Social deprivation index and lymphatic filariasis: a tool for mapping urban areas at risk in northeastern Brazil

Article Outline

• Summary
• 1. Introduction
• 2. Methods
  • 2.1. Study area
  • 2.2. Subjects and sample collection
  • 2.3. Parasitological test
  • 2.4. Ethical considerations and consent
  • 2.5. Formulation of the Social Deprivation Index (SDI)
• 3. Results
  • 3.1. Parasitological survey
  • 3.2. Social Deprivation Index (SDI)
• 4. Discussion
• Authors’ contributions
• Funding
• Conflicts of interest
• Ethical approval
• Acknowledgements
• References
• Copyright

Summary 

This paper describes the construction and application of a social deprivation index that was created to explore the relationship between lymphatic filariasis and socioenvironmental variables in the municipality of Jaboatão dos Guararapes, Pernambuco, Brazil, thereby contributing towards identifying priority areas for interventions. This indicator was obtained from principal-component factor analysis. Variables available from the national census representing socioenvironmental conditions, household characteristics and urban services were used. Epidemiological data came from a parasitological survey on lymphatic filariasis. 23673 individuals were examined and 323 were positive (1.4%). Two factors that together explained 80.61% of the total variance were selected. The social deprivation strata were capable of indicating a risk gradient, with 74.9% of the microfilaremia cases situated in the high-risk stratum. Principal-component factor analysis was shown to be sensitive for selecting indicators associated with the risk of lymphatic filariasis transmission and for detecting areas potentially at risk. The capacity of the social deprivation index for picking up social inequalities qualifies it as a new tool for use in planning interventions aimed at controlling lymphatic filariasis in urban spaces.

Keywords: Lymphatic Filariasis, Socioeconomic Factors, Poverty Areas, Health Inequalities, Factor Analysis, Control programs

Back to Article Outline

1. Introduction 

Health inequalities associated with socioeconomic factors are of great interest in public health. The relationship between socioeconomic level and health situation has already been established in the scientific literature. Socioeconomic factors are one of the principal determinants of health, although which socioeconomic component affects the health situation most and how this relationship should be measured still need to be better explored.¹

The concept of deprivation has frequently been used to characterize and study the impact of socioeconomic factors on health within a given geographical area.² It has been presumed that socioeconomic conditions in such areas influence the health situation.³

Townsend⁴ defined deprivation as an observable and demonstrable state of disadvantage in relation to the community, society or nation to which an individual, group or family belongs. Deprivation indexes have become practical tools for investigating health inequalities and socioeconomic conditions. Through such tools, interventions and resources can be directed towards areas or groups with greater social deprivation.², ⁵

Deprivation indexes are instruments used to measure socioeconomic conditions in specific geographical areas.¹ They are very simple and inexpensive because they are generally composed of variables from national censuses, combined by means of different types of statistical procedures.¹

Lymphatic filariasis is a disease associated with poverty and environmental conditions.⁶, ⁷, ⁸ Haddix and Kestler⁹ stated that lymphatic filariasis is a disease caused by poverty that perpetuates the poverty cycle. This is evident from the observation that 94% of the countries with the lowest human development index (HDI) are endemic for lymphatic filariasis.⁷ The HDI for Brazil in 2005 was 0.800, thus placing this country in the group of high human development.¹⁰ However, Brazil is a typical case of an extremely unequal society. The Gini index (an indicator that measures wealth distribution) is 0.59, thereby showing one of the greatest disparities between rich and poor people in the world.¹¹ Hotez¹² stated that one way in which these inequalities are manifested is through the weight given to neglected tropical diseases.

In Brazil, active transmission occurs only in the metropolitan region of Recife, State of Pernambuco. It has been estimated that 1.5 million people live in areas at risk and that 49000 are infected with Wuchereria bancrofti.¹³ In the municipality of Jaboatão dos Guararapes, 85% of the districts surveyed in the Jaboatão region presented cases of filarial infection.¹⁴ This municipality was selected for the present study because it forms part of an area that is recognized to be endemic for this disease and because the Global Program for the Elimination of Lymphatic Filariasis needs to have up-to-date information available on the spatial distribution and risk status of filarial infection. The aim of the present study was to construct and apply a social deprivation index that was created to explore the relationship between lymphatic filariasis and socioenvironmental variables in the municipality of Jaboatão dos Guararapes, Pernambuco, Brazil, thereby contributing towards identifying priority areas for interventions within this program.

