A neighborhood-level analysis of association between social vulnerability and COVID-19 in ahvaz, Iran

Abstract

Social vulnerability and society's resilience are two concepts frequently used to examine the capacity of social systems to prepare, absorb, and adapt to environmental hazards and shocks. With the emergence of the COVID-19 pandemic, the role of social vulnerability in dealing with risks has gained renewed attention. Assessing social vulnerability can help managers and planners prioritize budgets, develop prevention programs, and enhance risk preparedness. This study aimed to determine the association between social vulnerability and COVID-19 in the neighborhoods of Ahvaz, Iran. To assess the social vulnerability of Ahvaz neighborhoods, decision-making techniques (best-worst method and weighted aggregated sum product assessment method) and geographic information systems were applied. Moreover, to investigate the relationship between social vulnerability and COVID-19 cases, the Pearson correlation test was used. The results showed that the ‘20-meteri shahrdari’ neighborhood has the highest level of social vulnerability, and the lowest level of social vulnerability among the neighborhoods of Ahvaz belongs to the neighborhood of ‘Shahrak Naft’. There is a low inverse association between the integrated index of social vulnerability and the incidence of COVID-19 per 1000 people in Ahvaz. By revealing the most important details at the neighborhood level and levels of vulnerability, the results can inform effective planning actions at the neighborhood level.

1. Introduction

The novel coronavirus (COVID-19) has become one of the most critical global challenges due to its rapid and widespread spread. The United Nations has identified COVID- 19 as a human, social, and economic risk (United Nations, 2021). Although the COVID-19 outbreak has impacted many communities, there are weaker and marginalized groups, particularly in developing countries, that have been disproportionately affected. Such communities have experienced higher rates of COVID-19 cases and mortality and experienced significant damages as a result of poor socio-economic conditions and limited access to infrastructure and services (Pierce et al., 2021; [1,2] (see Fig. 4).

Fig. 4.

Social vulnerability index (Q).

In the 1970s, the concept of social vulnerability was established as a paradigm for disaster management, going beyond the physical and material components necessary to deal with a disaster. This entails identifying and quantifying the socio-economic aspects that influence a society's resilience. A frequently used approach to social assessment is the index of social vulnerability, which reveals the susceptibility of individuals or communities in light of certain demographic and economic parameters [3]. Additionally, it refers to social and demographic features such as levels of education and employment and demographic structure that have a direct impact on society's ability to adapt to disasters and recover from stressful situations such as the spread of pandemics and natural disasters [4,5]. Indeed, social vulnerability is rooted in social structures that could determine the extent of susceptibility to risks and social problems and influence the level of human and material losses [6].

Research on social vulnerability to natural hazards has increased in recent decades. Following the spread of COVID-19, various researchers attempted to examine the relationship between COVID-19 and social vulnerability from different perspectives and discovered a strong association, particularly in poor and vulnerable communities [[7], [8], [9], [10]]. The emphasis of these studies has mainly been on the city and county scales. Analysis at the neighborhood level is needed to gain a more detailed understanding of the vulnerabilities and differential impacts [[11], [12]]. Evidence suggests that the design and structure of the neighborhood environment affect a variety of health outcomes and could exacerbate health inequities [13]. Due to the subject's importance, some studies have investigated the relationship between COVID-19 and neighborhood social vulnerability.

Lak et al. (2021) discovered that factors such as population density, and accessibility to neighborhood centers, pharmacies, and chain stores are associated with vulnerability to the COVID-19 pandemic at the neighborhood level (Lak et al., 2021). Pierce et al. [12] expressed that COVID-19 incidence and mortality are higher in racially and economically disadvantaged neighborhoods with widespread education, health, and income inequalities [12]. Similarly, many studies in different countries including the United States and the United Kingdom indicate that ethnic and racial neighborhoods are more likely than white neighborhoods to contract COVID-19 disease [[12], [14], [15], [16]]. Whereas in developing countries such as Brazil [17], Pakistan [1], and Iran (Lak et al., 2021), population density and the state of neighborhood access to infrastructure and services are the most critical indicators of social vulnerability.

Although previous studies have helped us understand associations between social vulnerability and COVID-19, there are still gaps. The first gap is that the majority of research on the association between social vulnerability and COVID-19 has been undertaken at the city, county, and country levels (PEALBA, 2021; [10,18]. As mentioned, neighborhoods and their structure could have many health consequences. Moreover, analysis at the neighborhood level could facilitate gaining more detailed and context-sensitive insights. It should also be noted that there are limited studies in this regard in Iran, and no studies have been conducted in Ahvaz, a city that has been hit hard by the pandemic. The next significant gap is regarding the methodology and the use of analysis tools. Most research has focused on regression and correlation and tools such as geographically weighted regression models [15,16]; Lak et al., 2021; [19]. It seems necessary to use mixed methods, and we use a combination of decision-making methods to weigh the indicators and assess the vulnerability of neighborhoods.

Ahvaz is a major city in southwest Iran that has been highly impacted by the pandemic. Moreover, it faces various problems as a result of a deficiency in urban planning and management. Thus, our research aims to examine the relationship between social vulnerability and COVID-19 infection in Ahvaz neighborhoods. To this goal, the research begins by identifying the indicators of social vulnerability that are related to the neighborhood scale. We then attempt to use these indicators to identify the level of vulnerability across different neighborhoods of Ahvaz. Finally, the study seeks to determine the correlation between COVID-19 rates and the level of social vulnerability. The results of this study could be used to develop planning strategies for enhancing the capacity to deal with pandemics and similar stressors at the neighborhood level in Ahvaz. Additionally, the approach adopted in this study could be used to examine associations between social vulnerability and the pandemic in other contexts with similar conditions. Accordingly, this study pursues these two objectives: assessing social vulnerability in the neighborhoods of Ahvaz and investigating the relationship between social vulnerability and COVID-19 prevalence in the neighborhoods of Ahvaz.

