보건정보통계학회지제 39 권제 2 호 ISSN 2287-3708(Print) ISSN 2287-3716(Online) Journal of Health Informatics and Statistics (JHIS) 2014; 39(2): 77-89 사망원인과지역의관련성에관한연구 조성인 1), 김병수 1,2), 전민수 1) 1) 인제대학교데이터정보학과 2) 인제대학교통계정보연구소 A Study on the Relation between the Cause of Death and the Regions Seongin Cho 1), Byungsoo Kim 1,2), Minsoo Jeon 1) 1) Department of Data Science, Inje University 2) Institute of Statistical Information, Inje University Abstract Objectives: This study aims to examine the relation between the cause of death and the regions. Methods: It is examined what differences are in the cause of death based on regions and, if there are some differences, it is also examined in regard to regional characteristics focused on lift what the cause of death which has big differences is based on regions. Also, regional characteristics related to regional differences in the cause of death are traced out by using correlation coefficient. For the cause of death, the ones which are more than 10,000 among the ones which are applicable to the intermediate level for Korean Standard Classification of Diseases are examined and the regions are examined based on the unit of provincial governments and the one of municipal governments. Results: According to the result of analysis, lift for diabetes and suicide is the highest at Dongducheon-si and Siheung-si, respectively and it is also high at the regions where the rate of drinking alcohol and the one for early marriage and divorce are high. For the cause of death which has big differences based on regions, it is the death by heart disease as other systems and the lift is high at Gyeongnam regions. As a result of the correlation analysis performed for regional characteristics, correlation of variables for average erature, the lowest erature and the highest erature is high. Also, probability to die of disease of respiratory system is high at the regions whose rate for the oldest is high. It is low at the regions with a great deal of precipitation. Conclusions: We discovered the fact that there are causes to be discrepant on the basis of probability to die by region, and we can make a conclusion that a part of causes to die have a relationship to each regional characteristics such as atmospheric eratures. Keywords: Lift, Cause of death, Correlation analysis, Korean standard classification of diseases [ 접수일 : 2014 년 10 월 12 일, 수정일 : 2014 년 12 월 01 일, 게재확정일 : 2014 년 12 월 23 일 ] Corresponding Author: Byung Soo Kim, PhD Department of Data Science, Inje University / Institute of Statistical Information, 197 Inje-ro, Gimhae-si 621-749, Korea. Tel: +82-55-320-3300 E-mail: statkbs@inje.ac.kr
