InterCriteria Analysis of oncological data of the patients for the city of Burgas

The InterCriteria Analysis approach is applied to data connected with the field of oncology. The statistical data for newly registered patients with oncological diseases for 2018 in Burgas are studied. The results are commented from different points of view: relations between gender and age of the patients, relations between gender and marital status.


Introduction
The number of cancer patients in Bulgaria shows a steady upward trend in the last decades [20,36]. The trend has been marked by more than double the incidence of new cases compared to the 1980s. besides the absolute values, the newly formed cancers occupy an increasing share of the structure of the causes of hospitalization and mortality of the population, as they are second only to the diseases of the blood circulation. Considering the cost of treatment, traceability and prophylaxis, the economic effect on this trend is not to be underestimated either. In this respect, the fact that amongst people with temporary and permanent disability, tumors are the first to be considered, should not be neglected either.
With the advances in invasive cardiology, significant advances in cardiovascular disease prophylaxis and control, and increased public responsibility for this issue, the relative incidence of morbidity, mortality, hospitalization and disability in relation to these diseases declines considerably and yields room to neoplasms that take a significant lead.
In summary, these observations on the dynamics of disease causes of public expenditure and severe life-health events, oncological diseases are a major socio-economic problem, which should be a priority. Trend monitoring, dependence, correlation between causal and investigative factors would have a tremendous impact and help in resolving the problems arising from increased oncology morbidity.
For analysis of the data for the patients with oncological diseases, registered in Burgas for 2018, the InterCriteria Analysis (ICA) approach is applied. Iy was introduced by K. Atanassov, D. Mavrov and V. Atanassova in 2014, [6]. The concept of ICA is based on the apparatus of (two-dimensional) Index matrices (IMs, see [4]) and intuitionistic fuzzy sets (IFSs, see [5]).
The approach is especially designed for decision support of multicriteria decision making problems, where some of the criteria have a higher cost than others -for example the harder, more expensive, the ones where more human resource is employed or take more time to measure or evaluate. Although these criteria are seen as more disadvantageous, the method's objective is to discover sufficiently high levels of dependence or correlation between these criteria and others, which are easier, less costly or faster to measure or evaluate, in order to discard the disadvantageous ones from the future decision making process.
After applying the ICA method, we obtain an index matrix that gives the correlations of each pair of criteria presented in the form of intuitionistic fuzzy pairs of values [8].
The dependences between the criteria are called "positive consonance", "negative consonance" or "dissonance". Here we use the scale used in previous studies that is shown in [7].
In the present paper the ICA method is applied to studying some statistical data for newly registered patients with oncological diseases for 2018 in Burgas. In the next observations we will apply the ICA approach to patients with oncological diseases for 2014-2018 and to data connected to metastatic melanoma [24,25] and colorectal cancer [39].

Application of the ICA approach
The InterCriteria Analysis approach is applied to real data for 941 newly registered patients (511 mеn and 430 womеn) with oncological diseases for 2018 in Burgas. The data contains information about age of patients, name of the disease, according to International Statistical Classification of Diseases and Related Health Problems, gender, marital status, data of the registration of the patient, etc. In the observed data there are:

Applying ICA approach for age and gender data
We will use an index matrix that contains 16 rows (for age groups and gender) and 4 columns (for marital status). After the applying the ICA method we obtain index matrix (see Table 1) with intuitionistic fuzzy pairs that represents an intuitionistic fuzzy evaluation of the relations between every pair of criteria "age group and gender".

Applying ICA approach for marital status and gender data
We will use an index matrix that contains 8 rows (for marital status and gender) and 8 columns (for age groups). After the applying the ICA method we obtain index matrix (see Table 2) with intuitionistic fuzzy pairs that represents an intuitionistic fuzzy evaluation of the relations between every pair of criteria "marital status and gender". In both cases, the stronger the correlation between a given pair is, the more intense the color is.

Results and discussion
After applying the ICA approach for age and gender data the following conclusions can be made: • 9 pairs of criteria (age and gender groups) are in strong positive consonance. In the age groups "up to 20 years", "21-30", "51-60", "61-70" and "over 80" there is a strong positive consonance between gender groups -male and female 〈 〈 〈 〈1.000; 0.000〉. The age group "51-60" has a strong positive consonance with the age group "61-70", taking gender into account. This shows a constant trend of the patients with oncology diseases for some of the identical or the adjacent age groups, also taking gender into account.
• 16 pairs of criteria (age and gender groups) are in weak positive consonance. In 7 pairs of criteria males in 31-40, 41-50 and 71-80 age groups have similar behavior with females in the same age groups. Males in the 41-50 age group have similar behavior with following age groups: "men 51-60", "men 61-70", "women 51-60" and "women 61-70". In next 9 pairs of criteria the intuitionistic fuzzy pairs of age groups show similar behavior with the intuitionistic fuzzy pairs of the next in order adjacent age group. • 59 pairs of criteria (age and gender groups) are in dissonance. They are mostly made of age groups with a big difference in the years, the lower age groups and the highest age group, regardless of gender. • The lower and highest age groups have more unstable behavior, while the medium age groups show a higher consistency. After applying the ICA approach for marital status and gender data the following conclusions can be made: • The groups of married, widowed and divorced males and females are in positive consonance. This shows a stable tendency regardless of the age group. • 1 pair of criteria "Men Мarried -Women Мarried" is in a strong positive consonance -〈0.964; 0.000〉. This means that the married men and married women have a very similar tendency for oncological diseases.
• 1 pair of criteria "Men Widowed -Women Widowed" is in a positive consonance -〈0.929; 0.071〉. This means that the incidence of cancers in male widows is similar to that of female widows.  • 1 pair of criteria "Men Divorced -Women Divorced" is in a weak positive consonance -〈0.786; 0.000〉. This means that the trends for cancers in male widows is similar to that of female widows. • The married, widowed and divorced male and female groups show a clearly expressed tendency for consistency, while the groups of unmarried male and unmarried female have unstable behavior. The groups of unmarried male and unmarried female are in dissonance with each other. • There are no correlations between different marital status groups, regardless of gender.

Conclusion
The InterCriteria Analysis method is applied to statistical data for newly registered and dispensary patients with oncological diseases for 2018 in Burgas. The analyzed data are from regional databases.
The results are commented from different points of view: relations between gender and age of the patients, and relations between gender and marital status.
Thus predictors of disease can be generated and search for dependencies between determinants of multi-criteria decision-making in oncological diseases associated with limitations such as time and resources.
In the next paper, we will use the method proposed in [9] for constructing of three two-dimensional Index Matrices from the three-dimensional one aiming at detecting patterns and relationships across the data per oncology patients' profiles, per marital status, and per year. ICA approach will be applied into the next directions: for age and gender group data, for marital status data and by years with collected statistical data for registered patients with oncological diseases in Burgas for 2014-2018.