논문 상세보기

한국보건정보통계학회> 보건정보통계학회지> 변수 중요도를 이용한 설명 가능한 인공지능 기법의 시각화에 대한 고찰과 보건정보 자료에의 응용

KCI등재

변수 중요도를 이용한 설명 가능한 인공지능 기법의 시각화에 대한 고찰과 보건정보 자료에의 응용

Visualization of Explainable Artificial Intelligence Techniques Using Variable Importance with Its Applications to Health Information Data

정혜린 ( Hyerin Jeong ) , 박정훈 ( Junghoon Park ) , 이영섭 ( Yungseop Lee ) , 임창원 ( Changwon Lim )
  • : 한국보건정보통계학회
  • : 보건정보통계학회지 45권4호
  • : 연속간행물
  • : 2020년 11월
  • : 317-334(18pages)
보건정보통계학회지

DOI


목차

서 론
연구 방법
연구 방법
고찰 및 결론
REFERENCES

키워드 보기


초록 보기

Objectives: Deep learning techniques have been actively used in the medical field where precise diagnosis and results are very important. Deep neural network-based models utilizing big data from medical records are supporting medical opinions and are revolutionizing the medical industry. In addition, the convolutional neural network model shows excellent performance in analyzing image data and are used for image classification and X-ray/CT image reconstruction. Methods: In this paper, we conducted a visualization study using structured and unstructured data in the medical field. Results: In order to determine input variables affecting mortality and to evaluate their importance, a total of five techniques, namely, the augmented neural network model with multi-task learning, random forest, extra tree, gradient boosting and xgboost are applied to the intensive care unit data. Variable importance is calculated for each technique, and these indicators are all converted to ratios in consideration of the differences considering the patient group as a stratification variable. The converted values are shown in three graphs, a lollipop graph, a bubble chart graph, and a heat map graph. Through the visualization, it was easy to see which variables were relatively important for each technique and to what extent. InceptionResnetV2 was used as a classification model for skin cancer image data, and LIME and Grad-CAM were applied to the model to easily identify the characteristics of each cancer. Conclusions: Through this study, we apply several explainable artificial intelligence techniques to medical data to enhance understanding of the results of analysis and to help identify and visualize important input variables and features.

UCI(KEPA)

I410-ECN-0102-2021-000-001218604

간행물정보

  • : 의약학분야  > 예방의학및보건학
  • : KCI등재
  • :
  • : 계간
  • : 2465-8014
  • : 2465-8022
  • : 학술지
  • : 연속간행물
  • : 1976-2021
  • : 736


저작권 안내

한국학술정보㈜의 모든 학술 자료는 각 학회 및 기관과 저작권 계약을 통해 제공하고 있습니다.

이에 본 자료를 상업적 이용, 무단 배포 등 불법적으로 이용할 시에는 저작권법 및 관계법령에 따른 책임을 질 수 있습니다.

46권3호(2021년 08월) 수록논문
최근 권호 논문
| | | |

KCI등재

1한국인 당뇨병 환자에서 염증 표지자와 혈당 조절과의 관련성

저자 : 강민 ( Min Kang ) , 손석준 ( Seok-joon Sohn )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 46권 3호 발행 연도 : 2021 페이지 : pp. 247-256 (10 pages)

다운로드

(기관인증 필요)