78 조성인, 김병수, 전민수 1. 서론 최근세계경제협력개발기구 (Organization for Economic Co-operation and Development, OECD) 에의하면 34개국중한국이자살사망률 1위, 간질환사망률 1위, 결핵사망률 1위등으로높게나타나고있으며, 통계청에서최근발표한자료에따르면전국 16개시도지역별표준화사망률이계속적으로증가하는추세를보이고있다. Kim and Park [1] 에의하면, 조사망률역시 1990년부터점차높아지는추세를보여왔다. 따라서국민들의사망률을낮추기위해서는제한된자원을효율적으로사용할수있는정책을추진하여야한다. 이러한정책은사망원인에대한연구를통하여얻어진과학적근거에기초하여야한다. 즉, 현재국민들의사망원인과그비율이어느정도인가를파악하고그러한사망에영향을미치는인자들가운데사용가능한자원의범위내에서가장실행가능성이높고영향이큰인자들부터우선순위를정하여건강수준의향상을꾀하는정책을추진하여야할것이다. 이를위해서적절한연구를수행하고그결과를정책결정자들에게효과적으로전달할수있어야한다. 지역간에사망원인에차이가있을때그차이에영향을미치는인자들을파악하고그영향의크기가어느정도인지를파악하는과정이필요하게된다. 사망에영향을미치는요인들은지역, 연령, 질병, 날씨등다양하게나타날수있다. 이러한연구를통해지역에따른사망원인의차이를파악하고그영향요인을파악함으로써보건정책과정에서중요한부분인자원의지역배분과관련된결정을내리는데필요한기준을제공해줄수있으며각지역에서시행되고있는정책의효율성을측정하는데도움을줄수있을것으로기대된다. 지역별사망수준에관한연구로는 Hong et al. [2] 에서지역시군구지역의사회적배제개념과 물질결핍정도를고려하여통합결핍지수를산출하고지역별통합결핍지수에따른기대여명에대해연구하였고, Choi [3] 는순사망률과조사망률에대하여비교하며주요사망원인이외의다양한원인들도중요하다고하였으며지역별질병별사망자의차이에대한연구로는 Park [4] 과 Kim et al. [5] 이있다. Lee [6] 는 1992년과 1993년의대구직할시의인근지역인경산지역을시군으로사망원인별사망률을비교분석하였다. 최근지역간사망원인을연구한 Park [7] 은지역을자살사망률이높은지역과낮은지역으로나누어지역주민에게영향을미치는요인으로인구경제학적특성과건강행태, 정신건강수준에대한분석을하였다. Jung et al. [8] 에서는한국인암예방연구자료를이용하여흡연관련사망자수를계산하였고, 그외사망원인통계의시작이될수있는사망신고제도관련연구로 Kim [9] 의연구가있으며, 분석방법과관련된연구로는 Park [10] 의척도화리프트에대한연구가있다. 본연구의목적은우리나라사람들의사망원인이지역에따라차이가있는지에대해살펴보고, 그차이와관계된지역특성변수는어떤것이있는지알아보고자하는데있다. 이와관련된연구로는 Kim [11] 은 2002년 60살이상사망인구중허혈성심장질환의사망비율은부산이최다지역이었다는연구결과를보였으며, Shin [12] 에서는로지스틱회귀분석을통해대상자의사망시성별, 직업, 결혼여부, 연령등이사망과독립적으로작용하지않음을알수있었다. 또, Lee [13] 에서는지역별고온극한현상에대한사망에대해연구하였고, Jung et al. [14] 에서는사망과관련된사회적요인으로개인의생활양식을빼놓을수없다고하였다. 그러므로지역별사망원인과지역의관련성에대하여연구할때음주율, 흡연율등과같이각지역주민의생활양식과관련된사회적요인들에대하여파악하고지역별차이를알아볼필요가있다. 보건정보통계학회지제 39 권제 2 호
사망원인과지역의관련성에관한연구 79 본논문의구성은다음과같다. 2장에서는분석자료와분석방법에대해설명하고 3장에서는사망원인의연도별사망비율과사망원인과지역의리프트그리고지역특성변수와리프트의상관분석을통해사망원인에따라지역을비교한다. 4장에서는본연구의결론을제시한다. 2. 연구방법 1) 자료및변수설명 본연구는통계청운영의마이크로데이터서비스시스템의사망원인자료를사용하였고, 2008년 1월 1일부터 2012년 12월 31일까지 5년간사망한사람들의사망신고서에의한자료이다. 원자료는 2008년 247,757건, 2009년 248,609건, 2010년 257,068건, 2011년 258,994건, 2012년 268,980건으로총 1,281,408명에해당하는사망자에대한자료이며, 본연구에서는우리나라지역별사망원인의형태를알아보고자하는목적을고려하여주소지가국외인사망자들에대한자료 8,331건 과빈도가작은 2012년에출범한세종특별자치시의 2012년자료 820건을제외하여총 1,272,257건의자료를대상으로하였다. 사망원인코드는 1, 2 로구분되어있으며, 외인사에해당하는사망원인 2가있는경우에는이것을우선으로하여하나의사망원인변수를만들었다. 예를들어, 사망원인1 이 S02 ( 두개골및안면골의골절 ) 이고사망원인 2가 W13 ( 빌딩또는구조물에서의낙상 ) 인경우에는일차적인원인인사망원인2의값을사망원인변수의값으로하였다. 또한, 한국표준질병사인분류코드 (Korean Standard Classification of Diseases Code, KCD Code) 의대분류, 중분류, 그리고소분류에대한변수를각각생성하였다. 본논문에서는중분류에의한분석을주로다룰것이고사망자료에서고려하는변수는 Table 1과같다. 한국표준질병사인분류에서세분류로나눠져있는질병분류기호코드중중분류는종류가 266개이다. Table 2는사망자수가 10,000명이상으로중요하다고여겨지는중분류사망원인 23개를보여주고있으며, 이후분석에서는 Table 2에제시한 23개의중분류만다루겠다. Table 1. Variables description of the cause of death Variable name Variable description reg1 Address (provincial governments) reg2 Address (municipal governments) die_year Death year die_mon Death month die_day Death day death Cause of death KCD code Medium classification of Korean standard classification of diseases Journal of Health Informatics and Statistics (JHIS), Vol. 39, No. 2, 2014
