Prompts during trauma treatment sees major reduction in errors
The introduction of a computerised decision-support tool for treating trauma patients in the emergency department of Melbourne's The Alfred Hospital has seen the rate of clinical errors reduce by 21 per cent.
The system, known as Trauma Reception and Resuscitation (TRR), was devised by the head of The Alfred's Trauma Centre, Professor Mark Fitzgerald, with the software designed by a team led by Kon Mouzakis, head of the Software R&D Group at Swinburne University.
TRR has been trialled over almost three years in two of The Alfred's trauma bays, and in that time errors have reduced markedly, as has the amount of blood used in treating trauma patients.
TRR is a real-time computer-aided decision and action support tool that is designed to be used in the first 30 minutes of trauma management. It involves a scribe – usually the trauma nurse leader – entering patient data such as diagnosis and treatment via a touchscreen.
Vital sign data is captured from monitors and all of this information is displayed overhead on a large LCD screen.
At the heart of the system is a series of medical algorithms that takes the data and prompts the trauma team in real time to confirm the state of the patient, perform certain procedures and administer drugs. If the level of oxygen or heart rate goes below the threshold of the algorithm, for instance, alerts appear on the screen.
Mr Mouzakis said trauma doctors know what they are doing and the system in no way replaces their expertise. However, analysis of the data collected over the trial has shown that in first 20 or so minutes after a trauma patient arrives, a clinical decision has to be made every 72 seconds.
“Things do get missed,” he said. “This system is designed to prompt you and to ask if you have checked X or have you checked Y.”
Mr Mouzakis said the system was trialled in two of the four trauma bays at The Alfred, but due to its success in reducing errors it is now used in all four.
“In the study the patients were assigned to a trauma bay randomly, and they collected quite a substantial number of patient data over 30-odd months.
“There was an independent review done on the data – we provided the information and a video auditing tool – and there was a substantial reduction in human error. We thought a significant reduction would be about five per cent but it came to about 21 per cent.”
There was also a huge reduction of upwards of 80 per cent in the amount of blood administered to patients, he said.
“Administering blood into trauma patients was very common, whereas this walks you through it to make sure everything is on the right path. The reduction in blood came out of left field.”
Mr Mouzakis said Professor Fitzgerald had done the initial work on the system and had successfully bid for funding through the Transport Accident Commission to build the system. “His thoughts were that we know that clinicians make mistakes and we know how many mistakes, so we'd like to reduce the amount errors they make in the first 30 minutes they see a trauma patient.
“Mark and his team did all the clinical work and then they produced a spec to say this is what we think should happen, some sort of computer-aided piece of software. It went out to tender and [Swinburne was] lucky enough to be the awarding institution.
“We worked with them collectively for nine months to get a decision-support system up and running, and then after it was up and running we worked with them for 33 months to collect data to see if it made a difference. And it did.”
Initially the software was a research tool, but as it has since been validated and findings published in a peer-review journal, the software has moved out of research mode and into clinical use. The Alfred and Swinburne teams are currently talking to hospitals in Queensland that are interested, and a cut-down version has been designed for use in the field such as on a helicopter.
Mr Mouzakis and his team have installed a version of the software on a Windows 8 tablet device that will enable it to be used in the field, he said.
“We already have an interface developed and we'd like to trial that in the next couple of weeks. We plan to go up to Sydney and talk to the doctors there, who travel on the emergency helicopters, and then we'll have a prototype for them to trial.”
The teams are also very keen to see it in use in regional hospitals that don't have specialised trauma teams. “If we can get it into regional hospitals they might not need two hours to fly the patient to a trauma centre. A lot of cases of severe trauma happen on country roads, so to be able to get to a facility and use the technology would be great.”
The system being used at The Alfred is also being improved. A small trial was run to see if voice recognition technology could be used instead of a scribe, but the amount of noise generated when attending to a trauma patient meant the system wasn't accurately picking up the information.
“We're now investigating other ways of getting that input,” he said. “We are looking at ways of potentially not having a scribe sitting behind a desk but having some sort of tablet that they can use while moving around the trauma area.”
Posted in Australian eHealth