This study provides a spatial and multidimensional analysis of social inequalities in Ahvaz. These inequalities have increased the vulnerability of people to COVID-19. The findings of this study can contribute to effective decision-making and planning for the COVID-19 Management Center, urban management, support organizations, and facilitating institutions. Since this study examined an empirical example from Iran as a developing country using vulnerability indicators, its findings can contribute to extending the theoretical knowledge in this field. Besides, the techniques used in this study can be applied for statistical modeling at different neighborhood and suburban scales.

2. Materials and methods

2.1. Study area

Ahvaz, with an area of 19,494 ha, is located in southwestern Iran at an approximate altitude of 17 m (Fig. 1 ). According to the Statistical Center of Iran (2016), the population is about 1.4 million people. Ahvaz is separated into east and west parts by the Karun river. This city is one of the most polluted places in the world as a result of pollution generated by dust and the oil and gas industry. Ahvaz's climate is typically hot and dry in the winter and hot and humid in the summer. The average temperature ranges from 48 °C in summers to 4 °C in winters. Different ethnic groups, including Arabs, Persians, and Lors live in Ahvaz [20,21]; planning. ahvaz.ir, 2022; [22,23].

Fig. 1.

Location of ahvaz.

Passing five waves of the pandemic, Iran had 6,835,221 confirmed COVID-19 cases with 133,886 deaths as of February 15, 2022 [24]. Ahvaz is no exception given its large population and the daily influx of people from all over the province and the country. In a year)from October 2020 to September 2021 (, more than 75,000 people in Ahvaz have been infected with COVID-19.

2.2. Method

A mixed methods approach was adopted in this study. The general social vulnerability index was developed using multi-criteria decision-making methods. Moreover, correlation analysis was conducted to explore the association between the vulnerability index and the number of cases. The unit of analysis is the neighborhood. SPSS software and ArcGIS were used for statistical and spatial analysis, respectively.

2.3. Data collection method

Following the research design, various sources of data were used in this study to measure social vulnerability and the Covid-19 prevalence rate. To measure social vulnerability, existing and objective data and subjective data from experts’ opinions were used. The objective data were extracted from the statistical blocks of Ahvaz prepared by the Statistics Center. The raw data were converted into indexes and ratios. For example, the unemployment rate was calculated by dividing the unemployed population by the active population, or the dependency rate index was estimated by dividing the total of people over 65 years old and children under 14 years old by the active population. The data were used as an integrated dataset for Ahvaz and were calculated separately for each neighborhood using ArcGIS software. The subjective data were collected by surveying subject-matter experts and were used to weigh the research indicators.

The COVID-19 data were received from the Jundishapur University of Ahvaz as a center for managing and recording COVID-19 data in Khuzestan Province. The data were presented in raw form for each neighborhood. The correlation analysis was performed for the COVID-19 incidence rate per 1000 people. COVID-19 statistics have been collected for the period from September 2020 to August 2021.

2.4. The indicators

[25] stated that indicators help to understand, compare and improve a system; indicators are not fixed, and different researchers use various ones according to the purpose of the research [[26], [27], [28]]. In addition, indicators change with time and place due to their dynamic nature. However, even under the same temporal, spatial, and thematic conditions, the choice of indicators may differ based on the views of the researchers, criteria of analysis, etc. Concerning the same disasters and hazards, indicators may also change from region to region [26,29]. Since evaluating social vulnerability is dependent on the measurement of indicators, various indicators such as physical, demographic, economic, social, and health have been proposed for assessing vulnerability to COVID-19 (Lak et al., 2021; [13,15,16,30]. Based on what was said, we tried to select indicators that reflect the details of the relationship between social vulnerability and COVID-19 in the neighborhoods of Ahvaz. In the following, the selected indicators are briefly mentioned.

2.4.1. Unemployment

Unemployment is a major indicator of social vulnerability [[31], [32], [33], [34], [35]]. Employed persons typically have a higher standard of living and better access to health care, and given their stable employment conditions, they can cope better with economic restrictions. On the other hand, unemployed individuals are more vulnerable to disasters and risks due to a lack of financial assistance [36]. emphasized the significance of limited employment opportunities and unemployment as indicators of poor mental and community health, both of which contribute to increased social vulnerability. Moreover, at the neighborhood level, the presence of unemployed individuals and families without permanent jobs is an obstacle to social distancing and other health precautions such as the use of masks, sanitizers, etc., which in turn leads to the further spread of COVID-19 [15,26,37]; Pierce et al., 2021).

2.4.2. Literacy

Literacy is directly related to awareness of disasters and hazards. Enhanced education leads to better awareness and more effective coping strategies. Illiterate or less educated people are less likely to be adequately aware of measures needed to diagnose and prevent the disease. Moreover, people with low literacy levels have less life expectancy than educated people, which in turn leads to a reluctance to follow health practices. In addition, at the neighborhood level, families with low education cannot function effectively in terms of awareness and education of their children and usually do not want to use educational materials and books. This, in turn, could cause problems with disease control. Finally, people with low levels of education and literacy also lose the opportunity to use new ways of working from home, which in turn could lead to more exposure to the virus [15,37,38].

