nutrients

Article
Consumer Evaluation of the Role of Functional Food
Products in Disease Prevention and the
Characteristics of Target Groups
Brigitta Plasek 1 , Zoltán Lakner 1 , Gyula Kasza 2 and Ágoston Temesi 1, *
1 Department of Food Economics, Faculty of Food Science, Szent István University, Villányi str. 29-43,
1118 Budapest, Hungary; [email protected] (B.P.); [email protected] (Z.L.)
2 Directorate for Food Safety Risk Assessment, National Food Chain Safety Office, Keleti Károly str. 24,
1024 Budapest, Hungary; [email protected]
* Correspondence: [email protected]; Tel.: +36-1-305-7100 (ext. 6178)




Received: 2 December 2019; Accepted: 23 December 2019; Published: 26 December 2019


Abstract: Our research explores the methods consumers would consider using in order to prevent
non-communicable diseases, with consuming functional food products being one of these methods.
Previous research has pointed out the importance of missing information such as which diseases
worry consumers and what they would use to avoid them. We conducted a personal survey with
1027 people in Hungary about 13 diseases and four prevention methods. We analyzed the results
with descriptive statistical methods, binary logistic regression, and random forest. According to our
results, the highest proportion of worried respondents think it is justified to consume functional
foods for the prevention of digestive problems, a weakened immune system, and high cholesterol
level. Our results help to characterize the target group for these three diseases. Completed education
plays a key role in choosing functional foods to prevent a weakened immune system. Those with
tertiary education are the most likely to choose this prevention method. With the other two diseases,
age played a crucial role. All age groups over 36 would be more likely to choose functional foods to
prevent digestive problems, whereas in case of high cholesterol level, the 36–50 age group would be
more likely to apply this method.

Keywords: consumer research; willingness to purchase; non-communicable diseases; binary logistic

regression; random forest

1. Introduction
Changes in lifestyle, less physical activity, and improper nutrition have resulted in the spread of
non-communicable diseases to the extent that today they pose serious problems, and a significant ratio
of deaths are connected to them worldwide [1]. Globally, 41 million people die from non-communicable
diseases annually, which constitutes 71% of all deaths. Of these, about 15 million people are in the
30–69 age group. Cardiovascular diseases pose the biggest risk (17.9 million deaths) followed by
cancer-related diseases (9.0 million), respiratory diseases (3.9 million), and diabetes (1.6 million) [2].
In Hungary, non-communicable diseases present a host of problems for society. Based on the
statistical data of World Health Organization (WHO), it was concluded that in Hungary the average
number of years spent in good health was 67.4 years in 2015, and the average life expectancy was
75.9 years [3]. This means that the average Hungarian spends 8–9 years suffering from some disease.
Statistical data show that in Hungary, cardiovascular diseases, malignant tumors, chronic respiratory
diseases, diabetes, and some other non-communicable diseases accounted for 94% of all deaths in
2016 [4].

Nutrients 2020, 12, 69; doi:10.3390/nu12010069 www.mdpi.com/journal/nutrients

Nutrients 2020, 12, 69 2 of 19

Medicaments offer a classic solution to treat diseases. Recently, as a result of regulations,

medicaments have become more readily available for consumers, so today they have an array of
non-prescription products at their disposal that they can apply to cure themselves [5]. Dietary
supplements also offer a widely-used solution; however, these pills play a role in the prevention of
diseases already. A survey has shown that elderly consumers consider dietary supplements at least as
important as prescription medicine. The same research also concluded that the most common reasons
for taking dietary supplements were health protection, disease prevention, and the advice of a health
expert [6]. Over the last decades, consumption of dietary supplements has grown exponentially; as
Starr [7] demonstrated, in the United States, 4000 products were sold in 1994, and this number grew to
more than 90,000 by 2014.
Apart from lifestyle changes and the consumption of dietary supplements, the consumption of
functional foods can also reduce the risk of developing diseases. “A food can be regarded as ‘functional’
if it is satisfactorily demonstrated to affect beneficially one or more target functions in the body,
beyond adequate nutritional effects in a way that is relevant to either an improved state of health and
well-being and/or reduction of risk of disease.“ (p. S26, [8]). The different health claims manufacturers
often use to communicate the positive effects of functional foods on health to consumers fall under
European Food Safety Authority regulations. The specific regulation is included in regulation (EC) no.
1924/2006 [9]; the EFSA database makes all applicable health claims and nutrition claims available.
The EFSA distinguishes different types of claims in the category of health claims, such as “Function
health claims” (Article 13 claims), “Risk reduction claims” (Article 14(1)(a)), and “Claims referring to
children’s development” (Article 14(1)(b)) [10]. For the purposes of our research, one type of health
claim, namely, risk reduction claims, are worth highlighting as they communicate the effect on the
reduction of the risk of diseases [10].
In their research in Hungary, Szakály et al. came to the straightforward conclusion that “In the
current situation, there is no other choice than to bring the public’s attention to food products that
possess an extra nutritional advantage” (p. 16, [11]). Owing to their ingredients, functional food
products can help to prevent a variety of chronic diseases [12]. Encouragingly, several studies have
indicated that due to increasing consumer concerns related to health, consumers would be willing to
change their dietary habits and include products that would have a positive effect on their health in
one way or another [13,14]. From the results of their focus group interviews, Szakály et al. concluded
that if participants “have to choose between medication and functional food products, all of them
would vote for the latter” (p. 23 [15]).
Due to the complexities of developing functional foods, it is important to assess which diseases
consumers are concerned about [16]. However, apart from identifying these diseases, it is also important
to examine how consumers would prevent them. Without this information we cannot talk about
functional foods developed for, and communicated to, a wide range of consumers, and we cannot
expect functional foods to have a bigger role in diets. However, if we are aware of the specific diseases
that a wide circle of consumers would choose to prevent with the help of functional foods, then the
affected food manufacturers could concentrate on these diseases, and a more remarkable public health
effect could be achieved.
At the moment it is not yet known which diseases consumers would seek to reduce the risk
of through the consumption of health-protecting food products [17,18]. As Siró et al. put it, “It is
more interesting, however, to identify those diseases that consumers would make much account of
preventing them by nutrition” (p. 462 [17]). We believe that this missing piece of information is what
is needed to get a more complete picture of consumer behavior regarding functional foods. Kraus [19]
came closest to answering the question with his research where the potential health effects of functional
food products were ranked. Baba et al. [20] maintained that health-related consumer concerns may be
a crucial factor in food consumption, and thus also in developing new products. Consumer attitudes to
functional food products in light of worrisome diseases have already been investigated in general terms.
In accordance with previous research [21,22], Chen’s [23] research results showed that those who were
Nutrients 2020, 12, 69 3 of 19

