Search

Klaster analiza ima dugu istoriju i mada se primenjuje u mnogim oblastima i dalje ostaju značajni izazovi. U disertaciji je prikazan uvod u neglatki optimizacioni pristup u klasterovanju, sa osvrtom na problem klasterovanja velikih skupova podataka. Međutim, ovi optimizacioni algoritmi bolje funkcionišu u radu sa neprekidnim podacima. Jedan od glavnih izazova u klaster analizi je rad sa velikim skupovima podataka sa kategorijalnim i kombinovanim (numerički i kategorijalni) tipovima promenljivih. Rad sa velikim brojem instanci (objekata) i velikim brojem dimenzija (promenljivih), može predstavljati problem u klaster analizi, zbog vremenske složenosti. Jedan od načina rešavanja ovog problema je redukovanje broja instanci, bez gubitka informacija. Prvi cilj disertacije je bio upoređivanje rezultata klasterovanja na celom skupu i prostim slučajnim uzorcima sa kategorijalnim i kombinovanim podacima, za različite veličine uzorka i različit broj klastera. Nije utvrđena značajna razlika (p>0.05) u rezultatima klasterovanja na uzorcima obima 0.03m,0.05m,0.1m,0.3m (gde je m obim posmatranog skupa) i celom skupu. Drugi cilj disertacije je bio konstrukcija efikasnog postupka klasterovanja velikih skupova podataka sa kategorijalnim i kombinovanim tipovima promenljivih. Predloženi postupak se sastoji iz sledećih koraka: 1. klasterovanje na prostim slučajnim uzorcima određene kardinalnosti; 2. određivanje najboljeg klasterskog rešenja na uzorku, primenom odgovarajućeg kriterijuma validnosti; 3. dobijeni centri klastera iz ovog uzorka služe za klasterovanje ostatka skupa. Treći cilj disertacije predstavlja primenu klaster analize u definisanju klastera bihejvioralnih faktora rizika u populaciji odraslog stanovništva Srbije, kao i analizu sociodemografskih karakteristika dobijenih klastera. Klaster analiza je primenjena na velikom reprezentativnom uzorku odraslog stanovništva Srbije, starosti 20 i više godina. Izdvojeno je pet jasno odvojenih klastera sa karakterističnim kombinacija, The cluster analysis has a long history and a large number of clustering techniques have been developed in many areas, however, significant challenges still remain. In this thesis we have provided a introduction to nonsmooth optimization approach to clustering with reference to clustering large datasets. Nevertheless, these optimization clustering algorithms work much better when a dataset contains only vectors with continuous features. One of the main challenges is clustering of large datasets with categorical and mixed (numerical and categorical) data. Clustering deals with a large number of instances (objects) and a large number of dimensions (variables) can be problematic because of time complexity. One of the ways to solve this problem is by reducing the number of instances, without the loss of information. The first aim of this thesis was to compare the results of cluster algorithms on the whole dataset and on simple random samples with categorical and mixed data, in terms of validity, for different number of clusters and for different sample sizes. There were no significant differences (p>0.05) between the obtained results on the samples of the size of 0.03m,0.05m,0.1m,0.3m (where m is the size of the dataset) and the whole dataset. The second aim of this thesis was to develop an efficient clustering procedure for large datasets with categorical and mixed (numeric and categorical) values. The proposed procedure consists of the following steps: 1. clustering on simple random samples of a given cardinality; 2. finding the best cluster solution on a sample (by appropriate validity measure); 3. using cluster centers from this sample for clustering of the remaining data. The third aim of this thesis was to examine clustering of four lifestyle risk factors and to examine the variation across different socio-demographic groups in a Serbian adult population. Cluster analysis was carried out on a large representative sample of Serbian adults aged 20 and over. We ident

