손상중증도지수를 이용한 손상외인별 손상예측모형 개발

Abstract

Injury is being ranked 3rd following malignant cancer and cardiovascular disease as the recent major death causes in our country, thereby emerging as a serious health problem of threatening whole people's safety and health. Therefore, an effort is being desperately demanded for closely examining and preventing a cause for the occurrence of injury. Accordingly, the purpose of this study is to be conducive to a policy-based decision for preventing and managing injury by suggesting a prevention measure after examining distribution of injury characteristics in discharge injury patients, analyzing risk factors by external cause in injury and by injury severity, and developing injury prediction model. This study is a retrospective descriptive survey research for implementing injury prevention model. A researcher was allowed the inspection of medical records after explaining about the compliance with research ethics at 3 medical institution of operating Emergency Medical Center. Period of data collection was for 1 year from January 2010 to December 2010. It surveyed 4,047 cases of medical data for discharge injury patients by using a data collection tool, which was developed after passing through consideration and preliminary survey of documents at home and abroad. The data collection tool was composed of totally 42 items with 8 items for general characteristics, 7 items for hospital-visit characteristics, 7 items for health characteristics, 8 items for hospitalization characteristics, and 12 items for injury characteristics. The collected data were examined the distribution of injury characteristics by using SPSS 18.0. It analyzed injury risk factors by external cause in injury, thereby having implemented the injury prediction model with a decision tree. The following are the analytical results by injury characteristic in injury patients discharged from hospital. First, the subjects of this study were totally 4,047 people. The average age was 39.62(±18.94) years old. 40-59 years old accounted for 37.3%. The male-to-female ratio stood at about 6:4. Men accounted for 64.1%. As for insurance type, the health insurance accounted for 48.0%. Marital status was shown to be 59.5% in the married. Job was indicated to be 21.7% in production post. The largest distribution was shown by general-characteristic item. Second, the characteristics of injury diagnosis were including diagnoses from 1 pieces at the minimum to 9 pieces at the maximum by research subject. The total number of diagnosing injury, which was used, was indicated to be 431 types. 66.4% of patients with traffic accident were included diagnosis of sprain and strain of cervical spine(S13.4). 7.7% of patients with wipeout and fall were included diagnosis of fracture of lower end of radius(S52.5). 