concerned about modern lifestyle diseases displayed a more positive attitude towards functional food
products and had a larger likelihood of accepting and using them in order to prevent these diseases.
The spread of non-communicable diseases poses numerous societal and economic risks.
Nevertheless, it is still an unanswered question as to which prevention methods consumers would use
to avoid/treat specific diseases, so the aim of the research was to answer this question by asking them.

2. Materials and Methods

In our research, we conducted a survey with personal interviews (PPI) with 1027 participants
between 4–19 March 2013, at busy transport hubs of five big cities in Hungary, namely in Budapest,
Pécs, Sopron, Nyíregyháza, and Szeged. The respondents were rewarded with a small non-food gift
for participating in the survey. At the beginning of the survey, participants were informed about the
aim of the research; at the same time, the responses were anonymous. Before completing the survey,
respondents gave implicit consent to their answers being recorded and later analyzed; the study was
conducted in accordance with the Declaration of Helsinki. Respondents had the option to refuse to
answer any question or stop answering the survey at any point.
We employed quota sampling, as opposed to probability sampling, so our sample significantly
follows the segmentation of society. Among our respondents, compared to the population, the ratio of
people between ages 36–50 and >50 is smaller, and the ratio of those living in Budapest and having
tertiary education is higher.
In the survey, we asked participants questions in relation to 13 diseases: which are the ones they
are not concerned about; the ones they are concerned about but would not make a financial sacrifice to
prevent them; the ones they are concerned about and would make a financial sacrifice to prevent them.
Moreover, supposing these diseases can be avoided/prevented through various methods, we wanted
to find out exactly how much respondents would spend. Demographic variables included gender,
age, place of living, highest education, estimated income status of the household, and the presence of
underage children in the household. Finally, our questions included whether it was important for the
respondents to do something for their health, whether they paid attention to their diet, and who did
the grocery shopping in the household.
After collapsing the results, we used the statistics software SPSS 25.0 to analyze the answers.
The collected data were first analyzed through descriptive statistical methods, followed by binary
logistic regression to find the differences based on demographic variables.
In the course of the regression analysis, we considered only respondents who were concerned and
would make a financial sacrifice to avoid the disease under investigation. In the models employed,
the dependent variable was the financial expense of the prevention method (purchasing functional
food (yes/no or Y/N)) chosen in order to avoid the specified disease, and the independent variables
were the characteristics of the respondents, presented in Table 1. In our analyses, we applied the 5%
significance level (p < 0.05) and the forward Wald method.

Table 1. Grouping of respondents based on variables (n = 1027).

Characteristics of Survey Respondents %

Female 53.5%
Gender
Male 46.5%
<25 47.7%
25–35 18.8%
Age brackets
36–50 14.2%
50< 19.3%
Budapest 39%
Place of living Another city 48.6%
Village 12.4%
Nutrients 2020, 12, 69 4 of 19

Table 1. Cont.

Characteristics of Survey Respondents %

Primary school/vocational school 15.1%
Highest accomplished
Secondary technical
qualification 40.0%
school/secondary school diploma
Tertiary education 44.9%
Below average 35.5%
Subjective estimation of income
Average 53.6%
position
Above average 10.9%
Presence of a child under 18 in the Yes 79.5%
household No 20.5%
Undernourished 6.8%
Overweight or underweight status
Normal 61.3%
of the respondent (based on BMI) *
Obese/overweight 31.9%
It is important for me to do Important 88.9%
something for my health Less important/not important 11.1%
Important 72.8%
I pay attention to healthy eating
Less important/not important 27.2%
Respondent 25.7%
Grocery shopper Another person 22.7%
Shared 51.6%
* Based on World Health Organization (WHO) body mass index (BMI) [24].

The principle of triangulation provided another perspective from which our study approached
the research questions related to the identification of diseases that consumers would prevent with the
help of functional food products, and the socio-demographic features that have a stronger influence on
this choice.
As Meerza et al. (p. 29, [25]) summarized it: “binomial logistic regression generally works well
as a classifier when the relationship between dependent and independent variables is linear and the
data are relatively balanced between classes [26], the random forest (RF), a machine learning method,
outperforms the logistic regression when the non-linear characteristics and complex interactions among
predictor variables exist [27]”.
With the aim of forecasting consumer choice, we ran additional statistical analysis on the data.
Random forest is a method widely used for making predictions, with the advantages of being wieldy,
robust, and fast [28]. We used the random forest package of the R statistical software to examine the
extent to which consumer choice can be predicted. The analysis was conducted by including all the
examined variables. “Random forests are a combination of tree predictors such that each tree depends
on the values of a random vector sampled independently and with the same distribution for all trees
in the forest” (p. 5, [29]). During the analysis, we used the “OOB Estimate of Error Rate” value to
examine accuracy, and the “MeanDecreaseGini” values to examine the importance of the analyzed
variables. With each analysis, 70% of the total sample was used for training and 30% was used to
validate the models.

3. Results

3.1. Identifying the Diseases That Worry Respondents

One of the questions we searched an answer for in our research was to identify the
non-communicable diseases that concern consumers the most. Table 2 shows the ratios of the
respondents concerned about a specific disease and their propensity to make a monetary sacrifice to
avoid it.
Nutrients 2020, 12, 69 5 of 19

Table 2. The attitude of respondents towards different health problems and their propensity to spend
money to decrease the probability of their occurrence.

