Researchers from the European University of the Atlantic (UNEATLANTICO), including Santos Gracia, Silvia Aparicio, Rubén Calderón, Eduardo Silva and Luis Dzul, are participating in a study that develops clinical phenotypes based on vital signs and biomarkers of acutely ill patients to facilitate risk classification in pre-hospital care.
Emergency medical services (EMS) play a crucial role in the management of life-threatening acute illnesses. EMS physicians must make quick and accurate decisions in critical and dynamic situations. However, quickly identifying high-risk patients is a challenge in pre-hospital care. For this reason, strategies are being implemented to improve the timely recognition of these patients, such as the use of scoring, biomarkers, blood testing, risk modelling, phenotyping, among others.
Blood testing at the scene of the incident is one of the most effective strategies to detect hidden high-risk conditions in patients who do not have an acute, obviously life-threatening illness. It helps to detect conditions such as electrolyte imbalances, metabolic and endocrine diseases, respiratory failure, anaemia or renal failure. Point-of-care testing (POCT) is used to obtain rapid results from blood work, venous or arterial blood gas levels, renal profile, glucose, haemoglobin, among others. POCTs are usually only available in hospitals, however, they now also help to support the decision-making process at the scene of an incident.
In pre-hospital critical care, early warning scores, risk scales and predictive models are frequently used to detect time-dependent illnesses and their short- and long-term prognoses. Also, the use of phenotypes in hospitals to identify specific pathophysiological conditions is increasingly common. Phenotype refers to the set of observable morphological and physiological features of a person, the study of which is essential to understand the different ways in which a disease can present itself. However, studies and research on its use in pre-hospital care are very scarce.
In this context, the aim of the study was to develop clinical phenotypes based on vital signs and biomarkers collected by EMS physicians during initial emergency care in patients with acute life-threatening illnesses. The analysis was based on data from 7909 male and female patients aged 51 to 80 years.
Using unsupervised machine learning, three distinct clinical phenotypes, termed alpha, beta and gamma, were identified and associated with different levels of disease severity. The alpha phenotype is characterised by severe heart disease and other conditions associated with high short- and long-term morbidity and mortality. In addition, patients with these conditions showed a marked dependence on life-sustaining interventions at the scene. The beta phenotype conditions were highly heterogeneous. Patients were characterised by improved acid-base balance, increased blood oxygenation, mild hyperlactacidemia and mild hyperglycaemia. Finally, the gamma phenotype included a priori less severe diseases or non-specific conditions; patients presented results within the ranges considered normal.
Phenotyping, i.e. the classification of patients into different groups based on clinical characteristics and biomarkers, is becoming increasingly common in clinical practice. This methodology is already used in diseases such as sepsis, chronic obstructive pulmonary disease and heart failure. However, its application in the pre-hospital setting is still in its infancy. Further research is therefore recommended, as this methodology has important implications for emergency triage and pre-hospital critical care.
To learn more about this study, click here.
To read more research, consult the UNEATLANTICO repository.