MSc (Med), Population-Based Epidemiology, School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, 2009 / Introduction: Although socioeconomic inequality in health and mortality is currently on the top of the epidemiologic debate, studies however on the subject among adult population in Africa has been hampered due to the long absence of data in many countries. With the present reliable records of deaths from emerged demographic surveillance systems on the continent, adult mortality can now be accurately estimated. Objectives: The study‟s main objectives were 1. To calculate and show trend in adult mortality rate in ACDIS between 2001 and 2007. 2. To measure the association between mortality and individual‟s socio-economic status in the ACDIS. Methods: Individuals were selected based on age (15-64 yrs) and residency (Resident in the DSA on 1st January 2001). The total number of adults who met the criteria was 33,677; out of whom 4,058 died during the seven years follow up period. Mortality rates were computed using Kaplan-Meier survival estimates expressed per 1000 PYO. Household wealth index was constructed by use of PCA. The association between SES and adult mortality was assessed using Cox proportional Hazard model controlling for potential confounders such as age, sex and marital status. Results: The High group of the socioeconomic quintile had the highest mortality rate of 22.2 per 1000 PYO, 95% CI (20.7 - 23.7). There was no significant trend in the rates among the SES groups. After adjusting for the potential confounders; the effect of socioeconomic status in the highest SES category was 0.10 times less likelihood for death compared to the lowest SES group [HR=0.90, p=0.042, 95% CI (0.81 - 0.99)]. Conclusion: This study revealed that adult socioeconomic status is not significantly associated with adult mortality. Reducing the gap between the rich and the poor might not be the effective way in reducing adult mortality.

Research report in partial fulfillment for the degree of MSc (Med), Population Based Field Epidemiology, University of the Witwatersrand / Background Worldwide, injuries are ranked among the leading causes of death and disability, killing over 5 million people and injuring over 50 million others globally. Approximately 90% of these deaths occur in developing countries. The burden and pattern of injuries in low-income countries are poorly known and not well studied. Few studies have been conducted on injury mortality and therefore this study can add to the scientific literature. Analyzing injury mortality in rural Tanzania can assist African countries to develop intervention programmes and policy reform to reduce the burden caused by injuries. Objectives The objective of this study was to identify the risk factors and trend in injury mortality in the Rufiji Demographic Surveillance Area in rural Tanzania from 2002-2007. Specifically, the study would identify and describe the types and trends in injury mortality, calculate the crude death rates of injury mortality by gender, SES and age groups, describe the risks factors associated with injury mortality, and measure association between the risk factors and injury mortality. Methods Rufiji HDSS data used included people aged 1 year and older from 2002-2007. Verbal Autopsy data was used to determine the causes of death which was based on the tenth revision of the International Classification of Diseases (ICD 10) recommended by WHO. Injury Crude death rates (ICDR) were calculated by dividing number of deaths in each year by person years observed and multiplying by 100,000. Principal Component Analysis (PCA) was used to construct household wealth index using household characteristics and assets ownership. Also trend test analysis was done to assess a linear relationship in the injury mortality rates across the six year period. Poisson regression was used to investigate v association between risk factors and injury mortality and all tests for significant associations were based on p-values at 5% significance level and a 95% confidence interval. Results The overall injury crude mortality rate was 33.4 per 100,000 PYO. Injuries contributed 4% of total mortality burden with statistical significant association between gender, age and occupation. Mortality rate was higher for males [Adjusted IRR=3.04, P=0.001, 95% CI (2.22 - 4.17)]. The elderly (65+) were 2.8 times more likely to die from injuries compared to children [Adjusted IRR=2.83, P=0.048, 95% CI (1.01 - 7.93)]. The unemployed, casual workers, the retired, and farmers all had an increased risk of dying from injuries compared to students (P<0.005). Most injury deaths were due to road traffic accidents (28%), unspecified external injuries (20%), drowning (16%), burns (9%), accidental poisoning (8%), homicidal (8%) and animal attack (5%). Conclusion The contribution of injury to mortality burden in the Rufiji Demographic Surveillance Area was relatively low. However, there is the need to institute measures that would help prevent injuries. Life saving interventions such as road safety education, regular road maintenance, rapid response to accidents, use of life jackets for fishermen and recreational swimmers are very necessary in preventing injuries. Also, proper fishing practices should be imparted to the populace as precautionary measures to reduce the burden of injury mortality.