초록보기

목적: 이 연구는 한국 당뇨환자에서 저등급염증과 혈당조절 간의 관계를 평가하는 것을 목적으로 하였다.
방법: 2007-2018년까지의 국민건강영양조사를 바탕으로 19-80세까지의 4,186명의 당뇨환자를 대상으로 진행하였다. 잘못된 혈당조절은 HbA1C가 7미만일 때로 정의하였다. 저등급염증은 염증마커인 WBC counts와 hsCRP로 측정되었으며 모든 염증마커는 사분위수로 구분되었다. 사분위수의 염증마커와 혈당조절 불량은 연령, 교육수준, 소득, 결혼상태, 흡연, 음주, 운동, BMI, CAD, 고혈압, 빈혈, 고콜레스테롤혈증을 포함한 혼란변수가 적용된 다변량로지스틱회귀분석을 통해 분석되었다.
결과: 잠재적 혼란변수를 적용한 후에도 WBC counts와 hsCRP는 혈당조절불량과 직선관계를 나타내었다. WBC counts의 경우 첫 번째 4분위수를 기준으로 하였을 때 혈당조절불량에 대해 두번째 사분위수는 1.50 (95% confidence intervals, 1.26-1.78), 세번째 사분위수는 1.71 (95% CI, 1.43-2.05), 네번째 사분위수는 2.0의 OR 값을 가졌다(95% CI 1.67-2.40). HsCRP의 경우 첫 번째 4분위수를 기준으로 하였을 때 혈당조절불량에 대해 두 번째 사분위수는 1.15 (95% CI, 0.87-1.53), 세 번째 사분위수는 1.59 (95% CI, 1.20-2.11), 네 번째 사분위수는 1.69의 OR 값을 가졌다(95% CI, 1.27-2.26). 층화분석결과 hsCRP와 혈당조절 간의 관련성에 연령과 성별에 따른 차이점이 있었다. 65세 미만인 그룹과 남성 그룹은 HsCRP와 혈당조절에서 통계적으로 유의한 직선관계를 나타내었지만 65세 이상인 그룹과 여성그룹은 직선 관계를 보였으나 통계적으로 유의하지 않았다.
결론: WBC counts와 hsCRP는 혈당조절불량의 위험과 통계적으로 유의한 직선관계를 나타내었으나 hsCRP의 경우 연령과 성별에 따른 차이점이 있었다. 이러한 결과는 한국을 대표하는 한국인 당뇨병환자 표본에서 낮은 등급의 염증이 혈당조절불량과 유의하게 관련이 있음을 시사한다.


Objectives: This study evaluated the association between inflammatory markers and glycemic control in Korean patients with diabetes mellitus. Methods: Using data from the 2007-2018 Korean National Health and Nutritional Examination, 4,186 diabetic subjects aged between 19 and 80 years were included in this study. Poor glycemic control was defined as level of HbA1c higher than 7%. Inflammation was measured by inflammatory markers including white blood cell (WBC) count and high-sensitivity C-reactive protein (hsCRP), and those inflammatory markers were categorized into 4 groups by quartiles. A multiple logistic regression analysis was performed to evaluate the association between quartile of inflammatory markers and poor glycemic control after adjusting for demographic factors, lifestyle, and comorbidities. Results: Levels of WBC count and hsCRP were positively associated with the risk of poor glycemic control even after fully adjusting for potential confounders. For WBC count, compared with the first quartile, the second quartile was 1.50 (95% confidence interval, CI: 1.26-1.78), 1.71 (95 % CI:1.43-2.05) for the third, 2.00 (95% CI: 1.67-2.40) for the fourth. In case of hsCRP, compared with the first quartile, the second quartile was 1.15 (95% CI: 0.87-1.53), 1.59 (95% CI: 1.20-2.11) for the third, and 1.59 (95% CI: 1.19-2.12) for the fourth. There was significant age difference in the association between hsCRP and glycemic control. The association was more evident in the younger group than in the elderly group. Conclusions: High level of WBC and hsCRP were positively associated with the risk of poor glycemic control. These results suggest that in a Korean diabetic patient, inflammation was significantly associated with poor glycemic control.

KCI등재

2텍스트마이닝을 통한 노인장기요양보험 키워드 네트워크 분석

저자 : 서종근 ( Jonggeun Seo ) , 남정모 ( Chung Mo Nam ) , 김태현 ( Tae Hyun Kim ) , 박소희 ( So-hee Park )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 46권 3호 발행 연도 : 2021 페이지 : pp. 257-266 (10 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: This study conducted research using big data in order to overcome the limitations of existing qualitative research or analysis research. By analyzing keywords, the flow and role of long-term care insurance in society were analyzed. Methods: Issues were searched through text mining, one of the big data techniques, and the flow of agendas by period was examined by 3 time points (institutional settlement period, 1st basic plan, 2nd basic plan). Using R and NetMiner, Daum News (news.daum.net) and Naver News (news.naver.com) were web-scraped to collect 20,965 news articles, 4,994 articles were filtered for keyword extraction and analysis. Result: Looking at the characteristics of each data type, in all data types, long-term care institutions (including nursing homes) and care providers appear as the top keywords, and the keyword subgroup characteristics are ① grade/service, ② institution management, and ③ the employee group includes the keyword subgroup. Conclusions: This study is based on the subject of long-term care insurance for the elderly and applies big data analysis techniques, and can be used as a decision-making tool in establishing policies and systems.

