Associations between health factors and food insecurity. A multidimensional analysis with fuzzy set approach

doi:10.21203/rs.3.rs-2565852/v1

Background: Food insecurity, which is one of many factors to consider when studying loss of health, is an important problem in Mexico and many countries around the world. Much insight may be gained into how population health in Mexico is affected by studying these factors.

Methods: Two multidimensional indices were calculated, the first was the multidimensional health index, constructed using 6 factors: excess weight, diabetes, high blood pressure, symptoms of depression, kidney problems and heart problems, and the second was the food insecurity index with three levels, mild, moderate, and severe food insecurity. The multidimensional health index is constructed with the use of fuzzy sets, through the National Health and Nutrition Survey (ENSANUT 2018-19 - for its acronym in Spanish) in Mexico.

Results: The results show a close relationship between these indices and elucidate a specific profile of individuals in which there is a higher prevalence of health loss in relation to severe food insecurity.

Conclusions: This study aims to elucidate the current state of health of the population in Mexico using the most up-to-date data provided by the autonomous public body for statistics and geography to build a multidimensional panorama, as well as the close relationship between food security and the loss of health.

Public Health

Food safety

ENSANUT 2018-2019

Fuzzy sets

Multidimensional index

Food insecurity, generally characterized by factors such as poverty, inequality, hunger, malnutrition, lack of freedoms and human rights, among others, has been studied by multiple disciplines. Many different approaches, and a variety of methodological tools have been used for this undertaking, therefore, an abundance of knowledge is available on the subject. This has been used to implement projects to combat extreme poverty, malnutrition, and lack of access to food, as well as for the design of government strategies in the search for the reduction of food insecurity. Nevertheless, food insecurity continues to be a major social problem and continues to have a large impact on the population.

Food security (FS), popularized as a concept in the 1980s, with a focus on the contributing factors of poverty, has evolved into a point of reference, which can be used to improve food conditions for individuals. In this sense, food security is a situation that exists when all people have physical, social, and economic access to sufficient, safe and nutritious food that meets their dietary needs and food preferences for an active and healthy life, at all times [1].

The aim of the current study is to make the relationship between food insecurity (FI) and factors that enhance the deterioration of people's health, more evident. To achieve this objective, two specific indices were constructed: the multidimensional health index (MHI) and the food insecurity index (FII). This will facilitate the investigation of the current state of these two elements which afflict the Mexican population. In the past, the phenomenon of food insecurity has mainly been studied using food and/or nutritional factors [2].

The methodological analysis was constructed with two main objectives in mind: 1) To quantify the levels of food insecurity in Mexican households. This was achieved through the creation of a food insecurity index (FII) which shows the multidimensional behavior of the current level of food security 2) To present a multidimensional characterization between the socioeconomic factors through the use of a Multidimensional Health Index (MHI) and the Food Insecurity Index (FII).

Food Insecurity

The FII was created to quantify of the level of food insecurity of Mexican households using the Latin American and Caribbean Food Security Scale (ELCSA for its abbreviation in Spanish). This scale was created by characterizing the experiences of household members based on variables related to inadequate access to nutritional resources. The scale is represented through the food security module included in the ENSANUT 2018-19 survey, and comprises fifteen questions, it also distinguishes between households with children under 18 years of age and households in which all members are over 18 years of age.

The FII was adapted for this study to only include households with members over 18 years of age. Households were then categorized into three levels of food insecurity, according to the number of affirmative answers: Mild (1-3 answers), Moderate (4-6 answers), and Severe (7-9 responses).

Multidimensional Health Index

The MHI was presented by Flores, Hernández and González (2021) [3], it is based on the integration of a model based on fuzzy sets which were in turn created based on the involvement of four factors which impact upon people's health: diabetes, excess weight, high blood pressure and depressive symptoms. The MHI was adapted for the current study through the integration of six health related factors: 1) diabetes, 2) excess weight, 3) high blood pressure, 4) depressive symptoms, 5) heart problems and 6) kidney problems.