53.1% of burn-injured patients were included diagnosis of burns involving less than 10% of body surface(T31.0). As for the distribution of injury diagnosis according to general characteristics, a burn was the largest in less than 20 years old and over 60 years old. Head & face were indicated to be the largest in 20-59 years old. Thus, it was statistically significant(p<0.001). Third, the distribution of external causes in injury according to age was happened the most in a burn(39.9%) as for less than 20 years old, in traffic accident(49.3%) as for 20-39 years old, in other external cause(20.1%) as for 40-59 years old, and in a hurt from a fall(43.8%) as for over 60 years old. Thus, this was statistically significant(p<0.001). In the meantime, the distribution of injury severity according to age was happened the most in less than 20 years with 32.2% as for mild case, in 20-59 years old with 61.4% as for moderate case, and in over 60 years old with 24.3% as for serious case. Thus, this was statistically significant(p<0.001). Fourth, the distribution of injury severity(ICISS) according to external cause in injury was occurred the most in mild case(26.6%) as for bicycle accident out of traffic accident, in moderate case(73.9%) as for car accident, and in serious case(29.8%) as for motorcycle accident. Thus, this was statistically significant(p<0.001). A hurt from a fall was occurred the most in mild case(34.7%) as for a fall and in serious case(18.0%) as for a plunge. This was statistically significant(p<0.05). A burn was occurred the most in mild case(18.9%) as for boiling water and in serious case(40.8%) as for a flame. This was statistically significant(p<0.001). Fifth, as a result of analyzing injury risk factors by external cause in injury, a traffic accident is a risk factor of serious case on mild case. A risk that a serious case will occur was significantly higher by 3.1 times in 20-39 years old compared to less than 20 years old, by 6.4 times in 40-59 years old, and by 10.8 times in over 60 years old. Compared to the reference value, even hospital-visit passage, hospital-visit means, and causative agent were significantly higher by 3.9 times when having hospital-visit passage, by 2.3 times as for visiting hospital by ambulance, by 1.8 times as for pedestrian, and by 3.3 times as for motorcycle. As the risk factors of moderate case on mild case, a risk that serious case will occur, compared to under 20 years old, was significantly higher by 3.4 times in 20-39 years old, by 3.7 times in 40-59 years old, and by 4.0 times in over 60 years old. Compared to the reference value, even the hospital-visit passage and causative agent were higher by 1.9 times when having passage of visiting hospital and was significantly lower by 0.5 times, respectively, in pedestrian and bicycle. A hurt from a fall is risk factor of serious case on mild case. Compared to the reference value, the hospital-visit means and injury activity were significantly higher by 3.3 times in visiting hospital by ambulance and by 4.1 times during work. As a risk factor of moderate case on mild case, BMI, hospital-visit means, causative agent, and season, compared to the reference value, were significantly higher by 1.7 times as for obesity, 2.4 times as for visiting hospital by ambulance, by 2.1 times as for building/structure, and were significantly lower by 0.5 times as for winter. As a burn is risk factor of serious