Concerns, and the Concerns, but the

Respondent Is Ready to Respondent Does Not Does Not Concern
Disease
Spend Money to Want to Spend Money Respondents
Prevent It to Prevent It
Cancer 57% 21% 22%
Cardiovascular diseases 49% 28% 23%
Weakened immune
48% 24% 28%
system
Diabetes 42% 24% 34%
Osteoporosis 38% 26% 36%
Joint diseases 37% 31% 32%
Skin diseases 33% 28% 39%
Digestive problems 33% 33% 34%
High level of cholesterol 32% 33% 35%
Respiratory diseases 32% 32% 36%
Memory
lapse/concentration 32% 30% 38%
problems
Migraine 26% 27% 47%
Unbalanced
21% 35% 44%
mood/apnoea

According to our results, surveyed consumers are most concerned about cancer-related diseases
(78%). This same disease is the one with the highest ratio of respondents willing to make a financial
sacrifice. A high ratio of respondents would also make a monetary sacrifice to prevent cardiovascular
diseases, a weakened immune system, and diabetes. Surveyed consumers are least concerned about
migraines and unbalanced mood.

3.2. The Form of Financial Sacrifice Made Towards Preventing Diseases

Our research also aimed at finding out what exactly, if anything, worried respondents would spend
on. The four factors we examined were taking medicine, using dietary and nutritional supplements,
consuming functional foods, and lifestyle changes. Table 3 summarizes the results of the question
of what would worried respondents, who are willing to make a financial sacrifice, spend to avoid a
given disease. Respondents were free to indicate more than one answer. In the table we included only
the answers of respondents who were worried and willing to make a financial sacrifice to avoid the
disease. Thus, the displayed percentages for the prevention methods do not emerge from the whole
sample, but from a smaller part of it, which we consider to be 100% here.
Our results show that in order to prevent diseases, most people say that they would implement
lifestyle changes. Respondents ascribe the biggest importance to it in the case of cardiovascular
diseases, which means that consumers would apply this method to prevent the disease which carries
the biggest noncommunicable disease (NCD) risk in Hungary, based on data by WHO [4]. Thus, we
can also conclude that consumers are aware of the connection between lifestyle and the emergence of
cardiovascular diseases.
Nutrients 2020, 12, 69 6 of 19

Table 3. The form of financial sacrifice worried respondents are willing to make.

Concerned, Ready to Make Form of Financial Sacrifice

a Financial Sacrifice to Dietary Functional Lifestyle
Prevent It Medicine
Supplement Food Changes
n % % % % %
Digestive problems 327 100% 13% 29% 43% 55%
Weakened immune
473 100% 22% 36% 39% 56%
system
High cholesterol level 322 100% 12% 19% 37% 64%
Diabetes 418 100% 36% 17% 35% 59%
Cancer 552 100% 47% 19% 31% 48%
Osteoporosis 376 100% 38% 24% 30% 40%
Cardiovascular
480 100% 25% 15% 29% 70%
disease
Joint diseases 368 100% 30% 26% 27% 48%
Skin diseases/eczema 325 100% 47% 14% 26% 37%
Memory
317 100% 25% 26% 26% 46%
lapses/concentration
Mood
211 100% 12% 21% 24% 66%
disorders/apnoea
Respiratory diseases 311 100% 39% 15% 21% 47%
Migraine 253 100% 43% 11% 21% 44%

The main aim of our research, however, was to identify the diseases that most people are likely to
prevent with the consumption of functional food products. To the question raised by Siró et al. [17],
namely, what are the diseases that consumers would prevent by choosing functional foods, our results
yielded the following answer: digestive problems, a weakened immune system, and high cholesterol
level. These are the three diseases for which consumers would choose functional foods in the highest
proportion. If we combine this result with the disease that worries respondents the most, we can
conclude that functional food products can play the biggest role in the prevention of a weakened
immune system for a wide circle of consumers. It should also be added that functional foods can play
an important role in preventing many other diseases apart from the three highlighted in our study, in
the case of which a significant reluctance of consumers has to be overcome.
In his research, Kraus [19] ranked the potential health effects of functional food products according
to their importance. Consistent with our results, strengthening the immune system and lowering the
risk of cancer-related diseases also ranked on top of his list, and improving memory ranked lower.

3.3. Disease Prevention with Functional Foods—Consumers’ Characteristics

We considered it important to utilize the characteristics of respondents to identify the groups that
can be most associated with specific prevention methods. Table 4 displays the results derived from
the binary logistic regression. The table shows only the diseases and characteristics of respondents
for which significant differences emerged between the groups within a certain variable. The analyses
included all the variables from Table 1. The odds ratios received from the regression analysis show the
values compared to the first group of the variable, that is, we can see the results compared to females
in the case of gender; compared to the 18–25 age group in the case of age; to the primary/vocational
school educated in the case of education; and a different town or village compared to Budapest in the
case of place of living. Regarding the question about the person who does the shopping, the basis of
comparison is the case when the respondent does the shopping; to this we compare the case when
someone else does it or when it is a shared responsibility. In the case of the statement “I pay attention
to healthy eating”, we compare those for whom it is important to those for whom it is not.
Nutrients 2020, 12, 69 7 of 19

Table 4. Results of the binary logistic regression.

