Influence of Demographic Determinants on the Number of Deaths Caused by Circulatory System Diseases in Comparison to the Number of Deaths Caused by Neoplasms in Slovak Regions from 1996-2014

Abstract

Objectives: The objective of our study was to evaluate the influence of available demographic determinants on the number of deaths caused by circulatory system diseases as compared to deaths caused by neoplasms in Slovakia in 1996-2014.

Methods: Mortality data were kindly provided by the National Health Information Centre in Slovakia. The first method was trend curve fitting of death ratios caused by circulatory system diseases (Chapter IX) and of deaths caused by neoplasms (Chapter II) as a function of age for both sexes. The second method comprised a decision tree for classification between deaths caused by Chapter IX and Chapter II diseases. Input variables were available demographic indicators: age, sex, marital status, region, and calendar year of death. Statistical data analyses were performed by IBM SPSS version 19 statistical software.

Results: We found that the odds ratios of deaths caused by circulatory system diseases (Chapter IX) in comparison with deaths caused by neoplasms (Chapter II) were non-decreasing. At first, the values of odds ratios are constant until they reach a critical sex-dependent value with a subsequent steady increase. In the case of men the odds ratio was greater than in the 60 years age-group where the odds ratio value increased slowly (from 1.14 at age 60 to 7.25 at age 90 years). The relative increase was 6.36 (7.25/1.14). The odds ratio in the women group was smaller but increased more rapidly (from 0.81 at age 60 to 12.27 at age 90 years). The relative increase was 15.15 in women (12.27/0.81). Hence, the odds ratio of death caused by Chapter IX diseases vs. Chapter II was greater in the older women group (i.e. higher age values). Utilizing the decision tree model, we have found that the most significant demographic determinant of death counts in both ICD Chapters was the age of the deceased, followed by marital status and finally gender. The last two predictors (year and region) were relatively negligible though formally significant.

Conclusions: The proposed method could be useful for prognostic classification of patients and primarily beneficial for hospitals in human or financial resources planning.

Keywords: decision tree; diseases of the circulatory system; mortality; neoplasms.