Back to Article Outline

2. Methods 

2.1. Study area 

A parasitological survey was conducted in the municipality of Jaboatão dos Guararapes, State of Pernambuco, Brazil, between 2000 and 2002. This municipality covers an area of 256073km² and is located approximately 19km from the state capital (Recife). According to the 2000 census, the municipality has 581556 inhabitants, and 97.8% of them live in the urban area. It is administratively divided into 27 districts and 492 census tracts. In this study, two districts were excluded because of lack of information.

2.2. Subjects and sample collection 

The minimum sample size (5915 households) was estimated based on the total number of households recorded in the last census (111666), assuming that 6% of them would harbor at least one microfilaremic individual and that no more than 5% of the subjects within each 95% confidence interval would be misclassified as microfilaremic when they were not, or vice versa.⁸

Although only a 12-household sample was needed in each of the 484 census tracts to give the minimum sample, 14 households were registered in each census tract to allow for possible refusals to participate. These 14 households per census tract were selected based on the maps used in the last national census. A total of 28612 people were registered to participate in the survey, but 4939 of them (17.26%) did not undergo the test and thus a final sample of 23673 individuals was obtained.

2.3. Parasitological test 

A fingerprick sample of ‘night’ blood, of about 50μl, was then collected from the consenting members of each study household, between 23:00h and 01:00h. These samples were used to prepare thick smears, which were Giemsa-stained and then checked under an optical microscope for microfilariae of Wuchereria bancrofti.

2.4. Ethical considerations and consent 

The study was approved by the Research Ethics Committee of the Centro de Pesquisas Aggeu Magalhães, Fundação Oswaldo Cruz, Pernambuco, Brazil. Biological samples were collected only after the informed consent statement had been signed. Cases of microfilaremia were treated with diethylcarbamazine.

2.5. Formulation of the Social Deprivation Index (SDI) 

A set of variables available in the 2000 demographic census, representing socioenvironmental conditions, household characteristics and urban services, was used in the factor analysis (Table 1).

Table 1. Definitions of the indicators used in the Social Deprivation Index.

The conditions required for applying factor analysis were verified using the Kaiser-Mayer-Olkin (KMO) coefficient and Bartlett's Test of Sphericity. The latter determines the null hypothesis and thereby affirms that the correlation matrix is not significantly different from the identity matrix.

The SDI was constructed using the principal component analysis (PCA) method, with varimax orthogonal rotation, by means of the Statistical Package for the Social Sciences software, version 15.0 (SPSS, Chicago, IL, USA). In PCA, the weight attributed to each variable is not determined arbitrarily but results from the relationships between the statistical indicators within the geographical area selected. This type of method makes it possible to simplify the data through reducing the number of variables. These variables are called the principal components or factors and are obtained using linear combinations of the original variables. The relationships between each original variable and the new factors are measured according to the factor loadings in the components. These loadings range from -1 to 1 and the greater their absolute value is, the greater the representativeness of the original variable in the composite axis is.

The number of factors to be extracted was defined from the graph of variance vs. number of components (scree plot), in which points on the greatest slope indicate the appropriate number of components to retain. The reliability of the factors was evaluated by means of Cronbach's alpha coefficient, and an index ≥ 0.50 was considered acceptable.

From the set of factors extracted, those that presented individual values greater than one were used to construct the SDI. The indicator values were calculated by means of regression. Based on this indicator, the districts were grouped according to social deprivation, into three strata: low, medium and high social neediness. This last stratum was composed of the third and fourth quartiles.

Back to Article Outline

3. Results 

3.1. Parasitological survey 

A total of 23673 individuals were examined and 323 were found to be positive (1.4%). The sample consisted of 13386 women (56.6%) and 10287 men (43.4%). Significantly more men than women were positive (2% vs. 0.9%; p<0.001). Figure 1 presents the proportions of positive individuals according to age and sex. All age groups presented positive cases and the highest rate (2.2%) was found in the age group from 20 to 29 years. The magnitude of the microfilaremia prevalence rate per district ranged from 0 to 5.09% (Figure 2A).