80 조성인, 김병수, 전민수 Table 2. Medium classification of KCD code Variable name [A15-A19] [C15-C26] [C30-C39] [C51-C58] [C64-C68] [C81-C96] [E10-E14] [F00-F09] [G30-G32] [I10-I15] [I20-I25] [I30-I52] [I60-I69] [J09-J18] [J40-J47] [K70-K77] [N17-N19] [R50-R69] [R95-R99] [V01-V09] [W00-W19] [X60-X84] [Y10-Y34] Variable description Tuberculosis Malignant neoplasms of digestive organs Malignant neoplasms of respiratory and intrathoracic organs Malignant neoplasm of female genital organs Malignant neoplasm of urinary system Malignant neoplasms of lymphoid, hematopoietic and related tissue Diabetes mellitus Organic, including symptomatic mental disorde Other degenerative diseases of the nervous system Hypertensive diseases Ischemic heart diseases Other forms of heart disease Cerebrovascular diseases Influenza and pneumonia Chronic lower respiratory diseases Diseases of liver Renal failure General symptoms and signs Ill-defined and unknown causes of mortality Pedestrian injured in transport accident Falls Suicide Event of undetermined intent 2) 분석방법지역과사망원인의관련성에대해알아보기위해총 16개의시도, 231개의구시군에대하여사망원인별리프트를구하였다. 리프트를통해서각사망원인에해당하는사망자수가지역이라는조건에의해어느정도달라지는지에대하여볼수있다. 리프트는사건와사건에대해, 사건가일어났을때사건가발생할확률인와의조건부확률와사건에대한발생조건이없을때의확률의비로나타낸다. 조건이있을때와조건이없을때의확률이같다면리프트는 1 을가지며사건와는서로독립이라할수있 다. 본연구에서사건는특정지역에주소지를두고있는사람이사망한사건이되고, 사건는특정사망원인으로사망한사건이된다. 리프트를통해지역과사망원인간의연관성에대해설명하고지역마다다르게나타날수있는지역특성변수를이용하여지역에따른리프트와의상관관계를보고자한다. 사망원인은중요하다고여겨지는 23개의중분류에대해서고려하였고지역은시도와구시군각각에대해서다루었다. 보건정보통계학회지제 39 권제 2 호
사망원인과지역의관련성에관한연구 81 지역특성변수는 Lee [13] 와 Jung et al. [14] 에서언급되거나유사한변수를중심으로구성하였다. 사회적인변수로는음주율, 흡연율, 이혼율, 고령자비율을고려하였고, 경제적인변수로는 1인당개인소득, 1인당민간소비를고려하였고, 자연적인변수로는평균기온, 최저기온, 최고기온, 평균일교차, 강수량변수를사용하였다. 3. 연구결과 1) 사망에대한기초분석 Table 3 은사망원인의연도별사망비율에대하여 나타낸표이다. 소화기관의악성신생물 (C15-C26) 이 15.1%, 뇌혈관질환 (I60-I69) 이 10.3% 로높은비율을보이고있다. 호흡기및흉곽내기관의악성신생물 (C30-C39) 과인플루엔자및폐렴 (J09- J18), 비뇨기관의악성신생물 (C64-C68) 그리고심장과관련된허혈성심장질환 (I20-I25), 기타형태의심장병 (I30-I52) 은 2008년부터 2012년까지사망자비율이꾸준히증가하고있다. 신경계통의기타퇴행성질환 (G30-G32) 과불명확하고미상의사인 (R95-R99) 은 2008년에비해 2012년에각각약 0.5%p, 0.7%p가증가하였으며, 자살 (X60-X84) 은증가와감소를반복하고있다. Table 3. Death cause-specific ratio KCD code 2008 2009 2010 2011 2012 Total (freq.) Total (%) A15-A19 0.9 0.9 0.9 0.9 0.9 11,805 0.9 C15-C26 15.3** 15.5** 15.2** 14.9** 14.7** 192,119 15.1 C30-C39 6.3* 6.4* 6.4* 6.4* 6.5* 81,487 6.4 C51-C58 0.9 0.9 0.9 0.9 0.8 11,062 0.9 C64-C68 0.8 0.8 0.9 0.9 0.9 10,955 0.9 C81-C96 1.4 1.5 1.5 1.5 1.6 18,910 1.5 E10-E14 4.2 4.0 4.1 4.2 4.3 52,644 4.1 F00-F09 1.8 1.6 1.6 1.7 1.8 21,422 1.7 G30-G32 0.8 0.8 0.8 1.0 1.3 11,924 0.9 I10-I15 1.9 1.9 1.9 2.0 2.0 24,525 1.9 I20-I25 5.2* 5.2* 5.2* 5.3* 5.5* 67,050 5.3 I30-I52 3.4 3.7 3.8 4.2 4.3 49,427 3.9 I60-I69 11.4** 10.5** 10.4** 9.9* 9.6* 131,351 10.3 J09-J18 2.2 2.6 3.0 3.4 3.9 38,519 3.0 J40-J47 3.0 2.8 2.8 2.7 2.9 36,095 2.8 K70-K77 2.9 2.8 2.7 2.6 2.5 34,440 2.7 N17-N19 1.5 1.5 1.5 1.5 1.6 19,249 1.5 R50-R69 8.7* 7.6* 6.9* 6.3* 6.1* 90,277 7.1 R95-R99 2.1 2.4 2.7 2.8 2.8 32,627 2.6 V01-V09 1.0 1.0 1.0 0.9 0.9 12,094 1.0 W00-W19 1.0 0.9 0.8 0.8 0.8 10,996 0.9 X60-X84 5.2* 6.2* 6.1* 6.2* 5.3* 73,859 5.8 Y10-Y34 1.1 1.2 1.0 1.0 0.9 13,324 1.1 Total 83.1 82.6 81.9 81.9 81.7 1,046,161 82.2 *ratio 5.00, **ratio 10.00 Journal of Health Informatics and Statistics (JHIS), Vol. 39, No. 2, 2014
