Background: During the 2014 Ebola outbreak in Sierra Leone, widespread infections among healthcare workers at Kenema Government Hospital severely disrupted clinical operations. Whether these infections primarily arose from repeated patient exposures or from transmission among colleagues remained unclear, particularly as patient admissions exceeded the hospital's capacity for safe isolation and care. Methods: We revisited the outbreak using expanded viral surveillance and classical genomic analysis, Bayesian transmission inference with juniper, and firsthand epidemiologic accounts from hospital staff. We analyzed 279 Ebola virus genomes from Kenema District, including 35 from infected healthcare workers, and 514 genomes in the broader transmission reconstruction, explicitly accounting for unsampled cases. Results: The 35 healthcare worker genomes comprised 28 haplotypes, including five shared by multiple cases. In the reconstructed transmission network, 22 candidate healthcare worker-to-healthcare worker links exceeded a posterior probability threshold of 5%, representing 6.1-fold enrichment over expectation. Healthcare workers accounted for 33.2% of expected onward transmissions at their peak in the two-week window beginning June 23, 2014. Firsthand anecdotal accounts situated inferred relationships within specific wards, clinical procedures, and working relationships of KGH, including care of infected colleagues and infections among laboratory, pharmacy, and other personnel beyond routine bedside care. Conclusions: Ebola transmission involved interconnected groups of healthcare workers across multiple hospital departments during a period of severe resource constraints and expanding patient demand. These findings underscore the importance of protecting and surveilling healthcare workers as an interconnected workforce during emerging outbreaks.