Background: This study examines the association between place of delivery, delivery assistants and perinatal mortality in Nigeria. Previous studies have found these factors to be associated with the risk of perinatal mortality. This study therefore aims to determine the extent to which these two factors predict perinatal mortality in the Nigerian context as this information will be useful in informing health policy decisions and actions in so far as a desirable reduction in childhood mortality in Nigeria is concerned. Methods: This study uses cross sectional design through secondary analysis of the 2003 Nigerian Demographic and Health Survey (NDHS). The variables representing place of delivery and delivery assistants have been fitted into logistic regression models to determine their association with perinatal mortality. Several other known risk factors for perinatal mortality such as maternal education and birth weight, to mention a few, have also been investigated using the logistic regression analysis. Results: 5783 live singleton births were analyzed with 194 newborns dying within the first seven days of life giving an early neonatal mortality rate (ENMR) of 33.5 per 1000 and an estimated perinatal mortality rate (PNMR) of 72.4 per 1000 live births. The results also show that place of delivery [p=0.8777] and delivery assistants [p=0.3812] are not significantly associated with perinatal mortality even after disaggregating the analysis by rural and urban areas. However being small in size at birth [AOR= 2.13, CI=1.41 – 3.21], female [AOR=0.57, CI= 0.42 – 0.77] and having a mother who practiced traditional religion [AOR= 4.37, CI= 2.31 – 8.26], were all significantly associated with perinatal mortality. Conclusions: Place of delivery and delivery assistants are not good predictors of perinatal mortality in the Nigerian context. However various limitations of the study design used such as the issue of uncontrolled confounding may have affected the findings. Nonetheless, the increased risk of perinatal deaths in small babies and the decreased risk of death among female babies are consistent with other studies and have both been attributed elsewhere to biologic mechanisms.

Background: This study examines the socioeconomic and biodemographic determinants of infant, neonatal and post neonatal mortality in Egypt. It also tries to reach better understanding on the relative importance of these factors. Methods: Population-based cross-sectional secondary data analysis of the 2005 Egypt Demographic and Health Survey (EDHS). Cox’s proportional hazard models have been fitted to select the significant determinants of infant and post neonatal mortality, while logistic regression models were adopted for the neonatal period. Results: 13,851 live births were analyzed with 463 deaths before the first birthday; the total follow up time is 144,835 babymonths, giving an IMR of 32 per 10,000 baby-months. After adjusting for all socioeconomic and biodemographic factors, the analyses reveal strong association between infant mortality and biodemographic factors, while the only significant socioeconomic determinant is the mother’s education. Also it has been shown that mother’s education, child’s sex and place of delivery are time dependent covariates. Analyses of neonatal period indicate no association with any socioeconomic factor, while child’s sex and place of delivery are significant predictors. Exclusion of neonatal deaths shows that the risk of post neonatal mortality is inversely related to mother’s education, and not determined by sex of the child nor place of delivery. The risk of infant, neonatal and post neonatal death is consistently related to birth interval and birth size. Conclusions: Biodemographic characteristics represent the most substantial impacts on infant mortality. The only significant socioeconomic predictor (maternal education) has a modest impact, at best, on infant mortality, which appears at later stages of infancy period (namely post neonatal period); since the later proved to be a time varying covariate. Unlike neonatal period, analysis indicates lack of association between post neonatal mortality and child’s sex contradicting the biological knowledge, and supporting the hypothesis of gender discrimination and male sex preference.

