Pump Up the Volume

This blog is a little different. My hospital has asked me to produce a themed blog related to a series of incidents we have had across a number of different acute specialties. Some of the links point to local 'in house material' so I apologise in advance if you can't see them. Your hospital will have something similar or perhaps better.....if you do let me know.
However it doesn't really matter if you are an EM type or a PICU nurse the messages are still the same.

Pump Action

Modern hospital medicine requires the delivery of a number of, well lets face it potentially dangerous chemicals, directly into a patient's veins or tissues many thousands times of day. To help us do this more accurately we rely very heavily on variety of  syringe driver devices and pumps. This of course has been a great step forward in patient safety. 
To get to the position we enjoy now, highly technical pieces of kit and smart pumps, has not been an easy journey. Hands up who knew that the first described IV infusion into an animal was performed by Sir Christopher Wren in 1656 who developed

 ‘a way to convey liquid poison into the mass of blood’. 

He fashioned a quill and a pig’s bladder to instill wine, ale, opium and liver of antimony into a dog’s veins. An intoxicated dog was the 'successful' endpoint.

For those of you keen to know more about some of the dubious practices that have lead us to today's relative IV nirvana try this IVI History or driver inventor 

I promise you at least one Pope, quite a lot of death and several unfortunate animals. 

Sadly where humans go, errors follow, and even more sadly patients have come to harm as a result.

The National Picture

In the five years between 2005 and 2010 the Medicines and Healthcare products Regulatory Agency (MHRA) investigated 1,085 incidents involving infusion pumps alone in the UK. In 68% of the reports no cause was established (either due to reporting error or the device was working as intended), 11% of incidents were due to device error but 21% were attributed to user error.
MHRA and pumps

Locally

Below you will find real cases where patients have been damaged, or potentially damaged by pump or infusion errors. As always theses cases are not here to vilify the unfortunate health care professionals involved but to show you what can happen with real patients in real hospitals and what real harm can result.

If you know about what can go wrong, you are in a far better place to avoid making the same mistake. Learn from others mistakes, don't wait for your own.

Give Me Five

A patient with pre-existing vascular disease came into the ED with status epilepticus (continuous seizures). Standard therapy failed to stop the fitting activity. The correct call was made to rapidly paralyse and anaesthetise the patient (RSI) and place them on a breathing machine (ventilator). The on-call anesthetist was unfamiliar with the ED Resus room so correctly asked for support. An ODP came down from theatres bringing a syringe driver infusion pump with him, for which he drew up 50 ml of 2% propofol from the ED stock.

The patient was successfully anaesthetised and the syringe pump was connected to the patient with the propofol in place.  A verbal order was given by the anaesthetist to 'start the rate at 5mL/hr' and the rest of the patient's care continued.

Approximately 15 minutes later the patient became hypotensive and unstable. They required adrenaline boluses and further intravenous fluids to maintain their blood pressure. An ultrasound of their heart showed global hypokinesia (poor pumping). At this point the syringe pump alarmed and it was noted to have delivered the entire 50 ml of propofol in the preceding 30 minutes. The patient's blood pressure dropped to 50/? before a cardiac arrest occurred.  Life support protocols were followed and a pulse was restored after one cycle of CPR and further adrenaline.

In the following hours the patient was stabilised on ICU but did develop heart failure for which they were successfully treated. They subsequently were found to have suffered a Non STEMI (heart attack) almost certainly caused by the period of severe low blood pressure. Sadly they also developed an ischaemic limb which required amputation. This too was felt to have been contributed to by the  low BP and decreased heart pump function.

Unsurprisingly a Serious Incident report was generated.

So what happened?

The patient had lots of good care. The ambulance service followed their protocols correctly, the ED initial work up and management were correct. The plan to 'tube and vent' was correct. The initial RSI was safe and secure.  The maintenance of anaesthesia was the issue.  Additionally as soon as the error was discovered a serious incident was declared and the pump removed and stored for later investigation.

So what went wrong?

The theatre pump that had been connected to the patient was set to run at 5 mg/hr not 5 mL/ hr. 

