Spatial patterns of leprosy in an urban area of central Brazil.

Spatial patterns of leprosy in an urban area of central Brazil C.M.T. Martelli,1 O.L. Moraes Neto,1 A.L.S.S. Andrade,1 S.A. Silva,' I.M. Silva,2 & F. Zicker3 Reported is the spatial variation of leprosy in an urban area of Brazil and its correlation with socioeco- nomic indicators. From November 1991 to October 1992 a total of 752 newly diagnosed leprosy patients who were attending all outpatient clinics in Goiania city, central Brazil, were identified. A data- base of leprosy cases was set up linking patients' addresses to 64 urban districts. Leprosy cases were detected in 86% of the districts and three risk strata were identified. The highest-risk area for leprosy was in the outskirts of the city and detection rates increased on moving from more developed to poorer areas. The risk of detecting leprosy cases was 5.3-fold greater (95% Cl: 3.8-7.4) in the outskirts of the town than in the central zone. Discussed are the methodological issues related to leprosy case ascertainment, completeness and reliability of information, and the interpretation of the spatial distribution of leprosy per unit area. High- lighted also are the lack of leprosy control activities in primary health care units and the usefulness of geographical analysis in planning health services. Introduction The occurrence and distribution of many infectious and parasitic diseases has changed because of the rapid growth of cities, internal migration, and socio- economic changes in the least developed countries. Several studies have demonstrated the importance of rural-urban migration in disease pattems and the need to improve control strategies in urban settings (1-3). In more than 90 countries, whose total popula- tions amount to 2.4 x 109, leprosy is endemic, with a prevalence of at least 1 per 10 000.a Although the introduction of multidrug therapy (MDT) in the 1980s made a major contribution to treating leprosy in most of the endemic countries, the disease is still an important cause of morbidity and physical de- formities in Asia, Africa, and Latin America (4, 5). Examination of the records of patients attending the routine services of specialized leprosy pro- ' Associate Professor, Departamento de Saude Coletiva, Institu- to de Patologia Tropical e Saude Publica, Universidade Federal de Goias, Goiania, Brazil. Requests for reprints should be sent to Dr Martelli at the following address: rua 132-A n° 74, Setor Sul, 74093-220, GoiAna, Goias, Brazil. 2 Secretaria de Saude e Meio Ambiente, Goias State, Brazil. 3Communicable Disease Epidemiologist, Pan American Health Organization, Venezuela. a Strategic plan for applied field research in tropical diseases. WHO unpublished document CTD/TDR/AFR/93.1, 1993. Reprint No. 5603 grammes has permitted time-trend analysis and com- parisons between countries to be made (6-8). How- ever, few studies have addressed the dynamics of leprosy transmission in urban environments in South America, particularly Brazil. To the best of our knowledge, the only study that did address this issue showed a rather variable distribution of leprosy cases among different socioeconomic census tracts in a highly urbanized area of south-east Brazil.b The need to improve understanding about the pattems of trans- mission of leprosy to facilitate the planning of health care services has encouraged the development of analytical tools based on studies of the geographical pattems of the disease (9-11). This article reports the results of an ecological study designed to investigate the spatial variation of leprosy in an urban area and its relation to socioeco- nomic indicators, as a possible methodology to assist in the planning of health services. Methods Study area The study was carried out in Goiania, central Brazil, a city of population 900 000 (population density, 520 inhabitants/kM2), 99% of whom live in the urban b Andrade, VLG. [Epidemiological characteristics of Hansen's disease in a municipality of SJo Gongalo, Rio de Janeiro]. MSc thesis. Escola Nacional de Saude Publica, Rio de Janeiro, 1990 (in Portuguese). Bulletin of the World Health Organization, 1995, 73 (3): 315-319 © World Health Organization 1995 315 C.M.T. Martelli et al. area. A large proportion of the population migrated from rural areas during an intense migration influx in the mid-1960s. Over the last two decades the annual population growth rate has decreased from 6.5 to 2.3 per 100 due to an important decline in fertility rates and in internal migration waves (12). Goiania is divided administratively into 64 dis- tricts, which are aggregations of census tracts and cover areas with broadly similar socioeconomic characteristics. These spatial units were conceived by the city council to define a framework for socio- economic policy and research purposes. The study area has the second highest leprosy endemicity in Brazil, with a prevalence and annual detection rate of 3.7/1000 and 56.7/100 000 inhabi- tants, respectively. The detection rate has increased by 8% per annum over the last 20 years (7, 13). The leprosy control programme in the city is based on passive case detection and treatment, and BCG vaccination of household contacts. Multidrug therapy (MDT), delivered as an outpatient regimen, is the first therapeutic choice for leprosy (13, 14). The local notification system