Keywords

1 Introduction

By 30 January 2020, the world was shocked by the epidemic declaration of a new coronavirus called SARS-CoV-2 (Covid-19) as a public health emergency of international concern, causing severe acute respiratory syndrome (Grasselli et al. 2020). On 11 March 2020, the World Health Organization (WHO) declared the Covid-19 outbreak a pandemic due to the steady increase in number of cases outside China (WHO 2020). Covid-19 disease is highly contagious and the transmission can occur through direct contact by sucking droplets from secretions that an infected patient expels when coughing or sneezing, through indirect contact with surfaces in the immediate environment or with objects used by the infected person (Ong et al. 2020). Clinical symptoms of Covid-19 patients include fever, cough, fatigue and gastrointestinal infection (Guo et al. 2020). The elderly and people with underlying diseases are susceptible to infection and prone to serious outcomes (WHO 2020).

In Mexico, the first reported case of Covid-19 was on February 27 2020, through a person with history of having traveled to Italy causing a national alert, which initiated phase 1 of the epidemiological contingency (SSA 2020). On March 24 2020, the Mexican government instituted phase 2, called community contagion, establishing preventive and restrictive plans by applying the national day of healthy distance, suspension of mass events and concentrations of people in confined spaces, as well as the voluntary isolation of people. On April 21 2020, with the disease present throughout the country and 2,772 confirmed cases, phase 3 was announced as mandatory suspension of non-essential activities in public, private and social sectors (www.coronavirus.gob.mx). In this quarantine context, the entire country at this moment is experiencing a severe economic crisis with more than 12 million Mexicans without receiving economic income (SSA 2020).

The state of Quintana Roo, which is the world’s leading national tourist destination and the eighth most popular tourist destination, reported on March 13 2020, its first case of Covid-19. Air, sea and land mobilization and migration movements with the countries of Belize and Guatemala were suspended and health surveillance became an element of strict order, guarantee and safety. Despite the most restrictive and temporary closing conditions, the number of infections has grown rapidly with an incidence rate of 339 cases per 100,000 inhabitants exceeding, the national average of 128 and compromising hospital occupation (Galera et al. 2020). This has led the local governments to adopt extraordinary, successful, and contingency plans by closing streets and avenues, and to apply the use of periphone, vehicle restriction and prohibition of the sale of alcoholic beverages without being able to stop the transmission chain. It is then imperative for the government and society to have methodologies to incorporate criteria, identified and weighted to support the local epidemiological surveillance system by improving decision-making to mitigate and control the spread of Covid-19.

Many researchers have established the use of MCDA and GIS as effective tools to identify and map areas of vulnerability and epidemic control (Qayum et al. 2015; Ahmad et al. 2017; Aboal-Viñas et al. 1999; Zemri and Hamdadou 2014; Honogh 2011). Vulnerability, is defined by Cardona (2003) as a community’s economic, social or political predisposition to destabilization by an external phenomenon, natural or man-made, was defined in this study as the community’s susceptibility to the risk of Covid-19 contagion. Among the different MCDA techniques, the most used is the Hierarchical Analytical Process (AHP) method developed by Saaty (1980). The basis of this proposal is that it allows numerical values to be assigned to the judgments made by individuals by allowing to quantify the contribution of each element of the hierarchy to the immediately higher level from which it is apparent (Berumen and Llamazares 2007). The AHP is very popular and widely used due to its easy understanding, interpretation and implementation of the results (Saaty 1978; 1979; 1980; Arulbalaji et al. 2019; Dar et al. 2020; Kaliraj et al. 2014; Kumar et al. 2018; 2020). The AHP uses paired comparisons by incorporating actor preferences between elements and uses a fundamental scale of 1, 3, 5, 7, 9 to incorporate judgments or assessments of the decision-making. This strictly positive scale allows to eliminate the ambiguities that human beings have when comparing elements in the proximity of zero or infinity (Moreno 2002). In addition, it allows to represent a wide range of decisions considering and combining qualitative and quantitative data, transforming a multidimensional problem (multicriterion) into a one-dimensional problem (scale of priorities) in which global outputs are represented (Assefa et al. 2018; Opricovic 2011; Saaty 2008).

