Lassa fever is a viral hemorrhagic illness with a case fatality rate for hospitalized patients as high as 69%. Identifying cases before they progress to serious illness can lead to earlier treatment and improved clinical outcomes. Three existing clinical prediction tools were evaluated on their ability to predict the in-hospital mortality in Lassa fever: the quick Sequential Organ Failure Assessment (qSOFA), the Modified Early Warning System (MEWS), and the Universal Vital Assessment (UVA). This was a retrospective cohort study of patients admitted to the dedicated Lassa fever ward of the Kenema Government Hospital in Sierra Leone between May 2013 and December 2019. Data among three serology groups were analyzed: Lassa antigen-positive (Ag+) regardless of IgM status, Lassa Ag- and IgM+, and Lassa Ag- and IgM- cases. There were 123 cases of suspected Lassa fever included in this study. Abnormalities in respiratory rate, oxygenation status, mental status, and serum markers of kidney and liver dysfunction were more likely seen in the Ag+ group, which had an in-hospital mortality of 85.7%. For the Lassa Ag+ group, the sensitivity and positive predictive value of qSOFA ≥ 2 was 70.6% and 92.3%, MEWS ≥ 5 was 96.9% and 86.1%, and UVA ≥ 5 was 60.0% and 100.0%. The MEWS and UVA scores show potential for use in Lassa fever, but there is opportunity for future development of a tool that includes the clinical and laboratory markers specific to Lassa fever.