Construction of the six factors that made up the MHI:

Diabetes: Based on the question: Has a doctor told you that you have diabetes (or high blood sugar)? Answer options: a) yes, b) yes, during pregnancy (only women, also known as: “gestational diabetes”), c) no. Response #1 was used to establish prevalence, while responses 2 and 3 were discarded.
Excess weight: based on the question: Has a doctor/dietician/nutritionist ever told you that you are or were obese? Response options: a) yes, b) no - only answer 1 was considered. To complement the nutrition component of the survey, data on average weight and height were considered in the anthropometry section through the use of the body mass index (BMI) formula: kg/m2 allowing for individuals to be classified according to the World Health Organization (WHO) criteria into normal weight from 18 to 24.9 kg/m2, overweight from 25 to 29.9 kg/m2, grade 1 obesity from 30 to 34.9 kg/m2, grade 2 obesity from 35 to 39.9 kg/m2, and grade 3 obesity > 40 kg/m2.
High blood pressure: based on the question: Has a doctor told you that you have high blood pressure? Answer options: 1) yes, 2) no. Only answer 1 was considered.
Depressive symptoms: based on the question: Do you feel as if you cannot get rid of a sensation of sadness? Response options: a) rarely or never (less than a day per week), b) rarely or sometimes (1-2 days), c) a considerable number of times (3-4 days), d) all the time or most of the time (5-7 days). The prevalence was established by considering option 1 as not having depressive symptoms, while responses 2 to 4, were considered as having depressive symptoms.
Kidney Problems: based on the question: Has a doctor ever told you that you have a kidney disease? Answer options: a) yes, b) no. To establish the prevalence, response 1 was considered and response 2 was discarded.
Heart Problems: based on the question: Has your doctor told you that you have (or had) heart failure (weakening of the heart's pumping capacity, which causes edema in the feet, ankles and legs, tiredness and lack of air)?. Answer options: a) yes, b) no. To establish the prevalence, response 1 was considered and response 2 was discarded.

A fuzzy inference system (FIS) was designed for the construction of the MHI. This system is based on the theory of fuzzy logic and its mathematical materialization of fuzzy sets (Ragin, 2008) [4]. It was configured using the six factors being studied as input variables, and the state of health and/or loss of health was established as the quantifying and output variable, as shown in Figure 1. The construction of the MHI through the FIS integrated the fuzzy sets of the six factors being studied, and that of the output quantifying variable, as shown in figure 2. A system of rules and conditions was established for the correct operation of the FIS, which were validated through substantive knowledge. The MHI scale was determined by a range from 0-100 where the value 100 expresses the least favorable state and/or loss of health of each individual of the total being study.

The toolbox application of the MATLAB®️ Software was used for the processing of the data in the fuzzy system, and for the creation of the fuzzy sets.

Data source

The data source for the calculation of the FII and the MHI was the National Health and Nutrition Survey (ENSANUT for its abbreviation in Spanish) 2018-19 [5]. This survey is conducted at the national level in Mexico, it consists of a health and a nutrition component; A sample size of 50,000 households were included in the health component and 32,000 in the nutrition component, which represents 125.5 million inhabitants (under a probabilistic, stratified and conglomerate scheme). ENSANUT integrates demographic, socioeconomic, health and nutritional factors. A total of 9,426 records were used for the construction of the indices, with an expansion up to 6,765,390, given that these records represent the group of households in which all individuals are over the age of 18 and who had completed the survey for all the variables used.

Table 1 shows the socioeconomic and infrastructure characteristics of the households being studied, as well as geospatial conditions and the prevalence of FI divided into 3 levels: mild, moderate, and severe. Similarly, the study population is categorized in relation to alcohol and tobacco consumption habits, with the purpose of incorporating these elements into the discussion of calculating the MHI.

An average of 59.85% of individuals were categorized as having mild FI, while 23.94% had moderate FI and 16.21% were considered as having severe FI. Figure 3 shows the spatial distribution of individuals considered as having severe FI based on Federal State. The states of Tabasco and Sonora were found to have the highest percentages of severe FI, while San Luis Potosí had the lowest percentage of individuals with severe FI.