2.4.3. Population density

Some studies have argued that there is a positive correlation between population density and vulnerability, implying that a high population density results in a high level of vulnerability. This is explained by the fact that a higher concentration of people increases risk exposure. In this study population density is considered an indicator of vulnerability, as maintaining social distancing is expected to be more challenging in highly populated places, increasing the likelihood of pandemic spread. Some studies have argued that higher population density at the neighborhood level results in increased contact between individuals, which exacerbates vulnerability and Moreover, accelerates virus propagation [26,[39], [40], [41], [42], [43], [44]].

2.4.4. Household size

The household size index is another indicator of vulnerability to disasters and hazards. The larger the household size, the greater the likelihood of vulnerability. Large households usually have less access to social resources and lower per capita income. For this reason, if the number of people working in such households is small, they will not be able to cover the damages caused by disasters. In addition, large households are usually less likely to have access to quality housing due to their low economic capacity and live in crowded homes that share utilities such as water, sanitation, etc. This could lead to increased vulnerability because it is an obstacle to adhering to health protocols [15,16,26,37,45,46].

2.4.5. People aged over 65

Older people are more likely to have limited physical abilities. They may have hearing and vision problems and move slowly from one place to another. In addition, they lack advanced knowledge such as communicating with others and gaining awareness through the Internet. Although the exact age is not mentioned for the elderly, there is usually a consensus that people over 65 are more vulnerable. Additionally, evidence indicates that the elderly are more susceptible to infectious diseases due to their weakened immune systems. Thus, the high proportion of people aged 65 and over is considered an indicator of neighborhood vulnerability [26,[47], [48], [49]].

2.4.6. Female-headed households

Gender is a significant predictor of vulnerability. Women tend to have more restrictions, fewer job opportunities, and less access to resources. Such conditions, especially when women are heads of households, can lead to increased poverty and mortality due to increased economic, social, and psychological pressures. Thus, the high female population, especially female-headed households, is another indicator of vulnerability. Increasing the rate of female-headed households increases the level of social vulnerability [26,[50], [51], [52], [53]].

2.4.7. Dependency ratio

The dependency ratio is calculated based on the percentage of the population under the age of 16 and over the age of 65 and the physically disabled population relative to the population between the ages of 16 and 65. Families with a large number of dependents are likely to have limited financial opportunities. Therefore, they face many problems in dealing with risk and recovery. Moreover, dependents need more extra services for safety; thus, the presence of a large number of dependents in an area may put more constraints on existing services which in turn is a factor in increasing vulnerability [26,44,54].

2.4.8. Housing below 80 square meters

Small houses occupied by large families result in overcrowding and make it challenging to comply with social distancing measures. Thus, smaller houses may increase vulnerability and disease prevalence [55]; Lak et al., 2021 [41]; Credit, 2020).

2.4.9. Rental housing

People who live in rental housing do not have a stable situation or do not have the necessary financial ability to buy a house. In times of risk and disaster, such as COVID-19, rents usually rise, and such people face limited access to affordable housing. They either have to pay more or move to lower-quality houses. Moreover, in such circumstances, even if the exterior and interior of the house are unsuitable and unsanitary, they will not be able to repair the house without the landlord's permission. Thus, people who rent a home are more vulnerable than homeowners or those who have the financial means to rent quality housing [26,56]; Credit, 2020).

2.5. The analysis process

Following the objectives of the study, the research procedure was conducted in the following steps.

• Step 1The indicators for measuring vulnerability were developed based on a review of the literature and the context of the study. Then, the data related to each indicator were collected from the relevant organizations and the indicates were collected at the neighborhood level. The neighborhoods were considered as alternatives and indicators as criteria in the raw matrix for the WASPAS model.
• Step 2The best-worst method was used to calculate the weight of the indicators. To complete the best-worst questionnaire, 6 experts were surveyed. Three experts worked in the field of management and three experts were engaged in research activities. The first groups of experts worked in Ahvaz Municipality, Khuzestan Province Crisis Management, and Ahvaz County Health Department. The experts in the second group were from Shahid Chamran University of Ahvaz and Ahvaz Jundishapor University of Medical Sciences.

The BWM method was used as it provides consistent results and requires fewer pairwise comparisons [57]. Moreover, following the literature, we have used a combination of decision-making methods. The combination of decision-making methods (BWM and WASPAS) improves the accuracy of the results [58].

• Step 3The weights obtained in the previous step were used in the WASPAS model. The neighborhoods in the city were ranked using 9 indicators. Besides, the output index Q of the WASPAS model was calculated as a composite index of social vulnerability in each neighborhood. This index was used as a variable in the second step.
• Stage 2

  The second stage was performed in one step. The WASPAS output and the COVID-19 prevalence rate were entered into SPSS software and Pearson correlation was used to measure the correlations between the variables.

Pearson correlation was utilized in SPSS software to determine the correlation between the integrated social vulnerability index and the COVID-19 incidence rate in the second stage. The Pearson correlation has been widely used in the literature to evaluate the association between social vulnerability and COVID-19 [[59], [60], [61]]. If the scale of the variables is interval or ratio, and the distribution of variables is normal, parametric correlation tests such as the Pearson correlation test are used. The data collected at the neighborhood level are on a ratio scale. Using correlation analysis we explore whether the rate of COVID-19 is higher in neighborhoods with a high level of vulnerability. Fig. 2 shows the research flowchart (see Fig. 3 ).

Fig. 2.

Research flowchart.

Fig. 3.

Spatial representation of research indicators.

Multiple-criteria decision-making (MCDM) techniques assist decision-makers in challenging situations by making suitable and transparent decisions. These techniques also assist in determining the ranking of options and selecting the desired option considering different criteria [62]. Due to the nature of the social vulnerability, the use of MCDM is a popular method among researchers (Table 1 ).

Table 1.

A summary of social vulnerability studies that use MCDM techniques.

