Authored by Kachiashvilia KJ*
Abstract
By statistical processing of Georgian Cancer Registry data of 2015-2016, clustering (grouping) of Georgian populations was realized, according to the intensity of the cancer disease prevalence, for the purpose of priority distribution of existed resources and means in the country and for the reduction of the number of patients and improvement of the quality of treatment. Cluster analysis methods of mathematical statistics were used for the study, which was directly implemented using universal statistical software package SPSS. The concept of disease index was introduced for achieving the intruded purpose. Its several variants were determined. The study results using indexes showed that it is possible to group objectively populated areas and regions of the country by intensity of dissemination of cancer disease.
Keywords: Cancer; Disease; Disease Index; Cluster Analysis; Populated Area; Region
Introduction
It can be said that cancerous diseases are a consequence of a developed civilization. It has become especially relevant to mankind since the second half of the 20th century, when human impacts on the environment have become increasingly significant. The spread of cancerous diseases is becoming increasingly important for humanity and causing significant social, economic and environmental losses. To reduce these losses and improve the quality of life of people, many countries are increasing efforts to ensure that all cancer patients receive the best possible treatment for their ultimate well-being. Many valuable scientific researches, preventive, curative, rehabilitation and other activities are planned and implemented for this purpose. To this end, cancer registries [1- 7] are being set up and developed in many countries around the world, which provide computerized databases that collect detailed information on cancer patients, including patients’ treatments. These databases are then used to search for information to answer many questions related to disease and treatment. The fight against cancer is unanimously given priority worldwide. Around 14 million people worldwide get sick each year and 8 million die of cancer. The loss caused by this is alarming at present and the forecast of growing by 57% in 20 years is made. Continuous, comprehensive, and unbiased information on cancer populations is essential to monitor the spread of the disease, establish public health priorities, and evaluate the effectiveness of cancer control programs in the community. The main purpose of the population-based cancer registry is to make available individual data of all patients with cancer. The first population-based cancer registry was created in 1929 in Germany. There are several hundred such registries for today, which actively cover 21% of the world’s population. An international standard for registered data has been introduced for the protection of quality, comparison to each other and for uniformity of information. Originally, cancer registries included only descriptions of illness severity, tendency, and geographic comparison. Subsequently, several registries expanded the data retrieval area to obtain patient survival rates, to determine the effectiveness of the health system. The coverage area of data registration has recently been expanded by adding clinical indicators: correct treatment, variety of care and duration of treatment. The Cancer Registry takes information from a variety of sources and, therefore, it is considered a trusted database. Registry data is used for a variety of purposes, including control and prevention of disease outbreaks, for optimal distribution and management of financial, medicinal, human and technical resources. The purpose of the present paper is to group the populations and regions of Georgia by the intensity of the spread of cancer through statistical data processing of the Cancer Registry of Georgia, in order to prioritize the distribution of these resources and means throughout the country to reduce the overall number of patients and to improve the quality of treatment. Also, to prioritize nationwide disease prevention measures, to carry out studies for establishment of disease reasons for their further reduction, to make the research of the causal links between disease and causal factors, and so on. The proposed research methodology is of general importance as it is universal and can be used to achieve the goals set for any country or region. Many countries are working to determine the intensity of the spread of cancer and many related facts. Experts from different fields and specialties work to solve this problem. Including physicians, biologists, chemists, physicists, sociologists, specialists in mathematical statistics and computer science, and more. The results of their work are presented in numerous published reports, scientific papers, reports on international meetings, conferences, workshops and so on. Below is a brief annotation of some of this work to illustrate the problem and the actual results. Theoretical and practical results obtained for increasing the accuracy of the use of the unity of the methods “decision, discovery and classification” of artificial intelligence is considered in [8]. Of the specific examples discussed, we are interested in the problem of segmentation and classification of skin cancers. Using this as an example, authors have shown that developed by them “Topological-geometrical voting” (uses comparisons of proximity and distance) greatly improves the conventional arithmetical voting (i.e. weighted averages) method in many cases. Cancer data in India [6] were compared with life expectancy for smokers, alcohol drinkers, and overweight [9]. The association of these factors with the incidence of the disease was established by statistical methods. The paper [10] examines the median age of death of female cancer patients in the Indian city of Trivandrum. Data are taken from the Trivandrum Cancer Registry. The Kulbach-Leibler distance