INTRODUCTION
PROBLEM STATEMENTS
The agricultural revolution significantly increased food production and improved the health of laborers, which is crucial for human capital However, food safety issues pose a serious threat to sustainable development and national security in developing countries Recent outbreaks of food-related diseases have shifted the focus of many nations from merely increasing food quantity to prioritizing food safety.
The use of pesticides, chemical fertilizers, and feedstuff enhances agricultural productivity, but their overuse can compromise food quality Additionally, preservation methods, processing techniques, and the use of food additives can further reduce food safety According to the World Health Organization (2015), the primary contributors to food-borne diseases include bacteria, viruses, parasites, chemicals, and toxins, with children, pregnant women, and the elderly being the most vulnerable populations.
The advancement of transportation and international trade has transformed food safety into a global concern, affecting countries worldwide Notable incidents, such as the Chinese milk scandal and the New Zealand material milk crisis, have resulted in significant losses for manufacturers and posed serious risks to consumer health across borders.
1 World Health Organization (WHO) and Food and Agriculture Organization (FAO) defined:
“Food safety is the assurance that food will not cause harm to the consumer when it is prepared and eaten according to its intended use” (WHO and FAO, 2009, p 6).
Figure 1.1: The number of food-borne cases annually (WHO, 2015)
Foodborne diseases (FBD) are a global issue, but developing regions like Southeast Asia and Africa report the highest cases, while developed areas such as Europe and America have fewer incidents Interestingly, despite facing food scarcity and limited safety controls, African countries experience fewer foodborne illnesses than Southeast Asia This discrepancy may be attributed to the greater diversity of nutritious foods and the tropical climate in Southeast Asia, which create optimal conditions for harmful bacteria and other food safety risks.
According to the World Health Organization (WHO) in 2015, food-borne diseases affect nearly 10% of the global population annually, leading to approximately 420,000 deaths, with one-third of these fatalities being children Among various food-borne diseases, diarrheal diseases are the most prevalent illnesses associated with these infections.
Africa unsafe food, accounted for half of global burden of FBD and made 550 million people falling ill (including 220 million children), cause 230,000 deaths (96,000 children’s)
Figure 1.2: The number of death caused by FBD annually (WHO, 2015)
The trend in food-borne disease (FBD) deaths mirrors the number of cases, with the Southeast Asian and African regions reporting the highest fatalities, while Europe and America experience the lowest Notably, Africa has fewer deaths compared to Southeast Asia, likely due to greater international medical support in Africa and differences in the physical resilience of local populations between the two regions.
The FBD caused the burden about 33 million DALYs 2 Diarrheal diseases agents were the largest contributors, accounted for 18 million DALYs, 54% of total All three
Disability-adjusted life years (DALYs) serve as a crucial health metric, quantifying the overall burden of disease by combining years of life lost due to premature death (YLL) with years lived with disability (YLD) resulting from various health conditions.
Africa figures 1.1, 1.2, 1.3 both indicated that South East Asian and African region’s food safety issue is severe and these areas suffered an enormous burden from FBD
Despite having fewer death cases than Asia, Africa experiences a burden of foodborne diseases (FBD) that is nearly double that of Southeast Asia and significantly higher than other regions The total burden in Europe, the Western Pacific, and the Americas is comparable to that of Southeast Asia, yet only half of Africa's burden These statistics highlight the disparities in healthcare access and food safety regulations across these regions and their profound effects on population health outcomes.
The burden of foodborne diseases (FBD) varies in severity, with time serving as a common metric for measuring death and disability According to the World Health Organization (WHO, 2015), one Disability-Adjusted Life Year (DALY) represents one year of healthy life lost due to these diseases.
Vietnam is classified by the World Health Organization (WHO) as part of the Western Pacific region, which experiences a moderate level of food-borne illnesses compared to global standards This region encompasses a diverse range of countries, including developed nations like Australia, Japan, and South Korea, alongside developing countries such as Cambodia, the Philippines, and Vietnam As a result, the contributions to food safety and health outcomes vary significantly among these nations Detailed information regarding Vietnam's specific situation is presented in Chapter 4.
Food-borne diseases (FBD) affect both developing and developed countries, with the USA reporting 864 outbreaks in 2014, leading to 13,246 illnesses, 712 hospitalizations, and 21 deaths, according to the Centers for Disease Control and Prevention (CDC) Restaurants were the primary source of food poisoning, accounting for 65% of cases, while private homes contributed to 12% Bacterial infections were the leading cause, linked to 22% of total cases These statistics highlight that even nations with robust healthcare systems face significant challenges related to food safety.
CDC also predicted that food safety issue would continue emerge in the future due to:
- Changes in our food production and supply, including more imported foods
- Changes in the environment leading to food contamination
- Better detection of multistate outbreaks
- New and emerging bacteria, toxins, and antibiotic resistance
- Changes in consumer preferences and habits
- Changes in the tests that diagnose foodborne illness.
Food-borne diseases are a global issue that transcends national borders, exacerbated by increasing immigration and trade exchange in today's interconnected world.
RESEARCH OBJECTIVES AND RESEARCH QUESTIONS
Vietnam's diverse food market offers convenience for households, leading most Vietnamese families to prepare and consume at least one meal at home daily This cooking behavior is influenced not only by external factors but also by individual socio-economic status, personal characteristics, and living conditions Additionally, the tradition of home cooking in Vietnam, rooted in Asian customs and agricultural culture, significantly shapes these practices The transmission of culinary knowledge and food preservation techniques across generations highlights the profound impact of tradition and culture on household cooking behaviors.
Research conducted in Thua Thien Hue province (Duong, 2013) and Ho Chi Minh City (Nguyen, 2010) highlights a significant correlation between individuals' knowledge and attitudes towards food safety and their actual food safety practices in various settings, including food factories, restaurants, and households The "10 golden principles in food processing" (MOH, 2005) outline essential behaviors that promote food safety.
- Clean, tidy kitchen and the cooker surface
- Using waste basket with cover
- Use clean water to handle food
- Use clean tool to prepare and divide food
- Not use forbidden food additives or out of date food
- Washing hand before cooking and after toileting
- Not smoke, spit out or nail polished while cooking
This thesis builds on golden principles frequently utilized in various research studies in Vietnam However, many of these studies primarily focus on medical perspectives, often neglecting socio-economic factors and individual characteristics Consequently, the impact of these elements on food safety behaviors remains inadequately assessed.
The research findings highlight the existing food safety issues within the community but fail to analyze multiple factors that could confirm their impact Additionally, the study's participants did not take part in the annual survey, resulting in a lack of discussion regarding the effects of government policies and activities on food safety.
This thesis aims to identify and evaluate the factors influencing food safety behavior, including socio-economic status, knowledge of food safety, personal perceptions, and information sources By understanding these personal determinants, the government can formulate effective strategies to modify behaviors, mitigate risks, and reduce the incidence of food-borne diseases, ultimately lessening their impact on public health.
To analyze food safety behavior and health outcomes of household primary cook
- To determine the relationship between knowledge, perception of consumer and their food safety practice at household kitchen
- To evaluate the impact of individual food safety practice to their food-borne disease probability
(1) Do food safety knowledge and perception have impact on individual food safety practice?
(2) How food safety practice affect to individual food-borne poisoning risk?
SCOPE OF RESEARCH
The thesis utilized data from a 2013 survey on Individual Food Poisoning and the Knowledge, Attitude, and Practice of households in Ho Chi Minh City Conducted across 24 districts from March to April 2013, the survey targeted primary cooks in households, with local medical staff serving as data collectors.
The annual survey features a diverse range of participants each year, including householders in 2010 and 2013, and restaurant workers from 2012 to 2016 Furthermore, the specific respondents within the same category vary annually, resulting in the latest household data being from 2013, which is not linked to the 2010 survey Consequently, this inconsistency prevents the creation of panel data for analysis.
This research focuses on the behavior and incidence of acute food poisoning in the Ho Chi Minh City community from March to April 2013, aiming to assess the influencing factors The study employs descriptive statistics alongside econometric methods, including factor analysis, multivariate probit, and propensity score matching, to analyze the data effectively.
