Food Protection Trends - July/August 2017 - 258

as the first language, formal food preparation training,
and the number of years passed since initial food handler
certification. The second section of the survey consisted of 27
food safety questions to be answered within 30 minutes. The
survey was accessible from a link in the invitation E-mail sent
by TrainCan Inc. The food safety knowledge survey portion
was created from a variety of multiple choice questions from
four different Ontario MOHLTC food safety exams, which
are used by local public health agencies throughout Ontario
to assess food safety comprehension and to certify students
after food handler training courses.
Permission from the Ontario MOHLTC was obtained
to use their tests for this research. To create the food safety
knowledge survey, tests and questions were selected at
random until three questions from each of the 10 main topics
of concern were chosen, resulting in 30 questions. The Food
Handler Training Program Requirements: Public Health Units
guideline of the Ontario MOHLTC indicates the central
topics of study as the following (15):
1. The role of the board of health/public health
regulations and legislation
2. Food safety management principles, e.g.,
HACCP-based principles
3. Basic microbiology
4. Time and temperature
5. Storage
6. Preparation
7. Cross-contamination
8. Food handler hygiene
9. Food premises sanitation/pest control, design,
and maintenance
10. Prevention of food allergy reactions, incidents
and response.
The survey was intended to be sent only to previously
certified food handlers in Ontario, but was distributed
to past TrainCan Inc. food handler certification students
across Canada. As a result, questions regarding Ontario
food safety regulations were removed from the survey,
bringing the number of questions to 28. A question asking
for the definition of the HACCP (Hazard Analysis Critical
Control Point) acronym was removed from the survey,
as it was determined not be a reliable indicator of food
handler knowledge; this removal resulted in a final total
of 27 questions.
The questions appeared in random order for each participant, and the multiple choice answers were in a different
order as well, to protect the integrity of the tests. Copies
of the answers for the food safety tests were obtained from
the Ontario MOHLTC and were used to determine the
participants' scores. The food safety survey multiple choice
questions were weighted equally for grading. Question 20,
"Handling food immediately after coughing and sneezing into
hands is an example of:" had two acceptable answers that
could be considered correct, and both choices, "physical con-

258

Food Protection Trends July/August

tamination" and "cross-contamination," were marked correct.
One common version of the definition for cross-contamination is the physical transfer of pathogens or foreign substances from a person, such as a food handler, or things, such as
utensils, equipment and surfaces, to food items (9). The word
"physical" was used in defining "cross-contamination" which
may have been misleading and as a result, both options were
considered correct. The passing grade of the food safety
survey was 70%, the same passing grade as on the Ontario
MOHLTC's exams, which should correspond with safe food
handling techniques (17).
Statistical analysis
Stata 14 for Windows (StataCorp LP, College Station,
Texas, USA) was used to perform statistical analysis.
Survey responses were analyzed descriptively with
frequency tabulations. Multivariable logistic regression
modeling was conducted to test the association between
demographic variables in this study and participants' food
safety knowledge scores. The dependent variable was the
proportion of correct answers out of 27 total questions. The
model was developed using a generalized linear modelling
framework with a binomial distribution and logit link. The
following demographic variables were assessed as predictors:
gender (female vs. male); age (60-79, 40-59, 20-39, vs.
≤ 19 years old); years lived in Canada (≤ 5, 5 to < 10, vs.
≥ 10), highest education level completed (high school or
lower, some college/university, vs. completed college/
university or higher degree); years worked in the food service
industry (< 1, 1 to < 5, 5 to < 10, vs. ≥ 10); currently holding
a managerial or supervisor position (yes vs. no); speaking
English as a first language (yes vs. no); formal training in
food preparation (yes vs. no); and years since obtaining most
recent food handler certificate (≥ 5, 3 to < 5, vs. 1 to < 3).
Demographic variables were first assessed in univariable
regression models, and those with P values ≤ 0.20 were
included in multivariable model building. Multivariable
models were calculated using a backward stepwise approach.
This included adding all significant demographic variables
from univariable models and removing them one at a time
until all remaining variables were significant at P ≤ 0.05.
Pairwise interactions between all variables in the final
multivariable model were assessed. Model diagnostics were
evaluated, including visual examination of residuals and
application of the Hosmer-Lemeshow goodness-of-fit test.
RESULTS
Of the 2,257 former TrainCan Inc. students who were
sent the survey invitation E-mail, 197 or 8.7%, completed
the survey. The average food safety survey score among
participants was 23.6 out of 27 (88%). A total of 17
participants obtained the highest possible score, 27 out of
27 (100%), on the food safety survey, and one participant
obtained the lowest score, 14 out of 27 (52%). The average

Table of Contents for the Digital Edition of Food Protection Trends - July/August 2017