Would Buy Functional Food to Avoid the Specified Disease (Dependent Variable)
Characteristics of the
Number of
Respondent Level of Attribute 95% CI for EXP (B)
Cases B SE Wald Sig. Exp(B)
(Independent Variable)
Lower Upper
High cholesterol level
Constant −0.752 0.189 15.773 0.000 0.471
289
Age 36–50 0.919 0.346 7.056 0.008 2.508 1.273 4.942
Cancer-related diseases
Constant −0.745 0.371 4.024 0.045 0.475
Gender Men −0.469 0.208 5.108 0.024 0.626 0.416 0.940
498
Level of education College/university 0.992 0.356 7.749 0.005 2.697 1.341 5.423
Grocery shopper Another person −0.862 0.296 8.498 0.004 0.422 0.236 0.754
Skin diseases
Constant −0.455 0.251 3.301 0.069 0.634
293 Another person −0.743 0.362 4.205 0.040 0.476 0.234 0.968
Grocery shopper
Shared −0.797 0.321 6.170 0.013 0.451 0.240 0.845
Memory lapse/concentration problems
Constant −1.871 0.483 15.006 0.000 0.154
284 Age 25–35 0.875 0.359 5.933 0.015 2.399 1.186 4.852
Level of education College/university 1.004 0.494 4.132 0.042 2.728 1.037 7.181
Digestion problems
Constant −1.103 0.396 7.758 0.005 0.332
291 36–50 1.055 0.362 8.478 0.004 2.872 1.412 5.844
Age
50+ 0.940 0.332 8.013 0.005 2.559 1.335 4.906
Nutrients 2020, 12, 69 8 of 19

Table 4. Cont.

Would Buy Functional Food to Avoid the Specified Disease (Dependent Variable)
Characteristics of the
Number of
Respondent Level of Attribute 95% CI for EXP (B)
Cases B SE Wald Sig. Exp(B)
(Independent Variable)
Lower Upper
Migraine
Constant −0.927 0.197 22.183 0.000 0.396
224
Gender Men −1.129 0.376 9.022 0.003 0.323 0.155 0.675
Diabetes
Constant −1.247 0.240 26.923 0.000 0.287
Age 36–50 0.997 0.330 9.121 0.003 2.710 1.419 5.177
375
Another city 0.509 0.252 4.074 0.044 1.664 1.015 2.729
Place of living
Village 0.984 0.357 7.588 0.006 2.674 1.328 5.385
Joint diseases
Constant 0.118 0.402 0.086 0.770 1.125
Secondary technical
329 Level of education school/secondary −0.981 0.408 5.783 0.016 0.375 0.168 0.834
school diploma
Grocery shopper Another person −0.956 0.379 6.351 0.012 0.384 0.183 0.809
Osteoporosis
Constant −1.495 0.345 18.827 0.000 0.224
Gender Men −0.518 0.254 4.161 0.041 0.596 0.362 0.980
333
Age 36–50 1.087 0.347 9.816 0.002 2.966 1.502 5.856
I pay attention to healthy eating Important 0.875 0.345 18.827 0.012 2.398 1.209 4.758
Weakened immune system
Constant −1.881 0.439 18.325 0.000 0.152
Gender Men −0.472 0.213 4.925 0.026 0.624 0.411 0.946
431 Age 36–50 0.938 0.314 8.916 0.003 2.554 1.380 4.727
Level of education College/university 1.355 0.381 12.656 0.000 3.875 1.837 8.174
I pay attention to healthy eating Important 0.616 0.266 5.358 0.021 1.852 1.099 3.121
Nutrients 2020, 12, 69 9 of 19

The table in Appendix A includes the goodness of fit test of models and the indicators that
characterize the explanatory power of the models. According to the results of the Hosmer–Lemeshow
test, the goodness of fit of the models is appropriate, and apart from one instance, none of the values
is significant. However, based on the two other indicators (Cox–Snell R2 and Nagelkerke R2 ), the
considered independent demographic variables only partially explain the dependent variables. The
Cox–Snell R2 values were around 0.05 on average, and the Nagelkerke R2 values were around 0.07.
From the small values we can conclude that apart from the demographic variables, other factors
also influence whether a consumer chooses the prevention method under consideration in order to
avoid diseases.
Based on the results displayed in Table 4, males consider the consumption of functional food as
relevant prevention methods to a lesser extent in the prevention of cancer, migraine, osteoporosis, and
a weakened immune system. Thus, we can emphasize that in the case of these diseases, it is important
that companies address their prevention-related communication primarily at women. This result is
in line with previous literature [30–34], which state that women are more open to functional food; at
the same time, our results provide a more nuanced picture in that they identify the diseases for the
prevention of which functional food development should be targeted to women.
For six of the 13 examined diseases, we obtained a significant result related to the effect of age. The
results show that compared to the 18–25 age group, it was the 36–50 age group that yielded significant
differences in the highest number. This age group considered functional food products more important
in the prevention of high cholesterol level, diabetes, osteoporosis, a weakened immune system, and
digestion problems. For this last problem, buying and consuming functional foods can be an important
prevention method also for the 50+ age group, whereas for memory and concentration problems, the
25–35 age group may consider this method. This result of ours clarifies the statements of previous
literature [17,31,32,34,35], according to which it is the middle-aged and the elderly rather than the
young who would consume such products.
It is also important to highlight the aspect of the person who does the grocery shopping, as for
three diseases, we obtained a significant result in this respect. For all of the three diseases, namely,
cancer-related diseases, skin diseases, and joint problems, it is a shared finding that respondents in the
household of whom someone else does the shopping consider the consumption of functional foods a
less relevant prevention method than respondents who themselves do the shopping.
Table 4 also includes significant results related to the highest level of completed education
of respondents. The responses of those with a higher level of education were compared to those
that were primary/vocational school educated. In accordance with results of previous research
literature [17,30–33], those with a college/university education consider the consumption of functional
foods more suitable for the prevention of certain diseases. Those with a college/university education
consider the consumption of functional foods a suitable solution for the prevention of cancer-related
diseases, memory and concentration problems, and a weakened immune system, whereas for the
prevention of joint problems, it is rather those with a secondary technical school/secondary school
diploma who would choose this method.
For certain diseases, place of living or the emphasis on a healthy diet proved to be an important
factor. Place of living gained importance with diabetes. Respondents living in the capital consider the
potential functional foods less suitable for its prevention. The importance of a healthy diet may play a
main role in the choice of products aimed at preventing a weakened immune system or osteoporosis.
Of all the diseases, we obtained the highest exp (B) value related to a weakened immune system.
Our results show that for the prevention of this disease, it is respondents with a higher level of
education who consider the consumption of functional foods the most suitable method.
We can make sense of the results in Tables 2–4 collectively. Comparing Tables 2 and 3, we can
highlight a weakened immune system, diabetes, and digestive problems as diseases that worry a
significant number of respondents (Table 2), for the prevention of which a higher ratio of the worried
would also choose functional foods (Table 3). Adding in the results of the binary logistic regression
Nutrients 2020, 12, 69 10 of 19

