Death caused by undetected e-prescribing error

hospital staff wheel away empty gurney

A man’s death from opioid toxicity occurred due to “persistent failure of critical thinking” by those involved in his care, according to findings handed down by the NSW State Coroner

Following on from a hearing held in early February, Acting State Coroner Magistrate Teresa O’Sullivan has handed down her findings of the inquest into the death of Paul Lau.

On 18 June 2015, Paul Lau, 54, underwent day surgery for knee reconstruction at Macquarie University Hospital.

The surgery was uneventful and he was expected to be discharged home the following day.

However, at approximately 1am on 19 June 2015, despite attempts to resuscitate him, Mr Lau was pronounced dead.

During the five-day inquest, Magistrate O’Sullivan heard from eight witnesses on the events leading up to Mr Lau’s death.

An electronic prescribing error

A new electronic medical record system had been rolled out across the hospital for all new admissions just one month earlier on 2 May 2015.

The system provided the ability for ‘one touch’ prescribing; contained a complete medical record of a patient; recorded all medications administered to a patient within the hospital; and allowed pharmacists to review a patient’s medication history online from the pharmacy.

It was found that on the day of the surgery, the anaesthetist in charge had used ‘one touch’ and longhand prescribing to chart various medicines pre- and post-operatively for Mr Lau, including sedatives, anaesthesia, antibiotics, anti-inflammatories, paracetamol, oxycodone and celecoxib – all of which were routine and of appropriate doses.

At 1.52pm on 18 June, the anaesthetist was in theatre with another patient (Patient GS) when he remembered to chart a small amount of fluids for Mr Lau using the electronic system.

A few minutes later, the anaesthetist began prescribing further medication to Mr Lau’s electronic record, which were in fact intended for Patient GS.

He did not realise he still had Mr Lau’s record open and failed to close his file to open Patient GS’s file.

It was during this time that the anaesthetist ordered intravenous fentanyl (infusion of 20 mcg/1mL, 60mL) in Mr Lau’s chart, intending to chart the medication for Patient GS.

At 2pm, the anaesthetist also electronically ordered a fentanyl 100mcg/hour transdermal fentanyl patch, one patch every three days (total five patches) in Mr Lau’s chart, also intending this for Patient GS.

While inadvertently prescribing these medications on Mr Lau’s chart, a total of 22 different alerts were triggered by the electronic system – however the anaesthetist overrode all of them, populating the field by selecting “consultant’s decision” from the drop-down menu of reasons.

Pharmacist’s failure to act

The dispensing pharmacist working at the hospital saw the order for a fentanyl patch on Mr Lau’s electronic record, and allocated it for dispensing shortly after 2pm.

While the pharmacist was aware that the 100mcg fentanyl patch was the strongest available, and there was nothing in Mr Lau’s record to suggest he had ever taken opioids before, she continued to dispense the medication.

The dispensing pharmacist said in her statement that while it was normal practice to do clinical reviews post-operatively and discuss any issues with the doctor, she was “reluctant” to contact doctors while they were in theatre.

Acting State Coroner O’Sullivan said the pharmacist ought to have reviewed the medication to ensure it was suitable for the patient.

Separately, in August 2017, the Pharmacy Council of NSW determined the dispensing pharmacist had failed to adequately assess the appropriateness of the dose. She was found to have engaged in “unsatisfactory professional conduct” and was cautioned.

She has expressed remorse for Mr Lau’s death.

Further lapses of judgement

While Mr Lau was in recovery, it was noted by a nurse expert witness that the several nurses in charge of the patient had failed to confirm the charted medications with the anaesthetist during handover, or realise that Mr Lau was an “unlikely candidate” for fentanyl patient-controlled analgesia.

In addition, there were two occurrences where the anaesthetist could have picked up on his prescribing error:

1) He received two phone calls: the first was about a discovery that Patient GS’s file did not contain any medication orders; and the second stated that Mr Lau’s chart contained a double-up order of paracetamol and cefazolin. The anaesthetist said he assumed he had closed Patient GS’s electronic file incorrectly, and that the system had thus not saved the medication orders.

2) When the anaesthetist returned to the hospital at approximately 7.30pm to check on Mr Lau, he noticed the patient had a fentanyl patch and was also using a patient-controlled analgesia machine. However he incorrectly assumed Mr Lau had been on a fentanyl patch pre-operatively for his chronic knee pain – despite there being no such record. The anaesthetist also assumed someone else had prescribed the additional fentanyl analgesia.