The pump used in this case has a number of 'smart functions' to allow weight based calculations for drug delivery.  This requires a number of data points to be entered, but when the pump is turned on it defaults to its previous settings. In this case 5mg/hr of  50 micrograms/ml drug concentration, which gives  a rate of 100mL/hr.

Other factors?

The ED staff involved were not familiar with the 'smart pump' device only using devices with a default mL/hr setting.

The documentation around the anaesthetic and maintenance infusion was minimal at best, with no prescription being written or checked. (Clearly in urgent situations this can be unavoidable, but should be done retrospectively)

What did we learn?

The final responsibility for ensuring the correct drug is correctly prescribed, drawn up, connected and delivered remains the responsibility of the doctor in charge of that specific procedure/treatment.

A second independent check of all of these elements is mandatory.

However single handed clinicians in operating theatre environments this may not always be practical but is considered good practice

The injectable medicine code is being reviewed to reflect the emergency work carried out by anesthetists and others outside a theatre setting

If you haven’t been trained to use a specific pump, don’t touch one.

Switch Hitting

A patient was admitted to critical care following a very long and complex operation to repair an acute aortic dissection. The patient was on IV adrenaline, IV noradrenaline, IV dopamine, to support the patient’s blood pressure and also on IV aprotinin to help reduce post-operative bleeding. 

On the instruction of the Consultant Surgeon the patient’s blood pressure was to be no higher than 120 systolic. A couple of hours into their admission the patient had a surge in blood pressure to 200 systolic. All blood pressure raising drugs were stopped and a bolus of IV labetolol 5mgs was given to try to reduce this surge. The blood pressure reduced to 105 systolic. Further labetolol was given before an infusion of labetolol was commenced.

An hour or so later the patient’s blood pressure dropped to 45 systolic and they were noted have bleed 200mls into a surgical drain. Full CPR was started before the patient ‘s return to theatre for emergency re-sternotomy & exploration.

Subsequently it was discovered that the syringe containing adrenaline had been placed in the pump for the aprotinin.  The adrenaline was prescribed at 5mls an hour, the aprotinin at 50mls an hour. As a result they received 4.5mg of adrenaline not 0.45mg, resulting in the blood pressure spike and subsequent complications.

Good Care & What Went Wrong?

The drugs were drawn up, labelled and checked correctly.  Unfortunately the syringes were not checked into the correct driver or the rates confirmed by two trained staff.

When the patient became unwell the response of the clinical teams was appropriate. The incident was reported immediately and shared with the family. The pumps involved were all removed from clinical use and sent to Medical physics for scrutiny. Additionally an immediate critical care safety alert to prevent recurrence was sent out.

Learning Bite

The lesson to learn here is best summed up by the Critical Care alert;

‘two registered nurses must be involved in all aspects of the  infusion preparation and administration process up to and including insertion of the syringe into the syringe driver’

Aortic Dissection on Up-to-Date

The Trouble with Double

All medics and nurses will be familiar with the concept of double checking prescriptions and infusions prior to delivering them to the patient. It's a fail safe belt and braces right?

Well actually no.  There is evidence that it can sometimes contribute to drug errors IF it is not carried out properly and rigourously. Double checking medicines: defence against error or contributory factor?

 Here are four possible explanations;

-deference to authority

            ‘they know what they’re doing, I can relax’

-reduction of responsibility

            ‘I don’t need to read it properly, they are’

- automatic processing

            Eyes are open but no one is home

-lack of time

            Rushing

A safer way of avoiding error is 'independent checking' This consists of two individuals checking each element of the infusion independently and in isolation from each other. That is the two staff do not stand side by side checking each other checking the infusion ..... They check the infusion.

Further guidance to this can be found locally in the medicines code , here in the injectable medicine code

Theres also a good overview on independent checking from our American Cousins 

Unfortunately no system is fail safe if those carrying it do not do so with appropriate care

Frequent Little Things

There are several common themes in our reported pump errors….

Are You Concentrating?

A sick baby required blood pressure support with Dopamine at 10micrograms/kg/min. A syringe driver was drawn up containing 60mg of the drug in a 1,200 microgram/mL solution. However when prompted during set up, the concentration entered into the driver was 2,400microgram/mL. As a result the baby received half the dose that the team believed they were getting.

