Please activate JavaScript!
Please install Adobe Flash Player, click here for download

ISCB2014_abstract_book

ISCB 2014 Vienna, Austria • Abstracts - Poster Presentations 103Tuesday, 26th August 2014 • 10:30-11:00 Monday25thAugustTuesday26thAugustThursday28thAugustAuthorIndexPostersWednesday27thAugustSunday24thAugust ≥5 for GHQ-12. The predictive power for depression by comorbidities was relatively low, with an AUC of 0.69 and 0.63 for cancer and stroke respec- tively. Conclusion: The study confirmed the reliability and validity of both screening instruments. Agreement of the two scales is between moderate to good, highlighting fairly high rates of false negatives and false positives if assessment relies on either. Caution is necessary when one scale is to be used. P2.1.92 Investigating trend in the rate of suicide using regression methods in Hungary between 1963 and 2011 AM Laszlo1 , J Csicsman2 , F Bari1 , TA Nyari1 1 University of Szeged, Szeged, Hungary, 2 Budapest University of Technology and Economics, Budapest, Hungary   Introduction: Rates of suicide have been analyzed from multiple points of views in Hungary in the last decades. However, descriptive rates have only been reported. An epidemiological study was carried out to characterize the pattern of annual Hungarian suicide rates during a long period. Methods: Annual Hungarian suicide rates (per 100,000 population) were defined using gender, age group and suicide method based on tables published in Demographic Yearbooks and population data by gender and age group from Hungarian Central Statistical Office between 1963 and 2011. Trends were determined using joinpoint regression analyses to segment the investigated period. Beyond, trends and relative risks for suicide rates were examined using negative binomial regression models overall and by gender, age group and suicide method. Results: Significant peak in 1982 and nadir in 2006 were found in an- nual Hungarian suicide rates, which shifted to 1981 in females and 1983 in males using joinpoint regression in 1963-2011. Suicide rates remained constant for males after 1983. Different segmented patterns were ob- served for suicide rates in age groups. Overall 178,323 suicides were committed in Hungary during the investi- gated period (50,265 female and 128,058 male). Suicide rates were signifi- cantly higher in males than in females based on relative risks calculated by negative binomial regression methods overall, in all age groups and most suicide methods. Conclusion: Our findings support the results of joinpoint regression mod- els for suicide frequencies first time in Hungary and give internationally comparable results. Acknowledgement: This research was supported by grants TÁMOP- 4.2.4.A/ 2-11/1-2012-0001, TÁMOP-4.2.2.A-11/1/KONV-2012-0052.   P2.1.119 Health-related mortality predictors among Krakow older citizens. 25-year follow-up study A Pac1 , B Tobiasz-Adamczyk1 , M Florek1 , M Brzyska1 1 Chair of Epidemiology and Preventive Medicine JU MC, Kraków, Poland   The aim of this study was to assess the predictive ability of self-reported health status on 25-year all-cause mortality in community-dwelling older people in Krakow. Poland. The study population consisted of 2432 respondents, aged 65 and over, resident in Krakow city centre who were recruited in years 1986-1987. Mortality data were collected on regular basis from Vital Municipal Records up to September 31st , 2011. During the 25-year observation period 2210 deaths were observed. To test the associations between health status and all-cause mortality six parametric survival models and flexible Royston-Parmar survival models were tested. For the final analysis the Royston-Parmar proportional hazard model (df=5) with time-dependent covariates was choosen. The health status was assessed using the self-reported questionnaire data. All results were adjusted for known mortality predictors (demographic, life-style, functional activity). Among women, we have observed that higher mortality was related to diabetes (HR=1,82; 95%CI: 1,44-2,30), coronary heart disease (CHD) (HR=1,32; 95%CI: 1,11-1,58), hypertension (HR=1,31; 1,10-1,55) and asthma (HR=1,25; 95%CI: 1,00-1,57). No time-dependent effect of those predictors was observed. For men, the most important health-related predictors were diabetes mellitus (HR=1,59; 95%CI: 1,14-2,22) and asthma (HR=1,70; 95%CI: 1,24-2,34) as well as CHD (HR=1,36; 95%CI: 1,07-1,75) for this last predictor time-dependent effect was observed. The impact of health-related mortality predictors among elderly Polish re- spondents was stable over 25-year follow-up. The only exeption was CHD among men - its impact was strongly time-dependent and significant only during the first few years of observation.   P2.1.129 An IRT longitudinal model for graded repeated responses: IADL and ADL hierarchy and functional dependency trajectories in the elderly A Edjolo1,2 , C Proust-Lima1,2 , F Delva1,2 , J-F Dartigues1,2 , K Pérès1,2 1 ISPED - University of Bordeaux, Bordeaux, France, 2 INSERM U897, Bordeaux, France   This study aimed at describing the hierarchy of a combined Lawton and Katz Instrumental (IADL) and basic (ADL) activities of daily living scale, and the trajectories of functional dependency before death in the elderly population by using a longitudinal Item Response Theory (IRT) model. A 2-parameter probit IRT model combined with a mixed model for re- peated graded responses was performed on a sub-sample of 3238 dead community dwellers aged 65 years and over at baseline in 1988 from the Paquid prospective cohort on brain and functional ageing. IADL and ADL were collected at home every 2 to 3 years over 22 years on a 3-point rating quotation. The model investigated the 11-scale items sequence and the functional nonlinear trajectories adjusted for education and gender. A hierarchy of the combined scale was confirmed with early losses in half of IADL (shopping, partial transporting, finances and telephoning), last losses in ADL (total toileting, continence, eating and transferring), and an overlapping of concomitant IADL and ADL in the middle of the dependen- cy continuum.The more discriminating items were basic activities in bath- ing, toileting, dressing and eating. The findings on functional trajectories showed a persistent postponement of functional dependency brought by education in men, but not in women. IRT model shows that the ability to perform daily tasks is hierarchically af- fected from shopping to transferring disabilities. An in-depth understand- ing of this sequence provides an early warning of functional decline or a signal to continue further functional assessments.   P2.1.147 Factors associated with under five child mortality in mothers´ employed in agriculture, India R Singh1 1 HMC, Doha, Qatar   Objective: To examine factors associated with under five child mortality in mothers’s employed in agriculture. Design, Setting and Population: Data was retrieved from 2005-06 NFHS- 3 in India (2008). The study population constituted a national representa- tive cross-sectional sample of single children aged 0 to 59 months and born to mothers aged 15 to 49 years employed in agriculture from all 29 states of India. Demographic and socio-economic variables were consid- ered as covariates in the Cox Proportional Hazard model.

Pages Overview