(Table 4) with the characteristics of respondents, the consumer segment which is more likely to consume
functional foods in order to prevent specific diseases can be properly outlined.
Thus, for the prevention of a weakened immune system, functional foods are considered the
most suitable by women, by those with a higher level of education, by the 36–50 age group, and by
those for whom a healthy diet is important. If a product manufacturer would like to communicate the
prevention of a weakened immune system—an aim towards which, according to our results, 39% of
those concerned and willing to spend would purchase functional food—to its future consumers, then it
is important for them to assess how to reach consumers with these characteristics. For the prevention
of diabetes, it is the 36–50 age group and those living in a place other than the capital who would give
an important role to functional foods. In connection with digestive problems, the over 36 age group
can be highlighted, as they consider the consumption of functional foods a relevant prevention method
for this disease.

3.4. Random Forest with the Aim of Identifying Factors that Influence Consumer Decision
The results were also analyzed with the random forest method. Our aim was to identify the
demographic variables that most influence the decisions of consumers who would choose functional
food products to prevent diseases. The analysis was conducted by including all 10 variables.
In Table 5, only three variables are presented, namely those with the highest MeanDecreaseGini value.
The importance of the other variables is summarized in Appendix B.
The method worked with the lowest error rate in the case of mood disorders/apnoea, migraine,
joint diseases, and respiratory diseases for the grouping of respondents and the prediction of their
choice of functional food products to prevent diseases. During the statistical analysis, we examined
the importance of the 10 variables summarized in Table 1 with the help of MeanDecreaseGini. The age
and education of the consumer have a stronger influence during prediction. Age was more important
than the other factors for 12 of the 13 examined diseases, whereas in one case, education was more
important. Consumers’ place of living emerged among the first three most important influencing
factors for nine diseases, whereas in two cases, consumers’ BMI rating and perception of income, and
in one case, the identity of the primary grocery shopper, were among the most important influencing
factors. Contrary to previous results [30–33], which found that an important factor in the consumption
of functional foods is the gender of the consumer, in our analysis this factor was not among the top
three factors influencing prediction for any of the diseases.
During the triangulation process, we also compared the results obtained via the random forest
method with those of the regression analysis in order to arrive at a more accurate and justified result in
terms of the most relevant target group for disease prevention through functional foods. The ExpB
values of the regression analysis and the MeanDecreaseGini values of the random forest analysis
complete each other such that the value of importance indicates the most influential factor in the
prediction of consumer choice, while the odds ratio makes it more precise.
This information helps to identify the factors worth emphasizing for individual diseases. In the
case of weakened immune system, education is the most important factor for making predictions, and
based on the regression analysis, those with tertiary education can be an important target group.
In terms of the age of the consumer, five diseases can be highlighted among those for which
age was the most important factor when making predictions. During product development and the
preparation of marketing communication, companies should focus on the 36–50 age group for the
prevention of high cholesterol level, digestive problems, diabetes, and osteoporosis, the 25–35 age
group for the prevention of memory lapses and concentration problems and the 36 or older age groups
for digestive problems.
Nutrients 2020, 12, 69 11 of 19

Table 5. Results obtained with the help of random forest.

Relative Importance
Disease OOB Estimate of Error Rate
MeanDecreaseGini
age 23.52
Cardiovascular disease 33.66% grocery shopper 19.92
education 19.40
age 27.12

High cholesterol level education 22.02

40.89%
subjective estimation of income
21.54
position
age 21.88

Cancer 38.11% education 20.85

place of living 19.60
age 20.15
Mood disorders/Apnoea 26.15% place of living 17.57
education 17.19
age 20.74
Respiratory diseases 31.15% education 18.90
place of living 18.22
age 24.67
Skin diseases, eczema 33.38% education 19.89
place of living 19.89
age 22.18
Memory lapses, education 19.78
34.08%
concentration
place of living 18.56
age 28.10
Digestive problems 46.73% place of living 24.55
education 23.90
age 20.34

Migraine education 16.70

28.09%
BMI category 16.57
age 24.66
place of living 22.16
Diabetes 45.48%
subjective estimation of income
21.48
position
age 23.96
Joint diseases 30.88% education 21.71
BMI category 20.18
age 24.53
Osteoporosis 36.68% education 19.66
place of living 19.32
education 23.12
Weakened immune
39.64% age 23.00
system
place of living 21.40
Notes: 70% of the sample was used to train, 30% was used to validate the models; accuracy, sensitivity and specificity
values of models can be found in Appendix C.

4. Discussion
In our study we investigated which of the most dangerous non-communicable diseases worry
consumers the most and also what exactly people would spend on if they are concerned about
Nutrients 2020, 12, 69 12 of 19