Klaster analiza ima dugu istoriju i mada se primenjuje u mnogim oblastima i dalje ostaju značajni izazovi. U disertaciji je prikazan uvod u neglatki optimizacioni pristup u klasterovanju, sa osvrtom na problem klasterovanja velikih skupova podataka. Međutim, ovi optimizacioni algoritmi bolje funkcionišu u radu sa neprekidnim podacima. Jedan od glavnih izazova u klaster analizi je rad sa velikim skupovima podataka sa kategorijalnim i kombinovanim (numerički i kategorijalni) tipovima promenljivih. Rad sa velikim brojem instanci (objekata) i velikim brojem dimenzija (promenljivih), može predstavljati problem u klaster analizi, zbog vremenske složenosti. Jedan od načina rešavanja ovog problema je redukovanje broja instanci, bez gubitka informacija. Prvi cilj disertacije je bio upoređivanje rezultata klasterovanja na celom skupu i prostim slučajnim uzorcima sa kategorijalnim i kombinovanim podacima, za različite veličine uzorka i različit broj klastera. Nije utvrđena značajna razlika (p>0.05) u rezultatima klasterovanja na uzorcima obima 0.03m,0.05m,0.1m,0.3m (gde je m obim posmatranog skupa) i celom skupu. Drugi cilj disertacije je bio konstrukcija efikasnog postupka klasterovanja velikih skupova podataka sa kategorijalnim i kombinovanim tipovima promenljivih. Predloženi postupak se sastoji iz sledećih koraka: 1. klasterovanje na prostim slučajnim uzorcima određene kardinalnosti; 2. određivanje najboljeg klasterskog rešenja na uzorku, primenom odgovarajućeg kriterijuma validnosti; 3. dobijeni centri klastera iz ovog uzorka služe za klasterovanje ostatka skupa. Treći cilj disertacije predstavlja primenu klaster analize u definisanju klastera bihejvioralnih faktora rizika u populaciji odraslog stanovništva Srbije, kao i analizu sociodemografskih karakteristika dobijenih klastera. Klaster analiza je primenjena na velikom reprezentativnom uzorku odraslog stanovništva Srbije, starosti 20 i više godina. Izdvojeno je pet jasno odvojenih klastera sa karakterističnim kombinacija, The cluster analysis has a long history and a large number of clustering techniques have been developed in many areas, however, significant challenges still remain. In this thesis we have provided a introduction to nonsmooth optimization approach to clustering with reference to clustering large datasets. Nevertheless, these optimization clustering algorithms work much better when a dataset contains only vectors with continuous features. One of the main challenges is clustering of large datasets with categorical and mixed (numerical and categorical) data. Clustering deals with a large number of instances (objects) and a large number of dimensions (variables) can be problematic because of time complexity. One of the ways to solve this problem is by reducing the number of instances, without the loss of information. The first aim of this thesis was to compare the results of cluster algorithms on the whole dataset and on simple random samples with categorical and mixed data, in terms of validity, for different number of clusters and for different sample sizes. There were no significant differences (p>0.05) between the obtained results on the samples of the size of 0.03m,0.05m,0.1m,0.3m (where m is the size of the dataset) and the whole dataset. The second aim of this thesis was to develop an efficient clustering procedure for large datasets with categorical and mixed (numeric and categorical) values. The proposed procedure consists of the following steps: 1. clustering on simple random samples of a given cardinality; 2. finding the best cluster solution on a sample (by appropriate validity measure); 3. using cluster centers from this sample for clustering of the remaining data. The third aim of this thesis was to examine clustering of four lifestyle risk factors and to examine the variation across different socio-demographic groups in a Serbian adult population. Cluster analysis was carried out on a large representative sample of Serbian adults aged 20 and over. We ident

