- Email: [email protected]
P198 COMPARISON OF DIETARY HABITS ACCORDING TO FREQUENT DINNING OUT IN KOREAN WOMEN
108 variables as prescribed versus delivered volume, causes of interruptions in enteral infusions and calories intakes. Results: 11,638,986 calories were prescribed to all these patients during 2008, we found that 90% were delivered with success and only 1,182,733 calories were not delivered (10%). The causes of fasting founded were: surgical procedures 114,634, 6 calories (9.6%); gastrointestinal complications 82.763 calories (7%); Image diagnostic procedures 189,337, 6 calories (16%); hemodynamic instability 15,891 calories (1.3%); exchange feeding tube 71,975 calories (6%); respiratory tube procedures 30,938 calories (2.6%). No programmed pauses, as for a long bath, waiting time for a new diet pack and other common hospital routines were the reason for 63% of calories not received. Conclusion: We observe that the time for executed simple’s routines on the hospital can be the most important reasons for diet infusion pause. Educational programs and protocols are being organized to reduce fasting time and the lost of calories. Disclosure of Interest: None declared.
P197 SELECTION BIAS IN CROSS-SECTIONAL STUDIES: LENGTH BIAS IN THE NUTRITIONDAY STUDY E. Pernicka1 , P. Bauer1 , M. Mouhieddine2 , C. Schuh1 , K. Schindler3 , K. Steininger4 , R. Schlaffer4 , M. Hiesmayr2 . 1 Core Unit for Medical Statistics and Informatics, 2 Division of Cardiac-, Thoracic-, Vascular Anaesthesia and Intensive Care, 3 Division of Endocrinology and Metabolism, Medical University of Vienna, 4 Central Coordination Centre, nutritionDay in Europe, Austrian Society for Clinical Nutrition, Vienna, Austria Rationale: The one-day audit nutritionDay in Europe is a multinational cross-sectional study with a follow-up period of 30 days. This type of sampling causes length bias as patients with longer length of hospital stay (LOS) are more easily included in the study. This analysis investigated the extent of this bias in the nutritionDay sample. Methods: The nutritionDay was repeated three times and took place on 19.01.2006, 25.01.2007 and on 31.01.2008. Participation was open to any clinical unit that registered on the nutritionDay website and requested an anonymous centre and unit code. Every patient that was in hospital on the date of the nutritionDay had a chance to be included in the study. By nature of the cross-sectional study design, patients with longer LOS had higher probability to take part on nutritionDay. The length bias was compensated in the analysis by giving more weight to patients with shorter LOS (Una-Alvarez 2004) because the sampled patients may not represent the target population. Results: A total of 26,115 patients in 27 countries with available data on time spent in hospital before nutritionDay and available date on the total LOS were included in the study. In the unadjusted original sampled patients, the median (lower; upper quartile) of the LOS was 14 (8; 25) days. In the sample, adjusted for the length bias and censoring at 30 days after nutritionDay, the
Poster presentations median (lower; upper quartile) of the LOS was 8 (4; 13) days. Conclusion: The analysis showed that adjustment for length bias leads to very different estimates of LOS in such types of cross-sectional studies. This bias should always be taken into consideration when interpreting LOS data of cross-sectional study designs. References Una-Alvarez J. Nonparametric estimation under length-biased sampling and Type-I censoring: a moment based approach Ann Inst Statist Math 2004; 56: 667 81. Disclosure of Interest: All authors declare that they have no conflict of interest for this project.
P198 COMPARISON OF DIETARY HABITS ACCORDING TO FREQUENT DINNING OUT IN KOREAN WOMEN E. Lee1 , J. Park2 , B. Cho3 . 1 Nutrition, Seoul National University Hospital, 2 Family medicine, Healthcare System Gangnam Center of Seoul National University Hospital, 3 Family medicine, Seoul National University Hospital, Seoul, South Korea Rationale: The purpose of this study was to investigate dietary habits according to the frequency of dinning out Methods: This survey was conducted from January to March 2008, through questionnaires given to 780 Korean women aged ranging 35 49 years, who visited hospital for routine health check-up. The questionnaires were composed of general characteristics and dietary habits, and the results were analyzed using c2 -tests. The subjects were classified into 2 groups: frequent dinning out group (FD, more than 3 times a week, n = 314) and non frequent dinning out group (NFD, less than 3 times a week, n = 466). Results: The FD group were higher frequency of drinking, not-married and divorced, higher education and income (p < 0.05) except current smoking (NS). The NFD group ate breakfast more regularly, carbohydrate containing foods everyday, and chose low fat milk (p < 0.05). The FD group had more experience of overeating (p < 0.001). The FD group also ate more often the fatty meat like ribs and bacon (p < 0.001), internal organ (p < 0.05), chineseand fried-food (p < 0.001), and processed food (p < 0.05). Moreover, the frequency of bread and cake (p < 0.001) and soda intake (p < 0.05) as a snack was higher in the FD group than NFD group. There were no significant differences of salt, diversity of food group intake, and coffee drinking between 2 groups. Conclusion: The FD group had generally undesirable diet pattern, so nutritional education for healthy diet should be provided to encourage them to choose more desirable food and have healthier dietary habits. Disclosure of Interest: None declared.