82 조성인, 김병수, 전민수 2) 지역별사망원인에대한리프트각사망원인에의한사망비율이지역에따라어떻게달라지는지를보기위해다음의식을이용하여리프트를구하였다. Table 4와 Table 5는사망원인별로각지역의사망자수를가중치로하여계산한리프트의가중표준편차와리프트값이가장높은지역과낮은지역을보여주고있다. 리프트의가중표준편차가크다는것은지역에따라사망비율의차이가크다는것을의미한다. 특정지역의리프트가크다는것은그지역에서해당사망원인으로사망하는비율이높다는것을의미한다. Table 4에서불명확하고미상의사인 (R95- R99), 신경계통의기타퇴행성질환 (G30-G32), 기타형태의심장병 (I30-I52), 의도미확인사건 (Y10-Y34) 의가중표준편차가각각 0.40, 0.33, 0.23, 0.23으로지역별리프트의차이가크다. 불명확하고미상의사인 (R95-R99) 은제주도 (2.51) 에서가장높고부산 (0.27) 에서가장낮다. 즉, 제주도에서불명확하고미상의사인 (R95-R99) 의사망원인으로사망할확률이다른지역보다높다는것을알수있다. 알츠하이머병이포함되어있는신경계통의기타퇴행성질환 (G30-G32) 은전북 (1.57) 에서가장높으며강원 (0.45) 에서가장낮다. 심근염과심낭염등이포함되는기타형태의심장병 (I30-I52) 은경남 (1.50) 에서가장높고인천 (0.80) 이가장낮은리프트값을가지고있다. Table 5는구시군별사망원인에대한리프트이다. 불명확하고미상의사인 (R95-R99), 신경계통의기타퇴행성질환 (G30-G32), 전신증상및징후 (R50-R69) 에서리프트의가중표준편차가각각 0.50, 0.44, 0.33로지역별편차가크게나타났으며불명확하고미상의사인 (R95-R99) 의리프트는제주시 (2.85) 에서가장높고경남의령군 (0.11) 에서가장낮게나타났다. Table 4. Lift of death-cause between provincial governments KCD code Freq. Weighted SD The highest area The lowest area A15-A19 11,805 0.11 Gyeongsangbuk-do (1.23) Jeju-do (0.74) C15-C26 192,119 0.06 Busan (1.09) Gangwon-do (0.93) C30-C39 81,487 0.03 Gyeongsangbuk-do (1.07) Jeju-do (0.85) C51-C58 11,062 0.15 Seoul (1.17) Jeollanam-do (0.70) C64-C68 10,955 0.10 Seoul (1.15) Ulsan (0.86) C81-C96 18,910 0.14 Seoul (1.24) Chungcheongbuk-do (0.81) E10-E14 52,644 0.14 Gyeonggi-do (1.21) Jeju-do (0.63) F00-F09 21,422 0.16 Jeollanam-do (1.26) Jeju-do (0.60) G30-G32 11,924 0.33** Jeollabuk-do (1.58) Gangwon-do (0.45) I10-I15 24,525 0.17 Busan (1.43) Jeju-do (0.77) I20-I25 67,050 0.15 Busan (1.32) Jeju-do (0.70) I30-I52 49,427 0.23* Gyeongsangnam-do (1.50) Incheon (0.80) I60-I69 131,351 0.06 Jeollabuk-do (1.14) Jeju-do (0.81) J09-J18 38,519 0.14 Chungcheongbuk-do (1.44) Ulsan (0.78) J40-J47 36,095 0.21* Jeollabuk-do (1.35) Seoul (0.70) K70-K77 34,440 0.11 Busan (1.21) Daejeon (0.79) N17-N19 19,249 0.12 Daejeon (1.27) Chungcheongnam-do (0.72) R50-R69 90,277 0.22* Chungcheongbuk-do (1.38) Busan (0.56) R95-R99 32,627 0.40** Jeju-do (2.51) Busan (0.27) V01-V09 12,094 0.13 Jeju-do (1.48) Busan (0.83) W00-W19 10,996 0.11 Gwanju (1.24) Gyeongsangbuk-do (0.78) X60-X84 73,859 0.16 Incheon (1.20) Jeollanam-do (0.64) Y10-Y34 13,324 0.23* Jeju-do (1.68) Chungcheongbuk-do (0.75) *SD. 0.20, **SD. 0.30 SD: standard deviation 보건정보통계학회지제 39 권제 2 호