a disease and would make a monetary sacrifice to avoid it. We examined 13 diseases and four
prevention methods.
The conclusion of our research is that most consumers are aware of the risks of non-communicable
diseases and they are worried about them, and, depending on the disease, only a small group of the
concerned is not willing to make a financial sacrifice to avoid them.
Our results show that with most diseases, lifestyle change is the most favored prevention
method of consumers, followed by, depending on the disease, the consumption of functional foods or
dietary/nutritional supplements. Except for a few cases, medication as a treatment method is the least
sought-after, and the most divisive option for consumers.
Lifestyle changes are the most preferred methods for most of the diseases, such as for the
prevention of cardiovascular diseases and unbalanced mood/apnoea and a high level of cholesterol,
while dietary supplements are chosen to prevent a weakened immune system. It is important to point
out that with several diseases, the smallest ratio of respondents chose dietary supplements. Further
results indicate that consumers are the most concerned about cancer-related diseases. This is the only
disease where, besides lifestyle changes, consumers consider taking medicaments as the best option.
Consumers did not consider the consumption of health-protecting food products the best solution
in any of the cases. Even if not the best option when trying to prevent/treat diseases, it is still a
relevant choice when trying to avoid digestion problems, a weakened immune system, high level of
cholesterol, diabetes, and cancer. In his research, Kraus [19] drew up an order of importance regarding
the health effects of a product. Some of his results coincide with the order of importance we arrived at.
Strengthening the immune system and the prevention of cancer-related diseases as motivating factors
for the consumption of functional foods also appeared at the top position in that research, whereas
improving memory was ranked lower.
In order to utilize the characteristics of respondents to identify the groups that can be most
associated with the specific prevention methods, we conducted binary logistic regression. In the course
of the regression analysis, we investigated the 10 examined variables to identify the groups we can
emphasize within each variable for each disease. Altogether we obtained multiple significant results in
relation to age (the 36–50 age group), completed education (college/university), place of living, and the
primary grocery shopper of the household.
Based on the analysis done with the random forest statistical method, we can state that the choice
made by consumers participating in the survey can be predicted with an error rate of 26%–47%. Among
the examined variables, age and education showed greater importance; also, with some diseases, place
of living, BMI rating, and the identity of the primary grocery shopper in the household played an
important role in the prediction of decisions.
In accordance with the principle of triangulation, we examined our aim from several perspectives.
Comparing the results of the regression and random forest, we can more firmly state that the completed
education of respondents plays a bigger role in the prediction of the choice for functional food products
in the prevention of a weakened immune system; more specifically, this prevention method is relevant
for those with tertiary education. Age plays a more important role in predicting whether respondents
purchase functional foods or not to prevent high cholesterol level, digestive problems, diabetes,
osteoporosis, memory lapses and concentration problems. For the first four diseases, the 36–50 age
group is the relevant target, for memory lapses and concentration problems the 25–35 age group is the
relevant target and for digestive problems, the 36 or older age group is the relevant target.
Numerous literature sources [17,30–35] emphasize the important role of gender, age, and completed
education as key demographic factors influencing the consumption of functional foods. However, by
looking at such products from a different angle, apart from age and completed education, the results of
our research also confirm the effects of other influencing factors, such as place of living, BMI rating,
and the perception of income. Moreover, our results also show that in relation to disease prevention,
the gender of the consumer does not have a significant effect on choosing functional foods.
Nutrients 2020, 12, 69 13 of 19

These new pieces of information are filling a long-standing information gap. Our results help
functional food manufacturers with the identification of diseases for which the effects of their products
are worth emphasizing, and with the identification of consumer target groups. The results of the
research are useful for experts working in the field of product development and for marketing
communication professionals as well. We recommend that both functional food developmental
research and manufacturing companies’ research and development (R&D) departments, as well as
their marketing experts devote their resources to developing and communicating about products
helping to prevent diseases that have the highest consumer acceptance. It would also help to facilitate
all of this if this viewpoint—that the planned functional food development should aim at preventing
a disease that consumers would like to prevent by consuming functional foods—was included as a
basis for the assessment of international or national grants. This is how we can expect functional food
consumption to increase significantly and play a more important role in improving society’s health.

Limitations
This research is the very first step towards answering a complex question. Due to the limited
scope of the analysis, some of the steps in this research were aimed to be simple. Future research
should include the analysis of the threat of multi-collinearity as well.
Although the current health conditions of the participants and their family can influence their
answer to the given questions, we did not address this problem. Future research could focus on this
aspect as well.
The answer we have received is not a general one, as Hungary is a small country in Europe. For a
firmer, more widely accepted answer, international research might be useful.

Author Contributions: Conceptualization, Á.T.; methodology, Z.L., B.P., and Á.T.; formal analysis, B.P.;
investigation, G.K. and Á.T.; writing—original draft preparation, B.P. and Á.T.; visualization, B.P.; supervision,
Á.T. and Z.L.; funding acquisition, G.K. and Z.L. All authors have read and agreed to the published version of
the manuscript.
Funding: The project is supported by the European Union and co-financed by the European Social Fund (grant
agreement no. EFOP-3.6.3-VEKOP-16-2017-00005).
Conflicts of Interest: The authors declare no conflicts of interest.

Appendix A

Table A1. Goodness of fit tests of model.

Would Buy Functional Food to Hosmer–Lemeshow

Cox–Snell R2 Nagelkerke R2
Avoid the Specific Disease Test (p-Value)
High cholesterol level 1.000 0.027 0.036
Cancer-related diseases 0.880 0.059 0.082
Skin diseases 1.000 0.023 0.033
Memory lapse/concentration
0.892 0.064 0.093
problems
Digestion problems 0.951 0.074 0.099
Migraine - 0.044 0.069
Diabetes 0.014 0.052 0.071
Joint disease 0.991 0.040 0.057
Osteoporosis 0.997 0.073 0.102
Weakened immune system 0.913 0.102 0.138
Nutrients 2020, 12, 69 14 of 19

Appendix B

Table A2. Further results obtained with the help of random forest.

OOB Estimate of Relative Importance

Disease
Error Rate MeanDecreaseGini
Place of living 18.98
Gender 13.41
Presence of a child under 18 in the
9.98
household
Cardiovascular disease
Subjective estimation of income position 18.95
It is important for me to do something
7.73
for my health
I pay attention to healthy eating 10.20
BMI category 19.34
Place of living 20.95
Gender 14.04
Presence of a child under 18 in the
11.63
household
High cholesterol level
Grocery shopper 18.62
It is important for me to do something
8.50
for my health
I pay attention to healthy eating 12.51
BMI category 20.21
Gender 11.85
Presence of a child under 18 in the
11.74
household
Grocery shopper 18.59
Cancer
Subjective estimation of income position 19.83
It is important for me to do something
7.71
for my health
I pay attention to healthy eating 11.61
BMI category 18.90
Gender 9.29
Presence of a child under 18 in the
9.91
household
Grocery shopper 15.55
Mood disorders/apnoea
Subjective estimation of income position 14.61
It is important for me to do something
7.06
for my health
I pay attention to healthy eating 9.45
BMI category 16.66
Nutrients 2020, 12, 69 15 of 19

Table A2. Cont.