The anaesthetist agreed that if he had checked the dosage of fentanyl – which was easily readable on the patch – he would have seen the extremely powerful dosage of 100mcg/hour being worn by Mr Lau, an opioid naïve patient.

He also would have known there was no possibility further fentanyl would have been prescribed by another medical practitioner due to it being a highly inappropriate medication for an opioid naive patient.

During a counselling interview with the Medical Council of NSW in September 2016, Dr Kim admitted the prescribing error and admitted that he failed to recognise his prescribing error when he reviewed the patient on the ward post-operatively.

The Medical Council determined that this was a major clinical management error, however no further action was taken.

At the the time of the Medical Council of NSW proceedings, the precise nature of how the prescribing error occurred was unknown.

Patient’s deterioration

Just after midnight on 19 June 2015, Mr Lau was showing signs of oversedation, including shallow breathing; however a junior nurse failed to realise the urgency of these signs.

Other more senior nurses also failed to recognise that Mr Lau’s increasing need for oxygen was “a clear sign of deterioration” and “indicated imminent danger”, stated Acting State Coroner O’Sullivan.

Mr Lau’s condition deteriorated quickly and when the nurses checked on him again, his fingers were cold and nurses could not feel a pulse.

While an emergency team attempted to resuscitate Mr Lau, they were unable to do so and at 12.56am, he was pronounced dead.

Following autopsy, forensic pathologist Dr Theresa Ly concluded the Mr Lau most likely died from fentanyl toxicity, which would have caused severe respiratory depression, coma and hypotension producing a tendency to vomit and inhale gastric contents.

Dr Ly found fentanyl in his blood at a level of 8ug/L (potentially fatal level range 3-28ug/L).

Changes across the hospital

Acting State Coroner O’Sullivan found that while the electronic system did not cause Mr Lau’s death, “the initial prescription error was made easier due to a function of … great utility – the ability to open and close different patient records from a single terminal.

“Prior to the introduction of electronic medical records, it was much more difficult to charge medication on the wrong patient file.”

Nurses and pharmacy staff have since been disciplined, while a number of procedures at the hospital have reportedly been changed in the light of Mr Lau’s death.

This includes the addition of further alerts triggered in the electronic system when prescribing or administering opioid transdermal patches.

Significant training among staff members – particularly anaesthetists – has been implemented, while the hospital’s pharmacy has taken steps to improve training and encourage pharmacists to ask questions if they have concerns.

Despite the changes that have already been made, Acting State Coroner Magistrate O’Sullivan recommended that a working party be established by the Macquarie University Hospital to consider lessons learned as a result of Mr Lau’s death.

She offered her sincere condolences to Mr Lau’s family, and thanked them for enduring the inquiry “dissecting in great and laborious detail the death of this very real and very loved man”.

Previous Mandatory data breach notification laws
Next Clinical tips: premature babies

NOTICE: It can sometimes take awhile for comment submissions to go through, please be patient.


  1. Cogrady

    How many of us would have queried that Patch dose?Could we do a poll to show the norm please?

    • Jarrod McMaugh

      Some of the most common CIs we record is opioid-naive patients prescribed fentanyl, or patients in acute pain prescribed a patch delivery method.

      Both should have raised red flags.

      • Cogrady

        Where I work pharmacists don’t do medication histories before surgery.How would the pharmacist have realised the patient was new to the fentanyl patch when working downstairs dispensing?

        • Jarrod McMaugh

          I think you’ve answered your own question.

          Your director of pharmacy needs to have that policy changed.

          Why would any pharmacist supply medication without the ability to check prior history?

          • Cogrady

            So in the population of pharmacists ,only Jarrod and I think that we need more hospital pharmacists ,so that all wards are covered before we fill out MMP,emeds and dispense?

          • John Wilks

            Make that 3!!! And I know many many others who may not have read this tragic report who would also agree on the pressing need for more hosp pharmacists, notably but not limited to pre-admin med r/v

  2. Notachemist

    Reading this article I am reminded how close we all come to making this type of error. However a Fentanyl 100mcg patch is a very big dose and when supplying this (as with other high risk medications – think of the APINCH acronym) it is essential to check the medication history. If we don’t practice in this way a robot or a person with no knowledge of pharmacy might as well be doing the dispensing. In the same way that the prescriber had to override alerts from the prescribing program I would expect that pharmacists should have internal alerts even if they are not in the dispensing software.

    • Cogrady

      I would still like a poll please for every pharmacist. All our patients move around

Leave a reply