A research report submitted to the Faculty of Health Science, in partial fulfillment of the requirement for the award of a Master’s Degree in Nursing Sciences by the Witwatersrand University of Johannesburg, South Africa. / This study, entitled “Knowledge of and challenges of health workers managing maternity patients at PHC facilities of Yobe State, Nigeria,” had the following objectives:  To describe the socio-demographic characteristics of the health workers in the Primary health care (PHC) facilities in the state.  To determine the level of knowledge of the health workers regarding the management of maternity patients in Primary health care facilities in Yobe State.  To determine the challenges experienced by the health workers in the Primary health care facilities in the state. The study was primarily aimed at exploring the health workers level of knowledge, skills and competencies in the management of maternity patients and their contributions towards the control and reduction of maternal mortality in the state. METHODS: A cross sectional survey study in PHC clinics of six selected local government areas of the state, namely Gujba, Geidam, Fune, Fika, Nguru and Jakusko local government areas. A total of 221 health workers (n=221) were interviewed using structured questionnaires, whilst 46 facility managers (n=46) were given self-administered questionnaires while on duty, and data were collected concurrently from clinic records. The data were cleaned, entered in to Epi info statistical software, imported and analysed using STATA. Descriptive and inferential statistics were used to interpret the outcomes of the analysis. RESULTS: More than half of the respondents were female (65.61%) with an average age of 33 years (SD± 8.1). Categories of the health workers who participated in the study were SCHEW, JCHEW, SSCE, TBAS, EHA/EHO and others who were not trained in any form as health workers, but were found running the affairs of maternity patients. More than 80% of the clinics did not have functional ambulances and there was gross inadequacy of basic services in most of the clinics; only 14 out of 46 clinics had portable water supply and electricity. There was a large seasonal turnout of patients in the clinics but poor patronage by maternity clients. The health workers in the maternity unit were found to be deficient in the knowledge, skills and competencies to manage maternity patients. Protocols were not followed (68.78%), some significant information about ante-natal care was not given to v clients and maternity clients were not adequately informed of some of the danger signs of pregnancy. There was a poor standard of institutional deliveries and the majority of the Health Workers had no training on EOC or conducting a clean and safe delivery. CONCLUSION: Examining the background of the current health workers in the primary health care facilities, the attainment of MDG 5 by 2015 will not be a reality, unless urgent measures are put in place, including large recruitment and motivation midwives, and deployment of these midwives to the rural health clinics. There is a need for the existing health workers to be supported by government and enrolled in short course training in Colleges of Midwifery to acquire the much needed skills and competencies for the care of maternity patients.

o presente estudo teve como objetivo a identificação de fatores de risco relativo de morte em criancas de um a sessenta meses de idade residentes em áreas de pobreza da cidade de Recife. O delineamento utilizado é do tipo caso-controle. Estudou-se o papel de variáveis relacionadas com o saneamento básico, bem como a renda, a educacão dos pais, a densidade demográfica intra-familiar, a história reprodutiva da mãe, o aleitamento materno e a assistência à saúde. Foram estudados 149 óbitos - \"casos\", e número igual de sobreviventes - \"controles\", pareados segundo a idade e o local de residência. Estimaram-se riscos relativos de morrer em análises univariadas para o total da amostra e para estratos segundo faixas etárias e causas básicas de ocorrência do óbito. Mediante modelos de regressão logística foram conduzidas análises multivariadas para todas as causas e idades e para dois sub-qrupos: o de menores de um ano e o de óbitos por causas gastrointestinais e seus \"controles\". Foi constatado elevado risco relativo de morte nas criancas das famílias com mais de três pré-escolares no domicílio. Em todas as análises univariadas realizadas o risco dessas criancas foi superior a 3.6 (P < 0.01) cheqando a 7.8 (P < 0,01) na análise multivariada conduzida para estimar o risco de morte por causas gastrointestinais. Com relacio ao tratamento intra-domiciliar da água, foi observado risco relativo superior a 3.8 (P < 0.01) na análise univariada; porém, nos modelos ajustados pela condição de amamentação e alfabetismo materno o risco de óbito gastrointestinal tornou-se inexistente. Na maioria das análises, foi também elevado o risco relativo nas crianças das mães com alta fecundidade materna. Igualmente o acesso aos servicos de saúde - medidas através da falta de atenção pré-natal - constitui risco relativo de morte, confirmando-se nas análises multivariadas finais, referentes à morte por causas gastrointestinais (OR = 11.11; P < 0.05). O aleitamento também mostrou ser fator preditivo de morte pós-neonatal por causas gastrointestinais (OR = 3.3;P < 0,05). Com base nestes resultados recomenda-se: (a) realizar investimentos de base ampla destinados a melhorar as condições de saneamento básico da cidade; (b) fornecer uma infra-estrutura de apoio para os cuidados das criancas a nível comunitário e/ou nos lugares de trabalho; (c) promover o espaçamento adequado entre os