Another small child was on a midazolam infusion. At shift change safety check it was noted that the child had been receiving 66mg in 50ml rather than the prescribed 33mg in 50mL and as a result was receiving too much drug.

Neither child came to significant harm

Learning Bite

In both cases the infusions were drawn up and delivered with a signed off double check at each stage. This does not make those individuals ‘bad’. There was no deliberate error here.

The mistake was in believing that the act of double checking works well in itself. It doesn’t.

It only works if both parties concentrate and feel capable of questioning the other.

‘It’s Ok to ask’

A Matter of Rate x3

1- A stroke patient undergoing thrombolysis was transferred to the angio suite for clot retrieval. During

transfer the pump ran out of battery and turned itself off. In the angio suite the pump was restarted on mains power by a different member of staff who asked the Brain Attack Team what rate to run it at.

Once restarted the pump was then out of sight of the BAT team behind sterile drapes. Half way through the procedure the BAT nurse managed to look at the pump from a distance and noticed that the syringe instead of being nearly empty as expected was still nearly full.
The pump had been restarted at 0.53mls/hr not 53mls/hr.

2- A patient with abdominal pain required an infusion of C1-esterase inhibitor. Prior to transfer to a ward it was noted that the syringe seemed fuller than expected.  The pump had been set at 6mls/hr not 60mls/hr

3- A patient with heart failure on the ward required a furosemide infusion over 24 hours. After a few  hours the pump alarm sounded. On inspection the infusion was complete having been set to run at 41.7mLs/hr not 10.4mLs/hr

Again no serious harm resulted from any of these cases but this was more by luck than judgment and clearly we could have done things better.

Learning Bite

All rate changes to infusions should be checked, documented and signed.

If you haven’t been trained to use a specific pump, don’t even touch it.

If you are unfamiliar with the drug being given in pump, don’t change anything until you have found someone who is.

The Silence of the Alarms

A patient was on a post-operative Insulin-glucose sliding scale. During the ward round the insulin was noted to be running at the minimum ‘keep vein open’ rate of 0.5mL/hr, but the dextrose was left running at 80mLs/hr. The alarm on the insulin pump had been silenced.

Unsurprisingly the patient’s blood sugar was found to be higher than desirable. They required additional insulin and closer monitoring than planned but fortunately came to no serious harm

Learning Bite

Alarms are there for a reason.

When they go off, check why and act on the result.

Avoid ‘alarm fatigue’ by good cannula siting and infusion management

Summary

So pumps really are great. They make a massive contribution to delivering safe effective care to our patients every hour of every day.

However like all machines the still require humans to tell them what to do.

If the human bit goes wrong, so does the infusion...... and catastrophic harm can result.

Reduce your risk of harm by independent checking of all stages of pump/infusion prescription, solution mixing, rate setting and connection to the patient.

Learn from the mistakes of others, don't wait until next time, it could be you.... on either end of the infusion.

Here's a handful of some of the great people I get to work with each day

Think Pump. Think safety

Gratuitous Nonsense Link- For 1980's retronauts everywhere

The knee bone's connected to the thigh bone, the thigh bone's connected to the hip bone......or at least you hope it is

'Exactly what kind of a force Professor Röntgen has discovered he does not know. As will be seen below, he declines to call it a new kind of light, or a new form of electricity. He has given it the name of the X rays. Others speak of it as the Röntgen rays' 
'The New Marvel in Photography'

Almost all ED docs have missed a fracture.

Nope, all ED docs have missed a fracture. Doesn't mean they should all be clasped in irons and thrown in the hold. It does mean though that X-rays are always an area of risk, error.... and learning.

As it was, not that long ago

The author has fond memories of hard films being stuck to windows when all departmental viewing boxes have failed or feeling at risk of epilepsy as fluorescent tubes mercilessly flicker whilst you look for subtle signs of osteomyelitis (aren't all signs of osteomyelitis subtle) Osteomyeltis and Imaging.

 So you would have thought that digital imaging was a great step forward.

 Well of course it is, but errors still occur, sometimes because of the nature of digital media......and sometimes because you can never under estimate the ability for a human to muppet something up.

