Hospital drug alerts system would catch ‘nurse killer’
A specialised drug prescribing system could help prevent a repetition of patient murders, such as those by nurse Colin Norris, researchers claim.
Colin Norris, a nurse at the Leeds General Infirmary and St James Hospitals, was convicted in 2008 of murdering four patients using insulin to induce hypoglycaemia.
One question raised by the case was how likely it was for so many patients to experience hypoglycaemia.
The issue was again highlighted following the deaths of at least six patients at Stepping Hill between June and July 2011.
Researchers at Queen Elizabeth Hospital Birmingham and the University of Birmingham analysed information on 37,898 inpatients to establish how commonly hypoglycaemia occurred among non-diabetic patients on general wards.
To do this, they used the hospital’s Prescribing, Information and Communications System (PICS), a purpose-designed computer-based patient information system.
It identified 71 hypoglycaemic episodes at a cut-off of 3.3 mmol/l. Estimated frequency at 3.3 mmol/l was 50 per 10,000 admissions, the study authors found. Meanwhile, at 3.0 mmol/l it was 36 per 10,000 admissions, at 2.7 mmol/l it was 13, at 2.5 mmol/l it was 11 and at 2.2 mmol/l it was 8 per 10,000 admissions.
“Admissions of patients aged above 65 years were approximately 50% more likely to have an episode of hypoglycemia. Most were associated with important co-morbidities,” said the authors.
They said the findings confirmed that significant hypoglycaemia was rare in non-diabetic patients, but also identified the potential for the hospital’s PICS to spot unexplained clusters of the condition.
This could enable clinicians to spot the sort of misuse of insulin, which led to the deaths of patients under the care of Colin Norris in 2002 and other convicted killers, such as Beverly Allitt and Benjamin Green.
Lead author Krishnarajah Nirantharakumar, a clinical research fellow at Birmingham University, said: “We looked back at the rate of hypoglycaemia in non-diabetic patients outside critical care for 2010, and we found that it was very rare: only 13 in every 10,000 admissions below a blood glucose value of 2.7mmol/l.
“We then analysed the case notes of the patients whose blood glucose concentration was below this value and found most had plausible medical explanations in the case notes,” said Dr Niranthakumar.
“Then, because it was shown to be so rare, we questioned whether we could identify these patients using PICS while they’re with us. If so, then it would be a way of increasing vigilance for the misuse of insulin like in the Colin Norris case.”
Jamie Coleman, a consultant clinical pharmacologist at QEHB, said the potential for increasing patient safety was a further benefit.
“We have very good information systems that can provide retrospective surveillance data like this, but these systems also provide secondary benefits with automatic collection of data which may allow a greater degree of surveillance for patient safety,” he said.
“Within our organisation there were, ultimately, no unexplained cases but the potential is there to improve safety further.”
The research is published online in the free access journal PLoS One.