THESIS STRUCTURE
Due to the available of the data, thesis is composed as the structure below:
Chapter 1 introduces the research problem, highlighting the significance and scope of the study on FBD (Foodborne Diseases) It provides a comprehensive overview of the global impact of FBD and outlines the thesis objectives, setting the stage for the subsequent analysis.
Chapter 2: Literature Review examines the definitions of key concepts and explores prior research on the factors and models utilized in studies This foundational review supports the development of an analytical framework and methodology to assess the impact of each component effectively.
Chapter 3 outlines the research methodology, detailing the framework and econometric tools utilized in the study It also discusses the data sources and collection methods employed, along with a comprehensive description of the variables involved.
Chapter 4: Research Results presents a comprehensive analysis of the data, highlighting key findings and comparing them with existing literature This chapter also includes descriptive statistics for the variables under investigation, offering insights into the overall trends and patterns observed in the research.
- Chapter 5: Conclusion and policy implications This chapter concludes the research finding, provides implication, further suggestion as well as the research limitations
LITERATURE REVIEW
FOOD SAFETY AND FOOD-BORNE DISEASES
Food-borne diseases (FBD) are illnesses transmitted through ingested food, caused by enteric pathogens, parasites, chemical contaminants, and biotoxins (WHO, 2007) There are two primary methods for estimating the burden of food-borne diseases: the etiologic agent approach, which begins with exposure levels of food-related agents, and the syndromic approach, which starts with disease outcomes like gastroenteritis A thorough assessment of the disease burden necessitates the integration of both methodologies.
In 2015, the WHO reported approximately 600 million cases of food-borne illnesses and 420,000 related deaths, with 40% of the burden affecting children under five The data for this thesis was gathered through participant interviews using questionnaires, without any food testing Consequently, the diagnosis of food-borne diseases relied on the respondents' self-assessments, alongside evaluations by medical professionals based on reported symptoms.
Food safety is defined by the Vietnam Ministry of Health (2010) as the assurance that food will not harm human health or life This definition, while less specific than that of the WHO, encompasses the entire process of food production, including growing, harvesting, preserving, and processing, rather than just preparation and consumption Given its comprehensive scope and the increasing awareness of food safety regulations among the Vietnamese population, this definition serves as the foundation for understanding food safety in this thesis.
THE HEALTH BELIEF MODEL
The Health Belief Model (HBM), rooted in psychological theory, evaluates how individuals make decisions under uncertainty by assessing the "value-expectancy" of potential outcomes (Maiman and Becker, 1974) When applied to health, the model suggests that individuals carefully consider the benefits of illness prevention and health improvement, leading them to expect that their actions will help avert disease and enhance their well-being This expectation is influenced by their perceptions of susceptibility to illness, the seriousness of potential diseases, and the likelihood that their behaviors may lead to health issues Key components of the HBM are outlined by Glanz et al (2008, pp 47-48).
Table 2.1: The concepts of Health Belief Model
Perceived susceptibility Belief about the chances of experiencing a risk or getting a condition or disease
Define population(s) at risk, risk levels
Personalize risk based on a person’s characteristics or behavior
Make perceived susceptibility more consistent with individual’s actual risk
Perceived severity Belief about how serious a condition and its sequelae are
Specify consequences of risks and conditions
Perceived benefits Belief in efficacy of the advised action to reduce risk or seriousness of impact
Define action to take: how, where, when; clarify the positive effects to be expected
Perceived barriers Belief about the tangible and psychological costs of the advised action
Identify and reduce perceived barriers through reassurance, correction of misinformation, incentives, assistance
Cues to action Strategies to activate
Provide how-to information, promote awareness, use appropriate reminder systems
Self-efficacy Confidence in one’s ability to take action
Provide training and guidance in performing recommended action Use progressive goal setting
Give verbal reinforcement Demonstrate desired behaviors
The various components interact with individual characteristics to form three key groups of factors: modifying factors, individual beliefs, and actions The relationships and impacts of these groups are illustrated in the accompanying figure.
Figure 2.1: Health Belief Model Components and Linkages (Glanz et al, 2008)
Numerous studies have utilized the Health Belief Model (HBM) to examine food safety behaviors across various groups, including restaurant workers (Cho et al., 2010), primary food preparers in families with young children (Lum, 2013; Meysenburg et al., 2013), and older adults (Hanson and Benedict, 2002) These studies demonstrate that the components of HBM interact with one another and significantly influence individual behaviors, particularly highlighting that food safety knowledge plays a crucial role in shaping perceptions of food safety.
EMPIRICAL REVIEWS ON DRIVERS OF FOOD SAFETY PRACTICES
Participants with a college degree or higher demonstrated superior food safety knowledge and behavior compared to those with lower educational attainment (Meysenburg et al., 2013) The research team employed the Health Belief Model alongside mixed method analysis, examining a sample of 72 individuals through scripted interviews and group discussions.
Modifying factors Individual Beliefs Action
Age Gender Ethnicity Personality Socioeconomics Knowledge
Perceived susceptibility to and severity of disease
Unusan (2005) found that higher education levels positively influence confidence in food safety practices, leading to reduced risk behaviors compared to those with lower education However, the study revealed no correlation between socio-economic status and individual food safety practices Data was collected from Turkish households and analyzed using MANOVAs.
Unusan's research highlights the influence of gender and education level on food safety knowledge, noting that women, who are often the primary food preparers in households, tend to be more attentive to food safety information, especially those with higher education Supporting this, Byrd-Bredbenner et al (2007) and Mullan et al (2014) found similar results in their reviews, indicating that age also plays a role, with older individuals generally scoring higher in food knowledge Furthermore, the study by Jevsnik et al (2006) suggests that women are more likely to take responsibility for food safety compared to men These findings were derived from participant analysis using ANOVA.
A study conducted by Langiano et al (2012) revealed that married individuals exhibited healthier food behaviors compared to their single counterparts Additionally, the research found that the accuracy of food practices improved with the number of family members, indicating that the primary cook tended to adopt more precise cooking habits as family size increased.
Food preparers primarily acquire their knowledge about food processes from family members and relatives (Meysenburg et al., 2013) Research has shown that family serves as a significant resource for food safety knowledge, influencing individual behaviors (Kwon et al., 2008; Trepka et al., 2006) Kwon's study, which surveyed participants of the Special Supplemental Nutrition Program for Women, Infants, and Children, utilized a questionnaire to assess food knowledge and behavior, analyzing the results through ANOVA The findings indicated that respondents with strong food knowledge tend to engage in more appropriate food practices.
(T H Vo et al, 2015) These authors group investigated in the canteens’ and restaurant’s workers by using logistic regression model to get this finding
Research by Cho et al (2010) indicates that participants' food safety knowledge does not significantly affect their food practices This study specifically examined restaurant workers using multiple regression and maximum likelihood estimation Additionally, Roberts et al (2008) found that food workers exhibit only minimal changes in behavior despite receiving training and education in food safety These findings support the notion that knowledge has a limited influence on individual behavior change.
Individuals with high self-efficacy believe they can effectively prevent health threats and foodborne diseases (FBD) when they are in control of food handling (Meysenburg et al., 2013) However, this confidence diminishes when food is prepared by others Furthermore, those who have experienced foodborne illness or caused illness in family members due to improper food handling tend to have lower confidence in their food preparation abilities.
Research on food safety perception and behavior presents mixed findings Nesbitt et al (2013) revealed that many consumers believe food contamination occurs before it reaches their kitchens, with most individuals who experienced foodborne disease (FBD) attributing their illness to food prepared outside the home Conversely, Unusan's study (2007) indicated that consumers often do not recognize food poisoning or FBD as a significant health concern, viewing it instead as a normal occurrence This misleading perception results in a lack of motivation to modify food safety behaviors or prioritize food safety issues.
Jevsnik et al (2006) discovered that household cooks preferred products from farmers over those from industrial factories, believing that farmer-produced food is safer Despite this preference, consumers felt that food safety was the responsibility of food handlers—such as farmers, food factories, retailers, and caterers—as well as the government Additionally, research by Byrd-Bredbenner et al (2007) revealed that individuals under 30 often overestimated their ability to handle food safely, despite evidence to the contrary.