Table 1. Descriptive statistics of the households analyzed.

	n	Expanded N (X1000)	%(CI95%)
Household conditions
Water in the house	9081	6344	65.56 (64.20-66.91)
Electric light in the house	13054	9271	99.41 (99.19-99.63)
Drainage connected to the house	8653	6491	64.09 (62.72-65.46)
Head of household education
None	437	324	3.32 (3.02-3.63)
Basic Education	7648	5363	58.23 (57.36-5905)
Upper Secondary Education	2905	2120	22.11 (21.14-22.82)
Higher education	2004	1409	15.25 (14.64-1587)
Postgraduate	146	105	1.11 (0.93- 1.29)
Area
Urban	8721	6366	66.37 (65.56-67.18)
Rural	4419	2957	33.63 (32.82-34.44)
Region
North	2863	1674	21.79 (21.08-22.49)
Center	4861	3397	36.99 (36.17-37.82)
Mexico City	410	1025	3.12 (2.82-3-42)
South	5006	3225	38.10 (37.27-38.93)
Income by quartiles
Quartile 1	6386	4243	47.75-49.95)
Quartile 2	398	339	3.03 (2.7-3-32)
Quartile 3	60	42	0.46 (0.34-0.57)
Quartile 4	109	81	0.83 (0.67-0.98)
Food Insecurity Level
Mild	2820	2176	59.85 (58.65-61.25)
Moderate	1128	843	23.94 (22.72-25.16)
Severe	764	559	16.21 (15.16-17.27)

Source: own elaboration with data from ENSASUT 2018-2019

Table 2 shows the results of the multidimensional analysis for both the MHI and the FII. In the case of the MHI, a greater loss of health was evident among women compared to men. Furthermore, a relationship was found between increasing levels of education and lower MHI scores, that is, the higher the level of education, the lower the mean MHI value. The availability of access to socialized and private medical services is relevant, given that, in both cases, the average value of the MHI is higher than that among individuals with no access. On the other hand, a greater degree of health loss is also evident for individuals with alcohol and tobacco consumption, as well as for those located in the northern and southern regions of the country.

For its part, a relationship can be seen between the FII and MHI given that individuals who suffer from severe food insecurity tended to have higher MHI scores. Furthermore, severe food insecurity was more prevalent among woman than among men. An increased level of education was also found to inversely impact the prevalence of severe food insecurity.

Though most individuals living with food insecurity can be classified into the mild food insecurity category, the prevalence of severe food insecurity follows a distribution similar to that of the MHI, which suggests a relationship between these two indices.

Table 3 shows the relationship between the MHI and the FII, making evident the relationship between high levels of food insecurity with increasing MHI scores. This allows for the creation of arguments regarding the close relationship between factors impacting the loss of health and factors which impact the access to food, thus showing the importance of studying this relationship more broadly.

Table 2. Multidimensional characterization of the MHI and FII.

	MSI mean	95% confidence interval of the difference		Food Insecurity Index
		Lower	Upper	Mild FI	Moderate FI	Severe FI
Sex
Male	33.94	(33.49	34.27)	68.30%	19.10%	12.60%
Female	38.22	(37.72	38.71)	58.10%	24.90%	16.99%
Level of Studies
None	43.75	(41.40	29.49)	38.00%	33.10%	28.90%
Basic Education	40.53	(39.72	41.36)	54.80%	26.60%	18.60%
Upper Secondary Education	35.96	(35.41	36.50)	65.30%	20.80%	13.90%
Higher education	32.49	(31.82	33.16)	81.60%	13.30%	5.10%
Postgraduate	33.44	(30.96	35.92)	89.20%	10.80%	0.00%
Access to Medical Service
With Access	33.94	(33.34	34.55)	69.40%	17.80%	12.80%
Without Access	35.86	(35.33	36.35)	57.50%	24.50%	18.00%
Access to private health insurance
With Access	34.16	(32.60	35.72)	83.90%	8.40%	7.70%
Without Access	35.3	(34.89	35.72)	59.80%	23.40%	16.90%
Population Type
Urban	36.58	(36.15	37.00)	61.50%	22.50%	16.00%
Rural	35.92	(35.34	36.51)	56.90%	26.50%	16.60%
Regions
North	37.2	(36.43	37.97)	65.70%	19.60%	14.80%
Center	35.23	(34.70	35.75)	63.30%	22.90%	13.80%
Mexico City	36.94	(34.99	38.89)	62.40%	21.30%	16.30%
South	36.96	(36.39	37.53)	53.50%	27.30%	19.20%
Income by quartiles
Quartile 1	37.5	(36.45	37.98)	57.80%	24.60%	17.60%
Quartile 2	35.13	(34.21	35.15)	87.70%	7.90%	4.40%
Quartile 3	34.33	(32.72	35.62)	85.50%	10.50%	3.90%
Quartile 4	33.24	(33.14	34.35)	100.00%	0.00%	0.00%
Consumption of Alcoholic Beverages
No	37.24	(36.60	37.89)	64.20%	20.20%	15.60%
Yes	35.11	(34.58	35.64)	59.30%	24.10%	16.60%
Smoking
Some days	36.41	(35.12	37.69)	63.60%	15.20%	21.20%
Does not smoke	36.19	(35.74	36.64)	57.90%	23.90%	18.20%
Every Day	33.96	(33.19	34.80)	59.90%	24.20%	15.80%
Age
20-30	30.63	(30.19	31.57)	65.10%	20.70%	14.20%
31-40	33.87	(33.26	34.40)	59.60%	23.70%	16.70%
41-50	39.13	(38.32	39.85)	56.70%	27.20%	16.20%
Over 50	46.34	(35.34	47.34)	57.10%	23.60%	19.30%