Reference	Method	Study Scope
[63]	Fuzzy TOPSIS-Fuzzy analytic hierarchy process	Measuring social vulnerability to COVID-19
[64]	FISM	Social vulnerability assessment
[65]	Rough Analytic Hierarchy Process	Assessment of social vulnerability to earthquakes
[66]	fuzzy AHP, AHP	Assessing social vulnerability to floods
[67]	(TOPSIS) and the Differences in differences (DID)	Assessing social vulnerability to natural disasters
[68]	TOPSIS	Measuring social vulnerability to COVID-19
[69]	AHP	Analysis of vulnerability to natural disasters
[70]	AHP and VIKOR	Evaluation of social vulnerability to earthquakes

In COVID-19-related studies, MCDM methods such as analytic hierarchy process (AHP), analytical network process (ANP), BWM, and vlse kriterijumsk optimizacija kompromisno resenje (VIKOR) have been used (Table 2 ).

Table 2.

A summary of COVID-19 studies using MCDM techniques.

Researcher (Year)	Method	Study Scope
[62]	BWM	Prioritize COVID-19 coping strategies
[71]	AHP	Municipal health care capacity in Brazil
[72]	AHP-VIKOR	Prioritization of patients with COVID-19
[73]	MABAC WASPAS	COVID-19 effects on countries' sustainable development index
[74]	Grey theory and GDEMATEL	Unemployment problem affected by the COVID-19

2.6. The best worst method

BWM is a method for comparing the best criterion to all other criteria and all other criteria to the worst criterion in Multiple Criteria Decision Making (MCDM). The purpose is to determine the ideal weights and consistency ratios by utilizing the comparison system's basic linear optimization model [75]. Numerous studies have been published in the literature that made use of this MCDM technique.

The following are the steps taken in BWM to determine the weight of each criterion [76,77].

• 1)Specifying the set of decision criteria $\left\{c_1,c_2,\dots ,c_n\right\}$;
• 2)Identifying the ideal and optimal criteria for use in decision-making:

Using their personal preferences, decision-makers rank the best and worst criteria identified in Step 1 in this stage. The best criteria represent the most critical factors for the decision, whilst the worst criteria represent the least significant criteria for the decision.

• 3)Determining the best criteria's preference over all other criteria:

This value is denoted by a number between 1 and 9 (1 being equally critical, 9 being extremely critical). $A_B=\left(a_{B1},a_{B2},\dots ,a_{Bn}\right)$ would be the resulting Best-to-Others vector. Where $a_{B\mathrm{j}}$ indicates a preference for criteria B (best criteria) over criteria j and $a_{B\mathrm{B}}=1$;

• 4)Identifying the criteria that every other criterion prefers over the worst criteria:

This scenario is also allocated a number between 1 and 9. The Others-to-Worst vector would be as $A_W={\left(a_{1W},a_{2W},\dots ,a_{nW}\right)}^T$. Where, $a_{\mathrm{j}W}$ denotes the criteria j's preference over the worst criteria W and $a_{\mathrm{W}W}$;

• 5)Calculating the ideal weights ($w_1^{*},w_2^{*},\dots ,w_n^{*}$):

By resolving (1), the appropriate weights for the criterion will be determined. To calculate the best weights for the criteria, it is necessary to minimize the largest absolute differences $\left\{\left|w_B-a_{Bj}{w}_j\right|,\left|w_j-a_{jw}{w}_w\right|\right\}$ for all j.

$$\min {\max }_j\left\{\left|\frac{w_B}{w_j}-a_{Bj}\right|,\left|\frac{w_j}{w_w}-a_{jw}\right|\right\}\mathrm{s}.\mathrm{t}.\sum _j{w}_j=1w_j\ge 0,\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{a}\mathrm{l}\mathrm{l}j$$ (1)

It is possible to solve this model by converting it to a linear programming formulation (2) [78]:

$$\min \xi \mathrm{s}.\mathrm{t}.\left|w_B-a_{Bj}{w}_j\right|\le \xi ,\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{a}\mathrm{l}\mathrm{l}j\left|w_j-a_{jw}{w}_w\right|\le \xi ,\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{a}\mathrm{l}\mathrm{l}j\sum _j{w}_j=1w_j\ge 0,\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{a}\mathrm{l}\mathrm{l}j$$ (2)

The ideal weights ${(w}_1^{*},w_2^{*},\dots ,w_n^{*})$ and the optimal value of ${\xi }^{*}$ are obtained by solving this problem. ${\xi }^{*}$ is described as the comparison system's consistency ratio. This means that the closer ${\xi }^{*}$ is to zero, the more consistent the decision-makers' comparison system is. Formula (3) is applied to ensure that the comparisons are consistent [79]:

$$ConsistencyRatio=\frac{{\xi }^{*}}{ConsistencyIndex}$$ (3)

Table 3 shows the consistency index. The lower the consistency ratio, the more reliable the comparisons.

Table 3.

Consistency index table [76].

$a_{BW}$	1	2	3	4	5	6	7	8	9
Consistency index	0.00	0.44	1.00	1.63	2.30	3.00	3.73	4.47	5.23

2.7. WASPAS method

In 2012, Zavadskas et al. developed the WASPAS technique. It combines the outputs of the Weighted Sum Model (WSM) with the Weighted Product Model (WPM) [80]. Alternatives are ranked using the value of the combined optimality criterion. The approach can confirm that alternative ranks are consistent during operation by conducting sensitivity analysis. WASPAS is applied in the following steps.