Here are three X-ray related cases. All real films. all real patients. all with real harm..... But also all with learning

 All of the Films, All of the Time 

 A lady in her 50's, 'AB' presented with a several month history of non traumatic right hip and thigh pain, with extensive swelling and bruising of the thigh.
She was having great difficulty walking. The ED doctor sent AB for X-rays and then handed her over to another ED junior, leaving a clear plan in the notes. ' If X-rays abnormal- admit. If normal- home' 
The second doctor reviewed the films......

............declared them normal and sent her home with GP follow up.

In the following days she had a torrid time. She became entirely immobile, bed bound and unable to get to the toilet in time. Her GP, reassured by the recent ED visit prescribed her morphine and arranged urgent Musculoskeletal out patient review.
Hey, the X-rays are normal right? What could go wrong?

After ten days an Intermediate Care Team visit lead to an ambulance being called. On arrival at the ED for the second time her right leg was now externally rotated and short. An ED Reg saw her and sent her for repeat films which unsurprisingly showed a proximal femur fracture.

When the Reg reviewed AB's previous films taken 10 days before, he saw that the pelvic film had been red dotted by the radiographers and clearly showed a hip abnormality. This turned out to be a pathological fracture related to a disseminated malignancy. Bony Mets on X-rays

erosion of the cortex of the lesser torchanter

Sadly AB died 4 months later.

Whilst the eventual outcome for AB would not have been changed by the diagnosis being made ten days earlier the letter of complaint makes it clear how terrible those extra days at home were

 the pain was intolerable…she could do nothing but sit for hours on her landing until her neighbour could attend and carry her back to the bedroom. Her bed had to be changed due to incontinence, her leg had rotated and this was very frightening

 The quote is included not to sensationalise, but to remind us of how 'simple' decisions we make day to day can have hugely significant effects on lives

Learning Bite
The doctor who reviewed AB’s initial films missed the abnormality even though it was red dotted because they ONLY looked at the shaft of femur and knee films, which are indeed normal. They did not look at the pelvis films.

 PACS is a very useful tool BUT it can make it easy to miss all the films, or look at the wrong date or even the wrong patient’s images.

 So Triple Check- Is it the right patient? Is it the right date? Have I seen all the films?…..and finally if the films don’t seem to make sense for the patient in front of you. Double check the films and double check the patient.

Too obvious? 

LM a 20 something year old man, attended four days after an accident sustained on holiday. He had dived into a swimming pool and struck his head on the bottom. He had immediate neck and arm pain, and felt his arms were weak. He managed to climb out the pool after a few minutes.

A local doctor advised him that he had a muscular sprain. On return home he initially attended a MIU, who immobilised him and sent him to the local ED. On arrival he was log rolled off the board by a senior doctor who noted normal neurology and sent him for an x-ray of his neck.

On his return he was fully assessed by a different doctor who clearly documents the history and some possible right upper limb weakness. The doctor reviewed his x-rays including a swimmer’s view before making a diagnosis of neck sprain and discharging him without further senior review.

Six days later a radiologist contacted the ED to advise that LM had a fracture of C6 and the swimmer’s view did not show the C7-T1 junction.

 Unfortunately the ED had no patient contact phone number so attempts to recall him immediately proved futile. A letter was sent out to LM’s home address, but he did not receive that for a further four days.

Ten days after his first ED visit LM re-attended and was sent for CT, then MRI. This showed further minor fractures of C7 & T1 with evidence of ligamentous instability. He was admitted and underwent surgical repair the following day.



Learning Bite

LM had ‘the’ classic history for a forced flexion neck fracture. This should always make you raise your index of suspicion.  i.e ‘It’s such a good story that I am expecting a fracture but can’t see one- I had better talk to a senior.’

How do you know your contact details are correct in the event f recall?

How do you ensure radiology can feedback in real time?

Too likely? 

A man in his 70's fell down some stairs and hurt his back. He had suffered back pain for many years so was used to being sore but he thought this was worse than normal.

An ambulance was called and he ended up being seen in an ED. He was deemed to have no significant injury and discharged back home without imaging but with painkillers.

 Two weeks later he was still struggling. His analgesia requirement had increased and he was not able to mobilise as well as before the fall.