KCI등재

3한국 성인의 건강행위실천 군집현상 및 관련요인

저자 : 구혜민 ( Hyae Min Gu ) , 류소연 ( So Yeon Ryu ) , 박종 ( Jong Park ) , 최성우 ( Seong-woo Choi ) , 한미아 ( Mi Ah Han ) , 신준호 ( Jun Ho Shin )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 46권 3호 발행 연도 : 2021 페이지 : pp. 267-275 (9 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: This study was performed to identify and find the clustering of healthy behaviors and the factors related to the clustering among Korean adults aged 19-64. Methods: The data of 9,519 Korean adults aged between 19 and 64 years who participated in the 6th Korea National Health and Nutrition Examination Survey were used as the analysis target. Following the recommendation of the Health Plan 2020 for health-promoting lifestyle, four variables, including nonsmoking, moderate drinking, sufficient physical activity, and healthy diet, were used in the study. The multinomial logistic regression analysis was performed to analyze the relationships among demographic characteristics, health-related characteristics, and clustering of healthy behaviors. Results: The significant elements of healthy behaviors clustering included sufficient physical activities for males, and non-smoking, moderate drinking, and healthy diet for females. Clustering healthy behaviors were higher for “females, the elderly, people with higher level of education, people with higher household income, people with diabetes, and people who perceive their subjective health status is good. Conclusions: In this study, the factors related to the clustering of healthy behaviors were identified. In order to spread the practice of healthy behaviors, multiple interventions are more effective than individual health behaviors.

KCI등재

4표현적 글쓰기가 중환자실 간호사의 전문직 삶의 질과 극복력에 미치는 효과

저자 : 유단비 ( Danbi You ) , 박혜자 ( Hye-ja Park )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 46권 3호 발행 연도 : 2021 페이지 : pp. 276-284 (9 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: This study examined the effect of expressive writing on professional quality of life and resilience among intensive care unit nurses. Methods: Forty nurses stratified by the nursing proficiency career randomly assigned to an writing group (n=17) or a control group (n=23) from a university hospital-affiliated intensive care units. The 5-week expressive writing included a weekly 20 min of expressive writing and 10 min of reflection on the work-related traumatic events. Outcomes were assessed using questionnaires of a professional quality of life (compassion satisfaction, burnout and secondary traumatic stress) and a resilience. Data were analyzed by independent t-test and analysis of covariance. Results: The expressive writing program provided improvements in resilience measure, including measures of situational, philosophical, and professional resilience. Conclusions: Expressive writing might be a useful strategy for enhancing resilience from the stressful work-related traumatic events for nurses working at the intensive care units.

KCI등재

5다문화 청소년의 진로장벽이 우울에 미치는 영향: 부모의 양육태도, 다문화 청소년과 교사와의 관계의 매개효과를 중심으로

저자 : 오혜경 ( Hye-kyung Oh ) , 고미숙 ( Mi-suk Ko )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 46권 3호 발행 연도 : 2021 페이지 : pp. 285-292 (8 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: The purpose of this study was to investigate the effect of multicultural adolescents' career barriers on depression focused on mediating effect of parent's rearing attitudes and relationship between multicultural adolescents and teachers. Methods: The empirical analysis data of this study was extracted from the 8th Multicultural Adolescents Panel Survey. The subjects of this study were 1194 multicultural adolescents. The data were analyzed with Pearson's correlation coefficient and regression analysis. Results: The significant factor related to supervision (B=-0.26, p<0.001) and negligence (B=0.31, p<0.001) of parent's attitudes and relationship with teachers (B=-0.43, p<0.001) included perceived career barriers. Negligence of parent's rearing attitude (B=0.21, p<0.001) increased depression while supervision of parent's rearing attitude (B=-0.14, p<0.001) and relationship with teachers (B=-0.06, p<0.01) decreased depression. Parent's rearing attitudes and relationship with teachers were mediators between career barriers and depression. Conclusions: Based on the results of this study, it is desirable to reduce the level of depression by lowering the career barriers, parent's rearing attitude (negligence), and by increasing the parent's rearing attitude (supervision) and relationship with teachers (adaptation). It is suggested to develop effective educational programs to decrease the depression level of multicultural adolescents.