Step 1

The criteria (Cj) and alternatives (Ai) are established (j ¼ 1, …, n; i ¼ 1, …, m);

Step 2

The criteria weights are determined based on one of the MCDM methods;

Step 3

Eq. (4) and Eq. (5) are used to normalize the initial decision matrix, respectively, for benefit-based criteria (to be maximized) and cost-based criteria (to be minimized);

$$\overline{Xij}=\frac{Xij}{{\mathit{max}}_{i}Xij}$$ (4)

$$\overline{Xij}=\frac{{\mathit{min}}_{i}Xij}{Xij}$$ (5)

Step 4

Qi (1) is the initial total relative significance value, which is determined using the “Weighted Sum Model” in Eq. (6).

$$Q_i^{\left(1\right)}=\sum _{j=1}^n\overline{Xij}{w}_j$$ (6)

Step 5

Qi (2) is the second total relative significance value obtained through using the “Weighted Product Model” in Eq. (7).

$$Q_i^{\left(2\right)}=\prod _{j=1}^n{\left(\overline{Xij}\right)}^{wj}$$ (7)

Step 6

Qi is the total optimality value computed using Eq. (8). l is the combined optimality coefficient (λ $\in$ [0,1]). If the Weighted Sum and Weighted Product Model approaches have the same effect on the combined optimality criterion, then l is equal to 0.5.

$$Q_i=\lambda Q_i^{\left(1\right)}+\left(1-\lambda \right)Q_i^{\left(2\right)}$$ (8)

Step 7

Each alternative is ranked according to its combined optimality value (Qi). The alternative with the highest Qi value is the best one and is ranked first [81].

3. Findings

The outcomes of the indicator weighting, neighborhood ranking, and correlation analysis are discussed in this section. Table 4 shows the pairwise comparisons based on the views of six experts. The best pairwise comparisons based on the vector are shown in Table 4. According to experts, unemployment, dependency, and population density have the greatest effect.

Table 4.

Pairwise comparison for assessment criteria (Best to others).

Expert	BO	C1	C2	C3	C4	C5	C6	C7	C8	C9
Expert 1	Unemployment	1	3	3	5	2	5	3	8	9
Expert 2	Dependency ratio	3	4	5	5	3	5	1	5	8
Expert 3	Population density	5	7	1	3	4	5	2	7	9
Expert 4	Unemployment	1	5	7	5	3	2	2	9	5
Expert 5	Dependency ratio	3	7	5	5	6	7	1	7	9
Expert 6	Unemployment	1	6	7	6	3	5	2	7	5

Table 5 presents pairwise comparisons of the worst vector based on the views of six experts. The indices of housing less than 80 square meters, literacy, rental housing ratio, and population density have been determined to be the worst (lowest impact on social vulnerability). The indicators were then compared pairwise.

Table 5.

Comparison of assessment criteria on a pairwise basis.

Expert	Expert1	Expert 2	Expert 3	Expert 4	Expert 5	Expert 6
Wo	Housing below 80 square meters	Rental housing	Literacy	Housing below 80 square meters	Rental housing	Population density
C1	8	8	3	9	4	7
C2	5	6	1	5	5	4
C3	5	4	7	5	3	1
C4	7	4	3	7	5	4
C5	6	6	7	6	6	5
C6	4	3	5	7	4	6
C7	7	8	7	9	9	4
C8	1	3	2	1	4	3
C9	2	1	3	5	1	2

Table 6 indicates the weights of indicators based on the opinions of six experts. The unemployment rate index has the highest weight, while the rental housing rate index has the lowest (see Table 7 ).

Table 6.

Criteria weights.

Expert/Indicator	Expert1	Expert 2	Expert 3	Expert 4	Expert 5	Expert 6	AVERAGE weight
Unemployment rate	0.1178	0.3034	0.1954	0.1718	0.3608	0.1968	0.2243
Literacy	0.2667	0.13	0.0782	0.2659	0.1531	0.2983	0.1987
Households size	0.1767	0.13	0.0977	0.1145	0.0765	0.1312	0.1211
Rate of female-headed households	0.1178	0.078	0.2931	0.0491	0.0918	0.029	0.1098
Population density	0.0707	0.078	0.0782	0.1718	0.0656	0.0787	0.0905
Age over 65 years	0.0707	0.078	0.1303	0.0687	0.0918	0.0656	0.0842
Dependency ratio	0.1178	0.0975	0.0279	0.0687	0.0656	0.0656	0.0739
Rate of housing below 80 square meters	0.0225	0.078	0.0558	0.0209	0.0656	0.0562	0.0498
Rental housing rate	0.0393	0.0271	0.0434	0.0687	0.0292	0.0787	0.0477
E	0.087	0.087	0.098	0.078	0.098	0.095

Table 7.

Q Index and Ranking of Neighborhoods in terms of Social Vulnerability.