case on mild case, the hospital-visit means and the time from injury occurrence to hospital visit, compared to the reference value, were significantly higher by 3.1 times as for visiting hospital by ambulance and were significantly lower by 0.5 times as for visiting hospital after passage of over 6 hours following injury occurrence. As a risk factor of moderate case on mild case, the hospital-visit means and causative agent, compared to the reference value, were significantly higher by 1.9 times as for visiting hospital by ambulance, and by 3.2 times as for flame/fire heat. Sixthly, as a result of the injury prediction model by external cause in injury, the traffic accident was shown to be causative agent as the major cause of having influence upon injury severity. The prediction accuracy of a model for age, hospital-visit means, and hospital-visit passage as factors according to this were indicated to be 68.5%. A hurt from a fall was indicated to be hospital-visit means, which is the main factor of having influence upon injury severity. The prediction accuracy of a model for BMI, season, marital status, age, and injury place as factors according to this were indicated to be 55.6%. A burn was indicated to be the hospital-visit means, which is the main factor of having influence upon injury severity. The prediction accuracy of a model for causative agent, time from injury occurrence to visiting hospital, communication, and emotional status as factors according to this were indicated to be 59.7%. Three things can be largely arranged on the basis of the results, which were analyzed in this study. First, this study specifically grasped injury factors among general characteristics, hospital-visit characteristics, health characteristics, hospitalization characteristics, and injury characteristics of subjects that were occurred injury conventionally, implemented injury prediction model with the confirmed factors, and identified the formation of the link among injury risk factors. Accordingly, a path, which can be occurred injury, can be intercepted in advance by proceeding with removing some of the link for risk factors like domino that was confirmed in a model. Thus, it is considered to have significance in the aspect of safe health. Second, typical injury types were indicated to be traffic accident, a hurt from a fall, and a burn. Thus, injury could be confirmed to occur in the form of diverse causes. Therefore, the safe-health approach is demanded for establishing injury prevention measure with each different method according to injury type, severity level, and general characteristics. Third, understanding about characteristics of discharge injury patients is thought to lead to confirming scope of injury risk factor in advance in a site of being demanded injury treatment such as first aid treatment and emergency transportation for injury patients and to contributing to lowering injury severity with prompt application.;최근 우리나라 주요 사망원인은 악성 암, 순환기계 질환에 이어, 손상이 3위를 차지하고 있어 전 국민의 안전과 건강을 위협하는 심각한 보건문제로 대두되고 있다. 그러므로 손상발생의 원인을 규명하고, 예방하기 위한 노력이 절실히 요구된다. 