is considered to be a reliable source of information, and is believed to include the majority of cases in this urban setting. For the purposes of the study, a new case was defined as a person who showed clinical signs of lep- rosy and who had never been treated for the disease in the past. Case detection was carried out by trained doctors, and patients were classified on the basis of clinical signs and skin-smear results for allocation to MDT regimens, according to the Brazilian Leprosy Control Programme protocol. The date of birth, address, date of first diagno- sis, date of attendance, clinical type of leprosy, and the number of contacts were recorded for all newly diagnosed leprosy patients attending all seven out- patient clinics that provided treatment for leprosy patients in the city over the period November 1991 to October 1992. Statistical analysis of data A database of leprosy cases was set up linking patients' addresses to districts. Detection rates and 95% confidence intervals (95% CI) were calculated using the total number of leprosy cases notified by the sanitary authorities as numerator and the district population as denominator, at different geographical levels. Cases whose addresses could not be assigned to a specific geographical area were excluded from the analysis (5.5% of cases). The administrative dis- tricts' boundaries were drawn up from existing maps using computer-aided design (CAD) software. Five socioeconomically homogeneous areas were defined by grouping contiguous districts, taking into account income level and the proportion of a district that was paved, according to the following categories: area I (minimum wage = 1; <15% of district paved); area II (minimum wage = 2; >15% to <90% of district paved); area III (minimum wage = 3; >15% to <90% of district paved); area IV (minimum wage = 4; >90% of district paved); area V (minimum wage >4; >90% of district paved). Water supply and sewage were not suitable for discriminating socioeconomic status since there were small variations in these variables between the districts. The estimated risk ratio and 95% CI were calculated for the high-risk stratum by age group, using the low-risk stratum as the reference group. Results During the study period, 752 new cases of leprosy were diagnosed. Complete information was available for 711 (94.5%) of them and they were included in the analysis. All patients had residential addresses in the urban area. Almost 80% of all cases were diag- nosed in only one health centre located in the cen- tral zone, irrespective of the geographical distribu- tion of cases and whether there were nearby health units for leprosy treatment. The mean age (±SD) of leprosy patients was 36.1 ± 16.4 years, with males accounting for 51.3% of the cases. A total of 71% were multibacillary cases (lepromatous and border- line clinical forms), while tuberculoid and indeter- minate forms accounted for 14.1% and 14.8% of the cases, respectively. Approximately 85% of the patients reported that their illness had lasted for less than 1 year and the low rate of all disabilities (8.0%) suggests that most cases were of recent onset. Leprosy was detected in 86% of the districts, with the detection rates by district in the range 7.4/ 100 000 to 38 1.3/100 000. Based on the detection rates and confidence intervals, area I and area V correspond to the low- and high-risk stratum, respectively. The intermediate stratum was built by pooling areas II, III and IV, since there is overlap of the confidence intervals, indicating similar detection rates (Fig. 1). Table 1 shows the newly detected cases of lep- rosy, stratified by age groups in the three socioeco- nomic strata. For the high- and intermediate-risk areas the detection rates increased up to the 45-54- year age group, while the peak incidence in the low- risk areas was for the 55-64-year age group. The overall risk of detecting leprosy cases was 5.3-fold higher (95% CI: 3.8-7.4) in the outskirts of the town than in the central zone. Also, estimated risk ratios were greater than unity when they were stratified by age and were statistically significant for all age WHO Bulletin OMS. Vol 73 1995316 Leprosy in an urban area of Brazil Fig. 1. Detection rates of leprosy cases and 95% confi- dence Intervals according to risk strata. V 40 2IV 0 0 I H = 95% cordidence interval -Ii I_ - - ---- a .0 CC IH 0 25 50 75 100 125 150 175 200 Detection rate (per 100 000) nwosso groups except for the youngest and oldest. Approxi- mately 83% of the population lives in areas consid- ered to be of high or intermediate risk for leprosy (Fig. 2). Discussion Our findings identified the spread of leprosy cases in an urban area of Brazil. However, there was a clearly defined high-risk area in the outskirts of the city, with the belt of higher detection rates corresponding roughly to the areas of low socioeconomic status that were lacking in health facilities. The geographical distribution of leprosy was far from uniform, with some groups being at a much higher risk of infection than others (15). About 15% of the population of Goiania lives in areas where detection rates were high and socioeconomic status was low. The relationship between leprosy and low socioeconomic status has long been recognized. Higher prevalences of leprosy in areas of low socio- economic status have been reported by