OOB Estimate of Relative Importance

Disease
Error Rate MeanDecreaseGini
Gender 12.54
Presence of a child under 18 in the
11.13
household
Grocery shopper 17.81
Respiratory diseases
Subjective estimation of income position 17.54
It is important for me to do something
8.09
for my health
I pay attention to healthy eating 11.02
BMI category 16.42
Gender 13.10
Presence of a child under 18 in the
11.26
household
Grocery shopper 17.96
Skin diseases/eczema
Subjective estimation of income position 16.37
It is important for me to do something
7.68
for my health
I pay attention to healthy eating 11.05
BMI category 19.29
Gender 12.15
Presence of a child under 18 in the
11.65
household
Grocery shopper 17.54
Memory lapses/
concentration Subjective estimation of income position 17.64
It is important for me to do something
8.31
for my health
I pay attention to healthy eating 10.98
BMI category 17.87
Gender 15.24
Presence of a child under 18 in the
11.90
household
Grocery shopper 22.28
Digestive problems
Subjective estimation of income position 22.65
It is important for me to do something
9.65
for my health
I pay attention to healthy eating 11.22
BMI category 21.93
Nutrients 2020, 12, 69 16 of 19

Table A2. Cont.

OOB Estimate of Relative Importance

Disease
Error Rate MeanDecreaseGini
Place of living 16.51
Gender 12.04
Presence of a child under 18 in the
9.34
household
Migraine
Grocery shopper 15.10
Subjective estimation of income position 15.43
It is important for me to do something
7.33
for my health
I pay attention to healthy eating 10.01
Gender 13.82
Presence of a child under 18 in the
12.48
household
Grocery shopper 20.31
Diabetes
Subjective estimation of income position 21.48
It is important for me to do something
9.21
for my health
I pay attention to healthy eating 12.44
BMI category 19.70
Place of living 18.70
Gender 13.39
Presence of a child under 18 in the
11.81
household
Joint diseases
Grocery shopper 17.88
Subjective estimation of income position 17.87
It is important for me to do something
8.88
for my health
I pay attention to healthy eating 11.68
Gender 12.90
Presence of a child under 18 in the
11.75
household
Grocery shopper 18.58
Osteoporosis
Subjective estimation of income position 17.93
It is important for me to do something
8.80
for my health
I pay attention to healthy eating 10.15
BMI category 17.36
Gender 13.69
Presence of a child under 18 in the
12.29
household

Weakened immune Grocery shopper 20.62

system Subjective estimation of income position 20.45
It is important for me to do something
9.98
for my health
I pay attention to healthy eating 12.27
BMI category 19.13
Nutrients 2020, 12, 69 17 of 19

Appendix C

Table A3. Accuracy, sensitivity, and specificity of random forest models.

Disease Accuracy Sensitivity Specificity

Cardiovascular disease 0.6861 0.9227 0.1011
High cholesterol level 0.6246 0.8557 0.1944
Cancer 0.6214 0.88835 0.08738
Mood disorders/apnoea 0.7184 0.97778 0.02381
Respiratory diseases 0.7152 0.93191 0.02703
Skin diseases/eczema 0.7023 0.94619 0.06977
Memory lapses/concentration 0.6634 0.8711 0.1071
Digestive problems 0.5372 0.7596 0.2143
Migraine 0.7379 0.9619 0.0137
Diabetes 0.6375 0.8702 0.1584
Joint diseases 0.6828 0.9178 0.1111
Osteoporosis 0.6204 0.8820 0.1354
Weakened immune system 0.5761 0.8939 0.1385

References
1. Diepeveen, S.; Ling, T.; Suhrcke, M.; Roland, M.; Marteau, T.M. Public acceptability of government
intervention to change health-related behaviours: A systematic review and narrative synthesis. BMC Public
Health 2013, 13, 756. [CrossRef] [PubMed]
2. WHO. Noncommunicable Diseases. 2018. Available online: https://www.who.int/news-room/fact-sheets/
detail/noncommunicable-diseases (accessed on 14 November 2019).
3. WHO. World Health Statistics 2016: Monitoring Health for the SDGs, Sustainable Development Goals; WHO Press:
Geneva, Switzerland, 2016.
4. WHO. World Health Organization—Noncommunicable Diseases (NCD) Country Profiles 2018: Hungary.
Available online: https://www.who.int/nmh/countries/2018/hun_en.pdf?ua=1 (accessed on 14 November 2019).
5. Smith, F.; Newbould, J. Perspectives of community groups on the reclassification of medicines from
prescription to non-prescription. Int. J. Health Promot. Educ. 2002, 40, 21–30. [CrossRef]
6. Albright, C.L.; Schembre, S.M.; Steffen, A.D.; Wilkens, L.R.; Monroe, K.R.; Yonemori, K.M.; Murphy, S.P.
Differences by race/ethnicity in older adults’ beliefs about the relative importance of dietary supplements vs
prescription medications: Results from the SURE Study. J. Acad. Nutr. Diet. 2012, 112, 1223–1229. [CrossRef]
[PubMed]
7. Starr, R.R. Too little, too late: Ineffective regulation of dietary supplements in the United States. Am. J. Public
Health 2015, 105, 478–485. [CrossRef] [PubMed]
8. Diplock, A.T.; Aggett, P.J.; Ashwell, M.; Bornet, F.; Fern, E.B.; Roberfroid, M.B. Scientific concepts of functional
foods in Europe: Concensus document. Br. J. Nutr. 1999, 81 (Suppl. 1), S1–S27.
9. Regulation (EC) No 1924/2006 of the European Parliament and of the Council. Available online: https://eur-lex.
europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:02006R1924-20141213 (accessed on 18 November 2019).
10. European Commission. Health Claims. Available online: https://ec.europa.eu/food/safety/labelling_nutrition/
claims/health_claims_en (accessed on 18 November 2019).
Nutrients 2020, 12, 69 18 of 19

