• 응용통계연구
• /
• 제26권6호
• /
• pp.1019-1032
• /
• 2013

최근 급격한 기대수명의 증가에 따라 미래 복지정책 등에 커다란 영향을 주는 장래 사망력의 정확한 예측은 중요한 이슈가 되고 있다. 사망력의 정확한 예측을 위하여 최적의 추정모형의 선택도 중요하지만 사망력에 대한 시계열 적용기간도 매우 중요한 이슈다. 이는 우리나라의 사망률 시계열이 짧고, 특히 1982년 이전 자료가 다소 불완전해서 이에 대한 고려가 필수적이기 때문이다. 본 논문에서는 우리나라 사망력 시계열을 기간에 따라 2개의 그룹(1976~2005년, 1983~2005년)으로 나누어서, 남녀별로 LC모형과 LC 코호트효과 확장모형에 대한 모수 추정값, 사망력지수와 코호트지수의 모형화 및 예측, 장래 기대수명의 예측 적합력을 각각 분석한 후 향후에 장래 기대수명 추계시 고려할 시사점을 제시하고자 한다.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제9권10호
• /
• pp.309-316
• /
• 2020

우리는 응급실을 방문한 65세 이상 노인환자의 의료 데이터를 각각 피드 포워드 신경망과 합성곱 신경망에 학습하여 사망률을 예측하였다. 의료 데이터는 노인환자의 성별, 연령, 체온, 심박 수 등의 기초적인 정보뿐 아니라 과거 병력, 다양한 혈액 검사 및 배양 검사 결과 등 다양하고 복잡한 정보를 포함하여 총 99가지의 자질로 구성된다. 이 중 사망률 예측에 크게 기여하는 자질을 선택하기 위해 랜덤 포레스트를 이용하여 자질의 중요도를 계산하였고, 그 결과 중요도가 높은 상위 80개의 자질을 선택하였다. 선택된 자질을 각각 피드 포워드 신경망과 합성곱 신경망의 학습에 사용하여 두 신경망의 성능을 비교하였다. 합성곱 신경망 학습을 위해 의료 데이터를 고정된 크기의 이미지로 변환하였으며 합성곱 신경망이 피드 포워드 신경망을 이용한 것보다 성능이 좋았다. 합성곱 신경망의 사망률 예측 성능으로 테스트 데이터에 대해 F1 점수는 56.9, AUC는 92.1을 각각 얻었다.

• Journal of Chest Surgery
• /
• 제54권2호
• /
• pp.88-98
• /
• 2021

Background: This study aimed to develop a new risk prediction model for operative mortality in a Korean cohort undergoing heart valve surgery using the Korea Heart Valve Surgery Registry (KHVSR) database. Methods: We analyzed data from 4,742 patients registered in the KHVSR who underwent heart valve surgery at 9 institutions between 2017 and 2018. A risk prediction model was developed for operative mortality, defined as death within 30 days after surgery or during the same hospitalization. A statistical model was generated with a scoring system by multiple logistic regression analyses. The performance of the model was evaluated by its discrimination and calibration abilities. Results: Operative mortality occurred in 142 patients. The final regression models identified 13 risk variables. The risk prediction model showed good discrimination, with a c-statistic of 0.805 and calibration with Hosmer-Lemeshow goodness-of-fit p-value of 0.630. The risk scores ranged from -1 to 15, and were associated with an increase in predicted mortality. The predicted mortality across the risk scores ranged from 0.3% to 80.6%. Conclusion: This risk prediction model using a scoring system specific to heart valve surgery was developed from the KHVSR database. The risk prediction model showed that operative mortality could be predicted well in a Korean cohort.

• 대한화상학회지
• /
• 제23권1호
• /
• pp.1-6
• /
• 2020

Purpose: PL, creatinine and urine output are biomarkers of the suitability and prognosis of fluid therapy in severe burn patients. The purpose of this study is to evaluate the usefulness of predicting mortality by biomarkers and its change during initial fluid therapy for severe burn patients. Methods: A retrograde review was performed on 733 patients from January 2014 to December 2018 who were admitted as severe burn patients to our burn intensive care unit (BICU). Plasma lactate, serum creatinine and urine output were measured at the time of admission to the BICU and after 48 hours. ABSI score, Hangang score, APACHEII, revised Baux index and TBSA were collected after admission. Results: 733 patients were enrolled. PL was the most useful indicators for predicting mortality in burn patients at the time of admission (AUC: 0.813) and after 48 hours (AUC: 0.698). On the other hand, mortality prediction from initial fluid therapy for 48 hours showed different results. Only creatinine showed statistical differences (P<0.05) in mortality prediction. But there were no statistical differences in mortality prediction with PL and UO (P>0.05). Conclusion: In this study, PL was most useful predictor among biomarkers for predicting mortality. Improvement in creatinine levels during the first 48 hours is associated with improved mortality. Therefore, efforts are needed to improve creatinine levels.