• 
  • View Large Image
  • Download to PowerPoint

• Figure 1. 

  Prevalence of microfilaremia according to sex and age group in Jaboatão dos Guararapes, Pernambuco, Brazil.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 2. 

  (A) Distribution of microfilaremia prevalence rates per district. (B) Distribution of deprivation strata per district according to the SDI. Municipality of Jaboatão dos Guararapes, Pernambuco, Brazil.

3.2. Social Deprivation Index (SDI) 

Table 2 presents the basic descriptive statistics for all the indicators considered in this study. It was observed that the mean proportion of households with inadequate sewage disposal (65.42%) was higher than the proportions for garbage collection (35.14%) and water supply (18.49%). The mean proportion of heads of households with income between half and one minimum monthly salary was 25.13% (range 6.90% to 38.97%). Coefficients of variation close to 100% (water supply) and greater than 100% (basis of home occupation) were also observed, thus indicating that the variations between the districts were very large.

Table 2. Descriptive statistics on the variables used for constructing the Social Deprivation Index.

Bartlett's Test of Sphericity (χ² 127.22; p<0.001), the KMO coefficient (0.62) and the correlation matrix determinant (0.02) showed that the correlations between the items were adequate for performing factor analysis. The correlation matrix between the indicators is presented in Table 3. There were high correlations between the number of people in the household and the basis of home occupation (0.88); between inadequate garbage collection and water supply (0.85); and between the schooling level and income of the heads of households (0.74). The negative correlations ranged from -0.02 (between the schooling level of the heads of households and the number of people in the household) to -0.27 (between the schooling level of the heads of households and the basis of home occupation).

Table 3. Correlation matrixes* between all the indicators considered in the study.

*Correlation is significant at the 0.01 level (2-tailed).

To assess the number of factors to be extracted, eigenvalues (>1.5) and scree plots were used. These revealed the existence of two factors that together explained 80.61% of the total variance (Table 4). The first principal component (PC₁) was considered to be the marker of the ‘dimension’ of social deprivation and it explained 52.95% of the total variance. This was taken to be the SDI. Five of the variables were strongly correlated with the dimension of social deprivation, as indicated by PC₁, with a factor loading of more than 70% (Table 4). Factor 2 represented the socioeconomic dimension and explained 27.66% of the total variance. Cronbach's alpha coefficient was 0.78 and 0.56, respectively.

Table 4. Result from principal-component factor analysis on the socioenvironmental factors, according to district, Jaboatão dos Guararapes, Pernambuco, Brazil.

Using the SDI, strata of degree of social need were formed (low, medium and high), taking the distribution of this indicator into quartiles as the cutoff points (Figure 2B). To form the stratum of high social neediness, the third and fourth quartiles were grouped. In the first and second strata, the prevalence of microfilaremia was 0.6%. The third stratum presented a rate of 2.55% and accounted for around 75% of all of the positive cases identified (Figure 2).

The logistic regression model demonstrated that income of between half and one minimum monthly salary for heads of households, water supply only in one room of the home, water supply only in the yard of the property, other non-standard methods of water supply and the presence of 10 or more people in the household increased the prevalence of microfilaremia by 0.34%, 0.15%, 0.48%, 0.23% and 5.53%, respectively.

Back to Article Outline

4. Discussion 

The results from the parasitological survey indicate that filariasis is endemic in the municipality of Jaboatão dos Guararapes. All age groups presented positive cases, including children <10 years of age (Figure 1). In endemic areas for lymphatic filariasis, the microfilaremia rates are relatively low during the first decade of life and they gradually increase with age. It was observed that the prevalence of microfilaremia was higher among the men (Figure 1). Studies carried out in India,¹⁵ Haiti,¹⁶ and Ghana¹⁷ have found significantly greater prevalence among males. This has also been found in other Brazilian studies.¹⁸, ¹⁹