On the other hand, the use of GIS tools applied to the health field is increasingly being used to examine the impact of distance, along with non-spatial factors to access and use health care services (Higgs 2004; 2009). According to National Center for Geographic Information and Analysis (1989) a GIS is an information system composed of hardware, software and procedures for capturing, handling, manipulating, analysing, modelling and representing georeferenced data, with the aim of solving management and planning and health problems (Aronoff 1989; Fernández et al. 2006). GIS is very useful for the study of the association between environment, location and health due to its ability to analyse and visualize data. They bring numerous advantages to practice for public health planning and research. They increase the possibility of identifying disease hotspots and linking them to social, cultural or environmental variables. They also identify high-risk groups by locating the most vulnerable spaces, as regards the social composition of the population (Suárez et al. 2020).

This work aims to evaluate and map the susceptibility of covid-19 risk to contagion in Chetumal, Mexico, by calibrating and applying a vulnerability model using Geographic Information Systems (GIS) and Multi-Decision Criteria Analysis (MCDA). The results of this study are expected to provide important scientific information for decision makers on the use of GIS tools to take pandemic prevention and control measures and serve as a reference for implementation in other cities or regions.

2 Study Area

The study area for this research was the City of Chetumal (Fig. 1), Capital State of Quintana Roo and municipality of Othón P. Blanco, whose name comes from Chactemal, which in Mayan language means place where red trees grow. It is located southeast of the Yucatan peninsula, at the geographic coordinates of \( 18^{ \circ } \,30^{\prime}\,13^{\prime\prime} \) north latitude and \( 88^{ \circ } \,18^{\prime}\,19^{\prime\prime} \) west longitude, covering an area of 42 km2 (Hernández et al. 2014; Quivén 2016). Chetumal is the third most important city in the State with a population of 190,791 inhabitants and a growth rate above the national average (1.0%) (COESPO 2019). It borders the east with the Bay of Chetumal, to the south with the Río Hondo that marks the border with the country of Belize and to the west with the Bacalar lagoon system. The communication routes are to the north by federal highway No. 307 Chetumal to Cancún and to the west by federal highway No. 186 Chetumal to Escárcega (PDU 2018). Most of its population extends north and east, locating itself in areas prone to floods and unsuitable for human settlement (Romero-Mayo 2012). Economic activities belong to the services, tourism, government administration, and self-consumption agriculture sectors (Fragoso-Servón and Pereira 2018).

Fig. 1.
figure 1

Location map of the study area

Full size image

The climate is warm sub-humid with rains in summer (INEGI 2009). The average temperature is 26 °C, the warmest month is in July and the warmest in December. The average annual precipitation is 1,300 mm and the highest precipitation is in summer. The study area is hurricane prone. The gleysol-type soil predominates, characterized by poor drainage, with vegetation cover of the upper jungle and medium sub-evergreen (INEGI 2016). Physiographically, the city is characterized by descending stepped reliefs from west to east with little elevation above sea level (Quivén 2016). The upper part of the city reaches nine mamsl and the lower zone two mamsl. Limestone is found superficially, traditionally known as flagstone. The water flow is underground and the water table is shallow. The geology of the city is made up of Tertiary and Quaternary sedimentary rocks, both periods of the Cenozoic Era (Sánchez Crispín 1980).

3 Methodology

In this study, GIS-based techniques and the MCDA methodology (GIS-MCDA) were used to assess and map the susceptibility of the risk to Covid-19 contagion by calibrating and applying a vulnerability model (Fig. 2). The calculation of the GIS-MCDA was developed considering four main stages: 1) definition of the problem and selection of essential criteria, 2) acquisition of data and spatial analysis in GIS, 3) MCDA through the AHP technique and calculation of the ratio of consistency, 4) MCDA-GIS procedure.

Fig. 2.
figure 2

Geographic Information System and Multi-Criteria Decision Analysis (GIS-MCDA) developed in this study using the Analytical Hierarchy Process (AHP) technique.