A study by T H Vo et al (2015) established a link between food safety knowledge and individual attitudes towards food safety issues, although the relationship between attitude and food practices was found to be insignificant Similarly, Cho et al (2010) indicated that individuals with strong food safety knowledge were more likely to recognize the severity and likelihood of food poisoning Additionally, consumers with accurate food knowledge faced fewer challenges in practicing safe food handling However, the research did not demonstrate a direct influence of knowledge on perceptions of foodborne disease prevention or safe food practices Instead, it revealed that individuals who understood the benefits of avoiding food poisoning were more likely to engage in safe food behaviors consistently.
Hanson and Benedict (2002) found that a strong awareness of foodborne disease (FBD) severity can enhance individual behaviors, although the link between the perception of FBD hazards and food safety practices is weak Their findings were derived using nonparametric statistics, specifically Spearman rank correlation coefficients.
Research by Cho et al (2010) highlights a significant link between cues and individual food safety practices Individuals with a history of foodborne diseases (FBD) are more inclined to adopt safe food handling practices, as noted by Lum (2010) Nonetheless, Lum also points out that experiencing illness symptoms does not consistently result in improved food safety behaviors.
A similar result from Hanson and Benedict (2002) showed that the cue, content of
The study revealed that education has a lesser impact on males compared to females, while older individuals experience a stronger influence from educational attainment Additionally, the effect of education on individuals varies based on their frequency of food handling.
A study by Byrd-Bredbenner et al (2013) found that food labels conveying messages about risky food practices positively influenced individual behaviors The research revealed that consumers across different age groups were concerned about food safety knowledge, but their susceptibility and concern were heightened when information was tailored specifically to their demographic.
Mullan et al (2014) highlighted that previous habits significantly influence current behaviors These habits develop through the consistent repetition of actions in specific contexts or as responses to certain cues A lack of prompts or reminders may lead individuals to neglect food safety practices in their homes.
RESEARCH METHODOLOGY
ANALYTIC FRAMEWORK
The study utilized the Health Belief Model (HBM) framework to assess food safety behaviors, considering various individual and demographic factors of participants and their families It specifically focused on knowledge related to food safety issues, while also measuring individual beliefs through attitudes and awareness of respondents Furthermore, the evaluation of individual actions was based on a variety of food safety practices, with information sources serving as cues for these behaviors.
Due to the limitation of the secondary data, the components from HBM measured in several variables:
- Modifying factors: age, gender, residential location, occupation, education level, number of family’s member and the knowledge about food safety issue
- Individual belief: the awareness about the food safety problem, risky group and reason of food poisoning; attitude about food selection and processing; food source chosen
- Individual behavior: hygiene, process, preserve practice
- Cues to action: the food safety information source The relatives and interactions of those components illustrated in the figure below:
ECONOMETRIC MODELS
This study employs the multivariate probit model (MVP) to analyze the impact of independent variables on distinct behavior groups related to food safety The research focuses on three key aspects of food safety behavior: kitchen hygiene practices, food processing and preservation methods, and individual hygiene practices By utilizing the MVP with three equations, the paper aims to provide comprehensive insights into these behaviors.
- Food expenditure, Number of family’s member
+ TV, newspaper + Local food safety communicator
Perceived susceptibility to and severity of disease
- Attention about food safety problem
Individual beliefs can predict dependent variables, as demonstrated by the trivariate probit model proposed by Cappellari and Jenkins (2003) In this model, the latent variable \( y_{im}^* \) is defined as \( y_{im}^* = \beta_m X_{im} + \epsilon_{im} \) for \( m = 1, 2, 3 \), where \( y_{im} = 1 \) if \( y_{im}^* > 0 \) and \( y_{im} = 0 \) otherwise The error terms \( \epsilon_{im} \) for \( m = 1, \ldots, 3 \) follow a multivariate normal distribution with a mean of zero and a variance-covariance matrix \( V \), which has a leading diagonal of ones and correlations \( \rho_{jk} = \rho_{kj} \) as off-diagonal elements.
The log-likelihood function for the sample of N independent observation is given by:
Where ωi is an optional weight for observation i=1,…, N, and ϕ3 is the trivariate standard normal distribution with arguments ài and Ω, where
𝜇 𝑖 = (𝐾 𝑖1 𝛽 1 ′ 𝑋 𝑖1 , 𝐾 𝑖2 𝛽 2 ′ 𝑋 𝑖2 , 𝐾 𝑖3 𝛽 3 ′ 𝑋 𝑖3 ) With Kik=2yik – 1, for each I, k = 1,…,3 Matrix Ω has constituent elements Ωjk, where: Ωij = 2 for j =1,…,3 Ω21 = Ω12 = Ki1Ki2ρ21 Ω31 = Ω13 = Ki3Ki1ρ31 Ω32 = Ω23 = Ki3Ki2ρ32
The probability of every outcome is given by:
The study focuses on nine food safety practices categorized into three behavior groups: hygiene kitchen practices (kprac), process and preserve practices (pprac), and hygiene individual practices (iprac) Each practice is evaluated using a binary system, where a value of 1 indicates correct behavior and 0 indicates incorrect behavior Correct practice is defined as all behaviors within a group being precise The behaviors are based on "The 10 golden principles in food processing," with specific practices like maintaining a clean and tidy kitchen and properly separating cooked and raw foods further divided into four distinct behaviors for easier evaluation by the interviewer.
- “sex” is the dummy variable indicate the sexuality of participant, 0 for male and
1 for female The expected regression coefficient of this variable is predicted insignificant, due to the fact that most of the responders were females;
The variable "loc" serves as a dummy indicator for the responder's location, with a value of 0 representing suburban participants from 12 districts, including Binh Tan, Binh Chanh, Thu Duc, Go Vap, Districts 9, 12, 6, 8, Nha Be, Can Gio, Cu Chi, and Hoc Mon Conversely, a value of 1 denotes urban participants from another set of 12 districts.
The expected regression coefficient indicates that urban participants in areas such as Phu Nhuan, Tan Binh, Tan Phu, and Binh Thanh have a higher probability of accurate behavior compared to their suburban counterparts.
In this study, the variables "age," "exp," and "f_member" represent age in years, food expenditure in hundred thousand VND, and the number of family members, respectively The expected regression coefficients for these variables are anticipated to be positive To achieve a normal distribution, the thesis employs the natural logarithm of age (lnage) and expenditure (lnexp) for estimating the coefficients.
The article discusses the categorization of respondents based on their job types, such as office clerks, retirees, householders, physical laborers, and farmers, as well as their education levels, including primary, junior high, high school, college, and below primary To simplify analysis and interpretation, the study transforms the education variable into years of schooling and consolidates job categories into three groups: householders, common laborers, and others.
The "know" variable assesses participants' food safety knowledge, derived from a questionnaire based on WHO guidelines This knowledge is categorized into two groups: safe food selection and food processing and preservation Participants' scores, reflecting their understanding, are calculated using the difficulty index method as outlined by Collen (2006, pp 98-100).
𝑁 , where: ρ: difficulty index nc: the number of right answer
N: the total number of responders
The study utilized a questionnaire to assess participants' perceptions of food safety issues, focusing on four key areas: susceptibility to and severity of foodborne diseases (FBD), benefits, barriers, and self-efficacy Although the data was limited, Factor Analysis was employed to identify perception factors from three of the four groups, excluding the perception of benefits.
The study categorizes "cue" as the variable representing the information sources regarding food safety utilized by respondents, including TV, radio, newspapers, local medical staff, and food documentaries To facilitate analysis, the thesis employs dummy variables for these cues, namely TV, radio, news, local_staff, and food_doc A comprehensive description of all variables incorporated in the model is provided in the accompanying table.
