A risk-based restaurant inspection system in Los Angeles County.

The majority of local health departments perform routine restaurant inspections. In Los Angeles County (LAC), California, approximately $10 million/year is spent on restaurant inspections. However, data are limited as to whether or not certain characteristics of restaurants make them more likely to be associated with foodborne incident reports. We used data from the LAC Environmental Health Management Information System (EHMIS), which records the results of all routine restaurant inspections as well as data regarding all consumer-generated foodborne incidents that led to a special restaurant inspection by a sanitarian (investigated foodborne incidents [IFBIs]). We analyzed a cohort of 10,267 restaurants inspected from 1 July 1997 to 15 November 1997. We defined a "case restaurant" as any restaurant with a routine inspection from 1 July 1997 to 15 November 1997 and a subsequent IFBI from 1 July 1997 to 30 June 1998. Noncase restaurants did not have an IFBI from I July 1997 to 30 June 1998. We looked for specific characteristics of restaurants that might be associated with the restaurant subsequently having an IFBI, including the size of restaurant (assessed by number of seats), any previous IFBIs, the overall inspection score, and a set of 38 violation codes. We identified 158 case restaurants and 10,109 noncase restaurants. In univariate analysis, middle-sized restaurants (61 to 150 seats; n = 1,681) were 2.8 times (95% confidence interval [CI] = 2.0 to 4.0) and large restaurants (>150 seats; n = 621) were 4.6 times (95% CI = 3.0 to 7.0) more likely than small restaurants (< or =60 seats; n = 7,965) to become case restaurants. In addition, the likelihood of a restaurant becoming a case restaurant increased as the number of IFBIs in the prior year increased (chi2 for linear trend, P value = 0.0005). Other factors significantly associated with the occurrence of an IFBI included a lower overall inspection score, the incorrect storage of food, the reuse of food, the lack of employee hand washing, the lack of thermometers, and the presence of any food protection violation. In multivariate analysis, the size of restaurant, the incorrect storage of food, the reuse of food, and the presence of any food protection violation remained significant predictors for becoming a case restaurant. Our data suggest that routine restaurant inspections should concentrate on those establishments that have a large seating capacity or a poor inspection history. Evaluation of inspection data bases in individual local health departments and translation of those findings into inspection guidelines could lead to an increased efficiency and perhaps cost-effectiveness of local inspection programs.