Study Design, Area and Period
A cross-sectional study was conducted among food handlers who were participated in food preparation, dispatch and store of Adigrat University student’s cafeteria from March to July 2018. The annual rainfall ranges from 400-600mm and the minimum and maximum temperature range from 6-21.80c.Currently, the University enrolled more than 15,000 students who are getting dining services in the student cafeteria. There are six cafeterias and a total of 700 food handlers are working in the student’s cafeteria (Adigrat University human resource management and registrar office).
Sample size determination and sampling technique
Sample size determination
The sample size was determined by using a single population proportion formula.
n = (Zα/2)2 P (1-P)
The prevalence Salmonella among university food handlers taken from Arba Minch University, South Ethiopia (6.9%) which was done by Mama and Alemu  Then with a margin of error (5%), (d=0.03) and 95% level of confidence (z=1.96), the sample size was calculated as follows:
n= (1.96)2*0.069(0.931) = 274, with 10 % non response rate =301
Therefore, a total of 301 food handlers were included in the study from all cafeteria of the university. A simple random sampling technique was employed. The lottery method was used to select the study subjects after a complete list of food handlers was obtained from a roster of cafeteria office Adigrat University.
Food handlers working in Adigrat University student’s cafeterias were included in the study.
Food handlers who have taken antibiotics within one week, antihelminthics, and those with clinical signs of under- typhoid fever and were excluded from the study.
Data collection and Sample processing
Socio-demographic and Specimen Collection, Handling and Transportation
A structured questionnaire was used to collect the socio-demographic and risk factor related data. Questionnaires were checked for accuracy and completeness. After proper instruction, about 2 g of fresh Stool specimens were collected from food handlers with a labeled wide-mouthed plastic container and a clean wooden applicator stick. Specimens were immediately transported to the laboratory using icebox.
Isolation and identification
The stool specimen was collected and transported to Adigrat University Medical Microbiology laboratory within one hour of collection. Stool specimens were immediately inoculated in Selenite F enrichment broth and incubated at 37°C for 24 hours, and then subculture onto selective media of xylose-lysine desoxycholate agar (XLD) and Hektoen enteric medium agar incubate at 37°C for 18-24 Hrs. The isolated colonies were differentiated and identified based on gram stain, colonial morphology and pigmentation, hemolysis on blood agar, catalase test, oxidase test, carbohydrate fermentation, H2S production, motility, indole formation and urease production, citrate utilization and incubated for 24 to 48 hours at 37°C. Then colonies producing an alkaline slant with acid butt and hydrogen sulfide production on Triple Sugar Iron Agar, positive for lysine, negative for urea hydrolysis, negative for indole test, positive for citrate utilization and motility test were considered to be Salmonella. Colonies which were urease negative, indol positive/negative, in Triple Sugar Iron agar produce a pink-red slope and yellow butt with no blackening, Lysine decarboxylase negative and citrate negative is identified as Shigella species. Finally, all of the confirmed Salmonella and Shigella isolates were examined for antimicrobial susceptibility.
Antimicrobial susceptibility tests
Antimicrobial susceptibility testing was performed using the modified Kirby- Bauer disc diffusion method according to Clinical and Laboratory Standards Institute (CLSI) guidelines, 2016 . Using a sterile wire loop, 3-5 well-isolated colonies of the test organism was emulsified into a tube of 3-4 ml sterile physiological saline to get bacterial inoculums equivalent to 0.5 McFarland turbidity standards. Then the standardized suspension (test organisms) were uniformly swabbed within 15 minutes using a sterile cotton swab into Muller-Hinton agar and allowed to dry. After that, the antibiotic discs were placed manually on the medium and incubated at 37°C for about 18 hours and the zones of inhibition were measured using a caliper. The interpretation of the results was made based on the CLSI criteria as sensitive, intermediate and resistant . The following antimicrobials are prioritized by considering local prescription; gentamicin (10 μg),ampicillin (30 μg), amoxicillin (30 μg), ciprofloxacin (5 μg), clarithromycin (30 μg), chloramphenicol (30 μg), cotrimoxazole (25 μg), amoxicillin-clavulanic acid (30 μg), and ceftriaxone (30 μg) .
Data Quality Assurance
Data quality was ensured at various activities of the study by following a prepared standard operating procedure (SOP). Questionnaires were prepared in a clear and precise way and translated into local language and back-translated to English to ensure the consistency of the questionnaires. The pretest was done on 5% of food handlers and modifications were made accordingly. To ensure general safety; universal bio-safety precautions were followed. American Type Culture Collection (ATCC) strains P. aeruginosa (ATCC 27853), E. coli (ATCC-25922) were used as control strains for the culture and antimicrobial susceptibility testing.
After collection of socio-demographic characteristics, associated risk factors and laboratory data using a structured questionnaire and laboratory report format, Data were edited, cleaned, entered and analyzed using statistical package for social science (SPSS) version 22. Descriptive statistics, Bivariate, and multivariate logistic regression were performed. Bivariate logistic regression was employed to look association between the outcome variable and each independent variable. A binary logistic regression analysis was used to calculate the odds ratios (OR); Crude Odds Ratio (COR) and Adjusted Odds Ratio (AOR) to ascertain the degree of association between risk factors of culture-positive variable. In this study, multi-collinearity among independent variables was detected using the standard errors for regression coefficients. Finally, variables with p-value < 5% with a corresponding 95% confidence interval (CI) were considered as statistically significant.