The mean prevalence rate for the municipality was 1.4%, but when the analysis was done on smaller geographical units like districts, the prevalence reached 5.09%. According to the World Health Organization,²⁰ the prevalence rates can be classified as low (<5%), medium (5 to 10%) and high (>10%). The first two types were identified in this study. In addition, it needs to be emphasized that positive results were found in 80% of the district. Figure 2A presents the distribution of the prevalence rates for microfilaremia within the area covered by the municipality. Epidemiological maps showing the spatial distribution of infection are being given increasing value within lymphatic filariasis elimination programs.²¹, ²², ²³, ²⁴

To construct the SDI, variables available in the demographic census that are recognized as being associated with social conditions¹, ² and environmental conditions⁶, ²⁵ were used. The census data have the advantage of providing the same type of information for the whole spatial territory, as well as having periodic updates.² The study showed that there was a correlation between the variables selected and the risk strata established, thereby indication that the approach was congruent and valid.

It has been recognized that there are ecological attributes in geographically defined areas that affect whole groups. These attributes relate to the characteristics of socioeconomic conditions and the quality of the urban infrastructure and environment, which affect health beyond the contribution of individual characteristics.⁵, ⁸ Albuquerque²⁶ stated that an explanatory model for the endemicity of lymphatic filariasis in this region would need to presume that the infectious cycle is correlated with the population's living conditions, especially with regard to people living in dirty areas with precarious environmental sanitation.

In urban areas, precarious sanitation conditions (sewage disposal, wastewater management and water supply) favor the creation of artificial breeding sites for vectors.²⁷ In such situations, environmental factors interact strongly with socioeconomic factors, thereby contributing towards maintaining the endemicity of lymphatic filariasis.²⁸ In the present study, the indicators for garbage collection and inadequate water supply were highly correlated (Table 3). The proportion of households with inadequate sewage disposal was 65.42% (ranging from 21.40 to 99.26%) (Table 2). This precariousness presented by the sewage disposal system was translated into the presence of rudimentary cesspits, ditches and excrement left in the open air, thereby significantly favoring proliferation of the vector insect Culex quinquefasciatus.

The proposal to use the SDI to identify priority areas for interventions breaks with the approach traditionally used by endemic disease control programs in Brazil. These have been centered on identifying and treating postitive individuals, while ignoring the importance of the role played by socioeconomic conditions.²⁹ Thus, an approach centered on collective risk is proposed for the actions of the Global Program for the Elimination of Lymphatic Filariasis.

The special aggregation unit used in this study was the district and the risk strata formed by agglomeration of the districts. It is known that small areas present the advantage of greater internal homogeneity. On the other hand, aggregation of epidemiological and demographic data into larger units (districts, regions and municipalities, among others) reduces the effect of rate instability. Districts make it possible to identify differences in socioeconomic conditions (Figure 2B) and also form spatial units that are recognized both by the population and by administrative bodies, which facilitates their use in planning actions for health programs. In turn, the social deprivation strata were also capable of indicating risk gradients, such that 74.9% of the cases of microfilaremia (242/323) were situated in the stratum of high risk (Figure 2). It is emphasized that the purpose of the SDI is to identify areas with greater potential for transmission, which does not necessarily imply effective transmission. Also a logistic model indicated that poor social and environmental variables increases the probability of lymphatic filariasis in a house.

The results from this study make it possible to recognize that areas with similar socioeconomic characteristics may present different prevalence rates. In this respect, the influence of other factors needs to be considered, such as the proximity of water sources and intra-urban migration.⁶ Medeiros et al³⁰ reinforced the idea that circular migration is an important factor in the expansion of filariasis in the metropolitan region of Recife.

For areas in which no cases of filarial infection have been identified but the environmental conditions needed for transmission exist, a territorially based surveillance system needs to be created to detect new foci of transmission. The perspective to be borne in mind is the target of eliminating lymphatic filariasis as a worldwide public health problem by 2020.

Principal-component factor analysis is a useful statistical technique for selecting indicators associated with the risk of transmitting lymphatic filariasis and for detecting areas potentially at risk, such that most of the microfilaremia cases are situated in the high-risk stratum. The capacity of the SDI to pick up social inequalities has demonstrated its value as a new tool for planning interventions that aim towards controlling lymphatic filariasis in urban spaces.