사망원인과지역의관련성에관한연구 83 Table 5. Lift of death-cause between municipal governments KCD code Freq. Weighted SD The highest area The lowest area A15-A19 11,805 0.23* Bonghwa, Gyeongbuk (2.18) Yanggu, Gangwon (0.24) C15-C26 192,119 0.08 Dongnae, Busan (1.18) Gapyeong, Gyeonggi (0.78) C30-C39 81,487 0.08 Yeongyang, Gyeongbuk (1.31) Gyeryong, Chungnam (0.72) C51-C58 11,062 0.23* Gwacheon, Gyeonggi (1.56) Cheongdo, Gyeongbuk (0.20) C64-C68 10,955 0.19 Yangyang, Gangwon (1.78) Jeongseon, Gangwon (0.19) C81-C96 18,910 0.19 Seocho, Seoul (1.67) Danyang, Chungbuk (0.44) E10-E14 52,644 0.18 Dongducheo, Gyeonggi (1.57) Ulleung, Gyeongbuk (0.32) F00-F09 21,422 0.27* Hwasun, Jeonnam (2.15) Ulleung, Gyeongbuk (0.31) G30-G32 11,924 0.44** Jeongeup, Jeonbuk (2.53) Inje, Gangwon (0.09) I10-I15 24,525 0.26* Ganghwa, Incheon (2.09) Jeungpyeong, Chungbuk (0.45) I20-I25 67,050 0.21* Mungyeon, Gyeongbuk (1.59) Iksan, Jeonbuk (0.56) I30-I52 49,427 0.29* Pyeongchang, Gangwon (1.97) Taebaek, Gangwon (0.56) I60-I69 131,351 0.09 Jeongeup, Jeonbuk (1.33) Hadong, Gyeongnam (0.77) J09-J18 38,519 0.23* Jangsu, Jeonbuk (2.39) Ulleung, Gyeongbuk (0.35) J40-J47 36,095 0.27* Cheorwon, Gangwon (2.35) Ongjin, Incheon (0.51) K70-K77 34,440 0.17 Yeongdo, Busan (1.56) Seocho, Seoul (0.54) N17-N19 19,249 0.21* Jeonju, Jeonbuk (1.69) Dangjin, Chungnam (0.20) R50-R69 90,277 0.33** Ganghwa, Incheon (2.14) Miryang, Gyeongnam (0.35) R95-R99 32,627 0.50** Jeju, Jeju (2.85) Uiryeong, Gyeongnam (0.11) V01-V09 12,094 0.23* Seogwipo, Jeju (1.83) Ongjin, Incheon (0.28) W00-W19 10,996 0.22* Ulleun, Gyeongbuk (2.42) Gochang, Jeonbuk (0.27) X60-X84 73,859 0.22* Siheung, Gyeonggi (1.48) Goheung, Jeonnam (0.42) Y10-Y34 13,324 0.30** Ulleung, Gyeongbuk (3.75) Goesan, Chungbuk (0.50) *SD. 0.20, **SD. 0.30 SD: standard deviation 3) 지역특성변수와의상관분석 Lee [13] 에서개인의생활양식과관련된요인들도지역사회내에서관계를맺으며살아가므로사회적, 지역적요인들이개인의건강형태와관계가될수있고, Jung et al. [14] 의연구와같이기온에따라지역별로사망자수에영향이있다는것을고려하여지역특성변수를생성하였다. 생성된지역특성변수와사망원인에대한지역별리프트가관련이있는지알아보기위해상관분석을실시하였다. 시도별리프트의상관분석에사용된지역특성변수는 Table 6에나타내었으며구시군별상관분석에사용된지역특성변수는 Table 6에서 1인당개인소득과 1인당 민간소비변수를제외한나머지변수이다. 지역특성변수중에월간음주율, 흡연율, 조이혼율, 1인당개인소득, 1인당민간소비, 고령자비율변수는 2011년통계청자료를사용하였으며, 충남당진시와경남창원시에해당하는자료는없으므로결측치로처리하였다. 기온에관한변수는기상청의 2008년-2012년자료로부터평균을내어사용하였다. 기온과강수량자료는전국의 78 개관측소에서관측된값으로관측소가없는지역의자료는가장인접한지역의값으로대체하여사용하였다. 상관분석은사망원인에따라실시하였으며, 각지역의사망원인리프트와지역특성변수간 Journal of Health Informatics and Statistics (JHIS), Vol. 39, No. 2, 2014
84 조성인, 김병수, 전민수 의상관계수를구하였으며각지역의총사망자수를가중치로두었다. Table 7은시도별리프트와지역특성변수의상관계수와유의확률에대한정보를나타내고있다. 소화기관의악성신생물 (C15-C26), 여성생식기관의악성신생물 (C51-C58), 비뇨기관의악성신생물 (C64-C68) 및자살 (X60-X84) 은음주율과개인소득, 민간소비와양의상관관계가강하며신경계통의기타퇴행성질환 (G30-G32), 기타형태의심장병 (I30-I52), 전신증상및징후 (R50-R69) 는기온관련변수와양의상관관계가강하게나타나고있다. 강수량이많은지역은결핵 (A15-A19), 호흡기및흉곽내기관의악성신생물 (C30-C39), 뇌혈관질환 (I60-I69), 만성하기도질환 (J40- J47) 과음의관계가상대적으로강하며, 소화기관의악성신생물 (C15-C26), 여성생식기관의악성신생물 (C51-C58), 비뇨기관의악성신생물 (C64-C68), 림프, 조혈및관련조직의악성신생물 (C81-C96), 당뇨병 (E10-E14), 고혈압질환 (I10-I15), 자살 (X60-X84), 낙상 (W00- W19) 과양의상관관계에있다. 