Another ambulance was called and he ended up in a different ED. They decided that X-rays of his lumbar spine were indicated. Theses were reviewed and declared as showing no fracture. So...... he was sent home with a further change in his analgesia.

Ten days later the X-Ray department at hospital two contacted the ED and said that they thought there might be a fracture on the patients lumbar films but weren't sure. The ED phoned the patient who was still struggling away and sent out an ambulance.

On review he was sore throughout his spine but most notably in his lumbar region where he said he could feel some grinding, clicking and movement..... Movement ? So?...

Well the 'killer' fact missing from this review is that thus chap had a well established history of Ankylosing Spondylitis and had been fixed in the classical 'question mark' position for the past 15 years at least. Ankylosing Spondylitis

 His lumbar films do show a fracture.....

What fracture?

But the patients altered anatomy is so unusual that you would be forgiven for thinking 'Meh! That's a soft tissue shadow'......and that's kind of the point

there it is - right through L4 and all the posterior calcification

Question Mark Spine

The whole spine CT showed just how abnormal his whole spine is, and also clearly showed the fracture

 Learning Bite 

'Ank Spond' patients fall over.
 When they do they usually break something..... Assume, until you can be SURE that they don't, that they have a spinal fracture

Plain films in this group are abnormal before they have a fracture.
Their plains films are notoriously hard to get right. Consider CT as the investigation of choice

Finally, if an AS patient does have a fracture, then the long sections of fused vertebrae either side of the fracture mean that there is lot of leverage at the fracture site and neurological deficit is common.

Final Comment

When we send people home they trust what we say (amazing, but true)

This can mean that folk stay at home in the face of quite dire symptoms

It's really important to give them advice about when, where and how to re-access care.... and then document these 'triggers to return'

Irrelevant Nonsense

One for your kids

'Simple' Drug Errors- Tales of False Reassurance

'drink me'

Most senior health care professionals have a few drugs , their indications, interactions and doses so well embedded In their brains that they do not need to look them up.... But most of them have to turn to the trusted British National Formulary every now and again. More junior staff probably use the BNF more than senior docs, but as the author can attest 'tempus fugit' and having access to a BNF in hard or e-copy is an essential pre-requisite of safe EM.

But can using a BNF actually cause a drug error?...... course it can't. 
Can it?

Sadly the ability for the typical human to find the way to err (in medical parlance, 'cock something up') should never be under estimated, even when trying to do the right thing.





1-Route Cause?

A lady in her 40s presented with chest pain. A junior doctor saw her, made a full assessment and thought that her pain was most likely to be  gastro-oesphogeal in origin. He wanted to provide her with analgesia, but felt that her pain was not severe enough to warrant intra-venous morphine. He did not wish to provide her with a non-steroidal based drug, as he felt this might worsen her possible dyspepsia. 
She had already taken paracetamol at home within six hours. 
So he did the right thing, he asked advice.
'Do we have something not as strong as morphine but better than paracetamol?'
A senior nurse suggested codeine.
The junior doc was not familiar with codeine doses.
So he did the right thing, he went and looked it up in the BNF.
The patient had vomited once so the doctor thought he should use a parenteral route

Unfortunately he misread the BNF entry and did not note that codeine can only be given orally or into a muscle.
He prescribed the right dose.............but into a vein not a muscle.

Once prescribed he did the right thing.
He asked a nurse to give the patient the drug.
The junior nurse had heard the conversation between the doctor and the senior nurse and had also seen the junior doctor reading the BNF. She was also unfamiliar with codeine, but she had just seen the doc check the drug in the BNF, it was busy (isn't it always?)
The drug was delivered IV

The patient had an immediate 'rush' and felt 'awful', but this settled within a few minutes. This drug error led to 'no further harm', but that was serendipity not design.

Comments
Despite lots of really good, thoughtful care, a drug error still occurred. Completely well intentioned actions with the best motives resulted in harm. Good God man, they even read the BNF first.....

Someone must be to blame!

When the senior doctors were informed of this error, almost all of them immediately said
'Why do we have codeine solution in the ED I never use it?'
'IM analgesia, in an ED. Why?' 
(yes there are exceptions)

An immediate action was to remove all codeine solution vials form the ED within 24 hours.