was used for this purpose. Different methods of selecting variables for large dimensional data are compared on the basis of lung cancer data in work [11]. A spectralspatial classification method is proposed in work [12] for distinguishing cancer from normal tissue on a hyperspectral imaging. Tumor types are classified in paper [13], using artificial neural network based on brain imaging of astrocytoma type of different patients. A computerized decision-making system for the early detection of brain cancer is described in work [14]. In particular, the use of various statistical-based functions to calculate tissue structure is described. Based on these tissues, the segmentation of the brain tissue is classified into four categories based on the intensity of the histograms. The work [15] describes in detail the so-called “e-mail”. Using the Random Forest Algorithm for Cancer Prevention as an effective, reliable and optimal classifier among many possible algorithms. The work [15] describes the use in detail of the so-called “Random Forest Algorithm” for cancer prevention as an effective, reliable and optimal classifier among many possible algorithms. The use of the Monte Carlo method shows the existence of noted properties of the considered algorithm. Thus, the summaries of the reviewed papers provide the basis for concluding that mathematical statistics data grouping methods, by their classification, allow us to solve many practical problems, including the problem of determining the prevalence of cancer, for optimal planning and implementation of measures of the preventive and administrative-organizational nature of cancer illness. A set of methods, called cluster-analysis methods, is used in mathematical statistics for separation of homogenous groups from a given set of data by a certain sign or by a set of signs. The development of cluster analysis methods began in the seventies of the last century and is still developing with great intensity. In recent years, special attention has been devoted to the development of special classification methods for big data systems. Cluster analysis methods allow us to divide the investigated objects into groups of homogeneous objects. Such groups are called clusters. Cluster analysis methods are widely used to solve practical tasks in many fields, such as industry, economics, defense, medicine, biology, agriculture, ecology and others. Cluster analysis plays an important role in data mining, pattern recognition and machine learning [16- 21]. Many methods of cluster analysis and their use for solving different problems from different fields of human activity are discussed in variety of scientific works and their number is increasing day by day. As an example let’s introduce some of them. The method of identification of point’s clusters in multidimensional Euclidean space and its application in taxonomy is discussed in [22]. Two methods based on spatial dependencies between points are discussed: agglomerative (i.e. accumulative) and solvable (i.e. separable). The method is built on finding the nearest neighbor and then dividing it into clusters using the minimum inner sum criterion of a cluster. The procedure ensures effective reduction of the number of possible divisions. The method can be used for dichotomous dividing, but it is also well used for dividing into any number of clusters. The work [23] fundamentally addresses to the problem of cluster analysis and describes many divisive and heuristic methods. Programs developed for this purpose are also described. Monograph [24] gives a fundamental overview of the philosophy, essence, and existing methods of cluster-analysis methodologies. The application of these methods in various fields of science, including object classification, planning, engineering, and others. The appendices review books and articles on the problem under consideration, and many existing cluster-analysis software packages. The article [25,26] discusses the determination of the asymmetry of asthma using cluster analysis method. Many clinical, physiological and pathological parameters are associated with asthma. Therefore, multidimensional mathematical techniques – k means analysis, are used to identify distinct pheno-groups. In particular, k mean cluster analysis method for three different groups of asthma.
Basic Results of Investigation and their Consideration
As was mentioned above, the goal of this work is, by statistical processing of the cancer registry data, to group Georgian populated areas by intensity of cancer spread, for priority distribution of existed resources and means, with the purpose of the reduction of the total number of infected people and increasing the quality of the treatment.
Grouping of Georgian populations according to the incidence of cancer disease
The Cancer Registry data of Georgia was used to achieve this goal [7]. In particular, the study used the names of 961 settlements in Georgia with the reference to the number of population and the incidences of cancer in 2015-2016. It is clear that grouping settlements simply by the absolute number of infected people will not give the desired result to achieve the stated goal, since where there is a larger population there will always be a large number of infected people, and, in this case, small populated areas will be in unequal position in comparison with settlements with large population. To eliminate this obstacle, they use the disease intensity index to group the populations. Let us introduce the following denotations for computations of the disease intensity: ai- the number of the population in point, and bi- the number of patients. Then the number of infected, reduced to 100,000 inhabitants, or so called Incident Rate for i th settlement will be.
To read more about this article....Open access Journal of Advances in Cancer Research & Clinical Imaging
Please follow the URL to access more information about this article
To know more about our Journals....Iris Publishers
No comments:
Post a Comment