When Denmark got its first COVID-19 case, hospitals and politicians were very concerned. Meanwhile, the personnel behind the country's two emergency medical telephone lines were preparing to be overwhelmed in a situation they had not experienced before. New research indicates that the system was not overburdened.
When COVID-19 arrived in Denmark, the politicians and the population were especially concerned about the hospital intensive care units. They needed to be expanded, prepared and equipped with stocks of personal protective equipment and ventilators so that Denmark did not risk a situation such as that in Italy with insufficient beds or equipment available.
Meanwhile, another part of the emergency medical services was as busy as ever. At the regional 1813 medical helpline and the 112 general emergency number , nurses and doctors anticipated the many calls that might arise from people worried about COVID-19.
“Even before SARS-CoV-2 really reached Denmark, the volume of emergency calls had tripled in Italy – especially in the worst-hit regions. They desperately needed to solve this problem as events unfolded,” says Theo Jensen, a doctor and researcher at the Copenhagen Emergency Medical Services.
Studies from the United States, for example, have shown that long waiting times can lead to mental barriers that can make potential patients reconsider calling the emergency medical services the next time they have severe symptoms. Slow call response can also delay help to the patient, which can eventually lead to more deaths. Theo Jensen and several colleagues investigated how the Copenhagen Emergency Medical Services coped during the first wave of COVID-19 and how this knowledge can be used today and in the future.
Increasing personnel is insufficient
The first case of COVID-19 was confirmed in Denmark on 27 February 2020. One month later, Denmark had 430 people hospitalized, 109 individuals in intensive care departments and 52 deaths. About half the confirmed cases were in the Capital Region of Denmark.
Initially, people could call either 112 or 1813 if they had COVID-19 symptoms. Depending on how severe the symptoms were, they were recommended to self-isolate, be further investigated or be picked up by an ambulance. On 9 March, the system was changed so that callers with COVID-19 symptoms could click themselves to the separate COVID-19 support track on the initial menu of the 1813 medical helpline. In 3 days, experts developed a web-based self-triage system that was ready for use on 15 March.
“The emergency medical services act as a gateway to the hospitals, and we are therefore interested in ensuring that they are always ready to help people. Sometimes the lines have bottlenecks for various reasons, such as during festivals, when you can intervene and increase personnel. But an epidemic that affects such large parts of the population requires testing other measures. And we researchers need to maximize our efforts and learn quickly,” explains Theo Jensen.
He and his colleagues therefore dropped everything else to monitor the developments and challenges faced by the emergency medical services in the first month after Denmark got its first official COVID-19 case.
“Not everything works 100% perfectly during a pandemic. However, based on our starting-point, my immediate assessment is that the relief measures implemented have worked quite quickly and efficiently,” says Theo Jensen.
Emergency medical services inspired by example
The number of calls to either the nationwide 112 emergency number or the regional 1813 medical helpline number increased by 23% compared with 2019. The average waiting time for 1813 was 12.0 minutes versus 2.4 minutes in 2019. The 112 emergency number received about 8,000 calls and the Capital Region’s 1813 medical helpline about 84,000 calls during the first month of the pandemic, about the same number as 12 months earlier. When the special COVID-19 support track was established on 9 March, it received about 21,000 calls in just over 2 weeks, thus reducing the burden on the 1813 medical helpline, which would otherwise have handled calls about COVID-19.
“We wanted to determine whether this COVID-19 support track made any difference. The track was introduced based on considerable experience obtained globally when waiting times became much longer, such as in the United States, which has had some fairly severe influenza epidemics in which the vaccines have not always worked. They have had good experience with creating telephone lines staffed with nurses who could determine who should be referred and where to,” explains Theo Jensen.
In other countries, different types of platforms have also been introduced to ease the pressure on the classical emergency medical services, such as chatbots or various kinds of digital and dynamic questionnaires, through which people with mild symptoms can self-assess whether they should seek professional help or stay home. Something similar was introduced in Denmark 4 days after Denmark’s Prime Minister announced on 11 March that Denmark was being locked down.
“Something had to be done initially, because otherwise you would potentially have to bring in more and more support personnel. And many individuals with mild symptoms did not need to call and wait for a long time. Then the web-based self-triage system was established, which was quite simple and not yet evidence-based because we still did not know very much about COVID-19 and its symptoms,” says Theo Jensen.
Tools for future use
The self-triage system for assessing symptoms was available on the Capital Region’s website and asked potentially ill individuals various questions about their symptoms before finally advising them to either stay at home or seek professional help.
The researchers were interested in determining whether the opportunity to self-assess reduced the call volume to the emergency medical services. During the period Theo Jensen and his colleagues investigated, nearly 11,000 people used the web-based system, but they concluded that the amount of web triage was not correlated with changes in call volumes.
According to Theo Jensen, artificial intelligence could not be incorporated because the data were insufficient to base the web system on evidence. However, on 14 October, an updated version was launched based on artificial intelligence and on the knowledge acquired since the first version was launched.
“We have tried to learn from the experience of the web-based self-triage system and have collected a huge quantity of data. Now we gradually know when people will click further, give up or follow the self-triage to the end. We have also used artificial intelligence on both the standard 1813 medical helpline and COVID-19-related track and can see what happens in the different types of calls,” explains Theo Jensen.
Some people call about symptoms such as cough and fever, state their age and describe other health problems. The artificial intelligence continually learns from the information provided. The system is therefore getting better and better at sorting calls, assessing the risks of various patient profiles and determining who should talk to a doctor, be transported directly to a hospital or not immediately worry.
The researchers can further refine the intelligent software by comparing the artificial intelligence’s assessments with data for the individual’s hospital stay, such as when the system’s conclusions do not fit reality, including when people end up hospitalized 2 days later even though the opposite was predicted during the initial call.
“Based on this, you can design a rather ingenious tool that can ask the right people the right questions. We have already learned an incredible amount from these initiatives, and this knowledge will not be lost when COVID-19 is gone. We have a template that we can adapt if an epidemic arises with other symptoms,” concludes Theo Jensen.