Application of Factor Analysis and Principal Component Analysis in the Construction of Indicators

• DOI: 10.22533/at.ed.4222320091

• Palavras-chave: --

• Keywords: Statistical and Health; Indicators; Multivariate Analysis; Rondônia; Brazil

• Abstract: Abstract - The objective of the research: to identify the significance of epidemiological, socioeconomic and coverage indicators of health services in Rondônia, applying the multivariate statistical technique. The method: multivariate statistical technique, using Factor Analysis and Principal Component Analysis method. A sample with 121 variables was used, divided into two scenarios: Scenario 1, with 121 epidemiological, socioeconomic and health service coverage variables in the State of Rondônia; and Scenario 2, with 42 epidemiological, socioeconomic and health service coverage variables with an explanation greater than 95%. Results: The statistical inferences show the following findings: factor (1) called epidemiological indicators, presented 69 variables with explanation above 70%, factor (2) called socioeconomic indicators highlighted 4 variables with explanation above 70% and factor (3) indicators of coverage of health services highlighted only 2 (two) variables with explanation above 70%. Scenario 2, composed of 42 variables with explanation above 95% in the application of Factor Analysis and Principal Component Analysis obtained 34 variables with explanation above 70%. Of these, 25 are of the factor (1) epidemiological indicators and presented explanation above 95%. The variable with the greatest explanatory power is the first with 99.66% (referring to the % of children aged 10 to 14 years, with more than 1 year of school delay). The correlation between factors (1), (2) and (3) is explained by 66.53% for epidemiological indicators, 18.72% for socioeconomic indicators and 7.53% for service coverage indicators. From this study it is possible to infer the almost absolute predominance of epidemiological variables and, with less emphasis, of socioeconomic variables, which measure access to elementary education in the State of Rondônia. The variables of coverage of health services did not present significance. Conclusions: the study showed that of the 39 indicators agreed between the State Department of Health of Rondônia and the municipalities, 33 (thirty-three) are included among the indicators with statistical significance, which validates the choice of method.