We investigated an outbreak of 54 cases of Salmonella Typhimurium phage type 9 (STM9) with a specific antibiotic resistance pattern. We used sequential analytic studies: two retrospective cohort studies, a case-control study, and a modified case-control study. An outbreak of salmonellosis due to Salmonella Typhimurium PT9 SSu (resistant to streptomycin and sulphafurazole) was identified. Fifty-four cases had illness onset from November 1998 to March 1999. Notifications commenced following a restaurant birthday party in December 1998. An initial cohort and case control study found no association with consumption of custard cake. However, case follow-up identified another cohort of people who had attended a birthday party in February at which 8/27 people who consumed a continental custard cake were ill compared to 0/10 who did not (P = 0.07). A revised case control study found illness was strongly associated with consumption of a particular continental custard cake (Mantel–Haenszel matched OR ∞, P = 0.00004). This report highlights the epidemiological value of using sequential study types, and persisting with the investigation of apparently sporadic food-borne outbreaks.