Back to Article Outline

Authors’ contributions 

CB and ZM conceived and designed the study. AS was responsible for the clinical management of the study. DP was responsible for Census data use and interpretation. TC was responsible for data analysis and statistical procedures. JP was responsible for the mapping procedures. CO supervised the field data collection. CB and DP drafted the paper. ZM revised the draft of the paper. All authors approved the final version of the paper.

Back to Article Outline

Funding 

None.

Back to Article Outline

Conflicts of interest 

None declared.

Back to Article Outline

Ethical approval 

The study protocol and consent form were approved by the Committee of Ethics in Human Research of the Centro de Pesquisas Aggeu Magalhães of the Fundação Oswaldo Cruz, Recife-PE, Brazil (Certificate of Presentation for Ethical Consideration No. 0034.0.095.000-07).

Back to Article Outline

Acknowledgements 

We are grateful to the Secretary of Health of Jaboatão dos Guararapes, Pernambuco, for their cooperation and logistical support for the parasitological survey. Also, we are grateful to the data-collection team.

Back to Article Outline

References 

1. Testi A, Ivaldi E. Material versus social deprivation and health: a case study of an urban area. Eur J Health Econ. 2009;10:323–328
   • View In Article
   • CrossRef
2. Domínguez-Berjón MF, Borrell C, Cano-Serral G, Esnaola S, Nolasco A, Pasarín MI, et al. Constructing a deprivation index based on census data in large Spanish cities (the MEDEA project). Gac Sanit. 2008;22:179–187
   • View In Article
   • CrossRef
3. Sánchez-Cantalejo C, Ocana-Riola R, Fernández-Ajuria A. Deprivation index for small areas in Spain. Soc Indic Res. 2008;89:259–273
   • View In Article
4. Townsend P. Deprivation. J Social Policy. 1987;16:125–146
   • View In Article
5. Havard S, Deguen S, Bodin J, Louis K, Laurent O, Bard D. A small-area index of socioeconomic deprivation to capture health inequalities in France. Soc Sci Med. 2008;67:2007–2016
   • View In Article
   • CrossRef
6. Braga C, Ximenes RAA, Albuquerque MFPM, Souza WV, Miranda J, Brayner F, et al. Evaluation of a social and environmental indicator used in the identification of lymphatic filariasis transmission in urban centers. Cad Saúde Pública. 2001;17:1211–1218
   • View In Article
7. Durrheim DN, Wynd S, Liese B, Gyapong JO. Editorial: Lymphatic filariasis endemicity – an indicator of poverty?. Trop Med Int Health. 2004;9:843–845
   • View In Article
   • MEDLINE
   • CrossRef
8. Bonfim C, Evangelista Netto MJ, Pedroza Jr. D, Portugal JL, Medeiros ZA socioenvironmental composite index as a tool for identifying urban areas at risk of lymphatic filariasis. Trop Med Int Health 2009;14. in press.
   • View In Article
9. Haddix AC, Kestler A. Lymphatic filariasis: economic aspects of the disease and programmes for its elimination. Trans R Soc Trop Med Hyg. 2000;94:592–593
   • View In Article
   • Full-Text PDF (259 KB)
   • CrossRef
10. UNDP . Measuring Human Development: A Primer. New York: United Nations Development Programme; 2007;
    • View In Article
11. World Bank. Inequality and economic development in Brazil, a World Bank country Study. The International Bank for Reconstruction and Development. Executive Summary. Washington (D. C.): The World Bank. pp XVIII, 2004.
    • View In Article
12. Hotez PJ. The Giant Anteater in the Room: Brazil's Neglected Tropical Diseases Problem. PLoS Neglected Tropical Disease. 2008;2(1):e177
    • View In Article
13. Ministry of Health. Fundação Nacional de Saúde. Centro Nacional de Epidemiologia. Reunião de Avaliação do Programa de Controle da Filariose Linfática no Brasil. Brasília: Funasa; 2000.