Full size image

3.1 Definition of the Problem and Selection of Essential Criteria

The first stage consisted of analyzing the risk of Covid-19 contagion in urban establishments and public places to determine the elements involved and classify them into essential criteria. The problem was structured using the Analytical Hierarchy Process (AHP) technique described by Saaty (2008). To calculate the Covid-19 contagion risk vulnerability index, the variable with the greatest weight or value within each dimension obtained according to the AHP method was selected.

According to the multi-criteria approach and with the participation of a transdisciplinary team of doctors and GIS specialists, nine essential criteria were selected. This includes population density, distance to banks, distance to second-level health care units, distance to shopping malls, distance to convenience stores, distance to markets, distance to government agencies and first-level care units, distance to terminals and distance to infected people with Covid-19. The description and justification of each criterion is presented in Table 1.

Table 1. Description and justification of the criteria used in the AHP.
Full size table

3.2 Data Acquisition and Spatial Analysis in GIS

The second stage consisted in (1) the acquisition of spatial information on Chetumal’s demographic, pandemic and infrastructure aspects and 2) the use of spatial analysis tools (kernel density, Euclidean distance, and reclassification) for the later stage of the MCDA-GIS procedure.

For the acquisition of information, each establishment was located using a Magellan Mobile Mapper 6 brand Global Positioning System (GPS) using the Universal Transverse Mercator (UTM zone 16 N) coordinate system. In addition, estimated population data in 2020 of Chetumal were obtained from the Facebook Connectivity Laboratory and Center for the International Earth Science Information Network (CIESIN 2016). The kernel density tool was used only to calculate the population density criterion. Conceptually, the goal of kernel density estimation is to calculate the density of points in a given area using the distance between the points, if and only if, the points have the same weight (Roig-Tierno et al. 2013). The search radius was 564.19 m (1 km2 area). Subsequently, the kernel density was classified into five categories using the reclassification tool, with 1 being the category with the lowest population density and 5 being the highest population density (Fig. 3a).

Fig. 3.
figure 3

Reclassified Map with Euclidean distance and kernel density a Population density b Banks c Second Level Health Care Units d Malls e Convenience Stores f Markets g Government Departments and First Level Care Units h Ground transportation g Infected people (Color figure online)

Full size image

For the remaining eight criteria, the Euclidean distance tool was used to describe the relationship of each cell to a set of origins based on the distance of the straight line (ESRI 2019). Among the functionalities of the Euclidean distance algorithm, its usefulness for determining the similarity between two things or pairs of data stands out. It is calculated from the center of the source cell to the center of each of the surrounding cells. Conceptually, it works as follows: for each cell, the distance to each source cell is determined by calculating the hypotenuse with X_max and Y_max. Subsequently, each criterion was reclassified into five categories using proximity distances to represent the risk vulnerability to Covid-19 contagion; five being the nearest distance and one the furthest distance (Table 2).

Table 2. Rankin and weights of the criteria used in Weighted Overlay
Full size table

3.3 MCDA Through the AHP Technique and Calculation of the Consistency Ratio

The third stage used the AHP method developed by Saaty (1980; 2000; 2008) to assign the weights of importance to each criterion. In this method weights can be derived by taking the priority vector of a square reciprocal matrix of pairwise comparisons between criteria; comparisons deal with the relative importance of the two criteria involved. The scores in the comparison matrix were assigned on a continuous scale of 1 to 9, also considering the reciprocals of those scores (Table 2). Score 1 represented equal importance, 3 moderate importance, 5 strong importance, 7 very strong importance and 9 absolute importance. The consistency ratio (CR), which measures the inconsistency level of the pairwise comparison matrix (i.e. the probability that the factor weights were randomized), was calculated using the mathematical ratio

$$ {\text{CR}}\,\, = \,\,{\text{CI}}/{\text{RI}} $$
(1)

Where CR is the consistency ratio; CI Consistency Index; RI Random Index, which indicates the consistency of a random matrix.

$$ {\text{CI}}\,\, = \,\,\left( {\lambda_{ \hbox{max} } \, - \,{\text{n}}} \right)/\left( {{\text{n}}\, - \, 1} \right) $$
(2)

CI Consistency Index; λ max is the maximum eigenvalue; n the dimension of the decision matrix.