Depresija je jedan od najčešćih mentalnih poremećaja u opštoj populaciji širom sveta i prema podacima Svetske zdravstvene organizacije se nalazi na trećem mestu globalnog opterećenja bolestima. Cilj ovog istraživanja je bio da se utvrdi prevalencija depresije u populaciji odraslog stanovništva Vojvodine, da se identifikuju kategorije stanovništva pod najvećim rizikom i da se utvrdi povezanost depresije sa determinantama zdravlja, zdravstvenim stanjem i korišćenjem zdravstvene zaštite. Istraživanje je sprovedeno kao studija preseka na reprezentativnom uzorku stanovništva Vojvodine starosti 15 i više godina, kao deo nacionalne studije „Istraživanje zdravlja stanovništva Srbije“ koju je sprovelo Ministarstvo zdravlja Republike Srbije 2013. godine. Kao instrument istraživanja su korišćeni upitnici konstruisani u skladu sa upitnikom Evropskog istraživanja zdravlja. Prisustvo depresivnih simptoma je analizirano primenom PHQ-8 upitnika (Patient Health Questionnaire-8) a vrednost PHQ-8 skora 10 i više je ukazivala na prisustvo depresivne epizode (depresije). Istraživanjem je obuhvaćeno 3550 ispitanika prosečne starosti 50,2 godine. Prevalencija depresije je iznosila 4,8%. Primenom multivarijantne regresije je utvrđeno da žene imaju za 40% veću šansu za depresiju u odnosu na muškarce (OR=1,4), a kao prediktori depresije su se pokazali i starija životna dob, razvod/razlaz ili udovstvo, niži nivo obrazovanja, nezaposlenost, ekonomska neaktivnost i niži materijalni status. Osobe koje su u mesecu koji je prethodio istraživanju bile izložene stresu su imale oko sedam puta veću šansu za depresiju (OR=6,8), a kod ispitanika sa lošom socijalnom podrškom šansa za depresiju je bila oko četiri puta veća nego kod ispitanika sa jakom socijalnom podrškom (OR=3,7). Multimorbiditet je bio značajno prisutniji među ispitanicima sa depresijom, a šansa za prisustvo dve ili više hroničnih bolesti je kod njih bila skoro pet puta veća (OR=4,6) u odnosu na ispitanike koji nisu imali depresivne simp, Depression is one of the most common mental disorders in the general population worldwide and according to the World Health Organization it is the third leading cause of the global burden of disease. The aim of this study was to determine the prevalence of depression in the adult population of Vojvodina, to identify groups at the greatest risk of depression and to determine the relationship of depression with health determinants, health status and use of health care services. The research was carried out as a cross-sectional study on a representative sample of the population of Vojvodina, aged 15 and over as part of the National Health Survey of Serbia which was conducted by the Ministry of Health of the Republic of Serbia in 2013. Data was gathered by using questionnaires designed in accordance to the European Health Interview Survey. The presence of depressive symptoms was analyzed using the PHQ-8 questionnaire (Patient Health Questionnaire-8) and the score of 10 or above indicated a presence of a depressive episode (depression). The study included 3550 examinees with an average age of 50.2 years. The prevalence of depression was 4.8%. Multivariate regression showed that women had a 40% greater chance to suffer from depression as compared to men (OR=1.4). Also, factors such as older age, divorce/separation or widowhood, lower level of education, economic inactivity and poor financial situation were predictors of depression. People who were exposed to stress in the month prior to the survey had about seven times greater odds of depression (OR=6.8), and in patients with poor social support the chance of depression was about four times higher than among those with a strong social support (OR=3.7). Multimorbidity was significantly more prevalent among respondents with depression and they had five times higher odds of the presence of two or more chronic diseases (OR=4.6) in comparison with those who did not show depressive symptoms. Depression has been revealed as a si