• Asian Pacific Journal of Cancer Prevention
• /
• 제15권16호
• /
• pp.6929-6934
• /
• 2014

Background: To analyze the mortality distribution of esophageal cancer in China from 1991 to 2012, to forecast the mortality in the future five years, and to provide evidence for prevention and treatment of esophageal cancer. Materials and Methods: Mortality data for esophageal cancer in China from 1991 to 2012 were used to describe its epidemiological characteristics, such as the change of the standardized mortality rate, urban-rural differences, sex and age differences. Trend-surface analysis was used to study the geographical distribution of the mortality. Curve estimation, time series, gray modeling, and joinpoint regression were used to predict the mortality for the next five years in the future. Results: In China, the incidence rate of esophageal cancer from 2007 and the mortality rate of esophageal cancer from 2008 increased yearly, with males at $8.72/10^5$ being higher than females, and the countryside at $15.5/10^5$ being higher than in the city. The mortality rate increased from age 45. Geographical analysis showed the mortality rate increased from southern to eastern China, and from northeast to central China. Conclusions: The incidence rate and the standardized mortality rate of esophageal cancer are rising. The regional disease control for esophageal cancer should be focused on eastern, central and northern regions China, and the key targets for prevention and treatment are rural men more than 45 years old. The mortality of esophageal cancer will rise in the next five years.

• Asian Pacific Journal of Cancer Prevention
• /
• 제16권15호
• /
• pp.6729-6734
• /
• 2015

Background: To analyze the mortality distribution of nasopharyngeal carcinoma in China from 1991 to 2013, to predict the mortality in the ensuing five years, and to provide evidence for prevention and treatment of nasopharyngeal carcinoma. Materials and Methods: Mortality data for Nasopharyngeal Carcinoma in China from 1991 to 2013 were used to describe its epidemiological characteristics, such as the change of the standardized mortality rate, sex and age differences, urban-rural differences. Trend-surface analysis was used to study the geographical distribution of the mortality. Curve estimation, time series, gray modeling, and joinpoint regression were used to predict the mortality for the ensuing five years in the future. Results: In China, the standardized mortality rate of Nasopharyngeal Carcinoma increased with time from 1996, reaching the peak values of $1.45/10^5$ at the year of 2002, and decreased gradually afterwards. With males being 1.51 times higher than females, and the city had a higher rate than the rural during the past two decades. The mortality rate increased from age 40. Geographical analysis showed the mortality rate increased from middle to southern China. Conclusions: The standardized mortality rate of Nasopharyngeal Carcinoma is falling. The regional disease control for Nasopharyngeal Carcinoma should be focused on Guangdong province of China, and the key targets for prevention and treatment are rural men, especially after the age of 40. The mortality of Nasopharyngeal Carcinoma will decrease in the next five years.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제7권7호
• /
• pp.275-280
• /
• 2018

세계 인구의 고령화가 진행되는 오늘날 노인들을 위한 의료 서비스의 수요는 점차 증가할 것으로 보인다. 특히, 응급실을 방문하는 노인 환자는 일반 환자보다 다양한 질병을 갖고 있거나, 특이한 증상을 호소하는 등 복잡한 의학적, 사회적 및 신체적 문제를 가지고 있는 경우가 많다. 우리는 65세 이상의 응급실을 방문한 노인 환자의 사망률 예측을 위해 연령, 성별, 혈압, 체온, 혈액검사, 주증상명 등의 의료 데이터를 사용하였다. Feed Forward 신경망과 지지벡터기계를 각각 학습하여 사망률을 예측하고 그 성능을 비교하였다. 1개의 은닉층을 사용한 Feed Forward 신경망의 실험결과가 가장 좋았으며, 이 때 F1 점수는 52.0%, AUC는 88.6%이다. 좀 더 좋은 의료 자질을 추출하여 제안 시스템의 성능을 향상시킨다면 응급실에 방문한 노인 환자들을 위한 효과적이고 신속한 의료 자원 배분을 통해 더 좋은 의료 서비스를 제공할 수 있을 것이다.