The CR must have a value less than 0.1 (<10%) to indicate that the method is acceptable (Saaty 1980). The consistency ratio in this study was 0.06 (6%), being below the maximum consistency ratio recommended by Saaty.

3.4 MCDA-GIS Procedure

In the fourth stage, the Weighted Overlay tool was used to generate the vulnerability map model using the weights obtained from the calculation of the priority vector within the AHP. The value of the vulnerability model was classified into four categories: low vulnerability (value 1), moderate vulnerability (value 2), high vulnerability (value 3) and extremely high vulnerability (value 4).

The GIS through the ArcMap 10.7 ESRI (2019) software, were used as a tool for the development of this study. The maps were worked in raster format with a spatial resolution of 2 meters, allowing greater efficiency in the structure for overlapping maps.

4 Results and Discussions

Describing the first and second stages of the procedure, we achieve the geographic location and quantification of 699 establishments corresponding to eight essential criteria without considering the population density criterion. This includes 13 banks, 29 second-level health care units, 13 shopping centers, 121 convenience stores, 7 markets, 64 government agencies and first-level care units, 7 terminals and 414 infected people with Covid-19. The influence of the criteria involved in the propagation of the Covid-19 was quantified by reclassifying the essential criteria calculated with the kernel density and Euclidean distance into five categories using the ARCGIS 10.8 program.

Generally, urban density is defined by the building-to-land ratio or the floor area ratio (SAR et al. 2015). However, they are basically estimated through the use of data from cartographic servers that are not always available and are expensive. Therefore, in the present study, we have used the data from CIESIN (2016) with the help of GIS to map the population density or human settlements of Chetumal. According to Fig. 3a in the study area it is observed that the population density is high in the central north and west areas. The north central part is made up of popular colonies inhabited by people with a medium degree of marginalization, a relatively low socioeconomic level and workers employed in the informal sector (COESPO 2010). The western part with a high degree of marginalization was mostly characterized by irregular human settlements which are, in some cases, the only option for families in poverty to access land to build their homes. Population density is one of the most important indices because it not only indicates the construction density but also indirectly the daytime population corresponding to economic power. The raster layer resulting from the reclassification was integrated in stage four of this study.

Figure 3 a–i, illustrate the reclassification procedure using the results of the kernel density and Euclidean distance. In accordance with an exhaustive discussion by the transdisciplinary team and the support of Mexican Official Standards, 5 areas of influence were defined in each criterion, ranging from 0 to the maximum value, to establish restrictions on the risk of Covid-19 contagion. These restrictions were established by colors according to their magnitude of risk, with red being the most dangerous area and blue being the least risky area.

In addition, with the results of the GIS-MCDA procedure, we have performed a weighted overlay operation through the ArcGIS spatial analysis tool to obtain the areas of vulnerability to the risk of contagion of Covid-19 from Chetumal. In particular, nine weights were defined and calibrated using the multicriteria approach through the AHP technique (Table 3). The resulting influence was expressed as a percentage and the description of the results follows the classification parameters proposed in the first stage. Second Level Health Care Units (31%) including public and private hospitals and health centers was the most important criterion because they are the establishments most exposed to contagion due to the type of service they offer during the 24 h. In addition to the patient seeking late help, he comes when his symptoms are shortness of breath and oxygen, giving a possible case of Covid-19. This condition makes it more risky and increases the probability of being transmitted horizontally, to the point of causing saturation in hospital services, causing them to collapse due to the number of people seeking care without need, in addition to the relatives of hospitalized patients who are heavily exposed to contagion when meeting in the emergency and hospitalization areas without respecting prevention and hygiene protocols.