Rank	Neighborhood Name	Q	Rank	Neighborhood Name	Q	Rank	Neighborhood Name	Q
1	20-meteri shahrdari	0.364408	24	Kianababd	0.317589	47	Janbazan	0.285913
2	Sayahi	0.358403	25	Padshahr faz 1	0.312455	48	Rah-o-Tarabari	0.283907
3	Hasirabad	0.355865	26	-Kianpars GHarbi	0.310518	49	Kiashahr	0.282165
4	Aria-Shahr	0.355088	27	Cyrus faz 1 and 2	0.31036	50	Kuye-Taher	0.280743
5	Javaheri	0.351927	28	Piroozi	0.308765	51	Amaniyeh	0.278278
6	Moeinzadeh	0.34677	29	Bahonar	0.308605	52	Sepidar	0.276774
7	Kuye-Taleghni	0.344939	30	Sad Dastgah	0.308562	53	Melirah	0.275782
8	Kuye-Alavi	0.342287	31	Padshahr faz 2	0.305479	54	Goldasht	0.26917
9	Bagh-Sheikh	0.338712	32	Cyrus faz 4	0.304443	55	Zeiton-Karmandi	0.268944
10	Asiabad	0.334778	33	Padshahr	0.303473	56	Mohajerin	0.268291
11	Ameri	0.330007	34	Golestan shomali	0.303164	57	SHahrak Bargh	0.267057
12	Bagh-Moein	0.329732	35	Kuye-Modares	0.301941	58	Kuye-Mahdis	0.266176
13	Behzadshahr	0.328521	36	Mojahed	0.299816	59	Karoon	0.262993
14	Manba-e-Ab	0.328039	37	Nezam-Mohandesi	0.29956	60	Baharestan	0.261848
15	Khashyar (Kuye-Enghelab)	0.327324	38	Nabovat	0.29899	61	SHahrak Payam	0.260335
16	Zibashahr	0.324941	39	Aghajari	0.298909	62	Zooeyeh 1	0.255465
17	Zeiton-Kargari	0.323362	40	Farhanshahr	0.296969	63	Newsite	0.250736
18	Razmandegan SHahrak	0.323352	41	Akhar-Asfalt	0.296544	64	Foolad-Manazel	0.24672
19	Yousefi-Kuye-Fatemiyeh	0.322284	42	Pardis 1	0.292799	65	Boostan	0.237455
20	Lashkar	0.321887	43	Resalat GHarb	0.292772	66	Aliabad	0.23312
21	Kamplou Jenoubi	0.320201	44	Kamplou SHomali	0.291629	67	Manab-e-Tabiee	0.223994
22	Lashkarabad	0.319675	45	Kuye-Farhangian	0.290368	68	Salimababd	0.218929
23	Daneshgah SHahrak	0.317652	46	Kuye-Ramezanian	0.290194	69	SHahrak Naft	0.210284

The weights obtained in the first step are used in the WASPAS model. In the WASPAS model, indicators are divided into positive and negative categories. Positive indicators have an increasing role in identifying social vulnerability, i.e., with increasing their ratio, the level of vulnerability in the city increases. And negative indicators related to the display of vulnerability have a negative role and reduce vulnerability. Thus, the literacy rate is a negative indicator, and dependency rate, female-headed household rate, age over 65 years, household dimension, population density, unemployment rate, rate of housing below 80 square meters, and rental housing rate are positive indicators in the model. The results showed that the ‘20-meteri shahrdari’ neighborhood has the highest level of social vulnerability and the lowest level of social vulnerability among the neighborhoods of Ahvaz belongs to the neighborhood of ‘Shahrak Naft.

Fig. 5As stated in the research method section, the Q index (the final WASPAS) was entered as a variable in the correlation analysis. Additionally, a correlation was established between nine research indicators and the COVID-19 rate per 1000 people, as seen in Table 8. The results indicate that there is no positive and statistically significant correlation between literacy rate, unemployment rate, the rate of people over 65 years, dependency rate, rental housing rate, population density, and COVID-19 rate. There is a significant negative relationship between the rate of female-headed households and COVID-19 (0.023). That is, when the proportion of female-headed households increases, the COVID-19 rate declines. Furthermore, a significant negative association was found between the rate of housing below 80 square meters and the COVID-19 rate; as the rate increases, the COVID-19 rate goes down. However, this relationship is not very strong, and the inverse correlation rate is low. Given the significant level of 0.03 between the integrated vulnerability index (Q) and the COVID-19 rate, it can be said that this relationship is also inverse and significant; so, with the increase of social vulnerability rate in Ahvaz, the COVID-19 rate decreases. Fig. 6 shows the correlation between the general index of social vulnerability and the COVID-19 rate per 1000 in the neighborhoods of Ahvaz (see Fig. 7 ).

Fig 5.

Spatial representation of the incidence of Covid- 19 per 1000 people in the neighborhoods of ahvaz.

Table 8.

Correlation results between social vulnerability indices and COVID-19 rate per 1000 people in ahvaz neighborhoods.

		Q	Literacy rate	Unemployment rate	Households dimension	Age over 65	Female-headed households	Dependency ratio	Housing below 80 square meters	Rental housing	population density
COVID-19	Pearson Correlation	−.347**	.214	.006	−.040	−.201	−.274*	−.146	−.284*	−.124	−.115
	Sig	.003	.077	.960	.745	.098	.023	.230	.018	.309	.243

Fig. 6.

Association between integrated social vulnerability index (Q) and COVID-19 rate in ahvaz neighborhoods.

Fig. 7.

How patients get COVID-19 (percent) - Khuzestan Province (Ahvaz Jundishapur University of Medical Sciences).

4. Discussion

Academic interest in COVID-19 has expanded significantly as a result of the outbreak, and various studies have been undertaken on the diagnosis and treatment of the disease, as well as on its prevention and control [82]. Determining the influential factors on COVID-19 spread in spatial, social, health, and ecological dimensions have been of particular interest. The current study examined the relationship between social vulnerability indices and the COVID-19 rate in Ahvaz neighborhoods, taking into account both spatial and social dimensions. In the present study, a multi-criteria and spatial analysis approach was used. As indicated in the introduction, the first objective of the study was to measure social vulnerability in neighborhoods in Ahvaz.