그리고개인소득과민간소비는소화기관의악성신생물 (C15- C26), 여성생식기관의악성신생물 (C51-C58), 비뇨기관의악성신생물 (C64-C68), 림프, 조혈 및관련조직의악성신생물 (C81-C96), 고혈압질환 (I10-I15) 과상관계수 0.7이상으로강한양의상관관계를보이고있다. 구시군별사망원인의리프트와지역특성변수의상관분석결과를 Table 8에서보여주고있다. 음주율이높은지역에서는소화기관의악성신생물 (C15-C26), 여성생식기관의악성신생물 (C51-C58), 자살 (X60-X84) 로인한사망비율이높고만성하기도질환 (J40-J47), 의도미확인사건 (Y10-Y34) 의사망비율이낮음을알수있다. 최저기온은소화기관의악성신생물 (C15-C26) 과양의상관관계를보이며, 전신증상및징후 (R50-R69) 와는음의상관관계를보인다. 인플루엔자및폐렴 (J09-J18) 은고령자비율과양의상관관계를나타내며만성하기도질환 (J40-J47) 은고령자비율, 최고기온, 평균일교차와양의상관을, 강수량과는음의상관을보인다. 간의질환 (K70-K77) 은흡연율및조이혼율과양의상관관계를보이고있다. Figure1, Figure2, Figure3, Figure4는지역특성변수중유의한상관계수가많았던음주율, 고령화비율, 최고기온, 강수량에대해각각상관계수가가장높은 4개의사망원인과산점도, 상관계수를나타낸것이다. Table 6. Variables description of the regional characteristics data Variable name Variable description drk Monthly standardized drinking rate (%) smk Standardized smoking rate (%) dvc Crude divorce rate (%) income Personal income (10,000 won) consume Consumption of common people (10,000 won) aged Aged rate (over 65 years old) a_ Average erature l_ Minimum erature h_ Maximum erature d_ Average daily erature difference rain Rainfall 보건정보통계학회지제 39 권제 2 호
사망원인과지역의관련성에관한연구 85 Table 7. Correlation coefficient of regional characteristic variable with lift (provincial governments) KCD code drk smk dvc in come con sume aged a_ l_ h_ d_ rain A15-A19-0.28 0.17-0.32-0.37-0.33 0.55* -0.05-0.18 0.14 0.35-0.56* C15-C26 0.77** -0.37-0.43 0.74** 0.75** -0.54* 0.48 0.65** 0.14-0.66** 0.58* C30-C39-0.37 0.28-0.26-0.33-0.41 0.47-0.07-0.29 0.25 0.58* -0.69** C51-C58 0.79** -0.20-0.03 0.75** 0.83** -0.89** -0.12 0.13-0.44-0.55* 0.63** C64-C68 0.64** -0.47-0.12 0.73** 0.90** -0.59* -0.13 0.13-0.45-0.56* 0.82** C81-C96 0.52* -0.49-0.15 0.78** 0.88** -0.67** -0.22 0.03-0.51* -0.50* 0.64** E10-E14 0.32-0.13 0.37 0.21 0.42-0.56* -0.28-0.06-0.49-0.36 0.53* F00-F09-0.73** -0.11-0.03-0.56* -0.62** 0.55* 0.21-0.03 0.52* 0.50* -0.42 G30-G32-0.45-0.47-0.71 0.03-0.16 0.45 0.55* 0.35 0.69** 0.19-0.42 I10-I15 0.53* -0.50* -0.51* 0.81** 0.85** -0.22-0.07 0.14-0.35-0.48 0.51* I20-I25 0.38 0.33-0.29-0.03-0.05-0.04 0.52* 0.54* 0.36-0.33-0.25 I30-I52 0.13 0.19-0.26-0.05-0.2 0.09 0.64** 0.50* 0.69** 0.01-0.17 I60-I69-0.57* 0.17 0.25-0.53* -0.57* 0.33-0.20-0.27-0.06 0.27-0.50* J09-J18-0.32 0.35 0.41-0.59* -0.60* 0.41-0.11-0.16-0.01 0.19-0.28 J40-J47-0.83** 0.21 0.05-0.74** -0.86** 0.78** 0.09-0.20 0.49 0.68** -0.68** K70-K77 0.57* 0.44 0.11 0.03 0.09-0.24 0.38 0.48 0.14-0.46 0.16 N17-N19 0.20-0.40-0.34 0.43 0.34-0.33 0.22 0.25 0.16-0.17 0.17 R50-R69-0.28-0.25 0.25 0.00 0.10 0.18-0.59* -0.55* -0.51* 0.21 0.24 R95-R99 0.11-0.26 0.07 0.31 0.43-0.17-0.42-0.28-0.55* -0.15 0.36 V01-V09-0.62* 0.20 0.35-0.49-0.63** 0.27 0.03-0.18 0.29 0.48-0.33 W00-W19 0.57* -0.26 0.04 0.28 0.41-0.47 0.14 0.37-0.19-0.62* 0.54* X60-X84 0.63** 0.10 0.36 0.46 0.57* -0.88** -0.40-0.18-0.62* -0.33 0.54* Y10-Y34-0.60* -0.12-0.16-0.36-0.41 0.72** 0.18 0.05 0.29 0.20-0.37 *p-value<0.05, **p-value<0.01 Journal of Health Informatics and Statistics (JHIS), Vol. 39, No. 2, 2014