11. Szakály, Z.; Kiss, M.; Jasák, H. Funkcionális Élelmiszerek, Fogyasztói Attitűdök És Személyre Szabott
Táplálkozás (Functional Foods, Consumer Attitudes and Personalized Nutrition). Táplálkozásmarketing 2014,
1, 3–17.
12. Alkhatib, A.; Tsang, C.; Tiss, A.; Bahorun, T.; Arefanian, H.; Barake, R.; Khadir, A.; Tuomilehto, J. Functional
foods and lifestyle approaches for diabetes prevention and management. Nutrients 2017, 9, 1310. [CrossRef]
13. Prättälä, R. Dietary changes in Finland—Success stories and future challenges. Appetite 2003, 41, 245–249.
[CrossRef]
14. Niva, M. ‘All foods affect health’: Understandings of functional foods and healthy eating among
health-oriented Finns. Appetite 2007, 48, 384–393. [CrossRef]
15. Szakály, Z.; Berke, S.; Biacs, P.; Szakály, S.; Széles, G.; Szente, V.; Szigeti, O. Funkcionális élelmiszerek piaci
helyzetének vizsgálata= Market Analysis of Functional Foods. OTKA Kutatási Jelentések|OTKA Res. Rep. 2010.
Available online: http://real.mtak.hu/12130/1/69336_ZJ1.pdf (accessed on 23 December 2019).
16. Van Kleef, E.; Van Trijp, H.C.M.; Luning, P. Functional foods: Health claim-food product compatibility and
the impact of health claim framing on consumer evaluation. Appetite 2005, 44, 299–308. [CrossRef]
17. Siró, I.; Kápolna, E.; Kápolna, B.; Lugasi, A. Functional food. Product development, marketing and consumer
acceptance—A review. Appetite 2008, 51, 456–467. [CrossRef]
18. Bagchi, D.; Nair, S. (Eds.) Developing New Functional Food and Nutraceutical Products; Academic Press:
Cambridge, MA, USA, 2016.
19. Kraus, A. Development of functional food with the participation of the consumer. Motivators for consumption
of functional products. Int. J. Consum. Stud. 2015, 39, 2–11. [CrossRef]
20. Baba, Y.; Kallas, Z.; Costa-Font, M.; Gil, J.M.; Realini, C.E. Impact of hedonic evaluation on consumers’
preferences for beef attributes including its enrichment with n-3 and CLA fatty acids. Meat Sci. 2016, 111,
9–17. [CrossRef] [PubMed]
21. Devcich, D.A.; Pedersen, I.K.; Petrie, K.J. You eat what you are: Modern health worries and the acceptance of
natural and synthetic additives in functional foods. Appetite 2007, 48, 333–337. [CrossRef] [PubMed]
22. Petrie, K.J.; Sivertsen, B.; Hysing, M.; Broadbent, E.; Moss-Morris, R.; Eriksen, H.R.; Ursin, H. Thoroughly
modern worries: The relationship of worries about modernity to reported symptoms, health and medical
care utilization. J. Psychosom. Res. 2001, 51, 395–401. [CrossRef]
23. Chen, M.F. The mediating role of subjective health complaints on willingness to use selected functional
foods. Food Qual. Prefer. 2011, 22, 110–118. [CrossRef]
24. Body Mass Index—BMI. Available online: http://www.euro.who.int/en/health-topics/disease-prevention/
nutrition/a-healthy-lifestyle/body-mass-index-bmi (accessed on 11 January 2019).
25. Meerza, S.I.A.; Yiannaka, A.; Brooks, K.R.; Gustafson, C.R. Information avoidance behavior: Does ignorance
keep us uninformed? In Proceedings of the Annual Meeting—Agricultural and Applied Economics
Association, Atlanta, Georgia, 21–23 July 2019; Available online: https://ageconsearch.umn.edu/record/290757
(accessed on 23 December 2019). [CrossRef]
26. Muchlinski, D.; Siroky, D.; He, J.; Kocher, M. Comparing random forest with logistic regression for predicting
class-imbalanced civil war onset data. Political Anal. 2016, 24, 87–103. [CrossRef]
27. Cutler, D.R.; Edwards, T.C., Jr.; Beard, K.H.; Cutler, A.; Hess, K.T.; Gibson, J.; Lawler, J.J. Random forests for
classification in ecology. Ecology 2007, 88, 2783–2792. [CrossRef]
28. Gupta, A.; Gupta, G. Comparative study of random forest and neural network for prediction in direct
marketing. In Applications of Artificial Intelligence Techniques in Engineering; Springer: Singapore, 2019;
pp. 401–410. [CrossRef]
29. Breiman, L. Random forests. Mach. Learn. 2001, 45, 5–32. [CrossRef]
30. De Jong, N.; Ocke, M.C.; Branderhorst, H.A.; Friele, R. Demographic and lifestyle characteristics of functional
food consumers and dietary supplement users. Br. J. Nutr. 2003, 89, 273–281. [CrossRef]
31. Urala, N. Functional Foods in Finland: Consumers’ Views, Attitudes and Willingness to Use; VTT Publications:
Espoo, Finland, 2005.
32. Kraus, A.; Annunziata, A.; Vecchio, R. Sociodemographic factors differentiating the consumer and the
motivations for functional food consumption. J. Am. Coll. Nutr. 2017, 36, 116–126. [CrossRef] [PubMed]
33. Meyerding, S.; Kürzdörfer, A.; Gassler, B. Consumer preferences for superfood ingredients—The case of
bread in Germany. Sustainability 2018, 10, 4667. [CrossRef]
Nutrients 2020, 12, 69 19 of 19

34. Verneau, F.; La Barbera, F.; Furno, M. The role of health information in consumers’ willingness to pay for
canned crushed tomatoes enriched with Lycopene. Nutrients 2019, 11, 2173. [CrossRef] [PubMed]
35. Bimbo, F.; Bonanno, A.; Nocella, G.; Viscecchia, R.; Nardone, G.; De Devitiis, B.; Carlucci, D. Consumers’
acceptance and preferences for nutrition-modified and functional dairy products: A systematic review.
Appetite 2017, 113, 141–154. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).