This site is intended for health professionals only

Implementing change, improving patient safety


Christine Clark
BSc MSc PhD FRPharmS FCPP(Hon)

Understanding human factors is the key to tackling medication errors, according to Arpan Guha (Consultant, Critical Care and Anaesthesia, Royal Liverpool University Hospital, Liverpool). He said: “The final common pathway is the human element—often thought of as the weakest link in the chain.” Dr Guha has a special interest in decision making in acute situations, in trying to understand why people behave as they do and how errors with intravenous injections often occur in heated situations where circumstances collude to lead individuals to make certain decisions.

Errors have been defined as ‘occasions in which a planned sequence of mental or physical activities fails to achieve its intended outcome.’ Dr Guha said this definition is useful because it refers to mental activity of the individual involved, which is not amenable to a ‘technofix.’ He also noted that some intrinsically risky procedures, such as blood transfusions, had been made safer by the implementation of fail-safe systems.

Slips and lapses are unintended errors that can occur as a result of inattention or distractions whereas mistakes and violations, he explained, are often the result of a decision, for example, to take a short cut. In one study human factors were estimated to account for 66% of errors in anaesthetic practice. Another study showed that accidental interchange of syringes was one of the commonest errors. Such errors were usually attributed to inattention, haste, drug labelling or distraction.

The attitudes of doctors and nurses are critical to any error reduction scheme. A study of doctors and airline pilots showed that doctors generally believed that they performed effectively when fatigued whereas the majority of pilots believed that their performance was impaired. A study among nurses revealed the following views: “If it is not my fault is not an error” and “If you can put it right it is not an error.” In the face of such ingrained beliefs, exhortations to follow safe procedures do not work, explained Dr Guha.

A great many errors are caused by communication failures. Sometimes these could be fixed by something as simple as a feedback loop that can be very cost-effective. Task analysis research offers an insight into the ways in which errors occur. In a series of 47 simulated anaphylaxis scenarios, healthcare practitioners made medication errors in 21 – typically ‘wrong route’ or ‘wrong dose’ errors. Often these were due drawing up the wrong dose (for example, whole ampoule instead of dose required) or failure to check dose and route. Hierarchical task analysis breaks down each action into the individual steps required. For example, determining the dose of adrenaline involves finding a printed copy of the dosing algorithm and then using it to work out the dose required. Preparing and administering the dose involve many more steps. This type of analysis enables the design of bespoke solutions, said Dr Guha. Better procedures and checks coupled with improved communication can then be developed. “A double check must be a real check with a callback or feedback loop—all the other emergency services follow this procedure be we do not. It only takes one day to train people,” he added.

The value of simulation-based training was underlined by a study that compared the impact of didactic and simulation-based training for critical care nurses on medication error rates. After simulation-based training the error rate in the CCU fell from 30.8% to 4.0% and was sustained. However the error rate in the MICU (didactic group) did not change and subsequently rose.

Human factors engineering is about identifying what people really do and then providing fail-safe solutions. “All you need is someone who understands people,” concluded Dr Guha.

Best practice in IV therapy  
Lisa Dougherty (Nurse Consultant IV Therapy, Royal Marsden NHS Foundation Trust) told the audience that giving nurses more flexibility and educating them about the dangers of work-arounds would improve the standard of intravenous therapy.

In a recent survey in the USA 17,500 nurses had commented on the problems of complying with ‘30 minute rule’ that requires medicines to be given within 30 minutes of the prescribed time. The responses showed that it was impossible to comply for much of the time because of time pressures and having to queue to obtain medicines from electronic cabinets, amongst other reasons. Failure to comply with the rule is evident from the electronic administration record and this can lead to disciplinary proceedings. In order to avoid this many nurses admitted to taking short cuts such as removing medicines from the automated cabinet in advance and recording administration when doses had not been given.

A recent National Patient Safety Agency (NPSA) Rapid Response Alert showed that delayed and omitted medications were not uncommon in the UK. In addition, one survey of insulin treatment revealed that 10% of doses were missed or delayed and in many cases it was not even prescribed, said Dr Dougherty.

Double checks are routinely recommended for intravenous medicines but this is often done poorly in practice—it must be an independent double check, she emphasised.  Other measures that could improve the situation would be better education for nurses, in particular explaining the consequences of late dosing or giving intravenous doses too quickly, and educating doctors about realistic prescribing times. Dr Dougherty also recommended that patients who were accustomed to giving their own insulin should continue to do so in hospital because nurses were often unable to give it at the correct times.

IV standardisation
Standardisation of the intravenous medicines used in intensive care would reduce the risk of errors in preparation and administration of intravenous doses and improve the patient experience. Peter Keeling (Consultant Anaesthetist, Frimley Park NHS Trust) argued that consequently, it could also reduce litigation costs.

Errors during preparation and administration of intravenous doses are unlikely to be intercepted and can account for up to 60% of serious and life-threatening potential adverse events. A recent study shows that there is a high overall probability (0.73) of making at least one error in IV therapy.  If error checking were introduced, the rate could be reduced to 0.22. Moreover, if premixed or ready-to-use injections were introduced the overall error rate could be reduced yet further to 0.17.