Sex Gender of participant 0 for male, 1 for female
Location Residential place 0 for suburb, 1 for urban
Age The age of participant Years old
Expenditure Amount of money for food consumption Hundred thousand VND Family member Number of family member Person
Education Education level The number of schooling years
Job Occupation of participant Category variable: common labor, householder, other
Understanding food safety is crucial, as it encompasses knowledge about potential risks and practices to ensure safe consumption Awareness of food safety issues can be effectively measured using a 3-point Likert scale, which gauges individuals' perceptions Additionally, various sources of food safety information, such as television, radio, newspapers, local medical professionals, and food documentaries, serve as important cues to action, helping to educate the public and promote safer food handling practices.
Multicollinearity occurs when explanatory variables in a regression model exhibit a linear relationship It can be categorized into two types: perfect multicollinearity and imperfect multicollinearity In cases of perfect multicollinearity, the regression coefficients of the dependent variables become indeterminate, leading to infinite standard errors Conversely, with imperfect multicollinearity, although the regression coefficients are determinate, they have large standard errors relative to the coefficients, resulting in estimates that lack precision and accuracy (Gujarati, 2004).
The studies by Cho et al (2010) and T H Vo et al (2015) explored the relationship between knowledge and perceptions of food safety, utilizing a Likert scale to measure perception, which indicates trends rather than the extent of impact on behavior With over 1,000 observations, the effect of multicollinearity in the regression model was minimized Additionally, the research employed both the reduced and original forms of the MVP to estimate the regression coefficients effectively.
3.2.3 Propensity Score Matching (PSM) Method
Chow and Mullan (2009) emphasize that past behavior is a crucial predictor of food safety practices, suggesting that cues can help individuals adopt food safety habits to modify primary cooking behaviors Additionally, Jevsnik et al (2007) highlight that consumers often lack awareness of their role in the food safety chain, which is further supported by Byrd-Bredbenner et al in the context of preventing foodborne diseases (FBD).
2007) Due to these reason, the consumers could hardly change their behavior in a short time period after suffering food poisoning
Individuals demonstrate rationality when they understand the cause-and-effect relationship between proper behavior and health benefits (Mari et al., 2008) However, it can be challenging for household cooks to identify practices that contribute to foodborne disease (FBD) Additionally, the data on food poisoning only encompassed health statements for a two-week period, leading this paper to assume that the food poisoning incidents did not influence individual behavior Consequently, individual behavior is utilized to estimate the probability of food poisoning using the Propensity Score Matching (PSM) method.
DATA
The thesis utilized data from a 2013 investigation into individual food poisoning cases in Ho Chi Minh City, along with findings from the Knowledge, Attitude, and Practice (KAP) survey on food safety conducted among households in the same city.
- Data source: Safety Hygiene Food Branch of Ho Chi Minh city
- Data description: the data had two parts:
+ Part 1 is the investigation in individual food poisoning and the relevant symptoms, included: individual characteristic, demographic information, food poisoning statement and clinical symptoms
+ Part 2 is the KAP survey of primary cook of household, included: individual characteristic, food safety knowledge, food safety attitude and examining the food safety practicing of primary cook
- Size of samples: 1,174 households and 4,593 individual participate in the survey The households were chosen by Probability Proportional to Size sampling technique (PPS):
In a study conducted across 319 wards in Ho Chi Minh City, researchers randomly selected 30 wards for a comprehensive survey Each ward's investigation began with a random selection of the first household, followed by interviews with the next 39 households on the right side The survey included four key questionnaires: an acute food poisoning investigation, a food knowledge assessment, an attitude survey, and an evaluation of food practices While the food practice checklist was evaluated by the surveyor, the other questionnaires were completed by the respondents Local medical staff responsible for food safety in each selected ward conducted the research.
To participate in the survey, households were required to meet specific criteria, including having all members reside at the same location for a minimum of six months prior to the investigation Additionally, households needed to consent to participate, with members sharing the same address, having at least one meal together, and engaging in similar household responsibilities If a household was unapproachable after three attempts, it would be replaced with another household.
Participants in the survey were individuals without mental illness, deafness, or speech impairment Children included in the survey were at least 6 months old, as infants primarily consume breast milk Responses from children under 10 years old were verified by their mothers or primary caregivers.
Food poisoning can be diagnosed through various symptoms that manifest after a meal, including gastrointestinal issues such as colic, vomiting, and diarrhea, as well as neurological symptoms like a stiffened tongue, hallucinations, diminished vision, delirium, and convulsions It is crucial to note that food poisoning is considered valid only if the affected individuals consumed a meal prepared at home prior to the onset of symptoms A household is deemed to have experienced food poisoning when either the primary cook or any family member exhibits these symptoms following a home-cooked meal.
The thesis integrates two data components to assess the relationship between knowledge and perception of food safety and individual behaviors It also estimates how these food safety behaviors influence the likelihood of experiencing food poisoning.
RESEARCH RESULTS
FOOD SAFETY PROBLEMS IN VIETNAM
According to the World Health Organization (2016), food-borne diseases in Vietnam impose an estimated annual economic burden of approximately 1 billion USD, which accounts for 2% of the country's GDP This figure encompasses costs associated with lost work time, decreased productivity due to illness, and related market losses Additionally, statistics from the Vietnam Food Administration (VFA) indicate that between 2007 and 2015, there were 150 reported cases of food-borne illnesses.
Vietnam experiences approximately 250 mass food poisoning outbreaks annually, affecting over 5,000 individuals each year Despite the implementation of the National Strategy on Food Safety since 2006 and the validation of the strategy for 2011-2020, the incidence of food poisoning cases remains consistently around 5,000 per year.
The stagnation in food safety regulation can be attributed to the inefficiency of legacy institutions Although the Food Safety Law was enacted in 2011, accompanying regulatory documents are insufficient to address current needs, leaving the Food Safety Department ill-equipped for effective administration This gap results in manufacturers and consumers lacking essential information on food safety practices Additionally, the complex Food Safety Administration network, which involves multiple ministries and departments, complicates enforcement The tropical climate and climate change further exacerbate the risk of foodborne illnesses, while the diversity of Vietnamese cuisine combined with limited consumer knowledge increases the likelihood of foodborne diseases Despite a stable rate of food poisoning cases, fatalities have only gradually decreased, with a significant portion attributed to natural toxins (Nguyen, 2016).
Figure 4.2: The number of food poisoning outbreaks and death in Vietnam
Between 2012 and 2016, Ho Chi Minh City experienced a decline in food poisoning incidents, recording a total of 20 cases without any fatalities Of these incidents, 19 were attributed to bacterial sources, while one case had an unidentified cause.
Figure 4.3: The number of food poisoning cases in HCM city (FSBDH, 2016)
Food poisoning cases in Vietnam, particularly in Ho Chi Minh City, are underreported due to insufficient attention from the government and researchers, leading to a lack of educational programs on safe food preparation and cooking practices Most government resources are focused on managing manufacturers and merchandising rather than addressing individual incidents A 2013 survey revealed that the individual food poisoning rate stands at 2.18%, highlighting a significant risk of food-borne diseases within the population.
4.1.2 Problems with household’s cooking behavior
The Vietnamese government's insufficient focus on household food safety has led to poor practices among primary cooks A 2010 investigation by the Safety Hygiene Food Branch of Ho Chi Minh City revealed concerning statistics regarding food safety awareness and practices in households.
54.3% of consumers in Ho Chi Minh city behave accurately in cooking The figures for Lao Cai province and Dong Thap Province are 67.7% and 76%, respectively (Nguyen,
Researchers emphasize the need for targeted educational programs for household cooks to enhance their food safety practices; however, most communication and education efforts are directed towards food producers and workers While the risk of food poisoning at home is perceived as less critical than in schools and factories, the reality is that urban consumers often eat out due to time constraints, frequenting company canteens and food shops As a result, the significance of cooking behaviors in households is often overlooked, leading to minimal governmental initiatives aimed at improving these practices.
DESCRIPTIVE STATISTICS
The data includes 1,174 households primary cooks with the characteristics listed in Table 4.1 and Table 4.2
Table 4.1: Demographic characteristics of participants (category variables) Demographic characteristics Number of participants (n) %
Data indicates that 93.02% of primary cooks in households are female, while 69.78% of respondents identify as householders Other occupations, aside from common labor and householders, are minimal Consequently, the regression analysis categorizes occupations into three distinct values: householder, common labor, and others.
