Custom services
English (pdf)
Article in xml format
Article references
Automatic translation
Send this article by e-mail
Cited by SciELO 
Cited by Google 
Similars in
SciELO 
Similars in Google 
Permalink
Introduction
A substantial body of scientific evidence indicates that numerous synthetic environmental chemicals may increase the risk of some human diseases, and that for the general population food is often the main source of exposure to such compounds. Water and air pollution, personal care products, food packaging, and other consumer goods are also sources of exposure. Some types of persistent toxic substances were never intentionally released into the environment (e.g., dioxins), and the manufacture of others was halted decades ago in many countries (e.g., polychlorinated biphenyls, organochlorine pesticides). However, today many such persistent substances remain detectable in animal and human food webs, as well as in most human beings, both at low and high concentrations. Due to frequent exposure, food and human contamination by non-persistent toxic substances is also commonly observed.1-10
While research on the potential health effects of environmental chemicals is quite strong in Spain,6,11 only some in-depth studies exist on perceptions, attitudes, and behaviours of citizens related to food safety;12-17 and, to our knowledge, few studies have specifically addressed citizens’ views on the causes and consequences of the presence of synthetic toxic substances in food. Of course, there is a rich international scientific literature on such issues.10,12,18-24
Therefore, the main aim of the study was to explore factors influencing perceptions and views on the responsibility for the presence of toxic substances in food, on enforcement of laws and regulations that control human exposure to toxic substances in food, and on the effectiveness of such regulations.
Materials and methods
Study population and opinion survey
This is a cross-sectional study in which an online survey was the main methodological tool. Based on the theoretical framework, and findings from the first phase, of an ongoing research project on sociocultural dimensions of human contamination by persistent toxic substances and other environmental chemical agents,12,16,17 we designed the online survey “Presence of toxic substances and confidence in food products” (see Appendix 1 online). The Internet platform used was Surveymonkey,25 and the survey was available to be answered from 15th to 31st May 2013. A short informative cover letter was followed by the informed consent, seven sociodemographic questions, 16 questions regarding knowledge, sources of information, perceptions and opinions about chemicals in food, and a final section for optional comments (see Appendix 1 online). The invitation to participate in the survey was made by snowball sampling via email.26 The survey was initiated by 949 individuals, and 64 people (6.7%) did not proceed beyond informed consent (the first question, see Appendix 1 online). The present study is thus based on 885 individuals with at least sociodemographic information (Table 1). Of such 885 individuals, 740 (83.6%) answered all 24 questions, 94 (10.6%) only questions about personal characteristics, 38 (4.3%) responded until question 16, and 13 (1.5%) until question 20. People who did not answer all questions were older, had a lower level of formal education, and were more often pensioners or unemployed than the 740 subjects who completed the survey. As one of the purposes of the present study was to obtain information on the knowledge that participants had about the questions in the survey, we kept the maximum number of participants in each analysis.
Table 1 Sociodemographic characteristics by gender.
SD: standard deviation.
aUnless otherwise specified, p value derived from Fisher's exact test (two-tail).
bStudent's t-test (two-tail).
cMann-Whitney's U test (two-tail).
dGroups from the International Standard Classification of Occupations (ISCO-08).
Over 67% of 885 participants were women. The median age was 47 years, 69.9% had university studies, and 54.5% were employed by others (Table 1). Women were about two years younger, and had a slightly lower educational and occupational status than men. More than one third of respondents lived in the province of Barcelona, and 46% in parts of Spain other than those specifically mentioned in Table 1. For the classification of the occupation, the 10 categories of International Standard Classification of Occupations (ISCO)-08 (1996-2014) were used; in the analyses they were grouped into five categories (Table 1).
Main dependent and independent variables
The three main dependent variables of interest stemmed from three questions of the survey: question 14 (“The responsibility of the presence of toxic substances in food belongs to…”); question 15 (“In Spain, laws and regulations that control or should control human exposure to toxic substances are enforced in…”); and question 16 (“In Spain, laws and regulations that control or should control human exposure to toxic substances are effective”) (see Appendix 1 online). Responses to question 14 on responsibility were classified in three groups, according to whether respondent held responsible a) between 0 and 3 of the 8 social groups; b) 4 to 5 social groups; and c) respondents that held responsible 6 to 8 social groups. Responses to question 15 on enforcement were classified in four groups: a) ‘often’, when the responses were mostly always or often; b) ‘sometimes’; c) ‘never’, when they were mostly rarely or never; and d) ‘do not know’, when they were mostly “do not know”. Responses to question 16 on effectiveness were “yes”, “no” and “do not know” (see Appendix 1 online).