Klaster analiza ima dugu istoriju i mada se primenjuje u mnogim oblastima i dalje ostaju značajni izazovi. U disertaciji je prikazan uvod u neglatki optimizacioni pristup u klasterovanju, sa osvrtom na problem klasterovanja velikih skupova podataka. Međutim, ovi optimizacioni algoritmi bolje funkcionišu u radu sa neprekidnim podacima. Jedan od glavnih izazova u klaster analizi je rad sa velikim skupovima podataka sa kategorijalnim i kombinovanim (numerički i kategorijalni) tipovima promenljivih. Rad sa velikim brojem instanci (objekata) i velikim brojem dimenzija (promenljivih), može predstavljati problem u klaster analizi, zbog vremenske složenosti. Jedan od načina rešavanja ovog problema je redukovanje broja instanci, bez gubitka informacija. Prvi cilj disertacije je bio upoređivanje rezultata klasterovanja na celom skupu i prostim slučajnim uzorcima sa kategorijalnim i kombinovanim podacima, za različite veličine uzorka i različit broj klastera. Nije utvrđena značajna razlika (p>0.05) u rezultatima klasterovanja na uzorcima obima 0.03m,0.05m,0.1m,0.3m (gde je m obim posmatranog skupa) i celom skupu. Drugi cilj disertacije je bio konstrukcija efikasnog postupka klasterovanja velikih skupova podataka sa kategorijalnim i kombinovanim tipovima promenljivih. Predloženi postupak se sastoji iz sledećih koraka: 1. klasterovanje na prostim slučajnim uzorcima određene kardinalnosti; 2. određivanje najboljeg klasterskog rešenja na uzorku, primenom odgovarajućeg kriterijuma validnosti; 3. dobijeni centri klastera iz ovog uzorka služe za klasterovanje ostatka skupa. Treći cilj disertacije predstavlja primenu klaster analize u definisanju klastera bihejvioralnih faktora rizika u populaciji odraslog stanovništva Srbije, kao i analizu sociodemografskih karakteristika dobijenih klastera. Klaster analiza je primenjena na velikom reprezentativnom uzorku odraslog stanovništva Srbije, starosti 20 i više godina. Izdvojeno je pet jasno odvojenih klastera sa karakterističnim kombinacija, The cluster analysis has a long history and a large number of clustering techniques have been developed in many areas, however, significant challenges still remain. In this thesis we have provided a introduction to nonsmooth optimization approach to clustering with reference to clustering large datasets. Nevertheless, these optimization clustering algorithms work much better when a dataset contains only vectors with continuous features. One of the main challenges is clustering of large datasets with categorical and mixed (numerical and categorical) data. Clustering deals with a large number of instances (objects) and a large number of dimensions (variables) can be problematic because of time complexity. One of the ways to solve this problem is by reducing the number of instances, without the loss of information. The first aim of this thesis was to compare the results of cluster algorithms on the whole dataset and on simple random samples with categorical and mixed data, in terms of validity, for different number of clusters and for different sample sizes. There were no significant differences (p>0.05) between the obtained results on the samples of the size of 0.03m,0.05m,0.1m,0.3m (where m is the size of the dataset) and the whole dataset. The second aim of this thesis was to develop an efficient clustering procedure for large datasets with categorical and mixed (numeric and categorical) values. The proposed procedure consists of the following steps: 1. clustering on simple random samples of a given cardinality; 2. finding the best cluster solution on a sample (by appropriate validity measure); 3. using cluster centers from this sample for clustering of the remaining data. The third aim of this thesis was to examine clustering of four lifestyle risk factors and to examine the variation across different socio-demographic groups in a Serbian adult population. Cluster analysis was carried out on a large representative sample of Serbian adults aged 20 and over. We ident

Background Underweight and severe and morbid obesity are associated with highly elevated risks of adverse health outcomes. We estimated trends in mean body-mass index (BMI), which characterises its population distribution, and in the prevalences of a complete set of BMI categories for adults in all countries. Methods We analysed, with use of a consistent protocol, population-based studies that had measured height and weight in adults aged 18 years and older. We applied a Bayesian hierarchical model to these data to estimate trends from 1975 to 2014 in mean BMI and in the prevalences of BMI categories ( LT 18.5 kg/m(2) [underweight], 18.5 kg/m(2) to LT 20 kg/m(2), 20 kg/m(2) to LT 25 