The education levels of participants primarily ranged from junior to high school, with a nearly equal distribution between urban and suburban respondents (53.53% vs 46.47%) Television emerged as the leading source of information for households at 87.31%, followed by newspapers, while only 24.19% of participants relied on local medical staff for food safety information Additionally, the household food poisoning rate was reported at 5.11%, compared to 2.18% for individuals, highlighting that some households experienced multiple cases of food poisoning.
In contrast, the descriptive statistic of continuous variables is show on the table below:
Table 4.2: Demographic characteristics of participants (continuous variables)
Variables Mean SD Min Max
The average age of the primary cook in Vietnamese households is 47 years, indicating that food preparation is predominantly the responsibility of middle-aged women, in line with traditional family roles Households in Ho Chi Minh City average just over four members, which aligns with government population policies On average, these households spend nearly 100,000 VND daily on food, and the primary cooks have a mean knowledge score of 9.46 out of 14.04, reflecting a basic understanding of food safety However, the knowledge assessment from the 2010 survey was based on the percentage of correct answers, making it impossible to directly compare results across different periods.
The Health Belief Model identifies five key latent variables: perceived susceptibility, perceived severity, perceived benefits, perceived barriers, and perceived self-efficacy However, this thesis evaluated only four of these components due to data limitations, focusing on perceived susceptibility (3 questions), perceived benefits (3 questions), perceived barriers (1 question), and perceived self-efficacy (11 questions) The correlation coefficients, ranging from -0.0004 to 0.8587, indicate a relationship among these items Additionally, the Kaiser-Meyer-Olkin (KMO) value of 0.947 suggests a strong correlation, while the determinant of the correlation coefficients matrix is not equal to 0 (p-value < 0.01), and Barrett’s test confirms similar results (p-value < 0.01) These findings suggest that the data is well-suited for factor analysis.
The Cronbach alpha’s value of all components and the factor analysis result (after rotation) were showed in the table below
- Attention about food safety problem
- Hygiene in food processing Separate cooking tool
Hygiene hand before touching food Hygiene hand after touching food Use clean water
- Hygiene in food preserving Wash ingredient
Eating food right after processing Heating food after 2 hours
Among 18 factors (Appendix 2), there are three factors which Eigen value were more than one, accounted for a cumulative 66% of variance Therefore, the thesis showed all 3 factors and using only factor 1 (accounted for 50% of variance) as the indicator of perception value for the multivariate regression In summarize, the perception value varied from -3.01 to 2.36 with the average at 1 This figure implied the majority of participant perceived the threat of food safety issue
A recent survey revealed that individual hygiene practices ranked highest among three groups, with an accuracy rate of 81.48% Notably, hand washing was the most accurately practiced behavior, achieving a remarkable 97.77% among respondents Conversely, the processing and preserving behaviors were the least accurately practiced, with only 78.34% compliance Additionally, the practice of not smoking, spitting, or polishing nails while cooking was reported at a rate of 81.68%.
A significant portion of respondents demonstrated adherence to food safety practices; however, the percentage of individuals exhibiting complete compliance is notably lower within each group This finding reflects a slight increase from the 2010 survey, which reported a compliance rate of 54.3%, compared to the current figure of 61.87% Despite the 2010 evaluation not being conducted by medical professionals, the observed improvement in personal food safety practices holds considerable importance The detailed results are presented in the table below.
Table 4.4: Food safety practices Behavior (n68) Responder with right behavior
- Separate cooking tool for cooked and uncooked material
- Separate in preserving with enough facilities
- Use clean tool to prepare, divide food
- Not use forbidden food additives or out of date food
- Wash hand before cooking or after toileting
- Not smoke, spit out, nail polished while cooking
Figure 4.4: The nonparametric relationship between food safety practice and
Foo d s af et y K no w dle dg e v alu e
Foo d s af et y K no w dle dg e v alu e
Foo d s af et y K no w dle dg e v alu e
Figure 4.4 demonstrates the relationship between participants' practice evaluations, their perceived value, and knowledge scores The data indicates that respondents exhibiting appropriate behaviors in each practice group showed higher levels of perceived value and food safety knowledge compared to others, despite the explanatory variables falling within a similar range The subsequent regression analysis will further elucidate these relationships.
RESULTS FROM MULTIVARIATE PROBIT MODELS
The Health Belief Model suggests that behaviors are significantly influenced by perceived value and cues to action, while modifying factors like individual characteristics and food safety knowledge have an indirect effect This thesis employs both a reduced multivariate probit model, focusing on perception value and cues to action as independent variables, and the original multivariate probit model, which includes all relevant independent variables.
Table 4.5: MVP regression reduced form Variable Coefficient p-value Coefficient p-value Coefficient p-value
The table presents the coefficients, standard errors, and p-values from the reduced form of the multivariate probit analysis All estimated coefficients for the three pairs show positive and significant relationships, with values of 0.57 for Hygiene kitchen practice and Process/preserve practice, 0.46 for Hygiene individual practice and Hygiene kitchen practice, and 0.48 for Hygiene kitchen practice and Process/preserve practice These positive correlation coefficients indicate that a primary food preparer's likelihood of engaging in a specific set of behaviors increases when they exhibit proficiency in at least one of the other two behavior groups.
The correlation coefficients for perception value demonstrate a significant positive relationship with all three behaviors: hygiene kitchen practice (0.191), process/preserve practice (0.561), and hygiene individual practice (0.316) This indicates that perception positively influences food practice behavior.
Radio news significantly influences kitchen hygiene practices and food preservation, with positive correlations of 0.367 and 0.511, respectively, indicating that listeners tend to maintain cleaner kitchens and better food preservation habits Conversely, food documentaries and advice from local medical staff negatively impact food preparation practices, leading to poorer hygiene and preservation outcomes Regression analysis reveals that participants exposed to food documentaries or local medical staff information are likely to exhibit worse hygiene practices and food preservation behaviors Other factors assessed showed no significant effects.
Table 4.6: MVP regression original form
Variable Coefficient p-value Coefficient p-value Coefficient p-value
Hygiene individual practice Food safety knowledge
- Local staff (*) -0.160 0.194 -0.611 0.000 -0.378 0.002 ρkp ρik ρip
The regression analysis revealed consistent relationships among the variables in their reduced form, indicating that knowledge of food safety significantly influences participants' behaviors across all three groups Additionally, living in urban areas negatively impacts food processing and preservation practices (p-value < 0.05), suggesting that urban residents tend to engage in less precise food handling compared to those in suburban areas Furthermore, the number of household members correlates positively with kitchen hygiene practices, indicating that larger families tend to maintain cleaner kitchens However, the other variables examined did not show statistically significant effects The table below presents the marginal effects from the multivariate probit regression analysis.
Table 4.7: Marginal effect after MVP regression
Variable ME p-value ME p-value ME p-value
Hygiene individual practice Food safety knowledge
- Common labor (*) 0.017 0.722 0.015 0.722 0.016 0.722 Logarithm of Age 0.038 0.400 0.035 0.400 0.035 0.400 Logarithm of Food expenditure
Notes: (*) dummy variables, ME: marginal effect
The marginal effect of each variable highlights the specific impact of various factors on the independent variable For example, an increase of 1 point in a participant's knowledge score correlates with a 1.9% rise in the likelihood of practicing proper food safety in kitchen hygiene, and a 1.7% increase for both process/preservation and individual practices This trend is also observed with the number of family members Additionally, improved perceptions significantly enhance the likelihood of accurate practices across all three groups Notably, individuals who obtain food safety information from the radio demonstrate better behaviors, with a 10% improvement in kitchen hygiene practices and a 9.1% improvement in both process/preservation and individual practices compared to those who do not.
Local medical staff advice can reduce the likelihood of proper individual practices by 37.8% Additionally, food safety information from documentaries negatively impacts hygiene kitchen practices by 9.9% and food processing/preservation practices by 8.9%.