We created the variable ‘main means of information’ to consider combinations of responses to questions 9 (“work with or study toxic substances”), 10 (“frequency of consultation of information on toxic substances”), and 11 (“sources of information about toxic substances”). In this new variable, the category of ‘experience’ was for participants who worked with or studied toxic substances and, in addition, either a) consulted information often or sometimes from ≥2 different sources of information, except family and friends, or b) consulted information rarely but used three sources, except family and friends. The second category for the variable ‘main means of information’ was ‘interested’, and it included participants who did not work with or study toxic substances and a) consulted often or sometimes information about toxic substances from ≥2 sources of information, except family and friends, or b) consulted information rarely or never from 4 or 5 sources, family and friends included. The third category of ‘main means of information’ was ‘little or not interested’, and it grouped participants not classified in the other two categories.
Responses to question 12, about the degree of toxicity for human health of different substances at high concentrations, were categorized according to the number of correct responses (0-4, 5-7, and 8-9); except for vitamin C and folic acid, all the other substances were considered potentially toxic at high body concentrations. When asked about the two most important factors for a substance to be toxic for human health (question 13), 87.9% of participants indicated the concentration of the substance in the body; 77.4%, the toxicity of the substance itself; 77.0%, frequency of exposure to the substance; 32.9%, tolerance to the substance; 31.4%, origin or source of exposure to the substance; and 29.6%, the route of entry into the body. We grouped the responses to question 13 according to the most common pairs of responses, and we show the three most common pairs of responses: frequency of exposure and body concentration (60.9% of responses included both factors), body concentration and toxicity of the substance itself (60.6% of responses), and frequency of exposure and toxicity of the substance (52.4%). Question 18-about the presence of toxic substances in different phases of the food chain-was categorized into ‘food production, processing and distribution’, ‘only food production and processing’, and ‘other responses’. Question 19 inquired about the types of stores where the respondent expected to find foods containing higher concentrations of toxic substances; responses were categorized as: ‘no stores’ (no stores sell foods containing high concentrations of toxics), ‘1-2 food stores (neighbourhood shops included)’, ‘1-2 food stores (neighbourhood shops not included)’, and ‘3-4 food stores’. Answers to question 20 were categorized based on the number of health problems answered correctly (0-3, 4-5, and 6-7); all the 7 health problems were related with the presence of toxic substances in the food. Answers to question 21 were categorized in: 0-2, 3-4, and 5-6 factors answered correctly (all factors included in question 21 increased the likelihood of toxic substances being harmful to human health). Answers to question 22 on food groups were categorized in: 0-2, 3-5, and 6-7 food groups. Finally, question 23 on food packaging was categorized in ‘cans and plastic’ and ‘other responses’ (see Appendix 1 online).
Statistical analyses
Univariate statistics were computed as customary.27,28 To assess differences on age by gender and by specific answers to questions of interest, Student's t-test, ANOVA, Kruskal-Wallis, and Mann-Whitney's U tests were used. Fisher's exact test for homogeneity was applied to assess the relationship between two categorical variables.27,28 When a trend was observed, Mantel-Haenszel's χ2 test for linear trend was used.
To estimate the magnitude of the associations between socioeconomic factors and the survey questions, and our three main questions of interest (on responsibility, enforcement, and effectiveness, see above), multivariate-adjusted odds ratios (ORs) and their corresponding 95% confidence intervals (95%CI) were calculated by unconditional logistic regression with progressive degrees of adjustment.29 In the final analyses the perception of responsibility was dichotomously categorized (responsibility attributed to ≥6 vs. <6 social groups); enforcement was categorized in ‘often’ vs. other answers; and effectiveness in ‘yes’ vs. ‘no or do not know’. The main effects of all predictors were independently explored in the base models, and final models were adjusted for age and gender.28,29 Categorical ordinal variables were analysed for a linear dose-response relation through the multivariate analogue of Mantel's extension test; when a linear trend was not apparent, the probability test was used. The level of statistical significance was set at 0.05 and all tests are two tailed. Analyses were conducted using the Statistical Package for the Social Sciences (SPSS), version 18 (SPSS, Armonk, NY, USA, 2009), and R, version 3.2.1 (R Core Team, 2015).
