This site is intended for health professionals only!
Intravenous (IV) administration of medicines is a routine procedure in health care today. It is a quick way of delivering a dose of drug to the site of action via the bloodstream without the barrier of absorption from the gastrointestinal tract or from muscle tissue. It is generally assumed that 100% of the dose is delivered when this route of administration is used but recent research suggests that the introduction of new practices and devices over the past 10–15 years has resulted in a situation where up to 20% of doses administered in small volume IVs is routinely discarded and never reaches the patient. This occurs because when intermittent doses are administered, a separate administration set is used and discarded once the infusion bag is empty. At this stage the administration set is still filled with drug solution and so this is also discarded. Experts have warned that under-dosing in this way could put patients at risk of treatment failure in situations where dosage is critical, for example, antibiotic treatment and anticonvulsant treatment.1,2 Mark Santillo, who chairs the NHS Pharmacy Quality Assurance Committee, is concerned about the problem and believes it needs to be more widely understood. “Giving the full dose is important – not flushing can lead to very substantial losses in dosage”, he says.
This article describes the evidence for the problem, the contributory factors, the risks that arise as a result and the measures that are required to tackle the problem.
In the past, small volume IV infusions were commonly administered using a secondary set. This was commonly known as the ‘piggy-back’ system (Figure 1). A primary administration set attached to a large-volume, clear solution infusion, such as sodium chloride 0.9%, was set up. To administer a small volume IV, the bag was connected to a secondary set with a short tube, connected to the main IV line via a Y-connector or similar. The small-volume infusion was hung above the primary infusion (a wire hanger was often provided to enable staff to hang the primary infusion lower). This arrangement allowed the small volume infusion to run first, but also ensured that once the secondary infusion had been completed, the primary infusion automatically flushed any drug remaining in the primary infusion set into the patient. This is still the standard method in much of the rest of the world.
Current practice in the UK involves the use of a separate primary administration set for small-volume drug infusions. This practice has developed gradually over the past 10–15 years, possibly because it is cheaper than the ‘old’ way of using a secondary administration set. The separate primary administration sets are much longer than secondary sets and retain more drug infusion, so they need to be flushed with sodium chloride 0.9%, or a compatible solution, to administer the full prescribed dose. Clearly, failure to flush a separate primary set means that the patient does not receive the fraction of the dose that is left in the tubing. Research has shown that the residual volume of primary administration sets ranges from 10–22ml1 and so, if the initial dose volume was 100ml, this could represent a loss of 20% of the intended dose. Laura McLachlan (Lead Outpatient Antibiotic Therapy (OPAT) clinical nurse specialist, Hull University Teaching Hospital) confirms the scale of losses. “We had just accepted that there would be some losses but when we calculated it could be 10% of the total dose each time we realised it was serious. Some of our patients receive treatment for 12 weeks for serious, complex infections including brain abscesses, endocarditis, graft infections and prosthetic joint infections. The majority of doses are given in small volumes and we calculated that they could be missing the equivalent of 10–12 doses over this period”, she says.
In 2007, the then National Patient Safety Agency (NPSA) issued a Patient Safety Alert on the safer use of injectable medicines3 that highlighted the importance of “designing adequate and pragmatic procedures for IV safety”. However, a recent study of IV infusion administration in 16 hospital Trusts in England found widespread variations in practice.4 Focus groups revealed lack of policy awareness, ambiguous policies, safety and practicality concerns, different organisational priorities, and wide variation in policies and practice relating to prescribing and administration of IV flushes and double-checking. The authors described “a widening gap between work-as-imagined (policies) and work-as-done (practices)” and called for “more national guidance and standardisation … to help Trusts devise appropriate IV infusion policies”.
A recent survey conducted in one UK hospital showed that 74% of small volume IVs were not flushed resulting in losses of 5–21% of doses. They noted that flushing was routine in the oncology day ward but not in surgical and medical wards, including Critical Care and the High Dependency Unit (HDU).1
A service improvement initiative reported by Rachel Dixon (the former Infusion-Coordination Manager, Hampshire Hospital Foundation Trust) showed that the problem is far from unusual. She said, “An urgent issue came to my notice within the Trust regarding the under-dosing of patients with infusions of IV antibiotics by anything up to 50% of the prescribed dose. It involved the antibiotics given via an IV infusion (rather than a bolus push), and occurred because clinicians were not flushing the infusion lines when the bag containing antibiotics was empty.”5
Deborah Barry (Infection Control, Lead Nurse, St Bartholomew’s Hospital, London) found differing practices amongst a team of ten infection control nurses – recently-qualified nurses did not flush lines when giving antibiotics. When she conducted a spot-check across several wards, she found that most nurses did not flush after antibiotic administration unless a Y-line was in place. Moreover, many nurses thought that ‘line’ meant the part that was actually inside the patient rather than the exterior tubing.
The major risk of under-dosing is therapeutic failure or less-than-optimal response. Antibiotic treatment and anticonvulsant treatment are two areas where this could be critical. When antibiotic infusions are being administered to patients with sepsis, for example, under-dosing will lead to lower antibiotic blood levels and less time (or no time) above the minimum inhibitory concentration (MIC). Consequently, the antibiotic might not be as effective as expected and the risk of antibiotic resistance is increased. Balwinder Bolla (Consultant Antimicrobial Pharmacist, United Lincolnshire Hospitals Trust) says that IV under-dosing is not a well-recognised risk. At her hospital, checks of residual volumes after administration of IV antibiotics showed that more than 20% of the dose was lost on some occasions. “From the antimicrobial point of view we must hit the MIC”, she says. It is worthy of note that in the study by Cooper and colleagues,1 the two largest categories of drugs administered in small volume IV infusions were antibiotics and analgesics. Patients who are being treated for epileptic fits with phenytoin infusion will have lower blood levels and time to reach the loading dose will be longer. If paracetamol is being given for febrile convulsions, under-dosing could prolong seizures.