P197 SELECTION BIAS IN CROSS-SECTIONAL STUDIES: LENGTH BIAS IN THE NUTRITIONDAY STUDY E. Pernicka1 , P. Bauer1 , M. Mouhieddine2 , C. Schuh1 , K. Schindler3 , K. Steininger4 , R. Schlaffer4 , M. Hiesmayr2 . 1 Core Unit for Medical Statistics and Informatics, 2 Division of Cardiac-, Thoracic-, Vascular Anaesthesia and Intensive Care, 3 Division of Endocrinology and Metabolism, Medical University of Vienna, 4 Central Coordination Centre, nutritionDay in Europe, Austrian Society for Clinical Nutrition, Vienna, Austria Rationale: The one-day audit nutritionDay in Europe is a multinational cross-sectional study with a follow-up period of 30 days. This type of sampling causes length bias as patients with longer length of hospital stay (LOS) are more easily included in the study. This analysis investigated the extent of this bias in the nutritionDay sample. Methods: The nutritionDay was repeated three times and took place on 19.01.2006, 25.01.2007 and on 31.01.2008. Participation was open to any clinical unit that registered on the nutritionDay website and requested an anonymous centre and unit code. Every patient that was in hospital on the date of the nutritionDay had a chance to be included in the study. By nature of the cross-sectional study design, patients with longer LOS had higher probability to take part on nutritionDay. The length bias was compensated in the analysis by giving more weight to patients with shorter LOS (Una-Alvarez 2004) because the sampled patients may not represent the target population. Results: A total of 26,115 patients in 27 countries with available data on time spent in hospital before nutritionDay and available date on the total LOS were included in the study. In the unadjusted original sampled patients, the median (lower; upper quartile) of the LOS was 14 (8; 25) days. In the sample, adjusted for the length bias and censoring at 30 days after nutritionDay, the
Poster presentations median (lower; upper quartile) of the LOS was 8 (4; 13) days. Conclusion: The analysis showed that adjustment for length bias leads to very different estimates of LOS in such types of cross-sectional studies. This bias should always be taken into consideration when interpreting LOS data of cross-sectional study designs. References Una-Alvarez J. Nonparametric estimation under length-biased sampling and Type-I censoring: a moment based approach Ann Inst Statist Math 2004; 56: 667 81. Disclosure of Interest: All authors declare that they have no conflict of interest for this project.
P198 COMPARISON OF DIETARY HABITS ACCORDING TO FREQUENT DINNING OUT IN KOREAN WOMEN E. Lee1 , J. Park2 , B. Cho3 . 1 Nutrition, Seoul National University Hospital, 2 Family medicine, Healthcare System Gangnam Center of Seoul National University Hospital, 3 Family medicine, Seoul National University Hospital, Seoul, South Korea Rationale: The purpose of this study was to investigate dietary habits according to the frequency of dinning out Methods: This survey was conducted from January to March 2008, through questionnaires given to 780 Korean women aged ranging 35 49 years, who visited hospital for routine health check-up. The questionnaires were composed of general characteristics and dietary habits, and the results were analyzed using c2 -tests. The subjects were classified into 2 groups: frequent dinning out group (FD, more than 3 times a week, n = 314) and non frequent dinning out group (NFD, less than 3 times a week, n = 466). Results: The FD group were higher frequency of drinking, not-married and divorced, higher education and income (p < 0.05) except current smoking (NS). The NFD group ate breakfast more regularly, carbohydrate containing foods everyday, and chose low fat milk (p < 0.05). The FD group had more experience of overeating (p < 0.001). The FD group also ate more often the fatty meat like ribs and bacon (p < 0.001), internal organ (p < 0.05), chineseand fried-food (p < 0.001), and processed food (p < 0.05). Moreover, the frequency of bread and cake (p < 0.001) and soda intake (p < 0.05) as a snack was higher in the FD group than NFD group. There were no significant differences of salt, diversity of food group intake, and coffee drinking between 2 groups. Conclusion: The FD group had generally undesirable diet pattern, so nutritional education for healthy diet should be provided to encourage them to choose more desirable food and have healthier dietary habits. Disclosure of Interest: None declared.