Zuniga (16). More recently, Ulrich et al. found that low standards of living were closely related to infection with lepro- sy (17). While the relationship between poor health and lack of socioeconomic development is generally recognized, causal links have proved difficult to identify. Regionally, periurban populations are composed of rural migrants who are largely unskilled. The high concentration of leprosy cases in peripheral settings indicates an inverse association between the inci- dence of the disease and the quality of the environ- ment the individuals live in. In addition, an intense rural-urban migration in the 1960s and early 1970s may have contributed to this geographical distribu- tion of cases. It is interesting to note that in more developed settings the peak age group for leprosy is 10 years older than that elsewhere, perhaps because the age at infection is greater in the former settings. In our study, self-reported current residence was used to calculate detection rates by districts. This could be considered a relatively insensitive approach for the following reasons: leprosy has a long latent period; health facilities in the cities may attract patients from periurban and rural areas, leading to an overestimate of the leprosy distribution in urban areas; and leprosy patients could provide incorrect addresses because of the social stigma of the disease. The geographical mobility of cases also reduces the sensitivity to detect environmental hazards associated with leprosy (18). Lifetime residential histories are therefore suitable for analysing rural-urban migra- tion and different intervals of exposures for risk assessment purposes. Cases of leprosy detected by health services may have underestimated the real incidence, since they Table 1: Comparison of newly detected cases of leprosy in urban areas of Goiania city, central Brazil Low-risk stratum: Intermediate-risk stratum: High-risk stratum: Age groupa No. of Population No. of Population No. of Population Estimated (years) cases size Rateb cases size Rateb cases size Rateb risk ratioc 1-14 2 51 687 3.8 33 208 940 15.7 7 44 024 15.9 4.2 15-24 14 32 653 42.8 90 131 995 68.1 56 27 811 201.3 4.7 25-34 9 25 998 34.6 100 105 095 95.1 46 22 143 207.7 6.0 35-44 5 18 260 27.3 94 73 817 127.3 39 15 553 250.7 9.2 45-54 5 11 916 41.9 73 48 168 151.5 32 10 149 315.3 7.5 55-64 5 7 428 67.3 36 30 037 119.8 15 6 326 237.1 3.5 >65 4 6 809 58.7 36 27 525 130.7 5 5 799 86.2 1.5 Total 44 154 751 28.4 462 625 567 73.8 200 131 805 151.7 5.3 a Age data were missing for 5 cases. b Per 100 000 inhabitants. c With low-risk strata as the reference group. WHO Bulletin OMS. Vol 73 1995 317 C.M.T. Martelli et al. Fig. 2. Map showing spatial leprosy distribution pat- terns in Golinia city, central Brazil. represent only a proxy measure of all newly diag- nosed cases (19). Nevertheless, the use of such cases for correlation studies is preferable to the use of prev- alence data, since the nature of the association between socioeconomic conditions and leprosy may be influenced by the duration of the disease. The underlying assumption of ecological studies is that the variables describing groups of individuals corre- spond to the individuals themselves, which may be fallacious (20). Our interpretation of the spatial dis- tribution of leprosy based on the 845 census tracts (340 neighbourhoods) in Goiania may have some pitfalls because of the relatively small populations in each area. In such situations detection rates with large confidence intervals might be expected, imply- ing less precision in assessing leprosy indicators. We therefore consider the 64 study districts to be a rea- sonable scale for population samples; also, socioeco- nomic data were available for them. In contrast, if districts are merged to form larger regions, although interpretation of the data would have been easier, the ability to discriminate more precisely between risk areas and clusters of leprosy would have been more difficult (9). Public health services supplying primary health care were distributed throughout all the districts. However, few health units deliver leprosy treatment, and almost all cases were diagnosed and treated in only one unit located in the centre of the city. The health care coverage was clearly insufficient for lep- rosy surveillance purposes. Although the principle of using primary health care in leprosy control has been widely accepted in theory (21),c less developed areas still tend to concentrate their control activities in tra- ditional leprosy treatment centres. An effort should therefore be made to change this situation if leprosy is to be eliminated from underdeveloped regions. The spatial distribution of leprosy and the eco- logical approach of the study permitted the identifi- cation of different risk strata. Application of the method to epidemiological studies may provide use- ful data for health planning activities. Many of the issues regarding leprosy-demand versus need-based health service attendance as well as treatment com- pliance-still need to be tackled. The priority should focus on building stratified models using risk analy- sis to improve health care delivery in urban settings. Case-control studies are currently being carried out to determine the differences in case detection result- ing from migration and socioeconomic status. Also, the