KCI등재

6간호대학생의 소진과 행복감의 관계에서 경제적 스트레스와 사회적 자본의 매개효과

저자 : 김수진 ( Su Jin Kim ) , 정경순 ( Kyung Soon Jeong )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 46권 3호 발행 연도 : 2021 페이지 : pp. 293-301 (9 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: The purpose of this study was to examine the mediating effects of economic distress and social capital on the relationship between burnout and happiness of nursing students. Methods: This study used Maslach Burnout Inventory-Student Survey, The Oxford Happiness Questionnaire, Incharge Financial Distress/Financial Well-Being Scale, and Internet Social Capital Scales. The data was analyzed by using frequency, percentage, mean, standard deviation, t-test, one-way ANOVA, and Person's correlation coefficient. Baron and Kenny's method was used to analyze mediating effect between burnout and happiness. Results: Burnout, happiness, social capital, and economic distress had correlation with each other. There were the mediating effects of economic distress and social capital on the relationship between burnout and happiness of respondents. Conclusions: The result showed the mediating effects of economic distress and social capital on the relationship between burnout and happiness of nursing students. To increase happiness, it is necessary to apply a curriculum to strengthen the ability of nursing students.

KCI등재

7당뇨병 환자의 약물치료 관련 요인

저자 : 최선주 ( Sun Ju Choi ) , 류소연 ( So Yeon Ryu ) , 박종 ( Jong Park ) , 최성우 ( Seong Woo Choi )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 46권 3호 발행 연도 : 2021 페이지 : pp. 302-308 (7 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: The purpose of this study was to identify related factors related to the treatment of diabetic patients. Methods: The subjects of this study were 23,772 diabetic patients aged 30 years and older who participated in the 2016 Community Health Survey. The independent variables were sociodemographic factors, lifestyle and health-related characteristics and diabetes-related characteristics and the dependent variables were related to diabetes treatment. The statistical analysis methods used to find factors related to treatment in diabetic patients were the chi-square test and multiple logistic regression analysis. Results: The drug treatment rate of diabetic patients was 88.7%. Among sociodemographic characteristics, drug treatment for diabetes was significantly related to women, increasing age, higher educational level and living in rural area. Among the characteristics related to lifestyle and health status, walking practice and subjective health status were significantly related and among diabetes related characteristics, there were significant correlations in whether it was blood glucose level, self-cognitive blood glucose control, and diabetes management education. Conclusions: Conclusively, gender, age, education level, residential area, walking practice, self-rated health and knowledge of DM were associated with treatment of DM. An effective diabetes management program that reflects the relevant factors should be applied to the treatment of DM.

KCI등재

8생활 요인이 난임 여성의 2년 이내 출산 횟수에 미치는 영향

저자 : 김민영 ( Min Young Kim ) , 한규은 ( Kyueun Han )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 46권 3호 발행 연도 : 2021 페이지 : pp. 309-314 (6 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: Infertility has been increasing in Korea, but few studies have examined the effect of lifestyle factors on childbirth among infertile Korean women. This study was designed to identify whether the lifestyle factors can predict the childbirth of infertile women in 2 years. Methods: We selected women aged under 40 diagnosed with infertility in 2016 and retrieved their national health examination data from Korean National Health Insurance Service. Regression analysis was conducted to investigate whether infertile women's age, drinking, smoking, physical exercise, obesity, and sequela recorded in 2016 explained their childbirth in 2017 and 2018. Results: The results showed that age and obesity explained childbirth, respectively. In addition, the interaction of age and sequela predicted childbirth in 2 years. Conclusions: The findings suggest that the number of childbirth of infertile women can be inferred from lifestyle factors, which implies the importance of nonmedical factors in infertility.

KCI등재

9Patient Drug Database: 환자 생성 건강 데이터를 활용한 환자 주도적 약물 부작용 탐색을 위한 데이터베이스 구축

저자 : 이상민 ( Sang Min Lee ) , 이수현 ( Suehyun Lee ) , 김종엽 ( Jong Yeup Kim )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 46권 3호 발행 연도 : 2021 페이지 : pp. 315-325 (11 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: This study focuses on building a database for patient-led search on drug side effects using basic drug information, drug analysis results information, patient information, and patient-generated health data (PGHD). Methods: After collecting data from the Health Insurance Review and Assessment Institute, the Korean Pharmaceutical Information Center, the Ministry of Food and Drug Safety, and the Korean Pharmaceutical Association, basic drug information was created. By utilizing the Korea Average Event Reporting System (KAERS) side effect report data provided by the Korea Drug Safety Administration and MetaLAB, a drug side effect detection algorithm applied on the Konyang university hospital's real data, we designed and built a database using Oracle DB, which contains a table of patient information and PGHD. For drug information, a total of 49,553 drugs were mapped, and drug analysis results used KAERS and MetaLAB. Results: Based on the collected drug information, a total of 15 tables containing basic drug information (7 tables), drug analysis results (2 tables), patient information (1 table), and patient generation information (5 tables) were created using EDI codes, following mapping and normalization. Basic drug information included 49,553 EDI and 2,099 ATC codes. Drug analysis results included 2,046 KAERS ATC codes, 1,701 WHOART-ARRN (PT) that the result of 33 WHOART-SEQ (IT), 15,861 MetaLABEDI codes, and 101ATC codes. TheADR results were constructed using 62 DRUG_IDs and 73 MedDRA_PTI_IDs. Conclusions: The Patient Drug Database (PD2B) in this study was employed to allow patients to voluntarily report on their perception and drug side effects through application tools, which can provide quick measures against drug side effects and assist in the discovery of new ones.