To assess the robustness of the regression findings, the thesis employed a Poisson model to accurately predict the number of positive behaviors practiced by participants, assuming uniform effects on food poisoning The independent variables mirrored those in the multivariate probit model, while the dependent variable represented the count of correct behaviors performed by participants, with a maximum of 12 The regression results are presented in the table below.
Variable Coefficient p-value ME p-value
The Poisson regression analysis, like the MVP regression, demonstrates a significant relationship between behavior and factors such as food safety knowledge, perceived value, location, and various cues to action, including radio, food documents, and local medical staff However, unlike the MVP regression, the Poisson regression indicates that the number of family members does not influence the precise behaviors of the primary cook in the household.
RESULTS FROM PROPENSITY SCORE MATCHING MODEL
A survey of 4,593 respondents revealed 98 suspected cases of food poisoning across 79 households, with 60 of these cases linked to symptoms arising after consuming home-prepared meals.
In the MVP model, all variables serve as explanatory factors in the probit regression, with the addition of the independent variable "food_place" to clarify the causes of foodborne disease (FBD) symptoms This variable is assigned a value of "0" for food purchased from regulated markets and "1" for food from unregulated spontaneous markets, with an anticipated negative correlation to food poisoning incidents The expenditure variable has been excluded to maintain the balancing property, while the dependent variable in the probit regression is defined as whether a household experienced food poisoning, coded as "1" for affected households and "0" for those that were not It is important to note that the coefficients in this regression differ significantly in interpretation compared to other regressions due to the substantial variance in the two values of the dependent variable.
(60 versus 1115) The table below shows the result of the probit model:
However, after estimating the propensity score and choosing the control group by radius matching (caliper is 0.0001), there are only 33 observations in treated group and
In a study involving 96 participants in a control group, the expenditure variable was found to unbalance the probit regression To assess its impact, the thesis included this variable in the analysis, which did not influence the results of the Propensity Score Matching (PSM) method The mean differences of continuous variables between individuals who experienced foodborne disease (FBD) and those who did not were evaluated using a t-test, yielding the results detailed below.
Table 4.10: Differences of continuous variables
(Not suffered FBD – Suffered FBD) p-value
The analysis reveals that at the 5% significance level, the only notable differences between responders who experienced poisoning and those who did not are related to their food expenditure and years of schooling Specifically, individuals who suffered from poisoning spent approximately 21,000 VND more on food and had more years of education Additionally, there is a distinction in the behavioral accuracy between the two groups, with non-poisoned individuals demonstrating more precise behaviors; however, this difference is only significant at the 10% level.
On the other hand, the correlations between the binary variables and the FBD variable tested by the Pearson’s Chi square test:
Table 4.11: Correlations between binary variables and FBD
Variable Pearson’s Chi square value p-value
The analysis revealed that the only significant factor related to the possibility of Foodborne Disease (FBD) is the living environment, with a p-value of less than 0.05 Other factors did not demonstrate a meaningful relationship, indicating minimal differences between individuals who experienced FBD and those who did not.
DISCUSSION AND IMPLIED POLICY
DISCUSSIONS AND CONCLUSIONS
This study highlights the significant influence of knowledge and perception on participants' food safety practices The findings, derived from both the multivariate probit and Poisson models, reinforce the conclusions of T H.
Vo et al (2015), Hanson and Benedict (2002) although it is opposed to Roberts et al
Research by Cho et al (2010) and others has produced varying conclusions due to differences in survey instruments used to assess knowledge and perception values Notably, T H Vo et al (2015) employed a survey aligned with the World Health Organization and the regulations of the Vietnam Ministry of Health, while alternative studies utilized the FightBAC!™ questionnaire developed by the U.S Department of Agriculture in 2010.
Radio shows significantly influence both the type and number of positive behaviors, while local medical staff negatively affects the adoption of precise practices This detrimental impact may stem from the insufficient quantity and quality of local personnel, many of whom lack expertise in food safety and are also burdened with additional responsibilities.
Many demographic characteristics, including occupation and gender, do not significantly influence individual behavior, likely due to the predominance of female householders among survey participants However, an increase in family members can create pressure on the primary cook, despite the potential for additional support from others Consequently, this dynamic can enhance the hygiene skills of the cook, leading to a cleaner kitchen environment.
The PSM results reveal significant differences between responders affected by foodborne diseases (FBD) and those who are not, particularly in terms of food expenditure and years of schooling These disparities may stem from the fact that householders, who make up a large portion of participants, typically have lower educational attainment Additionally, food knowledge is often passed down within families rather than taught in schools, suggesting that individuals who spend more time at home may acquire greater food knowledge Moreover, higher food prices do not necessarily ensure consumer safety; the PSM findings indicate that non-sufferers of FBD often invest time in selecting or growing their own food, sourcing it from relatives, which allows them to obtain food at a lower cost while minimizing the risk of FBD.
The living environment of responders is a crucial factor that may contribute to foodborne disease (FBD) While healthy individuals tend to practice better food safety compared to those who experience food poisoning, individual behaviors do not clearly indicate a direct impact on the likelihood of food poisoning Additionally, the results from the PSM analysis do not provide sufficient evidence to determine the influence of environmental or external factors The study also reveals that food practice behaviors and purchasing locations between the two groups show minimal differences, with only a slight distinction in specific behaviors Therefore, further evidence is needed to draw any definitive conclusions.
To address the shortcomings of medical staff, the government must enhance their knowledge and develop institutions dedicated solely to food safety Improving communication channels is essential, particularly by increasing the quantity and quality of food safety information, focusing on proper food processing rather than solely airing warnings on national TV Radio has proven to be an effective medium for disseminating food safety information, especially in mountainous provinces where radio signals are more accessible than television The increase in food safety knowledge from 2010 to 2013 indicates a growing public interest in this issue, which the government can leverage to promote food safety education and encourage community involvement in administrative roles, such as freelancers or communicators.
The study indicates that there is no direct correlation between consumer behaviors and the places where food is purchased concerning the risk of foodborne diseases (FBD) Therefore, it is crucial to ensure high standards for food sources and restaurant quality to mitigate hazards Additionally, educational programs in schools should incorporate practical life skills to enhance students' understanding of food safety, while public communication strategies should prioritize outreach to the general public rather than focusing solely on manufacturers or restaurant staff Moreover, food sold at supermarkets or organized markets may not always be of superior quality compared to that from spontaneous markets, despite higher costs; thus, the government should establish specific standards and develop quality inspection institutions for these venues Lastly, environmental factors, particularly water sources, require regular monitoring and improvement, as a significant portion of Ho Chi Minh City's population lacks access to clean water, posing a serious threat to individual health outcomes.
5.3 LIMITATION AND IMPLICATIONS FOR FURTHER RESEARCH
The FBD survey relies on participants' self-reports rather than medical doctors' assessments, which may lead to inaccuracies Additionally, the data is primarily based on medical perspectives and lacks sufficient economic variables, particularly regarding perceptions To better understand the factors influencing FBD probability and the impact of living environments and food resources, further research is essential.
Foodborne diseases (FBD) are linked not only to food poisoning but also to chronic illnesses like cancer, making it challenging to directly attribute pathogens to food sources Consequently, further research is essential to uncover more evidence in this area.
The thesis highlights that the living environment significantly influences foodborne diseases (FBD), necessitating further evidence to clarify this relationship In particular, the distinction between water sources in urban and suburban areas is crucial, especially regarding water used for food consumption.
This research faces challenges in measuring variables, as many were assessed using different methods compared to similar studies, potentially hindering effective comparison of results Furthermore, the perception value was estimated through a questionnaire that lacked sufficient questions and effective techniques to encourage respondents to reveal their true perceptions.