Source: Self-made with data from ENSASUT 2018-2019.

Table 3. Relationship matrix between the MHI and the FII

	Mild FI	Moderate FI	Severe FI
IMS	20.32	23.92	27.62

Source: Self-made with data from ENSASUT 2018-2019.

One significant finding in this study is the relationship between the studied socioeconomic factors and levels of food insecurity and loss of health. The relationship between the "severe" level of food insecurity with a high level of health loss is also evident, that is, the greater the degree of food insecurity, the greater the loss of health, as expressed by the FII and the MHI, respectively.

One major contribution of this study is the multidimensional analysis of the food insecurity phenomenon, ensuring that it is studied and understood in such a manner as to be related to public health. Although food security has been analyzed from different approaches, studies such as those by Mundo-Rosas et, al. [6], Félix, Aboites and Castro [7], conclude with arguments which state the importance of analyzing food security in a comprehensive manner. This implies the consideration of the many different dimensions related to food insecurity, from access and availability to the consumption of adequate and healthy food.

The scientific literature related to the issue of food insecurity shows patterns that justify the findings of the current study. Food insecurity has steadily decreased between 2012 and 2018 with 14% of the Mexican population showing improved food security, increased wellbeing, and higher levels of education [8]. These results are similar to those found in the current study.

Therefore, if a profile were to be created which could encompass individuals for whom the FII would indicate a "severe" level of food insecurity, solely based on the characteristics of the heads of households, this would include: any individuals with less academic training, without access to medical services, whether public or private, living in rural areas, over the age of 50 and having an income within the first quartile. These characteristics were found to be associated with sever levels of food insecurity.

Some studies have investigated the phenomenon of food insecurity based upon its relationship to the limitation of access to food, that is, a component of poverty and the lack of economic resources for its acquisition. Studies such as those of Alvarado, Zunzunegui and Delisle [9]; Pico and Pachón [10], which investigate the elements that intervene in the prevalence of food insecurity, place a lack of economic resources among the main contributing factors.

Particular traits are evident, in the case of the FII a greater prevalence of "severe" FI is seen among rural households, though not for the MSI, this has been documented through studies which have found that the Covid-19 pandemic has aggravated the risk of food insecurity for rural residents [11].

The geospatial representation of the population is a relevant part of the FII, given that certain states at the northern and southern extremes of the country were found to have the highest prevalence of severe food insecurity. This increased prevalence of severe food insecurity in the southern states of the country has also been documented in previous studies such as those shown by Díaz, Sánchez and Díaz [12] and [13].

The integration of methodological strategies such as the construction of multidimensional indices built from fuzzy sets, as well as the approach around the correlation of health loss, measured using the MHI, with the increase in food insecurity, measured using the FII, is highly important. The relationship between these indices facilitates the creation of profiles for individuals and/or households to detect those at risk of food insecurity or health related issues.

Likewise, it becomes evident that people's economic standing is not the only factor, and perhaps,, is not the most relevant factor for classifying the level of food insecurity within a population. Food insecurity is a combination of socio-economic and consumption factors, along with other health related variables which all contribute to the overall level of food insecurity within a certain group of individuals.

There is great value in the analysis of the relationship between the factors that intervene in the loss of health and food insecurity. This analysis permits for the identification of factors and key elements for the design of institutional and/or governmental strategies which can be used for the improvement of people's health and nutritional states.

The relationship between the multidimensional health index and the food insecurity index is significant for future analyses, that is, the positive relationship between them allows us to understand the conditions in which the population in Mexico finds itself. Likewise, it is of great importance to highlight the use of alternative methodological tools and techniques, such as the use of fuzzy sets, in the multidimensional analysis of the impact that certain factors have upon an individual´s health.