kg/m(2), 25 kg/m(2) to LT 30 kg/m(2), 30 kg/m(2) to LT 35 kg/m(2), 35 kg/m(2) to LT 40 kg/m(2), = 40 kg/m(2) [morbid obesity]), by sex in 200 countries and territories, organised in 21 regions. We calculated the posterior probability of meeting the target of halting by 2025 the rise in obesity at its 2010 levels, if post-2000 trends continue. Findings We used 1698 population-based data sources, with more than 19.2 million adult participants (9.9 million men and 9.3 million women) in 186 of 200 countries for which estimates were made. Global age-standardised mean BMI increased from 21.7 kg/m(2) (95% credible interval 21.3-22.1) in 1975 to 24.2 kg/m(2) (24.0-24.4) in 2014 in men, and from 22.1 kg/m(2) (21.7-22.5) in 1975 to 24.4 kg/m(2) (24.2-24.6) in 2014 in women. Regional mean BMIs in 2014 for men ranged from 21.4 kg/m(2) in central Africa and south Asia to 29.2 kg/m(2) (28.6-29.8) in Polynesia and Micronesia; for women the range was from 21.8 kg/m(2) (21.4-22.3) in south Asia to 32.2 kg/m(2) (31.5-32.8) in Polynesia and Micronesia. Over these four decades, age-standardised global prevalence of underweight decreased from 13.8% (10.5-17.4) to 8.8% (7.4-10.3) in men and from 14.6% (11.6-17.9) to 9.7% (8.3-11.1) in women. South Asia had the highest prevalence of underweight in 2014, 23.4%

Background Underweight and severe and morbid obesity are associated with highly elevated risks of adverse health outcomes. We estimated trends in mean body-mass index (BMI), which characterises its population distribution, and in the prevalences of a complete set of BMI categories for adults in all countries. Methods We analysed, with use of a consistent protocol, population-based studies that had measured height and weight in adults aged 18 years and older. We applied a Bayesian hierarchical model to these data to estimate trends from 1975 to 2014 in mean BMI and in the prevalences of BMI categories ( LT 18.5 kg/m(2) [underweight], 18.5 kg/m(2) to LT 20 kg/m(2), 20 kg/m(2) to LT 25 kg/m(2), 25 kg/m(2) to LT 30 kg/m(2), 30 kg/m(2) to LT 35 kg/m(2), 35 kg/m(2) to LT 40 kg/m(2), = 40 kg/m(2) [morbid obesity]), by sex in 200 countries and territories, organised in 21 regions. We calculated the posterior probability of meeting the target of halting by 2025 the rise in obesity at its 2010 levels, if post-2000 trends continue. Findings We used 1698 population-based data sources, with more than 19.2 million adult participants (9.9 million men and 9.3 million women) in 186 of 200 countries for which estimates were made. Global age-standardised mean BMI increased from 21.7 kg/m(2) (95% credible interval 21.3-22.1) in 1975 to 24.2 kg/m(2) (24.0-24.4) in 2014 in men, and from 22.1 kg/m(2) (21.7-22.5) in 1975 to 24.4 kg/m(2) (24.2-24.6) in 2014 in women. Regional mean BMIs in 2014 for men ranged from 21.4 kg/m(2) in central Africa and south Asia to 29.2 kg/m(2) (28.6-29.8) in Polynesia and Micronesia; for women the range was from 21.8 kg/m(2) (21.4-22.3) in south Asia to 32.2 kg/m(2) (31.5-32.8) in Polynesia and Micronesia. Over these four decades, age-standardised global prevalence of underweight decreased from 13.8% (10.5-17.4) to 8.8% (7.4-10.3) in men and from 14.6% (11.6-17.9) to 9.7% (8.3-11.1) in women. South Asia had the highest prevalence of underweight in 2014, 23.4%

Svetska zdravstvena organizacija definiše kvalitet života kao percepciju pojedinca o sopstvenom položaju u životu u kontekstu kulture i sistema vrednosti u kojima živi kao i prema svojim ciljevima, očekivanjima, standardima i interesovanjima. To je širok koncept koga čine: fizičko zdravlje pojedinca, psihološki status, materijalna nezavisnost, socijalni odnosi i njihovi odnosi prema značajnim karakteristikama spoljašnje sredine. Procenjuje se da preko bilion ljudi živi sa nekim oblikom invaliditeta, što čini oko 15% svetske populacije. Prisustvo bilo kakvog telesnog invaliditeta kod osobe može značajno da utiče na njen psihofizički i socijalni razvoj. Procenjivanje uticaja fizičkog invaliditeta na svakodnevni život odslikava kakvo je funkcionisanje i blagostanje te osobe iz dana u dan i u različitim domenima života, što zapravo predstavlja procenu kvaliteta