementation of closed-loop administration processes using electronic prescribing and barcoding at the point of care.

The implementation of smart pumps in chemotherapy requires a different approach from that in other therapeutic areas, said Gerardo Cajaraville (Director of Pharmacy and Patient Safety, Onkologikoa Hospital, San Sebastian, Spain). Smart pumps were introduced at the Onkologikoa Hospital several years ago following an incident in which a patient received the wrong dose of 5-fluorouracil. The Onkologikoa Hospital is a private, not-for-profit cancer treatment centre that deals mainly with patients in the national health system. It has had electronic patient records since 2008 and operates as a ‘paperless’ hospital.

Chemotherapy presents many opportunities for error and even small mistakes can have devastating effects, said Dr Cajaraville. At his hospital all the processes from prescribing through to administration and toxicity monitoring are integrated and electronically controlled. Electronic prescribing has been in place since 2000, he added. Pharmacists have access to the electronic clinical records and add value to patient management through application of their own clinical skills. Compounding is electronically controlled and everything is checked using barcodes. A robotic system for chemotherapy compounding is now in use, he said.

An electronic medicines administration record (eMAR) is made using wireless computers on wheels. All the devices are IP65-rated, which means that they are effectively waterproof and therefore can easily be decontaminated without danger to the instruments or operators. The on-screen instructions lead the nurse through the process, he explained. The nurse first reads his or her personal bar code (on the ID badge), then checks the drug barcode and then the patient’s wristband or outpatient card. The barcode reading takes place at the time of administration, and it effectively records the intention to administer the drug but not the beginning and completion of administration, he emphasised.

Some years ago there was a serious error in the administration of 5-fluorouracil, as a result of a pump-programming error. Although the patient was unharmed, the incident highlighted the ease with which pump-programming errors could occur and prompted the adoption of smart pumps in the hospital. Smart pumps require the compilation of a drug library and agreement with clinicians about soft dose limits (which can be overridden) and hard dose limits (which cannot be overridden). Smart pumps can safeguard against dosage calculation errors but they do not prevent errors with the other four ‘rights’ – right patient, right medication, right route or right time unless they are integrated with electronic prescribing and administration systems and support the use of barcode technology, said Dr Cajaraville.

When used correctly, smart pumps reduce the frequency of errors. One study had reported a 22% reduction in adverse drug events after the introduction of smart pumps. An observational study of more than 600,000 paediatric infusions had shown that 92 errors were prevented. However, several earlier studies had shown that smart pumps made no difference to error rates. This was almost certainly due to the way in which the pumps had been used, he said.

The change from conventional pumps to smart pumps involves much more than the purchase of new pumps. It requires the compilation of drug libraries, compliance, systematic analysis of pump logs and updating of the system – something that is virtually impossible without wireless connectivity – and clinical leadership. Ideally smart pumps should be integrated with the electronic prescribing and administration systems. However, much like seatbelts in cars, the use of the smart pumps is optional, even though the value and effectiveness have been demonstrated, he added.

Turning to the use of smart pumps in chemotherapy, Dr Cajaraville said that a library was not useful because concentrations vary considerably and dose calculations for chemotherapy are rarely performed at the bedside.

Nevertheless, it is important to avoid programming errors with the infusion rates. It is also important to be able to accommodate complex administration protocols and, if possible, to monitor remotely. The risk of extravasation has to be minimised. The relationship between extravasation and infusion pressure needs to be understood and local policies should be in place to minimise risk, he said. The real-time connection is an important factor because it enables real-time monitoring of the pump and adjustment, if necessary, which can be crucial for the patient. Traceability is also critical – information from the pump log can be used to inform subsequent treatment or corrective measures.

At Onkologikoa, smart pumps are used to ensure that medicines are administered correctly making use of the real-time connection and alerts. The pumps currently in use are wirelessly connected to the chemotherapy software so that the administration parameters can be sent. At the point of administration, the nurse takes a fourth barcode reading – the pump. When all four barcodes have been read, the system displays the administration parameters on the pump screen so that the nurse can compare them with the prescription. The communication system is bidirectional and feedback is continuously sent from the pump. Critical events trigger warnings that are displayed on the computer screen. If there is a problem or discrepancy this must be resolved before administration of the drug can continue.