Results
About 23% of the participants worked with or studied toxic substances. Sixty-three percent often or sometimes consulted information about toxic substances, and 75% received this information through mass media, among another sources (Table 2). When asked about the degree of toxicity of different substances at high body concentrations, 39% of participants (significantly more men than women) identified correctly all or almost all options (Table 2). Over 71% of men and 66% of women (p value = 0.008) identified the body concentration and the toxicity of the substance as the two most important factors for a substance to be toxic (Table 2).
Table 2 Responses to the main study questions about toxic substances by gender.
aUnless otherwise specified, p-value derived from Fisher's exact test (two-tail).
bWork with/study toxic substances in two categories, “Yes” and “No or do not know”.
cExcluding the “do not know”.
dMain means of information in two categories, “Informed by experience or interest” and “Little or no interest”.
eNumber of correct answers about toxicity degree for human health of certain substances at high body concentrations.
fIndividuals who held responsible of the presence of toxic substances in food between 0 and 3 of the 8 social groups of the question.
gIndividuals who held responsible 4 to 5 social groups.
hIndividuals who held responsible 6 to 8 social groups.
iMantel-Haenszel's χ2 test for linear trend.
jIndividuals who consider that the laws and regulations are always or often enforced.
kIndividuals who consider that the laws and regulations are enforced occasionally.
lIndividuals who consider that the laws and regulations are rarely or never enforced.
mIndividuals who do not know if laws and regulations are enforced.
Women held responsible a larger number of groups than men for the presence of toxic substances in food (p for trend = 0.032) (Table 2). Only 16.5% of respondents (significantly more men than women) considered that laws and regulations to control human exposure to toxic substances are effective. The information about toxic substances in food provided by clinicians, technicians, scientists or associations was considered of good quality by 609 subjects (80.9% of respondents), while only 19.1% deemed the information provided by the mass media of good quality (Table 2). Over 27% of women and 35% of men considered of good quality the information provided in the workplace (p value = 0.009). Of the 753 participants who answered the question about substances in food that may cause health problems, 278 (37%) identified correctly 6 to 7 such health problems (40% of women and 31% of men, p value = 0.044) (Table 2). Over 68% of participants (73% of women and 58% of men, p value <0.001) considered that cans and plastic food packaging may have adverse health effects. Over 87% of subjects (with no gender difference) answered it was possible that throughout their life they could have accumulated in their body toxic substances potentially dangerous to their health in the mid- or long-term (Table 2).
Regarding occupational status, the percentage of subjects who considered that laws and regulations on exposure to toxic substances are effective ranged from 10.5% among self-employed participants to 20% among subjects employed by others (i.e., low figures, and a difference of 10 percentage points only) (see Table I in Appendix 2 online). Almost 19% of participants who worked with or studied toxic substances and 16% of participants who did not work with or study responded that laws are effective. Half of participants who never sought information about toxics responded ‘do not know’ when asked about the effectiveness of regulations, whereas 68% of participants who sought it ‘often or sometimes’ considered regulations ineffective. Regulations were also deemed ineffective by over 65% of participants who considered that foods with higher concentrations of toxic substances are available in almost all types of food stores, of participants who identified correctly 6 to 7 health problems caused by toxic substances in food, and of participants who identified 6 to 7 food groups with concentrations of toxic substances harmful to human health.
Only 38% of participants considered that laws and regulations about toxic substances in food were ‘often’ enforced in the different parts of the food production chain, whereas 17% indicated that they ‘never’ were (see Table II in Appendix 2 online). 47% of men, 46% of participants employed by others, 41% of subjects that often consulted information about toxic substances, and 48% of individuals that considered that no stores sell food with high concentrations of toxics answered that regulations about toxic substances in food are ‘often’ enforced (all p values <0.001). Over 42% of participants who worked with or studied toxic substances and 37% of subjects who did not also considered that regulations are ‘often’ enforced.