KCI등재

10체계적 문헌고찰과 메타분석을 통한 소방공무원의 외상 후 스트레스 장애 중재 프로그램 효과 검증 연구

저자 : 강민주 ( Min-ju Kang ) , 김영화 ( Young-hwa Kim ) , 이상구 ( Sang-goo Lee ) , 한승우 ( Seung-woo Han )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 46권 3호 발행 연도 : 2021 페이지 : pp. 326-336 (11 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: The purpose of this study is to provide basic data for the development of effective post-traumatic stress disorder (PTSD) intervention programs for firefighters. Methods: Four domestic and four abroad databases were used, and 7 domestic and 4 abroad studies that met the criteria of this study were selected for systematic review and meta-analysis. Results: The overall effect size of the PTSD intervention program was 0.40 (95% confidence interval, CI: 0.17-0.64, p<0.05) in domestic and 0.37 (95% CI: 0.02-0.71, p<0.05) in abroad. Conclusions: Various domestic and abroad intervention methods for PTSD of firefighters have low effects. Therefore, based on the results of this study, it is necessary to develop and apply a post-traumatic stress disorder intervention program for firefighters.

12
권호별 보기
같은 권호 다른 논문
| | | | 다운로드

KCI등재

1단태 후기 미숙아의 모유수유 영향 요인

저자 : 김수진 ( Su Jin Kim ) , 장군자 ( Gun Ja Jang )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 45권 4호 발행 연도 : 2020 페이지 : pp. 311-316 (6 pages)

다운로드

(기관인증 필요)

초록보기

목적: 본 연구는 단태 후기 미숙아의 병원 퇴원 후 1개월 시점에서 모유수유 영향요인을 규명하기 위함이다.
방법: 후기 미숙아의 모유수유 실천율과 관련된 2개 선행 연구에서 쌍태아를 제외하고 단태아 데이터를 활용한 이차분석연구이다. 후 기 미숙아 244명의 퇴원 1개월 시점 수유형태의 데이터를 활용하였고, 모유수유군과 조제유군 간 일반적 특성 및 수유관련 특성은 chi-square test와 t-test, 모유수유 영향 요인은 로지스틱 회귀 분석으로 분석하였다.
결과: 후기 미숙아의 퇴원 1개월 모유수유에 영향을 미치는 요인은 병원 입원 기간 동안의 수유형태였다. 즉, 병원에서 완전 모유수유를 하거나 혼합유를 통해 모유를 제공받은 경우 퇴원 후에도 모유수유를 지속하는 것으로 나타났다.
결론: 후기 미숙아의 입원 동안 임상에서 모유수유를 적극적으로 지지하고 권장하는 것이 단태 미숙아의 모유수유를 증진하는 중요 한 요인임을 밝히는 바이다.