There have been a number of challenges in this project. It has been important, but sometimes difficult, to explain that a drug library adds little value in this field and that the pump is effectively in charge of safe IV administration. Wireless connectivity was essential for the scheme to work and there were other technical difficulties, he acknowledged. However, there was little resistance to change, as such, at Onkologikoa hospital.

In conclusion, Dr Cajaraville said that the implementation of smart pumps in the chemotherapy process requires a new approach. Compliance is critical and must be assured by strong management. The benefits have been that real-time managing of messages sent by the pumps has improved patient safety and that the introduction of smart pumps has allowed the hospital to complete its paperless chemotherapy process and improve the monitoring and recording of chemotherapy administration.

Some injections are simply not fit for purpose, according to Clare Crowley, (Consultant Pharmacist, Medicines Safety, Oxford University Hospitals NHS Trust, UK) For example, gentamicin injection 80mg in 2ml is supplied to prepare a once-daily dose of 420mg. Six ampoules are required to prepare the dose but nurses have been taught that the threshold for alarm is when more than two ampoules are required. Another example is paracetamol 10mg/ml in a 50ml vial – a product that is good for administration of a 500mg dose but potentially dangerous when preparing a dose for a neonate for whom the dose could be as little as 7.5mg/kg, she explained.

The ideal injectable medicines would be supplied in a concentration that is suitable for administration and therefore require no calculations or preparation. The vial size would correspond to commonly used doses and the label could be clearly read when in use. Technical information would be provided in the package and the product would be licensed as a medicine and not as a device. Finally, the product could be stored at room temperature, “because we are running out of refrigerator space!” said Dr Crowley

A major advantage of RTU or RTA injections is that they provide fewer opportunities for error and so patient safety is enhanced. There is less risk of particulate and microbial contamination, more timely administration is possible and nursing time is released for other duties.

A simulation study had compared the concentrations of epinephrine and norepinephrine injections prepared de novo by physicians with doses prepared in advance by physicians or pharmacists and those prepared by the industry. In each case the injections prepared de novo varied widely from the expected or intended concentration.

The introduction of RTA insulin (1mg/ml) injections had saved an average of nine minutes per dose of nursing time and led to considerable satisfaction among users. Similarly, the introduction of noradrenaline 4mg in 50ml syringes had saved seven minutes per dose.

Despite the obvious advantages of pre-prepared injections there are also drawbacks, said Dr Crowley. At present there are relatively few licensed products available and they do not always mirror existing clinical practice. They usually have short shelf-lives and often have bulky packaging and require refrigerated storage. The direct costs can be higher but purchasers must also take account of the cost of disposables when comparing costs. “Don’t assume it costs more,” she advised. Many unlicensed products are available but these incur large quality assurance costs.

Currently there are a number of barriers to progress, including the lack of clinical consensus about the drugs, doses and concentrations that are required. There is also variability in dose expression with some practitioners using mg/ml and others preferring mmol/ml. On some occasions, national guidelines advocate the use of products that are not available, for example, Hartmann’s solution with potassium chloride. Technical problems, licensing issues and positioning in a global marketplace are all matters that the industry has to tackle. A viable business proposal is essential and pharmacists need to become cleverer at developing these, she suggested. In this respect, there may be a need for flexibility and recognition of the need to work differently, she said. For example, in the UK, paediatric specialists have clung to ‘the rule of six’ (which prevents the use of standardised concentrations for injections) but in the USA they have accepted change.

When it comes to purchasing for safety it is essential to understand how medicines are being used – and then standardise and rationalise where possible, said Dr Crowley. It is also advisable to work with clinical networks and professional bodies to embed the changes into routine practice. In the UK a list of 17 standardised concentrations for injections used in intensive care had been adopted and published by the Intensive Care Society, she noted.

Products needed in clinical emergencies and products that are difficult to mix safely, such as potassium and magnesium injections, should be prioritised when considering new RTU or RTA products. Other high-priority items would be products that are complex to prepare and those where there are high risks, for example, epidural infusion preparations. In addition, any product that is commonly associated with severe harm, such as insulin, opiates and anticoagulants, and products that pose a risk to operators should be prioritised for RTU or RTA presentation.

Dr Crowley said that RTU and RTA products were not a panacea but they would provide a robust defence against preparation errors. Strong leadership and collaboration between all the stakeholders were now required to carry this forward, she concluded.