study population's awareness and knowledge about leprosy should be evaluated in order to opti- mize the provision of health care. c Buchmann H. Leprosy control services as an integral part of primary health care programmes in developing countries. Wurz- burg, Germany, German Leprosy Relief Association, 1978. Acknowledgements We thank the Regional Leprosy Control Programme (Secretaria de Saude e Meio Ambiente) and FundaqAo Nacional de Saucde for administrative support. We are also grateful to J. B. Peres Junior for entering the data. Resume J R6partition spatiale de la lepre dans une zone urbaine du centre du Bresil Cet article d6crit la repartition spatiale des cas de lepre dans une zone urbaine du Br6sil et sa corr6- lation avec les indicateurs socio-6conomiques. Entre novembre 1991 et octobre 1992, 752 cas de lepre nouvellement diagnostiqu6s, vus dans 1'ensemble des dispensaires de Goiania, dans le centre du Bresil, ont ete identifies. Une base de donnees concernant les cas de lepre a ete eta- WHO Bulletin OMS. Vol 73 1995318 Leprosy in an urban area of Brazil blie, et l'on a reli6 l'adresse des malades a 64 dis- tricts urbains correspondant a des secteurs de recensement et couvrant des zones de caract6ris- tiques socio-economiques tres voisines. Les taux de d6tection des cas de lepre, avec leurs inter- valles de confiance a 95% (95% Cl), ont ete cal- cul6s par district et par zone socio-6conomique. Des cas de Iepre ont ete detectes dans 86% des districts, et trois niveaux de risque ont Wt6 identi- fies. La zone a risque le plus eleve se situait dans les faubourgs de la ville, et les taux de detection augmentaient des zones les plus developp6es vers les zones les plus pauvres. Le risque de detecter un cas de Ibpre 6tait 5,3 fois plus 6leve (95% Cl = 3,8 - 7,4) dans les faubourgs de la ville que dans le centre. L'article examine les questions m6thodolo- giques liees a la confirmation des cas de 1epre, a la fiabilitd et a l'exhaustivit6 des informations, et a l'interpretation de la repartition spatiale de la lepre par zone. L'absence d'activitds de lutte antilepreu- se au sein des services de soins de sante pri- maires et les caracteristiques d'utilisation des ser- vices antilepreux dans les regions en d6velop- pement ont ete soulignees, de meme que lutilite de I'analyse geographique dans la planification des services de sante. References 1. Zicker F et al. Seroprevalence for Trypanosoma cruzi infection among unskilled urban workers in Central Brazil. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1989, 83: 511-513. 2. Mott KE et al. Parasitic diseases and urban devel- opment. Bulletin of the World Health Organization, 1990, 68: 691-698. 3. Barreto ML. Geographical and socioeconomic fac- tors relating to the distribution of Schistosoma man- soni infection in an urban area of north-east Brazil. Bulletin of the World Health Organization, 1991, 69: 93-102. 4. Noordeen SK. Elimination of leprosy as a public health problem. Leprosy review, 1992, 63: 1-4. 5. Remme JHF et al. The burden of tropical diseases. Medical journal of Australia, 1993, 158: 465-469. 6. Lombardi C. Evaluation of leprosy epidemiology in 12 countries of the Americas, 1980-1983. Bulletin of the Pan American Health Organization, 1989, 23: 284-294. 7. Motta CP, Zuniga MG. Time trends of Hansen's disease in Brazil. International journal of leprosy, 1990, 58: 453-461. 8. Noordeen SK et al. Estimated number of leprosy cases in the world. Bulletin of the World Health Organization, 1992, 70: 7-10. 9. Meade MS. Geographic analysis of disease and care. Annual reviews of public health, 1986, 7: 315- 335. 10. Gesler W. The uses of spatial analysis in medical geography: a review. Social science and medicine, 1986, 10: 963-973. 11. Richards Jr, FO. Use of geographic information systems in control programs for onchocerciasis in Guatemala. Bulletin of the Pan American Health Organization, 1993, 27: 52-55. 12. IPLAN. [Plan for integrated development in Goiania], vol. II. Brasilia, 1992: 113-217 (in Portuguese). 13. Minist6rio da Saude. [Guide to the control of Hansen's disease]. Brasilia, 1993 (in Portuguese). 14. Rodrigues MLO et al. Protective effect of intrader- mal BCG against leprosy; a case-control study in Central Brazil. International journal of leprosy, 1992, 60: 335-339. 15. Fine PEM. Reflections of the elimination of leprosy. International journal of leprosy, 1992, 60: 71-80. 16. Zuniga M. [Advances in the epidemiology of lepro- sy]. Boletin dermatologica sanitaria, 1982, 18: 1 (in Spanish). 17. Ulrich M et al. Leprosy in women: characteristics and repercussions. Social science and medicine, 1993, 37: 445-456. 18. Polissar L. The effect of migration on comparison of disease rates in geographic studies in the United States. American journal of epidemiology, 1981, 111: 175-182. 19. Opromolla DV et al. [An estimate of the prevalence of leprosy by means of investigation into non-specif- ic demand for health services]. Revista de saude publica, Sao Paulo, 1990, 24: 178-185 (in Portu- guese). 20. Piantadosi S et al. The ecological fallacy. American journal of epidemiology, 1988, 127: 893-904. 21. McDougall AC. Leprosy control and the implemen- tation of multiple drug therapy: to what extent can the operational strategy be simplified for primary health care? Leprosy review, 1992, 63: 193-198. WHO Bulletin OMS. Vol 73 1995 319