A survey of 154 critical care units in the UK showed wide variations in the concentrations of common products including 18 different concentrations of noradrenaline and 19 of adrenaline in regular use. In a follow-up survey the majority of respondents agreed that they would use proposed standard concentrations of 16 commonly used injections, although there was no consensus on adrenaline, potassium and phosphate. Agreement is now needed on a national list, said Dr Keeling.

The use of standardised, ready-to-use products would allow staff to move from one hospital to another without having to learn new concentrations and new administration techniques. It would also reduce the risks of calculation errors and the amount of time spent preparing injections and result in overall risk reduction. In addition, it should be possible for licensed products to be developed and it should be easier to ensure availability of products. Dr Keeling concluded by suggesting that a business case for standardised injections could be made on the basis of reduced litigation costs to NHS trusts.

Safe use of insulin
Errors with insulin have caused 18 incidents of death or serious harm and 972 moderate harm incidents between 2003 and 2009. Anna Bischler (senior pharmacist, medication and medical devices team, NPSA) explained that there were two main problems—the inappropriate use of intravenous syringes, which are marked in mls and not in insulin units, and the use of abbreviations such as ‘U’ or ‘IU’ for units, resulting in 10-fold overdoses.  Some health care practitioners may not have had sufficient training in the use of insulin, she said.  Another problem is that diabetics can be well controlled at home before a hospital stay but are made ill in hospital. A survey in 2009 showed that one in four insulin-dependent diabetics experience a hypoglycaemic episode whilst in hospital and 34.4% are subject to at least one prescription error.

In order to improve the situation the NPSA has established an insulin safety forum to enable sharing of information and resources. An e-learning programme is also available at

Gentamicin in neonates
There were 507 reports of problems with gentamicin in neonates during 2008–9 and this prompted the issue of a safety alert by the NPSA February 2010. David Upton (chief pharmacist, Sheffield Children’s NHS Foundation Trust) described how his trust had responded.

The alert called for a local protocol for gentamicin use and the implementation of a care bundle. Dr Upton’s hospital already had a rigorous protocol in place and a care bundle that required use of the 24-hour clock for prescribing, minimisation of interruptions during preparation and administration by the wearing of a coloured apron, double checking of preparation and administration and administration within one hour of the prescribed time. However, there is wide inter-individual variation and it was still not unusual to see blood levels of gentamicin of 2mg/L at 20 hours after the dose, he said.

One step that the trust took was to avoid the use of gentamicin in neonates if possible. Thus cefotaxime is now first choice for meningitis and cefuroxime/metronidazole for abdominal surgery or sepsis. The second step involved the introduction of prefilled gentamicin syringes containing 4mg in 1ml, wearing of tabards to show that the nurse preparing and administering the dose should not be interrupted and the redesign of medicine charts with space to record blood level results.

Dr Upton emphasised that the preparation of intravenous doses should always be a non-interruptible process and should ideally be adopted for all high-risk products. Furthermore, he suggested that a national medicines chart (as already exists in Wales) should be introduced.

Dosing of anticancer drugs  
There are at least six different formulae for the calculation of body surface area (BSA) and they give surface areas ranging from 1.71–1.93 square metres for the same weight and height, explained Bruce Burnett (Consultant Pharmacist, Cancer Services, Glan Clwyd Hospital, Betsi Cadwaladr University Health Board). BSA predicts pharmacokinetics and pharmacodynamics for docetaxel but not for cisplatin, epirubicin, topotecan or carboplatin. For all other cytotoxics drugs the relationship is unknown.

For certain drugs genotyping can be useful for identifying likely responders, for example, identifying HER2 status for trastuzamab or KRAS for cetuximab. Measuring enzyme activity is important for some dosage adjustments, for example, CYP3A4 activity and docetaxel, thiopurine methyl transferase (TPMT) and six-mercaptopurine, dihydropyrimidine dehydrogenase and five-fluorouracil.

Dose banding has attracted interest because it has become impractical to provide individualised doses in a timely manner. Initially it involved providing a range of doses that were within 5% of the prescribed dose although in Wales the possibility of 10% variation is now under consideration, he noted.

Fixed dosing has also been advocated because for many drugs it is associated with comparable variations in response and toxicity to BSA-based dosing. It also offers the advantages of easy preparation, and reduced wastage. It is suitable for monoclonal antibodies such as rituximab that behave very differently from conventional drugs, he said. Moreover, simulations show that fixed dosing leads to relatively little under- or overdosing.

Metronomic dosing involves giving a low dose chronically on the basis that cancer is a chronic disease and the mechanism of action might be different from that with acute dosing, explained Mr Burnett. The benefits include reduced or different toxicity profiles and maintenance of tumour diminution over time. It also offers the potential for therapeutic drug monitoring. However, it needs new end points for monitoring and its place in therapy is not yet fully defined. Examples of drugs studied so far include vinorelbine, cyclophosphamide and capecitabine. It is not certain whether metronomic dosing should be considered to be a example of fixed dosing or something different, he added.

Mr Burnett concluded that dose banding and fixed dosing could help to improve patient safety and allow more efficient use of pharmacy aseptic facilities to prepare high-risk products. n

Implementing change, improving patient safety was a one-day patient safety seminar run by Aesculap Academia, a division of B Braun Medical Ltd. It was held in Birmingham on 13 October 2010. 

Latest Issue

Be in the know
Subscribe to Hospital Pharmacy Europe newsletter and magazine