KCI등재

2변수 중요도를 이용한 설명 가능한 인공지능 기법의 시각화에 대한 고찰과 보건정보 자료에의 응용

저자 : 정혜린 ( Hyerin Jeong ) , 박정훈 ( Junghoon Park ) , 이영섭 ( Yungseop Lee ) , 임창원 ( Changwon Lim )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 45권 4호 발행 연도 : 2020 페이지 : pp. 317-334 (18 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: Deep learning techniques have been actively used in the medical field where precise diagnosis and results are very important. Deep neural network-based models utilizing big data from medical records are supporting medical opinions and are revolutionizing the medical industry. In addition, the convolutional neural network model shows excellent performance in analyzing image data and are used for image classification and X-ray/CT image reconstruction. Methods: In this paper, we conducted a visualization study using structured and unstructured data in the medical field. Results: In order to determine input variables affecting mortality and to evaluate their importance, a total of five techniques, namely, the augmented neural network model with multi-task learning, random forest, extra tree, gradient boosting and xgboost are applied to the intensive care unit data. Variable importance is calculated for each technique, and these indicators are all converted to ratios in consideration of the differences considering the patient group as a stratification variable. The converted values are shown in three graphs, a lollipop graph, a bubble chart graph, and a heat map graph. Through the visualization, it was easy to see which variables were relatively important for each technique and to what extent. InceptionResnetV2 was used as a classification model for skin cancer image data, and LIME and Grad-CAM were applied to the model to easily identify the characteristics of each cancer. Conclusions: Through this study, we apply several explainable artificial intelligence techniques to medical data to enhance understanding of the results of analysis and to help identify and visualize important input variables and features.

KCI등재

3중등교사의 미세먼지 관련 건강행위 영향요인: 지식과 태도를 중심으로

저자 : 박용경 ( Yong-kyung Park ) , 김은휘 ( Eun-hwi Kim )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 45권 4호 발행 연도 : 2020 페이지 : pp. 335-340 (6 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: The purpose of this study was to investigate the factors affecting health behavior related to particulate matter (PM) in middle school teacher. Methods: The subjects of analysis were 192 middle school teachers. Data were analyzed with SPSS 22.0 program. Results: The greatest influential variable of heath behavior of PM were behavior attitude (β=0.38, p<0.001) and knowledge (β=0.13, p=0.034). These factors explained for 28.3% of health behavior related to PM. Conclusions: It will be necessary to develop a program to improve knowledge, attitude and health behavior for PM in middle school teacher.

KCI등재

4요양병원 전원 노인 환자의 종합병원 중환자실 재원기간 영향요인

저자 : 정이슬 ( Lee-seul Jeong ) , 이수정 ( Su Jung Lee ) , 신동수 ( Dong-soo Shin )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 45권 4호 발행 연도 : 2020 페이지 : pp. 341-348 (8 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: This study aimed to identify factors affecting the length of intensive care unit (ICU) and characteristics among older patients transferred from long-term care hospital. Methods: This is a retrospective study and we extracted data from electronic medical records. Study participants were patients aged 65 or older who were admitted to ICU after they had been transferred to emergency department from long-term care hospitals. Length of stay (LOS) in ICU was defined as the patients had stayed in ICU for 7 days or more. Factors influencing the LOS were identified through the binary logistic regression analyses based on demographic and clinical characteristics and laboratory parameters. Results: Of all screened patients, 243 subjects were transferred from long term care hospitals. Participants mean age was 80.3 (±7.0) years and female comprised 63%. Participants had stayed in ICU for 13.4 days in average. Factors affecting the LOS in ICU were presence of multidrug resistant organism after ICU admission (odds ratio, OR 5.66, 95% confidence interval, 95% CI=1.79-17.82, p<0.010), tracheostomy insertion (OR 3.35, 95% CI=1.28-8.74, p<0.014), mechanical ventilator application (OR 2.49, 95% CI=1.11-5.58, p<0.026), central venous catheterization (OR 2.13, 95% CI=1.09-4.17, p<0.027) and simplified acute physiology score III (OR 0.95, 95% CI=0.91-1.00, p<0.039) by multivariate binary logistic regression analysis. Conclusions: Older patients in long-term care hospitals stayed longer in ICU and were vulnerable to infection. It is important to develop strategies to reduce LOS in ICU as the number of older patients admitted from long-term care institutions increases.

KCI등재

5금연 시도자의 금연지원서비스 이용 관련 요인

저자 : 김현우 ( Hyun-woo Kim )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 45권 4호 발행 연도 : 2020 페이지 : pp. 349-355 (7 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: This study was executed to understand the factors related to utilization of smoking cessation support service among attempters for quitting smoking. Methods: The data of Korea National Health and Nutrition Examination Survey in 2016-2017 was used for the analysis in this study. Logistic regression analysis was executed to understand the factors related to utilization of smoking cessation support service among attempters for quitting smoking. Results: the factors related to utilization of smoking cessation support service are daily average smoking amount, current electronic cigarette usage status and subjective health status and the anti-smoking attempters use the smoking cessation support service more in case of having more daily average smoking amount, using the electronic cigarette currently and having worse subjective health level. Conclusions: To increase the smoking cessation success rate of the attempters for quitting smoking, the strategy to make the anti-smoking attempters having lower usage rate of smoking cessation support service participate the smoking cessation support service more should be established.