nascimentos; (d) aumentar a cobertura dos serviços de saúde, fomentando estratégias simplificadas de atenção primária com ativa participacão comunitária; (e) estimular ações permanentes e rotineiras de promoção do aleitamento ademais das campanhas eventuais. Estes resultados fornecem subsídios para a elaboração de políticas sociais destinadas a acelerar o processo de queda da mortalidade nos menores de cinco anos e identifica novos indicadores para entender os mecanismos que levam ao óbito nos estratos sócio-econômicos urbanos dos países em vias de desenvolvimento. / The aim of this case-control study was the identification of relative risk factors of mortality in children of 1 to 60 months of age resident in poor areas of the city of Recife. Sanitation, income, parent´s education, family density, mother´s reproductive history, breastfeeding and health care were studied and their roles as exposure factors were established. The sample consisted of 149 \"cases\" (dead children) and of the same number of \"controls\" (survivors) case-matched according to age and place of residence. Odds ratios (OR) were estimated through univariate analysis for the whole sample and for specific strata by age and cause of death. Multivariate modelling was then performed for infants and for gastrointestinal deaths and their \"controls\". High relative risk of death in families with more than three preschool children was verified. According to the univariate analysis the relative risk for multivariate analysis performed for gastrointestinal causes of death. Children whose families were not in the habit of treating water at home had a relative risk of gastrointestinal death of more than 3.8 (P < 0.01) in the univariate analysis. Whe adjusted for breast-feeding and mother´s literacy condition - through multivariate analysis - this risk was negligible. Most analysis showed important relative risk for children whose mothers had high fecundity. Likewise lack of access to health care - measured in terms of pre-natal care - was associated to a significant gastrointestinal causes (OR = 11,11; P< 0.05). Breastfeeding also shown to be a predictive factor of post-neonatal death due to gastrointestinal causes (OR = 3,3; P < 0.05). Based on these results it is recommended that: a) Sanitation be improved; b) an infrastructure for the day-care of children at community level be improved; c) an adequate intergestational interval be promoted; d) increase of health service coverage throught the strategy of primari health care with active comunity participation; e) the implementation of permanent activities (integrated to health service routine) with the aim of encouraging the practive of breastfeeding. It is, this considered that these results provide a basis for the formulation of social policies aiming at the acceleration of the decrease infant mortality. Also this study indicates new elements for understanding of the mechanism of death in poor urban areas.

Foram estudados 375 óbitos ocorridos no município de São Paulo, de pessoas de 15 anos ou mais residentes nessa cidade e em cujo atestado de óbito constava a tuberculose como causa básica ou associada. Desses 375 óbitos, foram localizadas 245 famílias dos falecidos e realizadas entrevistas domiciliárias. Houve 287 pacientes que faleceram em hospitais, e em dois desses casos não foi conseguido realizar a entrevista institucional. Dessas entrevistas, pôde ser concluído que os possíveis fatores que influíram nesses óbitos foram o baixo nível sócio-econômico e o alcoolismo; não influíram nestes óbitos a migração e o tabagismo. Foi observado que estes doentes entraram tardiamente no sistema de saúde e a maioria dos que chegaram a ser internados faleceram nos primeiros dias de internação. Foi observado também, que a maioria dos casos não estava notificada, nem como caso nem como óbito, e a Secretaria da Saúde tinha conhecimento de apenas 119 casos. Analisando esses atestados, notou-se que a tuberculose foi causa básica em 88,4 por cento deles e, após a revisão dos mesmos com base nas entrevistas domiciliárias e institucionais, a tuberculose passou a ser causa básica em 92,8 por cento , havendo uma concordância de 95,6 por cento entre o atestado original e o revisto. / Three-hundred and seventy-five deaths of persons aged 15 years or more, which ocurred in São Paulo city, have been studied taking into account the death certificates in which tuberculosis had been appointed as the basic or associate cause of death. From these 375 deaths, 245 families of the deceased have been located and then submitted to domiciliary interviews: it was found out that two-hundred and eighty-seven patients had died in hospitals; in two of the cases it was not possible to make an institutional interview. From these data, some factors which could possibly have influenced the deaths were the low-socioeconomic status and alcoholism; it was found out that neither migration nor tabaccoism had influenced the deaths. It was also noticed that these patients had entered the health system very late and the majority of those who had been duly hospitalized died within the first days of hospitalization. It was also observed that most of the cases had not been reported either, neither as a case nor as a death, being that the State Health Department only had knowledge of 119 cases. Analysing these certificates, it was observed that tuberculosis was the basic cause of death in 88.4 per cent of them, and that after reviewing them on the basis of domiciliary and institutionas interviews, tuberculosis came out to be the basic cause of death in 92.8 per cent 06 the causes, having been reached an agreement of 95.6 per cent between the original certificates and the reviewed ones.