Subjects who considered 6 to 8 social groups responsible for the presence of toxic substances in food answered in a lower proportion “do not know” to the question on the effectiveness of regulations (compared to subjects who considered responsible less than 6 groups). Subjects who held responsible less social groups also tended to consider that laws were effective (p value = 0.009) (Table 3). Regarding the relationship between law enforcement and effectiveness, 50% of subjects who answered that laws were always or often enforced and 76% of those who considered that sometimes they were, nevertheless answered that such laws were not effective. Over 65% of subjects who answered that laws were enforced rarely or never responded “do not know” when asked about the effectiveness of the laws (p value <0.001) (Table 3).
Table 3 Relationship between the attribution of responsibility and the enforcement of laws, and the effectiveness of laws.
Total N = 769.
aFisher's exact test (two-tail).
bIndividuals who held responsible of the presence of toxic substances in food between 0 and 3 of the 8 social groups of the question.
cIndividuals who held responsible 4 to 5 social groups.
dIndividuals who held responsible 6 to 8 social groups.
eIndividuals who consider that the laws and regulations are always or often enforced.
fIndividuals who consider that the laws and regulations are enforced sometimes.
gIndividuals who consider that the laws and regulations are rarely or never enforced.
hIndividuals who do not know if laws and regulations are enforced.
Multivariate analyses that adjusted for age and gender showed that consulting information about toxic substances was positively associated with attributing the responsibility for toxic substances in food to 6 or more social groups (vs. <6 social groups). Working with or studying toxic substances was not (Table 4). Subjects who often or sometimes consulted information were almost twice as likely to make such attribution as subjects who sought information rarely or never or did not know (OR = 1.92 for the adjusted model, p value = 0.001). Three other factors were even more strongly linked with attributing responsibility to 6 or more social groups: identifying more health problems that may be caused by such substances (OR for the upper category: 2.48); identifying correctly most factors that increase the likelihood of toxic substances in food being harmful to human health (OR: 2.86); and identifying more food groups with toxic concentrations harmful to human health (OR: 2.92). All p for trend <0.001 for these three factors. Such attribution was not more frequent among respondents who worked with or studied toxic substances (Table 4). Men without college education had an OR: 2.05 (95%CI: 1.03-4.09) (adjusted for age) of attributing the responsibility for toxic substances in food to 6 or more social groups (vs. <6 social groups) than men with college education.
Table 4 Associations between sociodemographic characteristics and questions about toxic substances, and attributing responsibility for the presence of toxic substances in food to 6 or more social groups (vs. <6 social groups).
OR: odds ratio; CI: confidence interval; Ref: reference category (OR: 1.00).
aAll models adjusted for age and gender.
bUnless otherwise specified, p-value derived from Wald's test.
cTest for linear trend (multivariate analogue of Mantel's extension test).
Women were 40% less likely than men (OR: 0.60, p value = 0.001) to respond ‘often’ (vs. other answers) to the question on the frequency of law enforcement; and so were participants who identified more food groups with toxic concentrations potentially harmful to human health (see Table III in Appendix 2 online). In gender-stratified analyses, the responses of both women and men to question 22 on food groups with toxic substances were very similarly unrelated to responding ‘often’ (vs. other answers) to the question about the frequency of law enforcement (i.e., there was no effect modification by gender). By contrast, among men, an answer different to “yes” to question 24 about lifelong accumulation of toxic substances (vs. an affirmative answer) was associated (OR: 2.36; 1.08-5.14) with responding ‘often’ (vs. other answers) to the question about the frequency of law enforcement, whereas for women there was no significant relation.
Women were 65% less likely than men to answer that regulations about toxics in food are effective (OR: 0.35; p-value <0.001) (Table 5). The corresponding age- (and in these instances, sex-) adjusted OR for participants employed by others (vs. self-employed) was 2.17 (1.10-4.27); for participants informed by experience or interest (vs. little or not interested), it was 1.53 (1.03-2.29); and for subjects who answered that laws were enforced sometimes or never (vs. always/often), they were 0.19 (0.11-0.32) and 0.13 (0.06-0.28), respectively. Considering that foods with higher concentrations of toxic substances are available in more types of food stores was also inversely associated with responding that laws were effective (Table 5). Such judgements about law effectiveness were generally not associated with working or studying toxic substances. The only exception was: the attribution of responsibility to more groups was inversely related to responding that laws were effective among subjects who worked or studied toxic substances (OR: 0.27; 0.08-0.89), whereas it was positively related to responding that laws were effective among subjects who did not work or study such substances (OR: 1.59; 0.87-2.90).