KCI등재

6만성신장질환의 동반질환이 말기신부전으로의 진행에 미치는 영향

저자 : 서화정 ( Hwa Jeong Seo )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 45권 4호 발행 연도 : 2020 페이지 : pp. 356-364 (9 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: Due to the aging of population and increase of chronic diseases worldwide, the prevalence of chronic kidney disease (CKD) is on the rise. The increase of CKD is connected to the development of end-stage renal diseases (ESRD) and the occurrence of death caused by ESRD. This study explored the risk factors affecting the progression of CKD patients to ESRD using the sample cohort database built by the Korean National Health Insurance Service. Methods: This study targeted 2,354 new patients to whom the CKD occurred between 2003 and 2012. The development of ESRD of the CKD patients was checked through the number of cases per 1,000 person years. Significance was confirmed through the t-test and Wilcoxon signed-rank test. For the factors affecting the occurrence of ESRD by risk factor of comorbidities, the Cox proportional hazard regression analysis was carried out. Results: The average routing period of the subjects of this study was 2.98±2.69 years. 105.69 cases of ESRD among CKD patients took place per 1,000 person years during the observation period. The proportional hazard of CKD patients' ESRD became higher as the subjects' age became younger (50-59, HR=0.76; 60-69, 0.64; 70-79, HR=0.63), and the proportional hazard was higher from the subjects having proteinuria (Yes, HR=2.16), high hypertension (Yes, HR=1.38), diabetes mellitus (Yes, HR=1.23), and anemia (Yes, HR=1.77) (All p<0.05). Conclusions: Risk factors for the progression of CKD to ESRD were low age group (less than 60), underweight (less than 23), diabetes, hypertension, and anemia patients. To control the occurrence of ESRD that increases social and economic burdens, along with death risk, it is necessary to establish individualized treatment plans depending on the characteristics and risks of patients based on each type of comorbidity.

KCI등재

7욕창예방간호의 임상적 의사결정지지 시스템의 컨텐츠 개발과 활용

저자 : 김명수 ( Myoung Soo Kim ) , 류정미 ( Jung Mi Ryu )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 45권 4호 발행 연도 : 2020 페이지 : pp. 365-372 (8 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: The purpose of this study was to develop the contents of a clinical decision support system and assess the actual degree of utilization in order to promote the performance of pressure ulcer prevention care. Methods: Data were collected from March to April 2020. The development of the clinical decision support system was carried out through the steps of according to the ADDIE (analysis -design-development-implementation and evaluation) model. The system was built using Android, iOS, and Safari 604.1 browsers. This study was assess the utilization of clinical decision support system developed over four weeks to 23 nurses. Collected data were analyzed using SPSS WIN 25.0 program. Results: Algorithm for calculation of risk of oral mucous pressure ulcer was developed in clinical decision support system, and information on determination of skin pressure ulcer and oral mucosa pressure ulcer stage and suitable nursing interventions were constructed. At the skin pressure ulcer stage, the deep tissue injury suspected stage and unclassified stage were used a lot. Information corresponding to the second stage of oral mucosa pressure ulcer and the risk level of 31-60% was used a lot. Conclusions: These results suggest that it has been developed as a clinical decision support system content applicable to practice for nurses at the pressure ulcer prevention care.

KCI등재

8소방공무원의 외상 후 성장 관련요인에 대한 체계적 문헌고찰 및 메타분석

저자 : 김명관 ( Myung-gwan Kim ) , 한승우 ( Seung-woo Han )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 45권 4호 발행 연도 : 2020 페이지 : pp. 373-384 (12 pages)

다운로드

(기관인증 필요)