2-Not all syringes are equal

A renal patient in his 30's came into the ED with symptoms of weakness, lethargy, low blood pressure and slow heart rate. 
He said, 'I've had exactly this before when my potassium levels have risen too high' 
The junior doctor who saw him established that hyperkalaemia was the indeed the correct diagnosis using a venous blood gas. The doctor then did the right thing, they immediately discussed the patient with a senior saying that they were moving the patient to the resuscitation room for standard hyperkalaemia treatment under cardiac monitoring.

Within 10 minutes the correct diagnosis & treatment plan were established with onward referral to the ward already made. All going well so far!

The junior doctor prescribed an insulin/dextrose infusion in line with the trust guidelines which they looked up on the hospital pharmacy website (they did the right thing). The doctor directed that 10 units of fast acting insulin should be used in the infusion, and did the right thing, by prescribing it all on an IV fluid chart with correct doses volumes and rates. They even wrote out 'ten international units of soluble insulin' in long hand.(see MPS link at end)

Resus room nursing staff drew up the correct volume of 100 units per ml soluble insulin, but did so in a standard 1ml syringe, not an insulin syringe. That volume is, 0.1ml. The doctor caring for the patient thought this volume looked wrong, so did the right thing.......... and checked.

Having searched and read the BNF the doctor recalculated the dose and decided that 1ml was the correct volume to give to achieve 10 units. 

Resus was busy (isn't it always?), so the nursing staff had been called away to a sick patient in another bay. The doctor was concerned about their patient so just wanted to 'crack on' with treating him...so despite never having made up an insulin infusion before drew up 1ml of actrapid (100 units) and used that to make the infusion and then filled in the additive label writing that it contained 'ten units of insulin'

The infusion was checked and signed by two nurses, one from the resus room

and one from another area of the ED (because, as you may remember, resus was busy), before being connected and given to the patient. The resus nurse knew that 1ml of insulin had been used not 0.1ml, but had seen the doctor check the BNF .............so assumed that her first calculation was wrong and that he was receiving ten units

The patient (unsurprisingly) almost immediately began to exhibit symptoms of hypoglycaemia and the error was equally rapidly noticed and corrected. An immediate explanation  and apology was  given to the patient and he was told that a formal review of the incident would be undertaken.

Fortunately the patient came to no harm except more frequent BM measurement, which was uncomfortable for the patient, but did him no long term harm, but again this was serendipitous.


Learning Bites

Insulin syringes are specifically designed to avoid dose errors by being graduated in units not millilitres.  Only use insulin syringes when drawing up insulin.(MPS on Insulin)  & Yorkshire Water Torture

All drug calculations must be checked by two trained health care professionals, as should the doses and volumes put into infusions prior to delivery to the patient

If you haven’t done something before, Resus is unlikely to be the place to start doing it

If you’re not sure about something ask.

If you think someone else has got something wrong it’s OK to ask

A culture of 'its ok to ask' is not a sign of weakness nor 'thickness' but of good safety processes


Commentary
In a system reliant on humans it is impossible to prevent all errors. The best you can do is try to maximise the barriers to error. Which is why I really object to 'aspiring to zero harm'. If you aspire  to the impossible, then you will always fall short and in the process mislead the patients and staff. 

From the patients perspective 'if harm is entirely avoidable then I should not come to harm. If I have, someone is to blame'

From the HCPs perspective 'if i make a mistake, I will be blamed. I could be disciplined. I could lose my job. I could lose my house. I'm not telling any about this and hoping they don't find out'

So who is to blame?

Well actually no individual is solely to blame. As with most healthcare incidents there was an alignment of system weaknesses that let both patient and the staff involved down. This is frequently described as the Swiss Cheese model.

It is rare that one person does something so enitrely inexpilcable that blame can be laid at their door alone. And anyway blame is such an emotive word. Once an incident has occured you cant undo it, but what you can do is reduce the chances of it re-occuring. 

We, the NHS, and EM more widley need to share learning and promote safety by embracing openness and honesty and avoiding blame.

Lets learn not blame

Lets share not hide


Random Link

bonus points available to anyone who knows what this chaps link to the above is

Leonard Thompson