MSc(Med), Population-based Field Epidemiology, Faculty of Health Sciences, University of the Witwatersrand, 2009

The purpose of this research project was to compare inpatient mortality rates for rural hospitals with mortality rates of urban hospitals of given sizes and ranges of service. Statistical adjustments for risk were made in the probability of death during hospitalization for 43,000 patients across 166 hospitals by age, gender, principal diagnosis, principal surgical procedure, characteristics of the secondary diagnoses, and whether or not cancer was a secondary diagnosis. Eighty-three small hospitals that had a relatively unspecialized range of services constituted the study group. Patient characteristics of this study group were moderately representative of the national population. A standardized score was calculated for each hospital using a formula based on the actual hospital death rate and the death rate expected for a given hospital with patients of the same demographic and medical characteristics. Patients admitted to hospitals in nonmetropolitan areas had a mortality rate of 0.41 percent compared with a mortality rate of 0.66 percent in peer hospitals in metropolitan areas. After mortality rates were risk-adjusted and converted to z scores, nonmetropolitan areas had an average z of +0.16, and metropolitan areas had an average z of -0.25, where positive z scores reflect a lower-than-average adjusted mortality rate. The metropolitan-nonmetropolitan (urban-rural) difference was not statistically significant, but it is meaningful in that rural hospitals tended to have a lower adjusted mortality rate than urban hospitals of the same size and type, indicating that rural hospitals had the same or lower adjusted mortality rates. The possibility of urban hospitals having riskier patients was minimized but could not be definitively ruled out. Taken together with other studies, the data are consistent with the view that small rural hospitals generally make appropriate transfer decisions for severely ill patients and provide quality care for retained patients.

Esta dissertação se baseia na exploração de técnicas para modelagem de óbitos provenientes de acidentes por automóvel no estado de São Paulo. A análise foi agregada por área, e utilizou a razão de óbitos por população, por área e por fluxo veicular como variáveis dependentes e as variáveis independentes foram características socioeconômicas, área, frota de veículos, IDHM, fluxo veicular anual e distâncias entre microrregiões. Os dados do ano 2000 foram utilizados na calibração e dados de 2010 na validação dos modelos, com a técnica de mineração de dados (algoritmos de Árvore de Decisão - AD: CART - Classification And Regression Tree e CHAID - Chi-squared Automatic Interaction Detection) e Regressão Linear Múltipla (RLM) para fins comparativos com os modelos de AD. A partir dos resultados verifica-se que a RLM foi a técnica que obteve melhores erro médio, erro médio absoluto e coeficiente de correlação, e o algoritmo CART da AD o menor erro médio normalizado. Ao comparar as taxas de óbitos, a relação por área apresentou melhor erro médio e coeficiente de correlação, já a relação por população obteve menor erro médio normalizado e erro médio absoluto. Vale ressaltar que os algoritmos de AD são técnicas adequadas para classificação de áreas segundo faixas de valores de variáveis explicativas e valores médios da variável objeto de estudo. Além disso, tais técnicas são mais flexíveis em relação a alguns pressupostos de modelos de regressão. Dessa forma, a principal contribuição deste trabalho consiste na exploração de tais algoritmos para previsão de acidentes e classificação de regiões. / This dissertation is based on techniques exploration for modeling of deaths from automobile accidents on the state of São Paulo. The analysis was aggregated by area, and used the ratio of deaths per population, by area and by vehicle flow as dependent variables and the independent variables were socioeconomic characteristics, area, vehicle fleet, Municipal Human Development Index (MHDI), annual vehicle flow and distances between micro-regions. The 2000 data were used for calibration and 2010 data to validate the models with data mining technique (decision tree - DT algorithms: CART - Classification And Regression Tree and CHAID - Chi-squared Automatic Interaction Detection) and Multiple Linear Regression (MLR) for comparative purposes with the DT models. From the results it appears that the RLM was the technique that achieved better mean error, mean absolute error and correlation coefficient values, while the CART algorithm presented the lowest value of mean normalized error. When comparing death rates, a relation by area showed better mean error and correlation coefficient values, as the ratio by population had lower mean normalized error and mean absolute error values. It is noteworthy that the DT algorithms are suitable techniques for classification of areas in accordance with explanatory variables of value ranges and average values of the variable object of study. Furthermore, such techniques are more flexible compared to some assumptions regression models. Thus, the main contribution of this study is the exploration of such algorithms for prediction of accidents and regions classification.