초록보기

목적: 본 연구는 소방공무원의 외상 후 성장에 관련된 요인들을 분석하기 위하여 선행논문들을 중심으로 한 체계적 문헌고찰 및 메타분석 연구이다.
방법: 최종 선정한 14편의 논문은 오픈 소스 통계 소프트웨어인 R (3.6.0)프로그램을 사용하여 상관계수를 활용한 효과크기 메타분석을 실시하였다. 4개의 국외 검색엔진과 4개의 국내 검색엔진으로 시행하였다. 국외 검색엔진으로는 EMBASE, CINAHL, PubMed, Cochrane Library를 통해 검색하였다. 반면 한국 검색엔진으로는 KISS, RISS, NDSL, KMbase를 사용하여 문헌검색을 하였다.
결과: 연구결과 내적요인의 자아탄력성과 외상 후 성장 간의 효과크기는 r= 0.37, 회복탄력성의 효과크기는 r= 0.34, 자존감의 효과크기는 r= 0.32로 나타났다. 외적요인에서 소명의식과 외상 후 성장 간의 효과크기는 r= 0.50, 직업소명의 효과크기는 r= 0.25로 나타났다. 대처기전의 사회적 지지와 외상 후 성장 간의 효과크기는 r= 0.36, 자기노출의 효과크기는 r= 0.20이었다. 대처전략의 문제중심대처와 외상후 성장 간의 효과크기는 r= 0.49, 의도적 반추의 효과크기는 r= 0.54로 나타났다. 마지막으로 외상 후 스트레스 장애와 외상 후 성장 간의 효과크기는 r= 0.23으로 나타났으며 의도적 반추는 외상 후 성장에 가장 크게 효과크기가 나타났다.
결론: 한국 소방공무원의 외상 후 성장에 영향을 미치는 관련요인으로 내적요인, 외적요인, 대처기전, 대처전략, 외상 경험에서 영향을 미치는 변인으로 확인되었다. 그 중 의도적 반추는 가장 큰 효과크기를 나타내었다. 따라서 의도적 반추를 촉진할 수 있는 전략을 개발하고 본 연구를 바탕으로 정신간호 중재 프로그램에 기초자료를 제공하고자 한다.

KCI등재

9댄스스포츠 프로그램이 농촌지역 여성노인의 신체기능과 인지, 우울 및 삶의 질에 미치는 영향

저자 : 안윤희 ( Yun-hee An ) , 홍남수 ( Nam-soo Hong ) , 윤희정 ( Hee-jung Yoon )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 45권 4호 발행 연도 : 2020 페이지 : pp. 385-393 (9 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: In this study, the effects of dance sports programs on the physical function, cognition, depression and quality of life of the elderly were verified for women aged 65 and older. Methods: The experimental group applied dance sports programs for once a week, 50 minutes/times and 12 weeks, and the comparison group was required to participate in existing sports programs without applying dance sports programs for 12 weeks. The experimental group and comparison group were randomly allocated with 33 subjects and 31 controls. Results: balance, depression, and quality of life were investigated in advance and post-production of the program to compare the changes between the experimental group and the control group. Conclusions: Dance sports programs targeting the elderly are life sports that help improve balance, depression, and quality of life.

KCI등재

10간호대학생의 간호전문직관, 돌봄효능감 및 윤리적 민감성이 돌봄행위에 미치는 영향

저자 : 진수희 ( Su Hee Jin ) , 김은희 ( Eun Hee Kim )

발행기관 : 한국보건정보통계학회 간행물 : 보건정보통계학회지 45권 4호 발행 연도 : 2020 페이지 : pp. 394-401 (8 pages)

다운로드

(기관인증 필요)

초록보기

Objectives: It is important for nursing students to properly establish the caring behaviors awareness in order to become a competent and professional nurse in the future. This study was to examine the effects of nursing professionalism, caring efficacy, and ethical sensitivity on caring behaviors in nursing students. Methods: The subjects of this study were total 205 nursing students in universities located in D Metropolitan City. Data were collected from September 2 to 30 in 2019 by using self-reported questionnaires. The data were analyzed by using statistical package the independent t-test, one-way ANOVA, Pearson's correlation coefficients, and stepwise multiple regression analysis with IBM SPSS 25.0 program. Results: The mean score of ethical sensitivity was 4.40±0.42, caring efficacy was 3.69±0.33, nursing professionalism was 3.57±0.40, and caring behaviors was 4.04±0.49. There were significant differences in biomedical ethics awareness by grade and satisfaction of nursing major. The influencing factors on caring behaviors in nursing students were ethical sensitivity (β=0.44, p<0.001), nursing professionalism (β=0.18, p=0.005), grade (β=0.18, p=0.006), and caring efficacy (β=0.16, p=0.006) which explained 31.1% of the variance. Conclusions: These results highlight the importance of developing caring behaviors in nursing students and the need to develop intervention programs that can enhance their nursing professionalism, caring efficacy, and ethical sensitivity should be considered.

12
발행기관 최신논문
자료제공: 네이버학술정보
발행기관 최신논문
자료제공: 네이버학술정보

내가 찾은 최근 검색어

최근 열람 자료

맞춤 논문

보관함

내 보관함
공유한 보관함

1:1문의

닫기