Results showed that the neighborhoods of Naft, Boostan, New site, and Shahrak Foolad have the lowest level of social vulnerability. In these neighborhoods, population density is lower, and their residents are people working in oil and steel companies and have good economic conditions. On the other hand, the 20-meteri Shahrdari and Hasirabad neighborhoods are among the densest neighborhoods in Ahvaz. Residents of neighborhoods with a higher level of social vulnerability have had more living and economic challenges in the last two years than residents of other neighborhoods, due to their reliance on temporary informal jobs. Even commercial and neighborhood-based businesses with low levels of vulnerability have stagnated. Residents with unstable jobs and lower-income groups have lower savings ratios to use during lockdown periods and need the help of government agencies and support packages. Therefore, in the first part, by measuring social vulnerability in the neighborhoods of Ahvaz, the status and ability of residents against the effects of COVID-19 were evaluated. If neighborhoods have lower COVID-19 rates but higher levels of social vulnerability, they have experienced worse conditions due to restrictions, health protocols, and market closures. Many vulnerable groups living in these neighborhoods, such as daily-paid workers, and poor people who are not under support organizations such as the relief foundation and welfare department cannot meet their basic living needs, let alone provide health items. As Fletcher et al. [83] stated, poor people may not have the necessary resources to prepare food for a longer period and are forced to leave their homes and attend markets and streets. Thus, they fail to comply with social restrictions and health protocols. These groups of people have less access to health care, health insurance, and paid sick leave.

The second part examined the correlation between social vulnerability indices and its integrated index (Q) with the COVID-19 rate per 1000 people. Contrary to our expectations, the results showed a significant inverse correlation between the integrated index of social vulnerability and the COVID-19 rate, but this ratio is weak. Neighborhoods with higher social vulnerability rates have lower COVID-19 rates.

The research findings are not consistent with the results of [9,10,[84], [85], [86]]; and [87]. They discovered that areas with a higher vulnerability had greater COVID-19 rates on average. Numerous studies have shown that urban slums with worse social and physical conditions are more vulnerable to COVID-19 infection and transmission, which contradicts this study's results [[88], [89], [90]]. This could be explained by differences in the context and the indicators used for the analysis. In the United States, blacks have historically had higher rates of morbidity and mortality [15]. According to Ref. [91]; the COVID-19 focal sections are concentrated in low-income neighborhoods with high rates of poverty and unemployment and high rates of the black population. Moreover, one of the developing countries where the issue of vulnerability and corona has received a lot of attention is India, where the high percentage of slums has increased the COVID-19 incidence, creating many problems for deprived groups [59,92]. However, such a feature does not exist in Iran or the city of Ahvaz, where residents of all ethnic groups and races live in areas with a significant level of social vulnerability. In Ahvaz, neighborhoods with more undesirable social and physical conditions, such as Sayahi, are Arab people, and the proportion of Lor people is higher in Hasirabad and Manba-e-Ab neighborhoods. Neighborhoods with less social vulnerability have greater ability and adaptability to cope with shocks such as COVID-19.

The findings of this study are consistent with the findings of [93,94]; and [95] in Iran and Tehran. They also found that neighborhoods with better economic and social conditions and more active economic and commercial sectors have a higher incidence.

Vulnerability and social vulnerability have been studied in pre- COVID-19 studies against natural hazards such as floods, earthquakes, and droughts [65,96] and have not been comprehensively studied against human risks, especially infectious diseases. In this study, COVID-19 was considered a human risk. By affecting the economy and closing markets and services, this risk affects social vulnerability and unemployment rates, and people's livelihoods [14,97]. This, in turn, could lead to severe economic effects and quarantine on the lives of the poor. COVID-19 had a more significant economic impact than a health impact on the poor [37]. It is also influenced by the community's socioeconomic qualities, such as the slums' poor condition [98].

COVID-19 human risk is displaced by human behavior and affects all countries, cities, and villages, as well as rich and poor people. This is distinct from natural hazards that may affect only a part of an area. COVID-19 has affected all neighborhoods of Ahvaz with suitable and poor economic conditions. COVID-19 rates are significantly higher in neighborhoods with better socio-economic conditions, perhaps as a result of increased social contact at parties, ceremonies, job centers, and markets [99]. discovered that employment is a critical feature in the COVID-19 pattern, which enhances social interaction and vulnerability [100]. also analyze the disease's spread in terms of human mobility. According to Figure, the highest percentage of cases is associated with contact with a positive family member, followed by frequent intra-city travel, interaction with other relatives at ceremonies, and occupational infection. Due to increased social mobility in neighborhoods with lower socioeconomic vulnerability, these occurrences are substantially more prevalent in the workplace and during ceremonies.

The rate of female-headed households and the rate of housing less than 80 square meters have an inverse and substantial association with the COVID-19 rate. That is, in neighborhoods where the rate of these indicators is higher, the COVID-19 rate is lower. Moreover, there is no significant relationship between literacy rates, unemployment rate, the rate of people over 65 years, dependency rate, rental housing rate, population density, and COVID-19 rate per 1000 people. This finding contradicts Neşe and Bakir's findings (2022) that population density and the rate of aging dependency have a significant association with the COVID-19 rate. COVID-19 rates in neighborhoods have also been affected by health attitudes and social behaviors, adherence to protocols, and referrals for testing and hospitalization. Residents of wealthy neighborhoods may have paid greater attention to their treatment and health and attended more testing. As [13] note, poor neighborhood circumstances have a direct effect on behavior and health outcomes. Furthermore [101,102], emphasize the importance of human behavior and the importance of a better knowledge of risky behaviors. In general, disease spread is complicated and multifaceted. Furthermore, numerous elements in the demographic, physical, cultural, economic, social, and managerial dimensions influence the development of COVID-19 (Lak et al., 2021), with the social dimension being the focus of this study. The COVID-19 prevalence was high in rich counties of Ahvaz, such as Padad or Zaytoun Karmandi, but in the areas with a lower economic, social and cultural status and less compliance with the health protocols, the COVID-19prevalence was lower. This shows that residents’ visits, awareness, and their sensitivity to the disease are different.