Appendix 1: The correlation matrix of perception’s factors a18new 0.1174 0.1153 0.0339 0.0901 0.1197 0.0665 0.1243 0.1093 0.1394 0.1417 0.1350 0.1319 0.1301 0.0782 -0.0087 0.1118 0.1200 1.0000 a17new 0.1610 0.2099 0.2406 0.1978 0.2501 0.1486 0.1548 0.1471 0.1862 0.2030 0.1775 0.2215 0.1583 0.1982 0.1153 0.5982 1.0000 a16new 0.1013 0.1286 0.1195 0.1334 0.1636 0.0604 0.0809 0.0925 0.1220 0.1296 0.1213 0.1917 0.1087 0.1246 0.1088 1.0000 a15new 0.0441 0.0741 0.1621 0.0148 0.0245 0.1149 0.0041 0.0378 0.0168 0.0130 0.0074 0.0338 -0.0004 0.0286 1.0000 a14new 0.5619 0.5268 0.3108 0.5817 0.5964 0.4942 0.5825 0.5709 0.6058 0.6225 0.5829 0.6567 0.6512 1.0000 a13new 0.6492 0.5873 0.3630 0.6400 0.6764 0.6181 0.7588 0.7078 0.7003 0.6969 0.7272 0.8068 1.0000 a12new 0.6458 0.6488 0.3511 0.7082 0.6909 0.6046 0.7204 0.6566 0.7490 0.7180 0.8083 1.0000 a11new 0.6593 0.6319 0.3071 0.7262 0.6331 0.6084 0.7228 0.6601 0.7992 0.7962 1.0000 a10new 0.6304 0.6304 0.2974 0.6964 0.6424 0.5571 0.6694 0.6450 0.8587 1.0000 a9new 0.6621 0.6442 0.3321 0.7376 0.6598 0.5705 0.6875 0.6632 1.0000 a8new 0.5867 0.5412 0.3792 0.6006 0.6047 0.5451 0.6806 1.0000 a7new 0.6580 0.6141 0.3697 0.6941 0.7063 0.6675 1.0000 a6new 0.6024 0.5311 0.4115 0.5939 0.6254 1.0000 a5new 0.6497 0.6180 0.3667 0.6981 1.0000 a4new 0.7168 0.6804 0.3928 1.0000 a3new 0.3972 0.3668 1.0000 a2new 0.7716 1.0000 a1new 1.0000 a1new a2new a3new a4new a5new a6new a7new a8new a9new a10new a11new a12new a13new a14new a15new a16new a17new a18new
Factor18 0.11946 0.0066 1.0000 Factor17 0.14848 0.02902 0.0082 0.9934 Factor16 0.18985 0.04138 0.0105 0.9851 Factor15 0.22501 0.03515 0.0125 0.9746 Factor14 0.24595 0.02094 0.0137 0.9621 Factor13 0.27260 0.02665 0.0151 0.9484 Factor12 0.33196 0.05936 0.0184 0.9333 Factor11 0.35120 0.01924 0.0195 0.9148 Factor10 0.38457 0.03337 0.0214 0.8953 Factor9 0.42785 0.04328 0.0238 0.8739 Factor8 0.49646 0.06861 0.0276 0.8502 Factor7 0.53475 0.03829 0.0297 0.8226 Factor6 0.63879 0.10405 0.0355 0.7929 Factor5 0.77367 0.13488 0.0430 0.7574 Factor4 0.92689 0.15322 0.0515 0.7144 Factor3 1.10618 0.17929 0.0615 0.6629 Factor2 1.61396 0.50777 0.0897 0.6015 Factor1 9.21238 7.59842 0.5118 0.5118 Factor Eigenvalue Difference Proportion Cumulative
Appendix 3: MVP regression (reduced form) chi2(3) = 233.196 Prob > chi2 = 0.0000 Likelihood ratio test of rho21 = rho31 = rho32 = 0: rho32 4790893 0463972 10.33 0.000 3831738 564783 rho31 4570518 0457674 9.99 0.000 3628716 5419888 rho21 5779966 0398654 14.50 0.000 4945877 650856 /atrho32 5218016 0602191 8.67 0.000 4037743 6398289 /atrho31 4935782 0578527 8.53 0.000 3801891 6069673 /atrho21 659449 0598651 11.02 0.000 5421155 7767824 _cons 1.115008 131358 8.49 0.000 8575512 1.372465 loc_staff -.2856251 1125314 -2.54 0.011 -.5061825 -.0650677 doc -.2218455 1445481 -1.53 0.125 -.5051546 0614635 news -.1736675 0932207 -1.86 0.062 -.3563769 0090418 radio 1360822 106534 1.28 0.201 -.0727206 344885 tv -.0526691 1323381 -0.40 0.691 -.3120469 2067088 perc_f1 3153075 0457522 6.89 0.000 2256348 4049802 indi
_cons 8597885 1280854 6.71 0.000 6087457 1.110831 loc_staff -.3910883 1129624 -3.46 0.001 -.6124905 -.1696861 doc 0380902 1532165 0.25 0.804 -.2622086 3383889 news -.1393973 0918987 -1.52 0.129 -.3195155 040721 radio 5054202 1102772 4.58 0.000 2892808 7215597 tv 01221 1287667 0.09 0.924 -.2401682 2645881 perc_f1 5595797 0538308 10.40 0.000 4540732 6650862 proc
_cons 8138421 1249069 6.52 0.000 5690291 1.058655 loc_staff -.041626 1137304 -0.37 0.714 -.2645335 1812815 doc -.3644403 1437517 -2.54 0.011 -.6461885 -.082692 news 0023601 0902982 0.03 0.979 -.1746211 1793413 radio 3546625 1059468 3.35 0.001 1470106 5623144 tv -.0520822 1268091 -0.41 0.681 -.3006235 1964591 perc_f1 1903142 0435442 4.37 0.000 1049691 2756593 kitc
Coef Std Err z P>|z| [95% Conf Interval]
Log likelihood = -1518.5305 Prob > chi2 = 0.0000 Wald chi2(18) = 178.29Multivariate probit (MSL, # draws = 5) Number of obs = 1168
Appendix 4: MVP regression (original form) rho21 6383753 0388324 16.44 0.000 5559593 7083616 /atrho32 5116009 0619345 8.26 0.000 3902115 6329902 /atrho31 4838027 060206 8.04 0.000 3658011 6018044 /atrho21 7554267 0655425 11.53 0.000 6269657 8838876 _cons -.043319 1.424135 -0.03 0.976 -2.834573 2.747935 loc_dum -.1004265 0978174 -1.03 0.305 -.292145 0912921 loc_staff -.3780484 1210943 -3.12 0.002 -.6153888 -.140708 doc -.2109011 1503647 -1.40 0.161 -.5056104 0838083 news -.2908288 1000615 -2.91 0.004 -.4869457 -.0947119 radio 1352087 1093668 1.24 0.216 -.0791462 3495636 tv -.119521 136583 -0.88 0.382 -.3872188 1481768 perc_f1 2590135 0495196 5.23 0.000 1619568 3560702 lnexp 0184813 1028084 0.18 0.857 -.1830195 2199822 lnage 0848241 1683457 0.50 0.614 -.2451273 4147756 job_new2 -.1420892 1721359 -0.83 0.409 -.4794695 195291 job_new1 -.1077489 1512042 -0.71 0.476 -.4041037 188606 edu_new 0176512 0135823 1.30 0.194 -.0089696 044272 sex 1051301 1846875 0.57 0.569 -.2568508 467111 f_mem 0219332 0236432 0.93 0.354 -.0244066 068273 know2 0663949 0219701 3.02 0.003 0233343 1094554 indi
