Foodborne Illness Outbreaks at Retail Food Establishments - National Environmental Assessment Reporting System, 25 State and Local Health Departments, 2017-2019.

Problem/Condition: Each year, state and local public health departments report hundreds of foodborne illness outbreaks associated with retail food establishments (e.g., restaurants or caterers) to CDC. Typically, investigations involve epidemiology, laboratory, and environmental health components. Health departments voluntarily report epidemiologic and laboratory data from their foodborne illness outbreak investigations to CDC through the National Outbreak Reporting System (NORS); however, minimal environmental health data from outbreak investigations are reported to NORS. This report summarizes environmental health data collected during outbreak investigations and reported to the National Environmental Assessment Reporting System (NEARS). Period Covered: 2017-2019. Description of System: In 2014, CDC launched NEARS to complement NORS surveillance and to use these data to enhance prevention efforts. State and local health departments voluntarily enter data from their foodborne illness outbreak investigations of retail food establishments into NEARS. These data include characteristics of foodborne illness outbreaks (e.g., etiologic agent and factors contributing to the outbreak), characteristics of establishments with outbreaks (e.g., number of meals served daily), and food safety policies in these establishments (e.g., ill worker policy requirements). NEARS is the only available data source that collects environmental characteristics of retail establishments with foodborne illness outbreaks. Results: During 2017-2019, a total of 800 foodborne illness outbreaks associated with 875 retail food establishments were reported to NEARS by 25 state and local health departments. Among outbreaks with a confirmed or suspected agent (555 of 800 [69.4%]), the most common pathogens were norovirus and Salmonella, accounting for 47.0% and 18.6% of outbreaks, respectively. Contributing factors were identified in 62.5% of outbreaks. Approximately 40% of outbreaks with identified contributing factors had at least one reported factor associated with food contamination by an ill or infectious food worker. Investigators conducted an interview with an establishment manager in 679 (84.9%) outbreaks. Of the 725 managers interviewed, most (91.7%) said their establishment had a policy requiring food workers to notify their manager when they were ill, and 66.0% also said these policies were written. Only 23.0% said their policy listed all five illness symptoms workers needed to notify managers about (i.e., vomiting, diarrhea, jaundice, sore throat with fever, and lesion with pus). Most (85.5%) said that their establishment had a policy restricting or excluding ill workers from working, and 62.4% said these policies were written. Only 17.8% said their policy listed all five illness symptoms that would require restriction or exclusion from work. Only 16.1% of establishments with outbreaks had policies addressing all four components relating to ill or infectious workers (i.e., policy requires workers to notify a manager when they are ill, policy specifies all five illness symptoms workers need to notify managers about, policy restricts or excludes ill workers from working, and policy specifies all five illness symptoms requiring restriction or exclusion from work). Interpretation: Norovirus was the most commonly identified cause of outbreaks reported to NEARS, and contamination of food by ill or infectious food workers contributed to approximately 40% of outbreaks with identified contributing factors. These findings are consistent with findings from other national outbreak data sets and highlight the role of ill workers in foodborne illness outbreaks. Although a majority of managers reported their establishment had an ill worker policy, often these policies were missing components intended to reduce foodborne illness risk. Contamination of food by ill or infectious food workers is an important cause of outbreaks; therefore, the content and enforcement of existing policies might need to be re-examined and refined. Public Health Action: Retail food establishments can reduce viral foodborne illness outbreaks by protecting food from contamination through proper hand hygiene and excluding ill or infectious workers from working. Development and implementation of policies that prevent contamination of food by workers are important to foodborne outbreak reduction. NEARS data can help identify gaps in food safety policies and practices, particularly those concerning ill workers. Future analyses of stratified data linking specific outbreak agents and foods with outbreak contributing factors can help guide the development of effective prevention approaches by describing how establishments' characteristics and food safety policies and practices relate to foodborne illness outbreaks.

Cost-effectiveness of Babesia microti antibody and nucleic acid blood donation screening using results from prospective investigational studies.

BACKGROUND: Babesia microti causes transfusion-transmitted babesiosis (TTB); currently, blood donor screening assays are unlicensed but used investigationally. STUDY DESIGN AND METHODS: We developed a decision tree model assessing the comparative- and cost-effectiveness of B. microti blood donation screening strategies in endemic areas compared to the status quo (question regarding a history of babesiosis), including testing by: (1) universal antibody (Ab), (2) universal polymerase chain reaction (PCR), (3) universal Ab/PCR, and (4) recipient risk-targeted Ab/PCR. The model predicted the number of TTB cases, complicated TTB cases, cases averted, and quality-adjusted life years (QALYs). Economic outcomes included each strategy's per-donation cost, waste (number of infection-free units incorrectly discarded), and waste index (number wasted units/number true positives). Sensitivity analyses examined uncertainty in transmission probabilities, prevalence rates, and other key model inputs. RESULTS: Universal PCR in four endemic states would prevent 24 to 31 TTB cases/100,000 units transfused (pht) at an incremental cost-effectiveness ratio (ICER) of $26,000 to $44,000/QALY (transmission probability dependent) and waste index of zero. Universal Ab/PCR would prevent 33 to 42 TTB cases pht at an ICER of $54,000 to $83,000/QALY and waste index of 0.05. The questionnaire is most wasteful (99.62 units wasted pht; 208.62 waste index), followed by the risk-targeted strategy (76.27 units wasted pht; 0.68 waste index). The model predicted zero cases of TTB or complicated TTB with universal Ab/PCR (versus [33, 42] and [13, 18] pht, respectively [no screening]). Results are highly sensitive to transmission probabilities. CONCLUSIONS: Universal PCR in endemic states is an effective blood donation screening strategy at a threshold of $50,000/QALY. Using a higher cost-effectiveness ratio, universal Ab/PCR is the most effective strategy.

Survival of blood transfusion recipients identified by a look-back investigation.

BACKGROUND: Survival of blood transfusion recipients is a critical consideration in assessing the outcomes of transfusion. Data from the USA on the short- and long-term survival of recipients are limited. MATERIALS AND METHODS: Blood product recipients were identified through a look-back study of Creutzfeldt-Jakob disease. Survival data were obtained from searches of the National Death Index or the Social Security Death Master File. Short- and long-term survival of recipients was analysed through descriptive statistics, Kaplan-Meier survival analysis, and stratified Cox proportional hazard modelling. RESULTS: This study includes data from 575 blood product recipients. One half of the recipients died within the first year of transfusion and the median time to death was 1.1 years. Survival rates at 5, 10, 15, 20, and 25 years after transfusion were 32%, 22%, 15%, 12%, and 9%, respectively. Survival rates varied with age at transfusion and type of component received, but not by gender. Survival after transfusion varied by year of transfusion, with recipients transfused in 1980-1989 having longer post-transfusion survival than those transfused in 2000-2010 (p=0.049). In multivariate models, the type of component transfused, but not the year of transfusion, was a significant predictor of survival among recipients; this effect varied by age. DISCUSSION: We provide an estimate of survival time from a geographically diverse sample of blood product recipients in the USA. Predictors of post-transfusion survival are numerous and complex, and may include year of transfusion and type of component transfused.