Most of the people who live in the marginal areas of Ahvaz are rural migrants. Indeed, all the marginal areas around and inside the city have been formed as a result of intra-provincial migrations, and poverty and social vulnerability have shifted from villages and small towns to big cities like Ahvaz. Thus, to control marginalization in Ahvaz and reduce social vulnerability, attention should be paid to the issues of social justice and spatial justice through territorial planning.

In other words, this study depicted the inequalities that already existed in the neighborhoods. The inequalities that exist between different, regions, cities, and villages in the province have caused the aggravation of deprivation and increased vulnerability to natural hazards such as floods, earthquakes, and drought. These inequalities have spread to cities and urban neighborhoods making residents vulnerable to human hazards such as COVID-19.

In general, the inequalities and deprivations that were revealed in terms of unemployment, literacy, and other indicators used are significant from two perspectives. First, these inequalities have some implications for community health management because people who have little literacy and have severe deprivation do not tend to comply with health protocols and social restrictions, endangering their health and that of others. Second, inequalities and deprivations affect the economic position and income of disadvantaged people and groups. Unemployed people have faced living problems and have limited access to food during hazards. Thus, coping approaches should be adopted to improve public health and income levels. Furthermore, awareness-raising and training activities and distribution of health items, and free access to health services can contribute to improving public and individual health. From an economic perspective, the support of the government and charitable organizations, and welfare institutions for disadvantaged groups should be increased, especially in marginal areas with a higher population density. During the COVID-19 outbreak, a short-term approach can be adopted with support measures, but it is necessary to form a fundamental and long-term prospect to reduce inequalities and poverty at the regional and local levels because Khuzestan Province and other areas in Iran are greatly affected by many natural hazards.

5. Conclusion

Understanding the factors that contribute to social vulnerability is an essential step in helping disadvantaged communities and individuals access the resources and adopt strategies needed to minimize the consequences of disasters.

Measuring social vulnerability and its association with COVID-19 has been deemed critical by academics, and this study is making a significant contribution to this field. The status of social vulnerability and its association with COVID-19 were evaluated in Ahvaz neighborhoods using MCDM approaches and correlation analysis. Nine indicators are used to rate the neighborhoods in Ahvaz. The integrated indicator of social vulnerability and COVID-19 in Ahvaz has a small inverse association.

Social vulnerability is assessed using a variety of indicators, which necessitates a comprehensive and multidimensional approach. Due to the nature of this study, multidisciplinary methods were used to analyze social vulnerability. For the first time, the BWM-WASPAS approach was employed to measure social vulnerability. These methods can be used to evaluate areas and neighborhoods of other cities worldwide. The methods used were a combination of mathematical models, and spatial and statistical analysis. Therefore, the theoretical and methodological contribution of the present study can be summarized in the following aspects.

• •Assessing social vulnerability and its relationship with COVID-19 in a new context and providing more information about developing countries (new experimental study);
• •Assessing social vulnerability in neighborhoods as an effective and vulnerable issue against human risk;
• •Considering the importance and differences in indicators of social vulnerability;
• •Simultaneous use of mathematical methods, spatial and statistical analysis;
• •Simultaneous use of subjective and objective data;
• •Calculation of integrated social vulnerability index;
• •Introducing and presenting new indicators (rental housing rate and housing below 80 square meters) according to the study context.

The research findings will aid managers and planners in developing new knowledge and insights about social vulnerability in Ahvaz and in planning to eliminate socioeconomic inequality and access to health services. Additionally, it will assist managers in providing health facilities and free testing in slums, as well as strengthening COVID-19 restrictions in neighborhoods with a high COVID-19 rate. Furthermore, arrangements should be established to ensure that during lockdown periods, informal workers, as well as day laborers, do not face difficult conditions. Therefore, it is necessary to provide special packages to support vulnerable groups.

The study has faced some limitations. Data on a neighborhood scale are not in the Statistics Center of Iran censuses, so data collection based on a neighborhood-scale has been difficult. Furthermore, COVID-19 statistics have been under the cultural and service conditions of the neighborhoods, and there is a mismatch between the neighborhoods that have the most social vulnerability and the level of access to health services and laboratories in the city of Ahvaz. Previous studies have reported such inconsistencies [103]. The lack of testing facilities is also one factor affecting the number of cases reported in neighborhoods [16]. Moreover, access to social vulnerability indicators such as car ownership rates and per capita income was complicated and such indicators were not used in this study. These two indicators are the most important ones that show the economic situation of the residents of the neighborhoods.

The indicators selected in this study may not be directly related to the COVID-19 rate. But they may be effective in influencing health attitudes and observance of protocols and social behaviors during lockdown periods. For example, the rate of people over 65 may not directly affect the increase in the COVID-19 rate. Nonetheless, their inability to work and their need for care and maintenance may complicate matters during COVID-19, increasing the mortality rate.

Moreover, it should be emphasized that the use of weights for indicators and the lack of access to some indicators may affect the neighborhood's position in comparison to what residents believe and the results of other studies. The population density index is highly correlated with the risk of COVID-19. While a neighborhood may rank last in terms of population density, other factors indicate a high level of social risk.

Future research could examine the health attitudes and social behaviors of residents in two neighborhoods with varying degrees of social vulnerability (suburban neighborhoods). Additionally, they can effectively assess social vulnerability by examining informal employment, ownership, and media income variables. Furthermore, access to services and laboratories in neighborhoods should be studied, as well as the effect on the accuracy of the infection rate. Thus, the population density is an index to show the vulnerability to COVID-19 but it is not a measure of deprivation.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.