_cons -1.336828 1.416824 -0.94 0.345 -4.113753 1.440096 loc_dum 7533018 1046908 7.20 0.000 5481116 9584921 loc_staff -.6107935 1252602 -4.88 0.000 -.856299 -.3652881 doc -.2434742 1594121 -1.53 0.127 -.5559162 0689678 news -.1172928 1002116 -1.17 0.242 -.3137039 0791183 radio 3673403 1156295 3.18 0.001 1407106 5939699 tv -.0984783 1385747 -0.71 0.477 -.3700797 1731231 perc_f1 4491142 0566682 7.93 0.000 3380466 5601818 lnexp -.0402706 1018324 -0.40 0.693 -.2398585 1593172 lnage 2859399 1696803 1.69 0.092 -.0466273 6185071 job_new2 -.1626987 1753452 -0.93 0.353 -.506369 1809716 job_new1 044128 1540247 0.29 0.774 -.2577549 3460109 edu_new 0365416 0138528 2.64 0.008 0093906 0636926 sex 0971573 1934886 0.50 0.616 -.2820734 476388 f_mem 0366151 0231059 1.58 0.113 -.0086717 0819019 know2 1033148 0223409 4.62 0.000 0595275 1471022 proc
_cons -.2533119 1.378816 -0.18 0.854 -2.955741 2.449118 loc_dum -.0398895 0942509 -0.42 0.672 -.2246178 1448388 loc_staff -.1597777 1228875 -1.30 0.194 -.4006328 0810774 doc -.3515957 1503444 -2.34 0.019 -.6462652 -.0569261 news -.0582825 0965137 -0.60 0.546 -.247446 130881 radio 3581531 1099266 3.26 0.001 142701 5736053 tv -.1535407 1309093 -1.17 0.241 -.4101183 1030368 perc_f1 1364628 046897 2.91 0.004 0445464 2283792 lnexp -.0294486 099663 -0.30 0.768 -.2247844 1658872 lnage 1361739 16177 0.84 0.400 -.1808894 4532373 job_new2 0605255 1698464 0.36 0.722 -.2723673 3934182 job_new1 -.1238851 1469004 -0.84 0.399 -.4118047 1640344 edu_new 0189987 0132713 1.43 0.152 -.0070125 04501 sex 021229 1788945 0.12 0.906 -.3293977 3718557 f_mem 0672752 0240019 2.80 0.005 0202324 114318 know2 0666814 0209181 3.19 0.001 0256827 1076801 kitc
Coef Std Err z P>|z| [95% Conf Interval]
Log likelihood = -1433.2637 Prob > chi2 = 0.0000 Wald chi2(45) = 273.03Multivariate probit (MSL, # draws = 5) Number of obs = 1147
_cons 2.258058 2828686 7.98 0.000 1.703646 2.81247 loc_staff -.0411455 0249034 -1.65 0.098 -.0899552 0076643 doc -.0341781 0305385 -1.12 0.263 -.0940326 0256763 news -.0023822 0203121 -0.12 0.907 -.0421931 0374287 radio 0403983 0212309 1.90 0.057 -.0012136 0820101 tv 0051544 0277787 0.19 0.853 -.0492909 0595996 perc_f1 0449872 009692 4.64 0.000 0259911 0639833 loc_dum 0327238 019411 1.69 0.092 -.005321 0707687 lnexp -.0051244 0205234 -0.25 0.803 -.0453495 0351006 lnage 015889 0337333 0.47 0.638 -.0502271 0820051 job_4 -.0224146 0341957 -0.66 0.512 -.089437 0446078 job_5 -.0173529 0296353 -0.59 0.558 -.0754371 0407313 edu 0097375 0097973 0.99 0.320 -.0094648 0289399 sex 0126925 0364342 0.35 0.728 -.0587172 0841023 f_mem 00365 0045554 0.80 0.423 -.0052784 0125783 know2 0098917 004481 2.21 0.027 0011092 0186742 count Coef Std Err z P>|z| [95% Conf Interval]
Log likelihood = -2627.7535 Pseudo R2 = 0.0100 Prob > chi2 = 0.0000
LR chi2(15) = 53.01Poisson regression Number of obs = 1143
BẢNG CÂU HỎI KIẾN THỨC VỀ VSATTP NGƯỜI DÂN
I ĐẶC ĐIỂM ĐỐI TƯỢNG KHẢO SÁT A1) Họ và tên người được phỏng vấn:
A2) Địa chỉ: A3) Giới 1 Nam 2 Nữ
A4) Tuổi người được phỏng vấn: ………
A5) Trình độ học vấn: Không Cấp 1 Cấp 2 Cấp 3 Đại học Khác
1 Cán bộ 1 [ ] Lao động phổ thông 4 [ ]
A7) Số người có trong hộ: …………
A8) Số tiền đi chợ trung bình 1 ngày : đ_/ người ăn (*) A9) Anh chị là người nấu ăn : Chính phụ trong gia đình
II KIẾN THỨC VỀ VSATTP: (4 Câu) B1) Anh chị có thường để ý đến vấn đề VSATTP không ? Có Không B2) Anh chị có được thông tin về VSATTP từ :
TV Đài PT Báo chí Sách vở CBYT Khác
B3) Theo anh chị tại sao bị ngộ độc thực phẩm?
- Thực phẩm nhiễm hóa chất
- Thực phẩm không vệ sinh , bị nhiễm vi sinh vật
B4) Theo anh chị, làm thế nào để phòng ngừa ngộ độc thực phẩm cho gia đình :
Mua những loại thực phẩm đã được chế biến an toàn
Rửa rau và thực phẩm kỹ
Nấu nướng thức ăn kỹ
Tránh đụng chạm giữa thực phẩm sống và chín
3 Sử dụng/Ăn uống sau khi nấu: Ăn ngay thức ăn vừa được nấu chín
Hâm nóng thức ăn trước khi ăn
4 Ngoài việc giữ vệ sinh thực phẩm, anh chị còn chú ý giữ vệ sinh cho những việc gì khác nữa:
Giữ vệ sinh nhà bếp
Bảo quản kỹ thức ăn đã nấu
Không để thực phẩm bị côn trùng , súc vật gặm nhấm
III LỰA CHỌN THỰC PHẨM (7 Câu) C1) Anh chị thường đi chợ nào ? vì sao ?
Thường xuyên Thỉnh thoảng Giá mắc Giá rẻ ATTP Tiện lợi (ghi rõ)
C2) Khi lựa chọn thực phẩm tươi sống, Anh chị dựa vào tiêu chuẩn nào là chính :
Màu Mùi Độ chắc Mắt Mang Da Mua người quen
C3) Khi lựa chọn rau quả tươi sống , Anh chị dựa vào tiêu chuẩn nào ?
Toàn vẹn (không bị trầy xướt , dập nát , gọt vỏ , xắt mỏng, )
C4) Khi chọn mua thực phẩm bao gói sẵn, đồ hộp, anh chị có đọc nhãn không? Có
C5) Nếu có , Anh chị thường xem nội dung gì trên nhãn ?
- Tên hàng hóa Tên cơ sở sản xuất
- Thành phần cấu tạo của sản phẩm Ngày sản xuất và hạn sử dụng
- Hướng dẫn bảo quản sử dụng Khối lượng
C6) Ngoài việc đọc nội dung nhãn , Anh chị còn để ý điều gì ?
- Bao bì còn nguyên vẹn, không bể, Hộp kim loại không bị phồng nắp, gĩ sét
- Nắp chai kín còn niêm phong, Khác
C7) Anh chị có sử dụng các loại phụ gia sau trong chế biến thức ăn cho gia đình
Có Không 1 lần/tuần Thỉnh thoảng
Bột nổi (làm bánh bông lan , làm mềm thịt )
Bột nổi nâu ( làm bánh mì, )
IV CHẾ BIẾN, SỬ DỤNG VÀ BẢO QUẢN THỨC ĂN (17 Câu):
E1) Mặt bếp nhà anh chị được xây dựng như thế nào và bếp sử dụng là loại gì?
- Gạch men Xi măng Gỗ Đất Bếp củi
- Bếp điện Bếp ga Bếp dầu Bếp than Khác
E2) Anh chị thường vệ sinh nhà bếp (mặt, vách bếp, bếp nấu) khi nào?
- Sau mỗi bữa nấu xong Cuối ngày/ lần
E3) Anh chị có mang tạp dề, găng tay khi nấu nướng không?
E4) Khi chế biến thức ăn, anh chị thường rửa tay lúc nào và rửa bằng gì?
Rửa nước sạch Rửa nước sạch với xà phòng
- Sau tiếp xúc với thực phẩm sống
E5) Nhà Anh chị có bao nhiêu cái thớt ? …… cái Có phân biệt sống, chín (*) Không E6) Anh chị có dùng khăn lau chén không? Có Không
