Type of publication:
Conference abstract
Author(s):
*Baker J.; *Day S.; *Marsh A.
Citation:
Emergency Medicine Journal. Conference: Royal College of Emergency Medicine Annual Scientific Conference, RCEM 2022. Belfast United Kingdom. 39(3) (pp 259-260), 2022. Date of Publication: March 2022.
Abstract:
Aims/Objectives/Background In the COVID-19 pandemic the Shrewsbury and Telford Hospital NHS Trust has isolated suspected cases in high and low suspicion cohort bays to reduce nosocomial infection. Before rapid PCR swabs were in routine use, we sought tools to aide identifying COVID-19 positive patients. Methods/Design We collected data from two cohorts in April and June 2020 totalling 317 patients, with positivity rates of 33% and 5% respectively. We retrospectively correlated neutrophil count, lymphocyte count, LDH and AST to positive and negative swab results. Predictive value of COVID-19 positivity was assessed via their receiver operator characteristic. Areas under the curve were as follows: Neutrophils 0.75, lymphocytes 0.64, combined neutrophil and lymphocyte count 0.82, AST 0.65 and LDH 0.7. We developed a diagnostic aide to assist in allocation of high and low suspicion based on parameters for neutrophil count, lymphocyte count and LDH, each of which was assigned red (higher probability) or green (lower probability) in a 'traffic light' system. Combined and applied retrospectively to 252 patients with suspected COVID-19, with a positivity prevalence of 5%, three green values generated a negative predictive value for COVID-19 of 100%, two greens 98% and three reds a positive predictive value for COVID-19 of 44%. Results/Conclusions This diagnostic aide was applied from August 2020 within the Trust Emergency Departments and Acute Medical Units to aide cohort decisions. A retrospective application to all 213 patients with positive swabs admitted from August to November 2020 demonstrated that 69% were highlighted as at least two 'red lights' and only 1.4% were erroneously highlighted as three 'green lights'. The aide is an example of a rapidly developed evidence based tool and, particularly if updated with data from other centres, could be widely employed in low-resource healthcare settings. (Figure Presented).