života. Istraživanje predstavlja studiju preseka na uzorku od 227 odraslih osoba sa motornim invaliditetom u Vojvodini, koji su članovi udruženja osoba sa invaliditetom. Kao instrument istraživanja korišćen je posebno kreiran upitnik za procenu kvaliteta života osoba sa invaliditetom. Osnovni cilj istraživanja je bio da se proceni kvalitet života osoba sa invaliditetom u odnosu na fizičko, socijalno i emocionalno funkcionisanje kod odraslih osoba sa motornim invaliditetom na teritoriji Vojvodine, kao i da se utvrdi postojanje razlike u kvalitetu života u odnosu na nivo obrazovanja, zaposlenost i bračno stanje. Podaci prikupljeni tokom ankete su kontrolisani na validnost, kodirani i unošeni u posebno kreiranu bazu podataka. Odabrana su pitanja i formirani domeni/skale (fizičkog, emocionalnog i socijalnog funkcionisanja i samoprocene zdravlja) na osnovu matrica korelacija, ICC i vrednosti Kronbah alfa. Aritmetička vrednost, mediana, standardna devijacija, minimalna i maksimalna vrednost i 95% inte

Istraživanje predstavlja studiju preseka na reprezentativnom uzorku od 672 deteta starosti od 7 do 19 godina u Vojvodini. Kao instrument istraživanja korišćeni su upitnici kojima su dobijeni podaci o demografskim i socijalno-ekonomskim karakteristikama, samoproceni zdravlja, zadovoljstvu životom, higijenskim navikama, navikama o ishrani, fizičkoj aktivnosti i slobodnom vremenu, mentalnom zdravlju, zastupljenosti rizika po zdravlje i korišćenju zdravstvene službe. Objektivni nalaz dobijen je merenjem telesne mase i visine, određivanjem indeksa telesne mase i merenjem krvnog pritiska. Osnovni ciljevi ovog rada bili su da se dobije ocena zdravstvenog stanja i utvrde vodeći zdravstveni rizici kao i razlike u odnosu na demografske i socijalne karakteristike, da se utvrde promene u zdravstvenom stanju školske dece i omladine u odnosu na retrospektivni deo istraživanja i da se testira validnost instrumenta. Rezultati su pokazali da 40,2% dece živi u domaćinstvima koja pripadaju grupi siromašnih, imaju jednog brata/sestru, 88% živi sa oba roditelja, a dve trećine ima svoju sobu. Redovne higijenske navike češće imaju devojčice, stariji uzast i deca iz bogatijih domaćinstava. Ishrana je neadekvatna kako sa aspekta redovnosti, tako i unosa pojedinih namirnica, nedovoljan je unos mleka i mlečnih proizvoda, voća i povrća i ribe, a prekomeran unos namirnica koje su izvor tzv. „praznih kalorija“, dok je 17% dece sa prekomernom telesnom masom i gojazno. Slobodno vreme karakterišu sedenterne aktivnosti i nedovoljna fizička aktivnost. Ponašanje u saobraćaju je rizično, većina ne koristi zaštitne mere. Povredu je imalo oko 15% dece, petina je bila izložena nekom obliku fizičkog ili psihičkog nasilja. Većina ocenjuje svoje zdravlje kao dobro, dve trećine svoju telesnu masu kao normalnu a tri četvrtine svoju fizičku aktivnost kao dobru. Higijenu i mogućnost rekreacije u školi petina ocenjuje kao lošu a svako deseto dete bezbesnost u školi kao lošu. Informacije o zdravlju najčešće dobija, The research represents a cross-sectional study of the representative sample of 672 schoolchildren 7-19 of age in Vojvodina. Instrument was specially constructed questionnaire providing data about demographic and socio-economic characteristics, self assessment of the health, life satisfaction, hygiene, nutrition, physical activity, leisure time, mental health, social support, presence of risk factors, and use of health care. Anthropometric measures included body mass, height, blood pressure measurement and calculating body mass index. The aim of the study was to assess health status of the children, identify health risks and determine demographic and socio-economic differences, as well as to determine differences in relation to the retrospective part of the study, and to determine validity of the instrument. Results showed that 40 percents of children live in households with low level of socio-economic status (SES), 88 percents live with both parents, have 1 sibling, and two thirds have their own room. Good hygiene practice is more likely present in girls, older age and children with better SES. Nutrition habits are inadequate regarding regularity