Data analaysis of non-communicable chronic diseases in Brazil

Abstract

Non-communicable chronic diseases (NCDs) include illness that are not transmitted from person to person such as high blood pressure, diabetes and cancer. NCDs are collectively responsible for the majority of deaths in Brazil and worldwide. This thesis aims to evaluate non-communicable chronic diseases in the Brazilian population, considering demographic, geographic, and socio-economic characteristics. The thesis methodology comprises a multi- paper approach composed by statistical and decision-making applications in the health management context focused in NCDs. Data analysis strategies of cleaning, organizing, and modeling data were explored to extract useful patterns to support data-driven decisions. Information from cross-sectional population-based researches in Brazil was used, considering the health supplements of the National Household Surveys (PNAD) and the National Health Survey (PNS). These databases have the potential to be increasingly used, justifying scientific researches to evaluate large amount of Brazilian health information. Initially, a systematic literature review was executed summarizing previous researches on NCDs in Brazil. Thereafter, it was investigated trends in the prevalence of NCDs related to age and sex from a temporal perspective showing that NCDs are highly associated with aging. Additionally, in a multi- period analysis, it was observed that women have been more affected than men have. Furthermore, evaluating NCDs in a geographic perspective, from the proposed Chronic Disease Index (gCDI) it was detected that the South of Brazil concentrated more chronic diseases compared with the North. It was executed an overtime analysis of multimorbidity, that is the co-occurrence of two or more chronic diseases in the same person. It was noticed that the risk of multimorbidity was higher for women and illiterate, increasing with age. People with multiple chronic diseases consider their health worse compared with people without chronic illnesses, demonstrating a greater need for health assistance and hospitalization. At the end, an integrated health assessment was proposed from an outranking multicriteria decision model using multiple periods of time, considering objective and subjective perspectives of health, that can be used as an indicator for monitoring the population overtime. Effectively, quantitative data analysis in different periods can facilitated trends observations to support health management. For instance, it was detected a tendency of increased risk of chronic diseases in groups like women and a higher need for health appointments and hospitalizations in individuals with multiple chronic diseases. PNAD and PNS researches demonstrated to be valuable data sources about chronic diseases in Brazil. A better understating of health patterns might support policy-makers to improve preventative actions that significantly affect groups of people with chronic conditions and multimorbidity.