Table 3. Pairwise comparison matrix for assessing the weights
Full size table

Public markets (20%) were the second strongest criterion due to the strong concentration of people who visit them to do their shopping, in addition to the small spaces that exist in them. The third strongest criterion was people infected with Covid-19 (18%) due to the ability to transmit the virus to other people such as family and friends who do not respect prevention and hygiene measures. The criteria Banks (9%), Malls (7%), Government Departments and First Level Care Units (6%), Ground Transportation (5%), Convenience Stores (2%) and Population Density (2%) were the criteria with lower weight (weight) of importance without discredit its value in the general analysis.

Criteria 1 Banks 2 s Level Health Care Units 3 Malls 4 Conveniencie Stores 5 Markets 6 Population Density 7 Government Departments and First Level Care Units 8 Ground Transportation 9 Infected people with Covid-19. λmax = 9.75, CI = 0.09, RI = 1.45 (Saaty, 2008) CR = 0.06 < 0.1 (consistency is acceptable).

4.1 Covid-19 Contagion Risk Vulnerability

To show the potential of the model, the proposed GIS-MCDA approach was used to obtain four areas of Covid-19 contagion risk vulnerability as depicted in Fig. 4. The extremely high vulnerability category represented 0.03%, the high vulnerability category 2.55%, moderate vulnerability 26.75% and low vulnerability 70.68%. The model allowed defining two areas with extremely high vulnerability, the first, located in the city center at the intersections of Benito Juárez and Efraín Aguilar avenues, where establishments for public use are located such as the Museum of the city of Chetumal, hospital Morelos, Potable Water Commission of Quintana Roo and the private university UNID, and a second, at the intersections of Avenue San Salvador and Calle Agustín Olachea, where the Campestre and Montecristo clinics, Banorte and HSBC banks, Super Akí self-service store, are located. Oxxo convenience store, and the integral friendship park. Both areas are characterized by high vehicle-pedestrian traffic. The areas with high vulnerability are delimited in the Central colonies, Italia, Guadalupe Victoria, Solidaridad, Proterritorio, Sian Kaan I and II, Fovissste II, Payo Obispo and Pacto Obrero. It is easily foreseeable that all of them encompass most of the geolocated establishments in stage two of this study. The zone of moderate vulnerability is located around the areas of high vulnerability, however, due to the conditions of the criteria, it is possible to observe large areas scattered throughout the city, covering different urban sectors.

Fig. 4.
figure 4

Contagion risk vulnerability map for Covid-19 in Chetumal.

Full size image

Finally, the areas of low vulnerability presented a more stable pattern within the city and despite having the highest percentage of surface, the fact of not being close to or immersed in points of high confluence mean that these areas are not risky and can be passable by applying prevention measures and security established by local governments. The results obtained suggest for local administrators to define specific useful strategies to safeguard low risk and keep the economy active at the same time.

5 Conclusion

The present study was related to the determination of the vulnerability and the risk of contagion of Covid-19 in the City of Chetumal by using a multi-criteria approach using AHP and GIS techniques. The use of AHP allowed reducing the subjectivity of the evaluation criteria by evaluating each of the criteria independently. GIS integrated with the multi-criteria approach are shown as a simple and powerful tool for decision-making regarding pandemic prevention and control.

The vulnerability model allowed to identify two areas with extremely high vulnerability (0.03%), located in the city center and in an area collected by various services. The areas with high vulnerability (2.55%) were delimited mainly in popular colonies with high and medium degrees of marginalization, the areas of moderate vulnerability (26.75%) were located around the areas of high vulnerability and without a defined pattern and the areas of low vulnerability (70.68%) presented a more stable pattern in the perimeter areas of the city.

Health officials, heads of local governments, the scientific community and the general population can easily use the results of this study while the Health Emergency in our country continues. Furthermore, under the MCDA-GIS approach proposed in this study, the data can be easily updated and published using interactive maps and web tools available now. Identifying the geographical areas with the highest vulnerability to contagion will allow the implementation of more rigorous measures in those places where the greatest number of people are concentrated, minimizing the incidence and mortality rates related to this disease.