and different food intake, there is low consumption of milk and dairy products, fruits, vegetables and fish, high intake of sweets, snacks and sweet beverages, and there is 17 percents of overweight and obese children. Leisure time is characterized with sedentary activities and lack of physical activity. Road traffic behavior is risky with inadequate safety measures. Injury had 15 percents of children, and the fifth was object of physical and psychological violence. Most of the children assessed their health as good and very good, two thirds their body mass as normal and three quarters their physical activity as good. School hygiene and recreational conditions were poorly rated by fifth of the children, and one in ten assessed school safety as bed. Health information are mostly provided by parents. Health is highly rated a

Uvod: Najveći doprinos nejednakostima u zdravlju, koje postoje kako unutar, tako i između zemalja širom sveta, pripisuje se socijalno-ekonomskim determinantama zdravlja. Što je socijalno-ekonomsko stanje ljudi gore, to je veća verovatnoća da sopstveno zdravlje procenjuju kao lošije uz češće prisustvo zdravstvenih simptoma i hroničnih stanja. Cilj ove disertacije bio je da se ispita povezanost demografskih i socijalno-ekonomskih determinanti sa obolevanjem stanovništva, da se konstruišu indeksi obolevanja uz procenu njihove pouzdanosti i validnosti, i da se izvrši analiza i poređenje socijalno-ekonomskih nejednakosti u zdravlju u 2000. i 2006. godini. Metod: U ovoj studiji korišćeni su podaci iz dva nacionalna ispitivanja zdravstvenog stanja stanovnika Republike Srbije, sprovedenih 2000. i 2006. godine po tipu studije preseka na reprezentativnom uzorku odraslih osoba uzrasta 20 i više godina (9921 u prvom i 14.522 u drugom ispitivanju). Kao instrument ispitivanja zdravlja stanovnika korišćeni su upitnici. Rezultati: Kao najučestaliji prediktori obolevanja stanovništva izdvojile su se sledeće varijable: pol, godine starosti, obrazovanje, radni status i materijalno stanje domaćinstva. Ispitanici sa nižim obrazovanjem (UO=2,69 u 2000. godini; UO=3,08 u 2006. godini), nezaposleni (UO=1,24 u 2000. godini; UO=1,53 u 2006. godini), neaktivno stanovništvo (UO=1,58 u 2000. godini; UO=1,96 u 2006. godini) i oni sa lošijim materijalnim stanjem domaćinstva (UO=4,34 u 2000. godini; UO=5,03 u 2006. godini) češće procenjuju svoje zdravlje kao loše. Vrednosti korišćenih indeksa obolevanja su veće u istraživanju iz 2000. godine, što ukazuje na bolje zdravstveno stanje stanovništva u 2006. godini..., Introduction: The greatest contribution to inequalities in health both within and between countries is attributable to the socioeconomic determinants of health. The lower the socioeconomic status, the higher the probability of reporting a poor self-perceived health and the higher the prevalence of health problems and cronic conditions. The aim of this doctoral dissertation was to examine the associations between demographic, socioeconomic determinants and morbidity of the population, to construct morbidity indices with assessment of their reliability and validity, and to analyze and compare socioeconomic inequalities in health between 2000 and 2006. Methods: Data from the 2000 and the 2006 National Health Surveys for Serbia were used, conducted as cross-sectional studies on a representative sample of adults aged 20 years and over (9921 in the first survey and 14,522 in the second survey). The information on population health was obtained from the questionnaires. Results: The most frequent predictors of morbidity of the population are: gender, age, education, employment status and financial situation of the household. Respondents with lower education (OR=2,69 in year 2000; OR=3,08 in year 2006), unemployed (OR=1,24 in year 2000; OR=1,53 in year 2006), inactive (OR=1,58 in year 2000; OR=1,96 in year 2006) and those with worse financial situation of the household (OR=4,34 in year 2000; OR=5,03 in year 2006) were more likely to report poor self-perceived health. The values of used morbidity indices were higher in 2000 National Health Survey, which implies better health status of the population in 2006...