86 조성인, 김병수, 전민수 Table 8. Correlation coefficient of regional characteristic variable with lift (municipal governments) KCD code drk smk dvc in come con sume aged a_ l_ h_ d_ rain A15-A19-0.08 0.18** 0.00 0.35** -0.06-0.11 0.05 0.15* -0.32** 0.35-0.56* C15-C26 0.41** -0.31** -0.30** -0.41** 0.34** 0.46** 0.03-0.47** 0.35** -0.66** 0.58* C30-C39-0.16*- 0.14* -0.09 0.45** -0.09-0.19** 0.12 0.29** -0.26** 0.58* -0.69** C51-C58 0.52** -0.13-0.02-0.68** -0.06 0.10-0.26** -0.29** 0.40** -0.55* 0.63** C64-C68 0.28** -0.21** -0.08-0.29** 0.00 0.15* -0.23** -0.32** 0.40** -0.56* 0.82** C81-C96 0.37** 0.32** -0.18** -0.53** -0.14* 0.04-0.39** -0.33** 0.44** -0.50* 0.64** E10-E14 0.32** 0.11 0.38** -0.43** -0.17** 0.01-0.37** -0.27** 0.34** -0.36 0.53* F00-F09-0.18** 0.07-0.02 0.17* 0.20** 0.10 0.28** 0.10-0.29** 0.50* -0.42 G30-G32 0.20** -0.29** -0.38** 0.14* 0.47** 0.35** 0.49** -0.02-0.33** 0.19-0.42 I10-I15-0.03 0.19** 0.04 0.15* 0.30** 0.25** 0.26** -0.08-0.27** -0.48 0.51* I20-I25 0.20** 0.21** -0.09 0.09 0.29** 0.27** 0.24** -0.11-0.17** -0.33-0.25 I30-I52 0.06 0.10-0.14* 0.11 0.47** 0.35** 0.50** -0.01-0.11 0.01-0.17 I60-I69-0.12 0.20** 0.15* 0.13* -0.06-0.10 0.05 0.15* -0.33** 0.27-0.50* J09-J18-0.26** 0.03-0.01 0.44** -0.13-0.21** 0.05 0.26** -0.18** 0.19-0.28 J40-J47-0.58** 0.12-0.03 0.69** 0.03-0.22** 0.41** 0.53** -0.49** 0.68** -0.68** K70-K77 0.29** 0.42** 0.33** -0.10 0.17* 0.22** 0.03-0.22** 0.07-0.46 0.16 N17-N19 0.19** -0.19** -0.09-0.34** 0.26** 0.33** 0.08-0.29** 0.06-0.17 0.17 R50-R69-0.34** -0.20** -0.08 0.35** -0.45** -0.44** -0.30** 0.25** 0.23** 0.21 0.24 R95-R99 0.10-0.17** 0.15* -0.30** -0.26* -0.12-0.42** -0.17** 0.27** -0.15 0.36 V01-V09-0.12-0.05 0.16* -0.21** 0.12 0.04 0.18** 0.09-0.18** 0.48-0.33 W00-W19 0.37** 0.05 0.08-0.38** 0.04 0.15* -0.12-0.25** 0.24** -0.62* 0.54* X60-X84 0.51** 0.09 0.29** -0.80** -0.24** -0.05-0.45** -0.27** 0.35** -0.33 0.54* Y10-Y34-0.44** -0.11-0.05 0.42** 0.13* 0.07 0.16* 0.04-0.25** 0.20-0.37 *p-value<0.05, **p-value<0.01 보건정보통계학회지제 39 권제 2 호
사망원인과지역의관련성에관한연구 87 0.2 0.6 1.0 1.4 0.4 0.8 1.2 0.0 1.0 2.0 0.5 1.5 2.5 0.5 1.0 1.5 2.0 0.4 0.8 1.2 0.5 1.5 2.5 3.5 0.6 1.0 1.4 1.8 0.2 0.6 1.0 1.4 0.0 1.0 2.0 35 45 55 65 12 14 16 18 20 35 45 55 65 0.5 1.0 1.5 2.0 0.5 1.5 2.5 3.5 Figure 1. Scatter-plot matrix of death cause with drk variable. 0.5 1.5 2.5 0.4 0.8 1.2 12 14 16 18 20 0.6 1.0 1.4 1.8 0.4 0.8 1.2 Figure 3. Scatter-plot matrix of death cause with h_ variable. 0.2 0.6 1.0 1.4 0.5 1.0 1.5 2.0 0.2 0.6 1.0 1.4 0.5 1.0 1.5 2.0 5 10 20 30 0.4 0.8 1.2 1.6 0.4 0.8 1.2 Figure 2. Scatter-plot matrix of death cause with aged variable. 