Table 5 Associations between sociodemographic characteristics and questions about toxic substances, and the probability of answering that laws are effective (vs. “no or do not know”).
OR: odds ratio; CI: confidence interval; Ref: reference category (OR: 1.00).
aAdjusted for age and gender.
bUnless otherwise specified, p-value derived from Wald's test.
cIndividuals who held responsible of the presence of toxic substances in food between 0 and 3 of the 8 social groups of the question.
dIndividuals who held responsible 4 to 5 social groups.
eIndividuals who held responsible 6 to 8 social groups.
fIndividuals who consider that the laws and regulations are always or often enforced.
gIndividuals who consider that the laws and regulations are enforced occasionally.
hIndividuals who consider that the laws and regulations are rarely or never enforced.
iIndividuals who do not know if laws and regulations are enforced.
jTest for linear trend (multivariate analogue of Mantel's extension test).
kExcluding the “do not know”.
Discussion
We found several statistically significant and socially relevant associations; notably, attributing the responsibility for toxic substances in food to a larger number of social groups was more frequent among subjects who consulted information about the problem more often, who correctly identified factors that increase the likelihood of toxic substances in food being harmful to human health, who better knew the health problems caused by such substances, and who recognized more food groups that tend to have concentrations of toxic substances harmful to human health; such attribution was not more frequent among respondents who worked or studied toxic substances. Women were significantly less likely than men to respond that laws and regulations are often enforced; and so were participants who identified more food groups with potentially toxic concentrations. Women were also significantly less likely than men to respond that regulations about toxics in food are effective. With due caution, it seems reasonable to suggest that some of the observed associations between factors might also exist in the general population, although probably with some differences in magnitude.
However, the study was not based on a representative sample of the general population due to the sampling design and, therefore, results should not be generalized to the general population of Spain; in particular, the univariate and bivariate results summarized in Table 2. Within the same theoretical framework, the study design and results complement our previous qualitative studies.12,16,17 Although we did not target for participation any specific group of people, a portion of respondents was probably comprised by citizens concerned about the study issues, and by subjects with college education. Yet, the majority of respondents were unlikely to be experts (e.g., when asked about the degree of toxicity for human health of different substances at high body concentrations, only 39% of participants answered correctly all or almost all options). Also, the educational level was not an important predictor of responses when age and gender were also considered. Associations were similar among participants who worked with or studied toxic substances than in participants who did not. The present study and project12,16,17 provide a basis for new and more complex studies, which could use representative samples of the general population or target specific subgroups through qualitative and quantitative methodologies.
Even if clearly not stemming from a representative sample, in our view results do provide ‘food for thought’: to authorities (notably, those whose policies influence public health, the safety of food and related consumer products, or chemical hazards), industries (agriculture, chemistry, food processing and packaging, distribution, marketing), social organisations, and other citizens. In the societal sectors reflected in the study, there seems to be a rather wide scepticism and distrust towards the enforcement and effectiveness of regulations that control or should control human exposure to toxic substances in food. Over 58% of participants responded that laws and regulations are not effective. And more than 87% said it was possible that throughout their life they could have accumulated substances potentially dangerous to their health; this proportion was identical in women and men.
We are not aware of the existence of survey questionnaires validated to address the specific study questions. The accuracy, reliability, reproducibility, interpretability, and other properties26 of the questionnaire used in the study were untested, and this is another study limitation. Nevertheless, no major inconsistencies were apparent; e.g., half of participants who never sought information about toxics responded ‘do not know’ when asked about the effectiveness of regulations, whereas most participants who sought it ‘often or sometimes’, but not all, considered regulations ineffective. It is also important to notice that some of the issues assessed in the survey are quite novel in Spain and other countries; thus, a sizable number of study participants (from about 0.1% to 38%) answered “do not know” to some questions. Over 45% of participants reported having little knowledge on synthetic toxic substances, their presence in food, and their possible health effects. Between 22% and 37% of participants identified correctly health problems that toxic substances in food may cause, and factors that increase the likelihood of toxic substances in food being harmful to human health.