Esta dissertação se baseia na exploração de técnicas para modelagem de óbitos provenientes de acidentes por automóvel no estado de São Paulo. A análise foi agregada por área, e utilizou a razão de óbitos por população, por área e por fluxo veicular como variáveis dependentes e as variáveis independentes foram características socioeconômicas, área, frota de veículos, IDHM, fluxo veicular anual e distâncias entre microrregiões. Os dados do ano 2000 foram utilizados na calibração e dados de 2010 na validação dos modelos, com a técnica de mineração de dados (algoritmos de Árvore de Decisão - AD: CART - Classification And Regression Tree e CHAID - Chi-squared Automatic Interaction Detection) e Regressão Linear Múltipla (RLM) para fins comparativos com os modelos de AD. A partir dos resultados verifica-se que a RLM foi a técnica que obteve melhores erro médio, erro médio absoluto e coeficiente de correlação, e o algoritmo CART da AD o menor erro médio normalizado. Ao comparar as taxas de óbitos, a relação por área apresentou melhor erro médio e coeficiente de correlação, já a relação por população obteve menor erro médio normalizado e erro médio absoluto. Vale ressaltar que os algoritmos de AD são técnicas adequadas para classificação de áreas segundo faixas de valores de variáveis explicativas e valores médios da variável objeto de estudo. Além disso, tais técnicas são mais flexíveis em relação a alguns pressupostos de modelos de regressão. Dessa forma, a principal contribuição deste trabalho consiste na exploração de tais algoritmos para previsão de acidentes e classificação de regiões. / This dissertation is based on techniques exploration for modeling of deaths from automobile accidents on the state of São Paulo. The analysis was aggregated by area, and used the ratio of deaths per population, by area and by vehicle flow as dependent variables and the independent variables were socioeconomic characteristics, area, vehicle fleet, Municipal Human Development Index (MHDI), annual vehicle flow and distances between micro-regions. The 2000 data were used for calibration and 2010 data to validate the models with data mining technique (decision tree - DT algorithms: CART - Classification And Regression Tree and CHAID - Chi-squared Automatic Interaction Detection) and Multiple Linear Regression (MLR) for comparative purposes with the DT models. From the results it appears that the RLM was the technique that achieved better mean error, mean absolute error and correlation coefficient values, while the CART algorithm presented the lowest value of mean normalized error. When comparing death rates, a relation by area showed better mean error and correlation coefficient values, as the ratio by population had lower mean normalized error and mean absolute error values. It is noteworthy that the DT algorithms are suitable techniques for classification of areas in accordance with explanatory variables of value ranges and average values of the variable object of study. Furthermore, such techniques are more flexible compared to some assumptions regression models. Thus, the main contribution of this study is the exploration of such algorithms for prediction of accidents and regions classification.

Master of Arts (Research) / This thesis examines a ledger which listed all the causes of death in Darlinghurst Gaol, Sydney’s main gaol, from 1867 to 1914 when the gaol was closed and all the prisoners were transferred to the new Long Bay Gaol at Maroubra. The ledger lists the name of the deceased prisoner, the date of their death, the age of the prisoner at the time of their death and the cause of death along with any special comments relevant to the death where necessary. This ledger was analysed in depth and the death rates and diseases causing the deaths were compared to the general population in New South Wales and Australia as well as to another similar institution namely Auburn Prison, the oldest existing prison in New York State and the general population of the United States of America (where possible). Auburn Prison was chosen because it was the only other prison in the English speaking world (British Empire and United States of America) that had a similar complete list of deaths of prisoners in the same time frame – in this case beginning in 1888. The comparison showed that the highest death rates were in the general population of the United States of America (statistics on New York State alone could not be found) followed by Auburn Prison followed by the general population of Australia then the general population of New South Wales (the latter two were very similar) and the lowest death rates were in Darlinghurst Gaol. The analysis showed that individuals were less likely to die in the main prison, compared to the relevant general population in New South Wales and New York State despite the fact that 8 – 9% of these prison deaths were due to executions, a cause of death not encountered in the general population. This thesis explores the reasons why mortality rates were lower in prison despite the popular perception was that Victorian era gaols were places of harshness, cruelty and death (think of the writings of Charles Dickens, the great moralist writer who was the conscience of the era) compared to the general free population.

Master of Arts (Research) / This thesis examines a ledger which listed all the causes of death in Darlinghurst Gaol, Sydney’s main gaol, from 1867 to 1914 when the gaol was closed and all the prisoners were transferred to the new Long Bay Gaol at Maroubra. The ledger lists the name of the deceased prisoner, the date of their death, the age of the prisoner at the time of their death and the cause of death along with any special comments relevant to the death where necessary. This ledger was analysed in depth and the death rates and diseases causing the deaths were compared to the general population in New South Wales and Australia as well as to another similar institution namely Auburn Prison, the oldest existing prison in New York State and the general population of the United States of America (where possible). Auburn Prison was chosen because it was the only other prison in the English speaking world (British Empire and United States of America) that had a similar complete list of deaths of prisoners in the same time frame – in this case beginning in 1888. The comparison showed that the highest death rates were in the general population of the United States of America (statistics on New York State alone could not be found) followed by Auburn Prison followed by the general population of Australia then the general population of New South Wales (the latter two were very similar) and the lowest death rates were in Darlinghurst Gaol. The analysis showed that individuals were less likely to die in the main prison, compared to the relevant general population in New South Wales and New York State despite the fact that 8 – 9% of these prison deaths were due to executions, a cause of death not encountered in the general population. This thesis explores the reasons why mortality rates were lower in prison despite the popular perception was that Victorian era gaols were places of harshness, cruelty and death (think of the writings of Charles Dickens, the great moralist writer who was the conscience of the era) compared to the general free population.

Background: In 1946, a proposal was made for a law on national pensions with the aim of reforming the pension system and giving pensioners better finances, which later was implemented at 1948. The theory of the social determinants of health suggests that income is an important factor that has an impact on health. The study aim was therefore to investigate whether the general pension reform implemented in 1948 could account for any differences in death rates for those aged 67 and older, and the research question was to examine if there was change in death rates for those aged 67 and older after 1948 that could be explained by an increase in income from the pension reform  Method: The study design was a quantitative inductive method. Data used in the study was death rates for those aged 67-90 years in Sweden between 1933-1962, as well as statistics from the Statistical Yearbook for Sweden to obtain data on income from the pension system. To investigate whether the pension reform has had any significant effects on the health of people of old ages, an interrupted time series analysis was used to measure changes in death rates for those aged 67-90 years between 1933 and 1962 in Sweden.  Results and conclusion: The results showed that there was an increase in the average pension by 461% for men and 442% for women between the years of 1947 and 1948. The study suggests that for women there was a significant change in death rates since the reform started, but that the reform of 1948 might not have had a significant effect on men’s death rates. Overall my study indicates that the changes in death rates for women in the pension age in 1948 and after could be explained by an increase in income from the pension reform, and that the pension reform seems to be a sustained policy effect that have accumulated over time.

No description available.