    • View In Article
14. Medeiros Z, Bonfim C, Alves A, Oliveira C, Netto MJE, Aguiar-Santos AM. The epidemiological delimitation of lymphatic filariasis in an endemic area of Brazil. 41 years after the first recorded case. Ann Trop Med Parasitol. 2008;102:509–519
    • View In Article
    • CrossRef
15. Pani SP, Balakrisnam N, Srividya A, Bundy DA, Grenfell BT. Clinical epidemiology of bancroftian filariasis: Effect of age and gender. Trans R Soc Trop Med Hyg. 1991;85:260–264
    • View In Article
    • Abstract
    • Full-Text PDF (532 KB)
    • CrossRef
16. Raccurt CP, Lowrie RC, Katz SP, Duverseau YT. Epidemiology of Wuchereria bancrofti in Leogane, Haiti. Trans R Soc Trop Med Hyg. 1988;82:721–725
    • View In Article
    • Abstract
    • Full-Text PDF (527 KB)
    • CrossRef
17. Dunyo SK, Appawu M, Nkrumah FK, Baffoe-Wilmont A, Pederson EM, Simonsen PE. Lymphatic filariasis of the coast of Ghana. Trans R Soc Trop Med Hyg. 1996;90:634–638
    • View In Article
    • Abstract
    • Full-Text PDF (649 KB)
    • CrossRef
18. Maciel A, Rocha A, Marzochi K, Medeiros Z, Carvalho AB, Regis L, et al. A Epidemiological study of bancroftian filariasis in Recife, northeastern Brazil. Mem Inst Oswaldo Cruz. 1996;91:449–455
    • View In Article
    • MEDLINE
    • CrossRef
19. Braga C, Dourado I, Ximenes R, Miranda J, Alexander N. Bancroftian filariasis in an endemic area of Brazil: differences between genders during puberty. Rev Soc Bras Med Trop. 2005;38:224–228
    • View In Article
    • MEDLINE
20. World Health Organization . Lucha Contra la Filariasis Linfática. Geneva: OMS; 1988;
    • View In Article
21. Hassan AN, Beck LR, Dister S. Spatial analysis of lymphatic filariasis distribution in the Nile Delta in relation to some environmental variables using geographic information system technology. J Egypt Soc Parasitol. 1998;28:119–131
    • View In Article
    • MEDLINE
22. Sabesan S, Palaniyandi M, Das PK, Michael E. Mapping of lymphatic filariasis in India. Ann Trop Med Parasitol. 2000;94:591–606
    • View In Article
    • MEDLINE
23. Gyapong JO, Kyelem D, Kleinschmidt I, Agbo K, Ahouandogbo F, Gaba J, et al. The use of spatial analysis in mapping the distribution of bancroftian filariasis in four West African countries. Ann Trop Med Parasitol. 2002;96:695–705
    • View In Article
    • MEDLINE
    • CrossRef
24. Ngwira BMM, Tambala P, Perez AM, Bowie C, Molyneux DH. The geographical distribution of lymphatic filariasis infection in Malawi. Filaria J. 2007;6:12
    • View In Article
25. Albuquerque MFPM, Morais HMM. Decentralization of endemic disease control: an intervention model for combating bancroftian filariasis. Pan Am J Public Health. 1997;1:157–163
    • View In Article
26. Albuquerque MFPM. Urbanization, slums, and endemics: the production of filariasis in Recife, Brazil. Cad Saúde Pública. 1993;9:487–497
    • View In Article
27. Mott KE, Desjeux P, Moncayo A, Ranque Ph , Raadt P. Parasitic diseases and urban development. Bull World Health Organ. 1990;68:691–698
    • View In Article
    • MEDLINE
28. Erlanger TE, Keiser J, Castro MC, Bos R, Singer BH, Tanner M, et al. Effect of water resource development and management on Lymphatic filariasis, and estimates of populations at risk. Am J Trop Med Hyg. 2005;73:523–533
    • View In Article
    • MEDLINE
29. Franco O, Lima DMS. Alguns aspectos das atividades contra a filariose bancroftiana no Brasil. Revista Brasileira de Malariologia e Doenças Tropicais. 1967;19:73–89
    • View In Article
30. Medeiros Z, Oliveira C, Quaresma J, Barbosa E, Aguiar-Santos AM, Bonfim C, et al. Lymphatic filariasis in Moreno, Northeast Brazil. Rev. Bras. Epidemiol. 2004;7:73–79
    • View In Article