When assessing the limitations and strengths of the present study it should be noted that it is part of a research project that has already published a synthesis of its theoretical framework and initial results from a qualitative ethnoepidemiologic study.12,16,17 Social processes of perceiving, facing, attributing meaning, and coping with risk are diverse and multifaceted; and more complex than enabled by some frameworks, as Beck's ‘risk society’.12,16,19,30-35 While the survey was conducted some three years ago, we think the results reflect long-term and profound underlying processes in Spain's society, which did not materially change in recent years.
Women were significantly less likely than men to respond that laws and regulations about toxics in food are enforced and effective. Women also tended to attribute the responsibility of the presence of toxic substances in food to more groups than men, and they more frequently considered that the responsibility belonged to groups not involved in the food production chain. Men were less likely to identify correctly health problems that may be caused by toxic substances in food, but more likely to answer correctly the degree of toxicity of certain substances. Gender differences in risk perception have been described in numerous studies, with women more often reporting higher levels of concern, higher sensitivity and lower tolerance to risk than men.18,20 The analysis of gender differences ought to take into account the sociocultural context and processes involved in this perception.20
Until a few decades ago, in the Western world most food problems were basically related to malnourishment, hunger, famine and the difficulties of achieving adequate nutrition due to food scarcity, plagues, and economic constraints.21 Modern progress in feeding populations has been immense. However, this has not happened at zero cost: adverse health, social, and environmental consequences must be acknowledged.11,36-38 Thus, our ‘risk societies’ have also long endured problems and reasonable concerns related to food and chemical risks, partly due to the vast industrialisation of food processing and the distance between food production and consumption.12 Some studies focusing on social representations of food risks have unveiled rather negative perceptions of citizens about the dangers of technological interventions in food production at an industrial scale.13 Some authors consider chemical hazards to human health as elements of risk safety.7,22 Such hazards are partly seen to stem from the presence of synthetic chemical substances (as pesticides, growth control hormones, feed conversion enhancers and feed contaminants, or antibiotics) that pollute food during the processes of production, storage, packaging and distribution.22 The literature also reflects positive views about agricultural successes, as well as social concerns about global poverty and undernutrition, or about social policies and individual behaviours to prevent obesity. Such views coexist - often, in a quite balanced and nuanced way - with the fact that for many people the presence of synthetic chemical compounds in processed food is also a problem.23 For example, some studies show how people who experience chronic illnesses express a greater concern about man-made chemicals and toxic residues in food.23
Therefore, the scientific and social (un)certainties about the effects that some food contaminants can have on human health are part of contemporary debates on food (in)security; e.g., on the strengths and weaknesses of glocal11,38-41 food systems. Such debates are also clearly related to economically- and culturally-shaped processes of risk perception and tolerance, the industrial scale of food processing, or a dysregulated globalisation of markets.11-17 Discussions on social media and networks thus abound on highly relevant (both material and symbolic) issues as, for example, the socio-political causes of ‘mad cow disease’ (bovine spongiform encephalitis) or the effects of genetically modified organisms on human health and food sovereignty.13,21,36-38,42 The frequent distrust of citizens in the production chain and its managers, or in regulatory agencies and institutions, is also partly due to actual shortcomings (e.g., failure to regulate conflicts of interests, ‘revolving doors’ and lobbying), and to the systems of risk communication, which have all significant impacts on the perception that consumers have about food benefits and risks.14-17,22,24,36,43
Hence, there currently is a broad shift from concerns basically focused on access to food, with quite a high level of tolerance to short-term microbiological risks and little regard for long-term risks, to other views which (though still worried about access to food) are more attentive to the possible negative health consequences of eating chemically contaminated foods.11,13,24,44 In influential segments of many Western societies there is thus increasing awareness of such non-microbiological, multifactorial, cumulative, and long-term risks.11,37,44 Such realities also imply changes in the relationships of trust on food governance of citizens and societal organisations towards institutions and companies.19,24,44 Food safety and public health authorities also ought to ponder the study findings.
Conclusions
Among study participants there was a rather wide concern about chemical contaminants in food, and distrust towards the enforcement and effectiveness of regulations that in Spain should control exposure to toxic substances in food. Respondents had a low level of trust on the results of the work performed by the different societal actors, companies, and institutions that must protect citizens from the potential adverse health effects of synthetic chemicals that are often present in food products.