Mark Santillo says, “When you get reports of anti-epileptics not working, you wonder whether they got the full dose”. It is often claimed that under-dosing in this way makes no difference. “How do they know? Under-dosing could be contributing to extended stays in hospital”, he adds.
It is almost impossible to demonstrate a reduced effect in these situations and poorer than expected results might be dismissed as ‘normal biological variation’. It is not surprising that no published evidence exists to indicate whether under-dosing causes harm as research is not usually undertaken to determine if medicines can be under-dosed. Moreover, it is unlikely that any funding body would look favourably on a study designed to show the effects of under-dosing. Clearly, a prospective study would be out of the question because it would be unethical to under-dose patients deliberately (as opposed to casually under-dosing by default as currently happens in practice). A cohort study might be possible although the practical difficulties would be considerable and funding bodies might reasonably ask, why spend money on such a futile question when the technology exists to give the correct dose in the first place?
There is also a legal consideration. When a nurse administers a dose in accordance with a prescription, in which the dose and volume are specified, there is an expectation that the whole of the dose will be given. This cannot be achieved with the method described if the line is not flushed, and the patient will receive only 80–90% of the prescribed dose. This practice could be viewed as a breach of the Medicines Act 1968, which requires medicines to be administered in accordance with a prescription. Moreover, it conflicts with professional standards for nurses, midwives and pharmacists, which also require medicines to be administered in accordance with prescriptions. It is argued that if a pharmacist knowingly dispensed only 80–90% of a drug dose, it would be seen as professional negligence. It is no more acceptable for a nurse to administer a dose using a system that only delivers 80–90% of the dose.
Furthermore, using a system that routinely delivers lower doses than those specified in the Summary of Product Characteristics (SPC) constitutes an ‘off label’ use of the product and liability passes to the hospital.
It can also be argued that under-dosing could lead to wasted resources – equipment and drugs – in addition to wasted clinician time, longer recovery times and delayed discharges.
The existing guidance on how to administer small volume infusions accurately and safely is sketchy and incomplete. The matter of dealing with the drug solution left in the tubing is usually not mentioned.2 When flushing is mentioned, it usually refers to flushing of the cannula (see Box 1) rather than as a measure to ensure full infusion doses. Deborah Barry found that the intravenous policy at St Bartholomew’s Hospital “just says ‘flush’ but is vague about exactly what should be flushed”. She says they plan redraft the policy to be more precise and not to leave room for different interpretations. The Marsden Manual of Clinical Nursing Procedures advises that “a ‘minibag’ may be used to flush the drug through the tubing” but says that “the cost implications of this as well as the risk to patients on restricted [fluid] intake should be considered before this is adopted routinely”. The rationale behind this practice is “to flush any remaining irritating solution away from the cannula”.6 There is no mention of the need to administer the full dose prescribed as a rationale for flushing the administration set. Arguably, under-dosing in this way could lead to longer stays in hospital and costs that far outweigh the cost of giving the full dose of antibiotics.
To ignore the problem is effectively to accept that a system that administers 80–90% of the prescribed dose must be satisfactory because it has been done that way for several years without apparent ill effects and that administering the whole dose would incur unnecessary additional cost. This puts the UK out of step with other countries (for example, US and Germany)
In order to administer small volume IV infusions accurately, in accordance with prescriber instructions and with dosing statements in the SPCs, hospitals need to find ways to make this happen at reasonable cost. One option is to use a flushable administration set, for example, Intrafix SafeSet Flush. This administration set allows a flush solution, such as 0.9% sodium chloride, to be administered via the Caresite needle-free valve using a prefilled syringe, ensuring that a closed system is maintained. This guarantees that only the active drug is administered, with no risk of fluid overload. The AirStop filter prevents air entry.
The SafeSet Flush system is used by the OPAT service at Hull University Hospital. Laura MacLachlan says, “We knew that we were losing about 10% of the dose by not flushing and as a specialist IV service we felt that we should lead the way. We also considered other ways to tackle the problem. We could put up another bag [to flush the line] but this comes with risks – air embolism, the risk of infection arising from another manipulation and the risk of fluid overload. The SafeSet Flush system is closed so it eliminates the risk of air embolism – it is a safer way of doing it”,
Three possible options and the additional costs for a typical district general hospital (DGH) are set out in Table 1. It would be possible to administer 100% of antibiotic doses in the average DGH for an additional cost of just over £21,000 per year.
When adverse drug events are discussed, prescribing errors and administration errors – such as calculation errors, drawing up the wrong dose, using incorrect diluent – feature commonly. However, administering only 80–90% of the dose because of the default administration system could also be considered to be a ‘wrong dose’ event. Pharmacists should take steps to find out about practice in their own hospitals and ensure that there is awareness of this issue and that measures are in place to ensure that intravenous drug administration is safe and accurate.
If hospital Trusts choose not to implement systems to ensure that 100% of critical doses are administered, the risks should be recorded in the hospital’s risk register, in particular, the off-label use of antibiotics/medicines. After all, at some point in the future Big Data might show that hospitals that routinely under-dose patients with IV drugs achieve poorer clinical outcomes than those that give full doses.
The paucity of clear, up-to-date guidance in the UK on how to administer small volume infusions accurately and safely is a serious gap in safe medication practice. Pharmacists should take the lead in raising awareness of the issue and developing guidance to ensure that patients consistently receive 100% of the intended treatment. As Mark Santillo argues, “the pharmacy profession [now] needs to engage with the nursing profession to address the problem.”