Através de levantamento amostral do peso ao nascer dos nascimentos ocorridos nas maternidades do município em 1976, estimou-se em 9,7 por cento a incidência de recém-nascidos de baixo peso em são Paulo. O ajuste dos coeficientes de mortalidade infantil de São Paulo segundo a distribuição do peso ao nascer da população branca americana determinou que em 1976 apenas 31 por cento do excesso de óbitos infantis de São Paulo poderia ser atribuído à performance do peso ao nascer verificada no município. O mesmo ajuste segundo a distribuição do peso ao nascer da Califórnia atribuiu ao peso ao nascer no período 1968-70, 15 por cento do excesso de mortalidade e revelou excesso de mortalidade particularmente notável não para os recém-nascidos de baixo peso, mas para os recém-nascidos com mais de 3.000 gramas. O ajuste simultâneo dos coeficientes de mortalidade infantil das áreas central, intermediária e periférica do município a uma idêntica distribuição de pesos de nascimento revelou uma participação decisiva do peso ao nascer no pequeno gradiente de mortalidade existente entre a área intermediária e a área central, porém descartou totalmente a possibilidade do peso ao nascer explicar o grande excesso de mortalidade infantil da área periférica em relação às duas outras. A consideração dos demais fatores que determinam no seu conjunto a probabilidade de sobrevida do recém-nascido no primeiro ano de vida, demonstrou ser capaz de complementar os achados relativos à participação do peso ao nascer no excesso de mortalidade infantil verificado no município. Assim foi para a disponibilidade de serviços públicos de pré-natal, insuficiente para o município como um todo, progressivamente menor na periferia. Assim foi para a disponibilidade de assistência médico-hospitalar ao parto e ao recém-nascido, suficiente em termos quantitativos para o município, porém insuficiente e precária para a população da periferia. Assim foi para a disponibilidade do saneamento do meio, razoável apenas na área central, praticamente inexistente na periferia. Associadamente, a verificação da distribuição de renda em São Paulo demonstrou que a distribuição desigual da disponibilidade de serviços públicos fundamentais à manutenção da saúde infantil superpõe-se exatamente à desigual distribuição de riquezas. A conclusão final deste estudo foi a de que são principalmente as precárias condições de vida que recepcionam o recém-nascido sobretudo na periferia, e não a vitalidade daquele, medida pelo seu peso ao nascer, que explicam o excesso de mortalidade infantil verificado em são Paulo. / By means of the sample research survey of the birth-weight of births which took place in the Maternity Hospitals of the city in 1976, it was calculated that the incidence of low birth-weight in São Paulo was of the order of 9,7 per cent . The adjustment of the infant mortality rates of São Paulo according to the birth-weight distribution of the white American population led to the conclusion that in 1976 only 31 per cent of the excess of infant mortality of São Paulo could be attributed to the birth-weight distribution found to exist in the city. The same adjustment according to the birth-weight distribution in California attributed 15 per cent of mortality excess to birth-weight in the period 1968-70 and a particularly clear mortality excess not for the infants of low birth-weight, but for those with a birth-weight of more than 3000 grams. The infant mortality coefficients of the central, intermediate and outlying areas of the city when seen against a hypothetical distribution of birth-weights for all areas revealed a decisive effect of the birth weight only in the small difference between the mortality rates of the intermediate and the central areas, but eliminated altogether the possibility of the birth-weight explaining the excess of infant mortality in the outlying areas in relation to the other two Cintermediate and central) areas. The study of the other factors wich determine, together, the probability of an infantis survival in the first year of life, was seen to corroborate the conclusions relative to the part played by birth-weight in the infant mortality excess of the city. This was true in relation