The multidimensional exploration of the relationship between loss of health and lack of access to basic nutrition (i.e. food insecurity) in specific regions of the country allow for future case studies to investigate this phenomenon.

ENSANUT : National Health and Nutrition Survey
FS: food security
FI: food insecurity
MHI: multidimensional health index
FII: food insecurity index

Ethics approval and consent to participate: Not applicable

Consent for publication: Not applicable

Availability of data and materials: The datasets generated and/or analysed during the current study are available in National Institute of Statistic and Geography (México) https://www.inegi.org.mx/programas/ensanut/2018/

Competing interests: The authors declare that they have no competing interests

Sources of funding: Not applicable

Author’s Constributions: LFP as sole author contributed to the design of the study, was also in charge of performing the statistical analysis and read, reviewed and approved the final manuscript.

Acknowledgements: I thank Professor Diana Mercedes Hernandez Corona for her advice on public health. We thank Assen Ognianov Iantchoulev, Executive Editor for Scientific Communications, for the English editorial assistance.

Author´s information: Not applicable

Sandhu A. National Food Security Actand Food Security Outcomes in India. Vision: The Journal of Business Perspective. 2014; 18(4), 365–370. https://doi.org/10.1177/0972262914552174
Food and Agriculture Organization of the United Nations (FAO). The state of food insecurity in the world. 2013. Retrieved from: http://www.fao.org/docrep/018/i3434e/i3434e.pdf/
Flores L, Hernández, D, González T. Multidimensional analysis of health in Mexico: implementation of fuzzy sets. BMC Public Health, 2021; 21(1), 1–13. https://doi.org/10.1186/S12889-021-10988-2/FIGURES/6
Ragin C. Redesigning Social Inquiry. Fuzzy Sets and Beyond. Chicago: University of Chicago Press. 2008.
INEGI INPS. Encuesta Nacional de Salud y Nutrición INEGI. 2018. Retreived from: https://www.inegi.org.mx/programas/ensanut/2018/.
Mundo V, Unar M, Hernández F, Pérez R, Shamah T. Food security in Mexican households in poverty, and its association with access, availability and consumption. Salud Publica de Mexico, 2019; 61(6), 866–875. https://doi.org/10.21149/10579.
Félix G, Aboites G, Castro D. Food security and its relation to availability and uncertainty of income: an analysis of household incomes. Acta Universitaria, 2018; 28(4), 74–86. https://doi.org/10.15174/au.2018.1757
Shamah T, Humarán I, Mundo V, Rodríguez S, Gaona EB. Factors associated with the change in food insecurity in Mexico: Ensanut 2012 y 2018-19. Salud Publica Mex. 2021 May 1;63(3):350–8.
Alvarado B, Zunzunegui, M, Delisle H. Food security and social support scale validation in an afro-Colombian population: application in a prevalence study of the nutritional state of children aged 6 to 18 months. Cadernos de Saúde Pública, 2005,21(3), 724–736. https://doi.org/10.1590/S0102-311X2005000300006
Fonseca S, Pachón H. Factors associated with food security in the rural northern municipality of Cauca, Colombia. Arch Latinoam Nutr. 2012;62(3):227–33.
Banna M, Sayeed A, Kundu S, Kagstrom, A, Sultana M, Begum M, et al. Factors associated with household food insecurity and dietary diversity among day laborers amid the COVID-19 pandemic in Bangladesh. BMC Nutrition, 2022, 8(1), 1–11. https://doi.org/10.1186/s40795-022-00517-8
Ángel M, Sánchez M, Díaz A. Food insecurity in Mexican states: a study on their major determinants. Economía, Sociedad y Territorio, 2016; 459–483. http://www.scielo.org.mx/pdf/est/v16n51/2448-6183-est-16-51-00459.pdf
Factors associated with food security in the rural northern municipality of Cauca, Colombia. (n.d.). Retrieved June 2, 2022, from http://ve.scielo.org/scielo.php?pid=S0004-06222012000300004&script=sci_arttext.

No competing interests reported.

Associations between health factors and food insecurity. A multidimensional analysis with fuzzy set approach

Status:

Version 1

Abstract

Figures

Introduction

Methods

Food Insecurity

Multidimensional Health Index

Data source

Results

Discussion

Conclusions

Abbreviations

Declarations

References

Additional Declarations

Status:

Version 1