• 대한응급의학회지
• /
• 제29권5호
• /
• pp.455-464
• /
• 2018

Objective: Machine learning is not yet widely used in the medical field. Therefore, this study was conducted to compare the performance of preexisting severity prediction models and machine learning based models (random forest [RF], gradient boosting [GB]) for mortality prediction in pneumonia patients. Methods: We retrospectively collected data from patients who visited the emergency department of a tertiary training hospital in Seoul, Korea from January to March of 2015. The Pneumonia Severity Index (PSI) and Sequential Organ Failure Assessment (SOFA) scores were calculated for both groups and the area under the curve (AUC) for mortality prediction was computed. For the RF and GB models, data were divided into a test set and a validation set by the random split method. The training set was learned in RF and GB models and the AUC was obtained from the validation set. The mean AUC was compared with the other two AUCs. Results: Of the 536 investigated patients, 395 were enrolled and 41 of them died. The AUC values of PSI and SOFA scores were 0.799 (0.737-0.862) and 0.865 (0.811-0.918), respectively. The mean AUC values obtained by the RF and GB models were 0.928 (0.899-0.957) and 0.919 (0.886-0.952), respectively. There were significant differences between preexisting severity prediction models and machine learning based models (P<0.001). Conclusion: Classification through machine learning may help predict the mortality of pneumonia patients visiting the emergency department.

• 한국산림과학회지
• /
• 제100권1호
• /
• pp.1-7
• /
• 2011

본 연구는 기존의 잣나무임분 거리독립 개체목 생장모델을 개선하기 위해 수행되었다. 직경생장함수 및 고사율함수의 매개변수들을 고정표본점의 정기평균생장량을 토대로 추정하고, 이 함수들의 특성을 기존의 총평균생장량을 토대로 추정한 모델과 비교하였다. 여기서 생장함수는 수관율함수, 잠재직경생장함수 및 임분을 구성하는 임목간 경쟁효과를 고려하기 위한 수정율함수를 의미한다. 고사율예측함수의 경우에는 고정표본점 자료의 한계로 인해 정기평균생장량 측정값을 구할 수 없어 대신 총평균생장량과의 관계식을 추정하여 대체하여 적용하였다. 연구결과 정기평균생장량을 토대로 하는 직경생장함수가 총평균생장량을 토대로 추정한 함수에 비해 개체목의 생장특성을 보다 현실적으로 반영하는 것을 보여주었다. 고사율함수의 경우, 총평균생장량을 적용하여 개발한 경우 고사율이 과대한 것으로 나타나는 문제가 있었으나 새로운 모델에서는 이 문제가 개선된 것으로 나타났다.

• Asian Pacific Journal of Cancer Prevention
• /
• 제15권6호
• /
• pp.2785-2791
• /
• 2014

Aims: To analyze time-dependent changes in female breast cancer (BC) mortality in China, forecast the trend in the ensuing 5 years, and provide recommendations for prevention and management. Materials and Methods: Mortality data of breast cancer in China from 1991 to 2011 was used to describe characteristics and distribution, such as the changes of the standardized mortality rate, urban-rural differences and age differences. Trend-surface analysis was used to study the geographical distribution of mortality. In addition, curve estimation, time series modeling, Gray modeling (GM) and joinpoint regression were performed to estimate and predict future trends. Results: In China, the mortality rate of breast cancer has increased yearly since 1991. In addition, our data predicted that the trend will continue to increase in the ensuing 5 years. Rates in urban areas are higher than those in rural areas. Over the past decade, all peak ages for death by breast cancer have been delayed, with the first death peak occurring at 55 to 65 years of age in urban and rural areas. Geographical analysis indicated that mortality rates increased from Southwest to Northeast and from West to East. Conclusions: The standardized mortality rate of breast cancer in China is rising and the upward trend is predicted to continue for the next 5 years. Since this can cause an enormous health impact in China, much better prevention and management of breast cancer is needed. Consequently, disease control centers in China should place more focus on the northeastern, eastern and southeastern parts of China for breast cancer prevention and management, and the key population should be among women between ages 55 to 65, especially those in urban communities.