이에 본 연구는 퇴원손상환자의 손상특성의 분포를 살펴보고, 손상외인 및 손상중증도별 위험요인을 분석하고 손상예측모형을 개발함으로써 보건학적 차원의 예방대책을 제시하여 손상예방 및 관리를 위한 정책적 결정에 도움이 되고자 한다. 본 연구는 손상예방모형을 구축하기 위한 후향적 서술조사 연구로서 연구자는 3곳의 응급의료센터를 운영하는 의료기관에서 연구윤리를 준수할 것을 설명하고 의무기록 열람을 허가받았다. 자료수집 기간은 2010년 1월부터 2010년 12월까지 1년간 이었으며, 국내외 문헌을 고찰하고, 예비조사를 거쳐 개발한 자료수집도구를 이용하여 퇴원손상환자의 의무기록 4,047건을 조사하였다. 자료수집도구는 일반적 특성, 8항목, 내원특성 7항목, 건강특성 7항목, 입원특성 8항목, 손상특성 12항목으로 총 42항목으로 구성되었고, 수집된 자료는 SPSS 18.0을 이용하여 손상특성 분포를 살펴보고, 외인별 손상위험요인을 분석하여 의사결정나무로 손상예측모형 구축하였다. 연구대상자의 손상관련 특성 분포 및 손상예측모형 결과는 다음과 같다. 첫째, 본 연구의 대상자는 총 4,047명으로 평균 연령 39.62(±18.94)세로 40-59세가 37.3%이며, 남녀의 비율은 6:4 정도로 남자가 64.1%였고, 보험유형은 건강보험이 48.0%이고, 결혼상태는 기혼 59.5%, 직업은 생산직이 21.7%로 일반적 특성 항목별로 가장 많은 분포를 보였다. 둘째, 손상진단 분포는 대상자별로 최소 1개부터 최대 9개까지 진단을 포함하고 있었으며, 총 사용된 손상진단 개수는 431개였다. 교통사고환자 66.4%가 경추의 염좌 및 긴장(S13.4) 진단을 포함하였고, 낙상환자는 7.7%가 요골하단의 골절(S52.5) 진단을 포함하였으며, 화상환자 53.1%가 신체표면의 10%미만을 포함한 화상(T31.0) 진단을 포함하였다. 일반적 특성에 따른 손상진단 분포는 20세 미만과 60세 이상에서는 화상진단이 가장 많았고, 20-59세 사이는 머리얼굴진단이 가장 많이 나타나 통계적으로 유의하였다(p<0.001). 셋째, 연령에 따른 손상외인 분포는 20세 미만에서는 화상(39.9%), 20-39세 사이는 교통사고(49.3%), 40-59세 사이는 기타외인(20.1%), 60세 이상은 넘어짐/추락(43.8%)이 가장 많이 발생하여 이는 통계적으로 유의하였다(p<0.001). 한편, 연령에 따른 손상중증도 분포는 경증은 20세 미만에서 32.2%, 중등증은 20-59세 사이에서 61.4%, 중증은 60세 이상에서 24.3%로 가장 많이 발생하였고 이는 통계적으로 유의하였다(p<0.001). 넷째, 손상외인에 따른 손상중증도 분포는 교통사고 중 자전거사고는 경증(26.6%)이 자동차사고는 중등증(73.9%), 오토바이사고는 중증(29.8%)이 가장 많이 발생하였고, 이는 통계적으로 유의하였다(p<0.001). 낙상 중 넘어짐은 경증(34.7%)이 추락은 중증(18.0%)이 가장 많이 발생하였고, 이는 통계적으로 유의하였다(p<0.05). 화상 중 열탕은 경증(18.9%)이 화염은 중증(40.8%)이 가장 많이 발생하였고, 이는 통계적으로 유의하였다(p<0.001). 다섯째, 손상외인별 손상위험요인을 분석한 결과 교통사고는 경증에 대한 중증의 위험요인으로 20세 미만에 비해 20-39세는 중증이 발생할 위험이 3.1배, 40-59세는 6.4배, 60세 이상은 10.8배로 유의하게 높았으며, 내원경유와 내원수단, 유발매체도 기준 값에 비해 내원경유가 있을 때 3.9배, 구급차 내원은 2.3배, 보행자는 1.8배, 오토바이는 3.3배로 유의하게 높았다. 경증에 대한 중등증의 위험요인으로 20세 미만에 비해 20-39세는 중등증이 발생할 위험이 3.4배, 40-59세는 3.7배, 60세 이상은 4.0배로 유의하게 높았으며, 내원경유와 유발매체도 기준 값에 비해 내원경유가 있을 때 1.9배로 높았고, 보행자와 자전거는 각각 0.5배로 유의하게 낮았다. 낙상은 경증에 대한 중증의 위험요인으로 내원수단, 손상활동은 기준 값에 비해 구급차 내원은 3.3배, 근무 중은 4.1배로 유의하게 높았다. 경증에 대한 중등증의 위험요인으로 BMI, 내원수단, 유발매체, 계절은 기준 값에 비해 비만은 1.7배, 구급차 내원은 2.4배, 건물/구조물은 2.1배로 유의하게 높았고, 겨울은 0.5배로 유의하게 낮았다. 화상은 경증에 대한 중증의 위험요인으로 내원수단, 손상발생부터 내원까지의 시간은 기준 값에 비해 구급차 내원은 3.1배로 유의하게 높았고, 손상발생 후 6시간이상 경과 후 내원은 0.5배로 유의하게 낮았다. 경증에 대한 중등증의 위험요인으로 내원수단, 유발매체는 기준 값에 비해 구급차 내원은 1.9배, 불꽃/화기는 3.2배로 유의하게 높았다. 여섯째, 손상외인별 손상예측모형 결과 교통사고는 손상중증도에 영향을 미치는 주 요인이 유발매체로 나타났고, 유발매체와 연결된 위험요인으로 연령, 내원수단, 내원경유로 모형의 예측정확도는 68.5%로 나타났다. 낙상은 손상중증도에 영향을 미치는 주 요인이 내원수단으로 나타났고, 내원수단과 연결된 위험요인으로 BMI, 계절, 결혼상태, 연령, 손상장소로 모형의 예측정확도는 55.6%로 나타났다. 화상은 손상중증도에 영향을 미치는 주 요인이 내원수단으로 나타났고, 내원수단에 연결된 위험요인으로 유발매체, 손상발생부터 내원까지의 시간, 의사소통, 정서상태로 모형의 예측정확도는 59.7%로 나타났다. 본 연구에서 분석되어진 결과를 토대로 크게 세 가지로 정리할 수 있다. 첫째, 본 연구는 손상을 경험한 대상자들의 일반적 특성, 내원특성, 건강특성, 입원특성, 손상특성 속에서 손상에 영향을 미친 요인을 찾아 확인된 요인으로 손상예측모형을 구축한 결과 손상위험요인간의 연결고리가 형성되어 있음을 확인하였다. 따라서 모형에서 확인된 도미노와 같은 위험요인 연결고리의 일부를 제거해 나감으로써 손상이 발생될 수 있는 경로를 사전에 차단할 수 있으므로 안전보건학적 의의가 크다고 본다. 둘째, 대표적인 손상유형은 교통사고, 낙상, 화상으로 나타나 손상은 다양한 원인과 형태로 발생하고 있음을 확인할 수 있었고, 손상의 유형과 중증도, 일반적 특성 등에 따라 각각 다른 보건학적 접근의 손상예방대책이 요구된다. 셋째, 퇴원손상환자의 특성을 파악함으로써 향후 손상환자의 응급처치 및 응급 후송 등 손상치료가 요구되는 현장에서 손상위험요인의 범위를 사전에 확인하고 신속한 적용으로 손상중증도를 낮추는데 기여할 것으로 생각된다.