A statistical analysis of clinical presentations and outcomes of COVID-19 patients in Sierra Leone: case study of admissions at 34 Military Hospital for COVID-19 patients

Abstract

Background: COVID-19 outcomes vary widely across settings and are shaped by patient factors and health-system capacity. Evidence from Sierra Leone is limited, particularly on how age, comorbidities and presenting features relate to in-hospital admission outcomes. To describe the demographic, comorbidity and symptoms presentation of COVID-19 patients admitted at 34 Military Hospital, Freetown, assess bivariate associations between patient characteristics and discharge disposition; and estimate independent predictors of death using multinomial logistic regression.

Methods: Retrospective analysis of routinely collected data for 524 laboratory-confirmed COVID-19 patients admitted at the said facility. Primary outcome was discharge disposition (Discharged/Recovered, Deceased, Transferred). Descriptives summarized cohort characteristics. Pearson’s χ² (Fisher’s exact as needed) tested bivariate associations (α=0.05). Multinomial logistic regression (reference: Discharged) included age bands, sex, diabetes and difficulty breathing; results reported as odds ratios (OR) with 95% CIs and p-values.

Results: Patients were 63.2% male; the largest age bands were 22–32 years (21.8%) and 33–43 years (21.9%). Outcomes: 96.6% discharged, 2.7% died, 0.8% transferred. Comorbidities: hypertension 15.3%, diabetes 5.0%; other chronic conditions were rare. Symptoms were sparsely recorded; difficulty breathing 5.2%, malaise 2.9%, ARDS 1.5%. In bivariate analyses, age was associated with outcome (χ² p=0.002); sex (p=0.794) and diabetes (p=0.242) were not. In multivariable models, odds of death increased steeply with age versus 33–43 years (e.g., 66–76 years OR≈7.3×10⁷, p<0.001). Diabetes (OR 1.41, 95% CI 0.26–7.66, p=0.688) and difficulty breathing (OR 2.73, 95% CI 0.65–11.43, p=0.170) were not independent predictors; transfers showed no significant predictors. Wide CIs reflected sparse events.

Conclusions: Mortality was low overall but rose sharply with age, identifying older adults as the principal risk group in this setting. Comorbidity and symptom variables, at their observed prevalences and documentation quality, did not independently predict death cause. Embedding age-based early escalation and strengthening routine documentation are pragmatic priorities to improve outcomes and future analytics.