Presenting complaint and mortality in non-surgical emergency medicine patients

Abstract: In 1995 and 2000 a total of 29 886 non surgical ED visits at Uppsala University Hospital were registered. Presenting complaint, admittance to a ward, length of stay, in-hospital mortality, discharge diagnoses, 30-day and long-term mortality were registered. The presenting complaints were sorted into 33 presenting complaint groups (PCGs).For different PCGs there was different in-hospital fatality rate. Compared to the largest PCG, chest pain, the gender and age adjusted OR was 2.12 (95% CI 1.01 – 4.44) for the miscellaneous complaint group and 2.04 (95 % CI 1.35 – 3.08) for the stroke–like symptom group. Within a given PCG the in-hospital mortality could vary depending on discharge diagnoses. By relating PCG and long term mortality to the expected mortality in the population, the Standardized Mortality Ratio (SMR) could be calculated. The SMR was found to be highest in seizure 2.62 (95 % CI 2.13 – 3.22), intoxication 2.51 (95% CI 2.11-2.98) and symptoms of asthma 1.8 (1.65 – 2.06). For the same discharge diagnoses the long term mortality could differ considerably depending on PCG at ED arrival (p<0.001).Between 1995 and 2000 there was a 30 % increase in ED visits at the non surgical ED. PCGs representing lesser severe conditions had increased. Demographic changes could account for 45 % of the increment and the remaining increase could be ascribed to change in visiting pattern.In the 2000 cohort 41.0 % of all visits were performed by re-visitors. The number of revisits and five-year mortality had an inversed u-shaped relationship were patients with three re-visits within the same year had an increased mortality compared to patients with more or less visits.Conclusion: It is possible to define presenting complaint groups (PCGs) that are robust and consistent over time and useful as a tool for epidemiological studies in the ED.