to public pre-natal services, insufficient for the city as a whole, and becoming progressively still less numerically adequate in the outlying areas. It was also true in relation to the medical and hospital attendance at the birth and to the new-born infant, wich were adequate in quantitative terms for the city as a whole, but inadequate and unreliable for the population of the outlying areas. It was true for the availability of water supply and sanitation systems, which are reasonable only for the central area but practically non-existent in the outlying areas. At the same time, research on income distribution in são Paulo showed that the unequal distribution of the availability of those public services essential to the maintenance of infant coincides exactly with the unequal distribution of wealth. The final conclusion of this study was that it is mainly the inadequate living conditions wich await the new-born infant, especially in the outlying areas, and not his/her vitality as measured by birth-weight, which explain the infant mortality excess found in são Paulo.

A mortalidade por câncer no Estado de São Paulo vem crescendo em importância à medida em que se processam as transições demográfica e epidemiológica. Atualmente, cerca de 13 por cento das mortes no Estado se devem ao câncer, prevendo-se aumento futuro nessa proporção. As neoplasias malignas mais frequentes como causa de morte foram, em 1993, as de pulmão, estômago, próstata, leucemias e linfomas e de boca e faringe, nos homens, enquanto, nas mulheres, foram as de mama, colo do útero e útero não especificado, estômago, cólon-reto e leucemias e linfomas. Houve diferenças regionais na mortalidade por câncer dentro do Estado de São Paulo, destacando-se a região administrativa de Santos, cujas taxas padronizadas foram as maiores, em ambos os sexos. Houve decréscimo das taxas padronizadas de mortalidade pelo conjunto dos cânceres entre 1972 e 1992, maior nas mulheres (-7,1 por cento), que nos homens (-3,7 por cento). As tendências temporais idade-específicas para o conjunto dos cânceres foram examinadas por análise de regressão que utiliza a distribuição de Poisson, com resultados significativos indicando quedas da mortalidade nas idades até 69 anos, nos dois sexos, estabilidade nas mulheres acima de 70 anos e aumento nos homens dessa idade. A maior redução de taxas, tanto padronizadas quanto idade-específicas, ocorreu para o câncer de estômago, em ambos os sexos, com resultados significativos em todas as idades. Ao contrário do câncer gástrico, as neoplasias malignas do pulmão, mama feminina e próstata apresentaram aumento das taxas padronizadas, entre 1970 e 1992; no entanto, para algumas faixas etárias mais jovens, o câncer de pulmão registrou quedas significativas de taxas em ambos os sexos. Enquanto isso, o câncer do colo do útero e as leucemias permaneceram estáveis, mas nestas últimas notou-se queda de mortalidade nas crianças até 9 anos. Tratando-se de doenças distintas, cuja etiologia é complexa. e geralmente não conhecida, exceto por alguns fatores de risco e mesmo estes, apenas para algumas delas, as tendências declinantes observadas nos mais jovens não têm uma explicação única, mas, ao contrário, devem-se, com maior probabilidade, a uma combinação peculiar de fatores específicos para cada câncer, não se podendo desprezar a contribuição advinda dos progressos materiais, sociais e tecnológicos ocorridos nas últimas décadas. / Cancer mortality has becoming progressively more important in the State of São Paulo as the demographic and epidemiologic transitions progress. Currently, around 13 percent of ali deaths in the State are due to the malignant neoplasms, and it is reasonable to anticipate a further increase in that proportion. The most frequent cancers causing deaths were, in 1993, those of the lungs, stomach, prostate, the leukemias and lymphomas and cancers of the mouth and pharinx, in males, and those of breast, uterus, stomach, colon and rectum and the leukemias and lymphomas, in females. There were regional differentials in cancer mortality within the State: the administrative region of Santos presented the highest age-adjusted mortality rates, both for males and