Achieving and maintaining good health and prosperity requires a clear and harmonized environment. Studying the environment to protect human health aims to determine the type of pollutants and to monitor their impact on humans. Pollutants include the chemical, physical and microbiological pollutants. It is believed that the best way of studying the interaction between the population and the environment is the method that allows selection of representative, easily measurable and routinely controlled environmental indicators and health status. The World Health Organization recommends "DPSEEA" (Driving Force, Pressure, State, Exposure, Effect and Action) system. "DPSEEA" describes a closed circuit system of interdependent factors of the environment and human health which can be defined as a cause _ effect model. Application of "DPSEEA" system to assess the influence factors of the environment on human health is defined by studying the air quality of the environment, water and sanitation, noise and health, housing, traffic accidents, chemical accidents, and radiation. The use of complex indicators by "DPSEEA" system in our country so far has been used to assess the environmental situation, but not for systematic assessment of the impact factors of the environment on human health. AIM: Aims of the research are based on determining the type of microbiological, physical and chemical hazards in the environment of Novi Sad, exposure of the population and assess the impact of identified hazards to human health, as well as determining the significance of the application of environmental factors as indicators of impacts on human health. METHODOLOGY: Identifying hazards present in the environment of the City of Novi Sad (air quality, drinking water quality, noise) in 2006 were carried out based on the statutory methodology for sampling, microbiological, physical and chemical analysis and preparation of expert opinions on the hazards present in the environment. There were sampled

Kvalitet života u vezi sa zdravljem kod starih je najznačajniji aspekt kvaliteta života uopšte, a merenje kvaliteta života ima značajno mesto u praksi zdravstvene zaštite i naučnoistraživačkom radu. Cilj istraživanja je bio da se primenom modifikovanog McMaster Health Index Questionnaire utvrde fizička, socijalna i emotivna dimenzija kvaliteta života starih, kao i da se definiše validni instrument za procenu kvaliteta života u vezi sa zdravljem kod starih u našim uslovima. Istraživanje je sprovedeno u periodu 1998-2000 u obliku studije preseka, putem anketnog upitnika na uzorku stanovništva Novog Sada starosti 60 i više godina (n=373). Inicijalni anketni upitnik je poslužio za analizu pojedinih domena zdravlja a iz njega su korelacionim analizama izdvojeni ajtemi za šest skala: fizičko funkcionisanje, samozbrinjavanje, osećanje u vezi korisnosti u društvu, osećanja u vezi života, socijalno funkcionisanje i samoprocena. Za svaku skalu i za ukupni indeks kvaliteta života (QOL-OLD indeks) je izračunat transformisani skor (ranga 0-100) kao i uticaj pojedinih skala na ukupni indeks kvaliteta života u odnosu na zdravlje. Indeks kvaliteta života QOL-OLD (ranga 0-100) predstavlja sumarnu numeričku vrednost dobijenu iz svih šest domena. Za svaku skalu kao i za sumarni Indeks utvrđena je pouzdanost i validnost pomoću svetski prihvaćenih kriterijuma za Cronbach Alpha vrednost. Analizom je izdvojeno šest skala koje imaju uticaj na kvalitet života starih i utvrđeno je da je uticaj fizičkog zdravlja, posmatran kroz skalu fizičkog funkcionisanja i skalu samozbrinjavanja, dominantan i čini 39,3% u globalnom kreiranju indeksa kvaliteta života. Svaki domen kao i ukupni QOL-OLD indeks je analiziran u odnosu na neke sociodemografske determinante za koje se pokazalo da imaju uticaj na pojedine domene kao i na ukupni kvalitet života u odnosu na zdravlje kod starih. Izdvajanjem ajtema iz modifikovanog McMaster Health Index Questionnaire dobijen je skraćeni i validan upitnik za procenu kva, Health related quality of life of the elderly is the most important aspect of quality of life concept in general, and it's measurement has an important place in health care practice and scientific work. The aim of this study was to determine physical, social and emotional dimension of health related quality of life among elderly, using modified McMaster Health Index questionnaire, as well as to construct valid instrument for health related quality of life assessment among elderly for our theritory. Cross sectional study was performed during 1998-2000. through poll, on the Novi Sad population sample aged 60 and above (n=373). Initial questionnaire was used for the purpose of analysis of certain health domains, and by using correlation analysis to extract items for six scales: physical functioning, selfcare, social usefulness, emotions about life, social functioning and self perception. Transformed score (range 0-100) was calculated for each scale separatly and for total health related quality of life index (QOL-OLD) as well as percentage influence of each scale on QOL-OLD index. Health related quality of life index (QOL-OLD) presents summar numeric value calculated from all six domains. For each scale and for QOL-OLD validity and reliability was determined using Cronbach alpha value. Analysis extracted six scales which influence quality of life among elderly, and results showed that physical health, covered with physical functioning scale and selfcare scale, has dominant influence (39,3%) in creating global QOL-OLD index. Each domain and QOL-OLD index was analyzed considering some socio-demographic determinants which showed impact on single domains and on total QOL-OLD index. Extracting items from initial modified McMaster Health Index Questionnaire enabled construction of shortened and valid questionnaire for assessment health related quality of life among elderly in the population but also for individuals of same age.