OBJECTIVES: To evaluate whether a computer-based decision support system could be useful for the emergency medical system when identifying patients with acute myocardial infarction (AMI) or life-threatening conditions and thereby improve the allocation of life support level. METHODS: Patients in the Municipality of Göteborg who dialled the dispatch centre due to chest pain during a period of 3 months. To analyse the relationship between patient characteristics (according to a case record form used during an interview) and the response variables (AMI or life-threatening condition), multivariate logistic regression was used. For each patient, the probability of AMI/life-threatening condition was estimated by the model. We used these probabilities retrospectively to allocate advanced life support or basic life support. This model allocation was then compared with the true allocation made by the dispatchers. RESULTS: The sensitivity, that is, the percentage of AMI patients allocated to advanced life support, was 85.7% in relation to the true allocation made by the dispatchers. The corresponding sensitivity regarding allocation made by the model was 92.4% (P=0.17). The specificity was also slightly higher for the model allocation than the dispatcher allocation. Among the 15 patients with AMI who were allocated to basic life support by the dispatchers, nine died (eight during and one after hospitalization). Among the eight patients with AMI allocated to basic life support by the model, only one patient died (in hospital) (P=0.02). CONCLUSION: A computer-based decision support system including a prevalence function could be a valuable tool for allocating the level of life support. The case record form, however, used for the interview can be refined and a model based on a larger sample and confirmed in a prospective study is recommended.
OBJECTIVES: To evaluate whether a computer-based decision support system could be useful for the emergency medical system when identifying patients with acute myocardial infarction (AMI) or life-threatening conditions and thereby improve the allocation of life support level. METHODS:Patients in the Municipality of Göteborg who dialled the dispatch centre due to chest pain during a period of 3 months. To analyse the relationship between patient characteristics (according to a case record form used during an interview) and the response variables (AMI or life-threatening condition), multivariate logistic regression was used. For each patient, the probability of AMI/life-threatening condition was estimated by the model. We used these probabilities retrospectively to allocate advanced life support or basic life support. This model allocation was then compared with the true allocation made by the dispatchers. RESULTS: The sensitivity, that is, the percentage of AMI patients allocated to advanced life support, was 85.7% in relation to the true allocation made by the dispatchers. The corresponding sensitivity regarding allocation made by the model was 92.4% (P=0.17). The specificity was also slightly higher for the model allocation than the dispatcher allocation. Among the 15 patients with AMI who were allocated to basic life support by the dispatchers, nine died (eight during and one after hospitalization). Among the eight patients with AMI allocated to basic life support by the model, only one patient died (in hospital) (P=0.02). CONCLUSION: A computer-based decision support system including a prevalence function could be a valuable tool for allocating the level of life support. The case record form, however, used for the interview can be refined and a model based on a larger sample and confirmed in a prospective study is recommended.
Authors: Charles Richard Knoery; Janet Heaton; Rob Polson; Raymond Bond; Aleeha Iftikhar; Khaled Rjoob; Victoria McGilligan; Aaron Peace; Stephen James Leslie Journal: Crit Pathw Cardiol Date: 2020-09