females. There has been a decrease in the age-adjusted death rates between 1970 and 1992, with rates for females decreasing more (less 7,1 per cent) than those for males (less 3,7 per cent). Age-specific time trends for ali cancer mortality were evaluated by means of the technique of regression analysis using the Poisson distribution. Significant decreases in rates were found throughout ali age-groups up to 69 years, for both sexes. For the eldest, stable mortality rates were found among women, whereas a trend for an increase was observed among men. The largest reductions in rates were observed for stomach cancer, for both sexes, with decreases both in age-standardized as in age specific death rates; in the latter, significant decreases in rates were observed for all age-groups. lnversely, cancers of the lung, breast and prostate ali experienced increases in their age-adjusted rates between 1970 and 1992, while cervical cancer and the leukemias remained fairly stable, but, for the latter, a significant fali in mortality was observed in younger ages. As different cancers are distinct diseases, whose causation is complex and usually not known in full detail, except for a number of risk factors and only for some of them, reasons for the declining mortality in the youngest should be due to a number of factors peculiar to each cancer considered, although the contribution of material, social and technological progresses that happened in the last decades should not be omitted.

A partir da metodologia utilizada por PUFFER e GRIFFTH, 1967, PUFFER e SERRANO, 1973, GUIMARÃES, 1979, LAURENTI e cal. 1988 e MELLO JORGE, 1988, foi possível caracterizar o óbito das mulheres de Diadema/SP. A maioria da população alvo era de mulheres idosas (64 por cento acima de 50 anos), oriundas principalmente do Nordeste, com baixa escolaridade (51,4 por cento de analfabetas), com baixo nível sócio-econômico (66,7 por cento com até um salário mínimo/pessoa na casa). Das que tinham assistência em saúde (86,3 por cento ), quase 70 por cento eram seguradas pelo INAMPS. Encontrou-se alto índice de alcoolismo nessa população. As principais causas de óbito foram as Doenças do aparelho circulatório, Neoplasias e Causas externas. No grupo de mulheres em idade fértil as principais causas foram Causas externas e Complicações da gravidez, do parto e do puerpério. As principais queixas pré-existentes foram a hipertensão e o diabetes. O coeficiente de mortalidade materna foi o de 289, 32/cem mil nascidos vivos, considerado muito alto entre os países aos quais foi comparado. A metodologia permitiu também alteração nas causas de óbito, assim sendo, houve aumento das Doenças das glândulas endócrinas, das Complicações da gravidez, do parto e do puerpério; o aparecimento das causas Transtornos mentais; e a diminuição da Doenças do aparelho respiratório e das causas mal definidas. / Starting with the methodology used by PUFFER & GRIFFTH (1967), PUFFER & SERRANO (1973), GUIMARÃES (1979), LAURENTI et al (1988), and MELLO JORGE (1988), it was possible to characterize the deads of Diadema/SP women on 1985/6. This population as being formed of old women (64 per cent over 50 years), mainly from the Northeastern Brazil, poor literacy (51. 4 per cent illiterate), with low social-economic level (66.7 per cent with up to one minimum salary earned by each person in the house). Almost 70 per cent of those who had medical and hospital assistance (86.3 per cent ) were insured by INAMPS. A high rate of alcoholism was found in the population. The main death causes were circulatory system Diseases, Neoplasias and external Causes. In the prolific age group of women, the main causes were external Causes and pregnancy, delivery, and puerperium Complications. The main pre-existing complains were hypertension and diabetes. The mortality death rate was 289.32/º born alive, which was found to be very high in comparison to other countries. The methodology also enabled an al teration in death causes. Thus, there was an increase of endocrinous gland Diseases, pregnancy, delivery, and puerperium Complications; an emergence of mental perturbations Causes, and a decrease in the respiratory system Diseases and badly defined Causes.