0.4 0.8 1.2 0.4 0.8 1.2 1.6 5 10 20 30 0.4 0.8 1.2 1.6 0.2 0.6 1.0 1.4 0.2 0.6 1.0 1.4 0.4 0.8 1.2 1.6 10000 14000 18000 0.5 1.0 1.5 0.5 1.0 1.5 2.0 Figure 4. Scatter-plot matrix of death cause with rain variable. 0.5 1.0 1.5 2.0 0.5 1.0 1.5 10000 14000 18000 4. 결론 본논문에서우리나라사람들의사망원인이지역에따라차이가있는지에대해살펴보았다. 먼저리프트를통하여지역별사망원인의차이를살펴본결과, 신생물에해당하는지역별리프트의편차가가장작았고불명확하고미상의사 인과신경계통의기타퇴행성질환의리프트가지역별편차가가장크게나타났다. 불명확하고미상의사인과의도미확인사건은제주도에서해당사망원인으로사망할확률이다른지역에비해높았다. 불명확하고미상의사인의리프트는최고기온이낮은지역에서높았으며의도미확인사건은음주율이낮은지역과고령자비율 Journal of Health Informatics and Statistics (JHIS), Vol. 39, No. 2, 2014
88 조성인, 김병수, 전민수 이높은지역에서리프트가높게나타났다. 한국표준질병사인분류상에노쇠 ( 자연사 ) 가포함되어있는사망원인인전신증상및징후의사망확률은충청북도에서가장높고부산에서가장낮았으며구시군별로는인천강화군에서가장높고경남밀양시에서가장낮게나타났다. 그리고이사망원인과지역특성변수의상관관계로는평균기온, 최저기온, 최고기온이높은지역에서사망확률이낮다는것을알수있었다. 당뇨병과자살에대한리프트는경기도에서가장높았으며두사망원인모두음주율과유의한양의상관관계를보였고고령자비율과음의상관관계를보였다. 치매가속하는사망원인인증상성을포함하는기질성정신장애는전남화순군에서, 신경계통의기타퇴행성질환과뇌혈관질환은전북정읍시에서다른지역에비해사망할확률이높게나타났다. 허혈성심장질환과기타형태의심장병의리프트는시도단위로보았을때부산과경남지역에서높았다. 상관분석결과로는기온관련변수와양의관계를보였다. 호흡계통의질환을이루는인플루엔자및폐렴과만성하기도질환을보면각각충북과전북에서가장높은리프트를, 울산과서울에서가장낮은리프트를보이는데울산과서울은 1인당개인소득과민간소비가가장높은지역이다. 그리고호흡계통의질환에의한사망과개인소득, 민간소비, 강수량은음의상관관계를띄고있는것을알수있었다. 부산영도구에서간의질환으로사망할확률은부산영도구라는조건이없는경우에비해약 1.56배높게나타났다. 상관분석결과로는음주율, 흡연율, 조이혼율이높은지역일수록간의질환으로사망할리프트가높다는것을알수있었다. 본연구는몇가지한계를가지고있다. 자료를횡적으로만살펴보았으며시간의흐름에따른종적인연구도필요하다고본다. 리프트와상관계수를중심으로하는상관관계에대해주로 살펴보았으며인과관계에대한연구는이후의과제로남겨둔다. 리프트의특성상빈도의영향을많이받기때문에빈도가낮은사망원인에대해서는설명할수없었다는점과지역별특성을나타낼수있는변수가다양하지못했다는점도본논문의한계이다. 하지만본논문은특정지역과특정사망원인에국한되어있지않다는것과분류별사망원인들이어느지역에서두드러지는지를보여주었다는것으로의미가있다고본다. References [1] Kim D, Park H. Regional differentials in mortality in Korea, 1990-2000. Korea Journal of Population Studies 2003; 26(1): 1-30. (Korean) [2] Hong J, Lee S, Shin H. Variation in Life Expectancy by Area Deprivation : Magnitude and Trends in Korea, 2000-2011. Korean Association of Health and Medical Sociology 2013; 34: 273-298. (Korean) [3] Choi G. A study of the difference between crude failure and net failure probabilities by cause of death. Journal of the Korea Society of Health Informatics and Statistics 1999; 24(1): 1-6. (Korean) [4] Park K. Cause-specific mortality at the provincial level. Korea Journal of Population Studies 2003; 26(2): 1-32. (Korean) [5] Kim K, Noh M, Ha I. A study using HGLM on regional difference of the dead due to injuries. Journal of the Korean Data & Information Science Society 2011; 22(2): 137-148. (Korean) [6] Lee Y. A study on the causes of death in Kyungsan area [dissertation]. Kyungpook National University Graduate School of Public Health; 보건정보통계학회지제 39 권제 2 호
사망원인과지역의관련성에관한연구 89 Korea, 1995. [7] Park Y. Distributions of socioeconomic and health-related factors between regions with high and regions with low suicide mortality rates [dissertation]. Department of Public Health, the Graduate School of Public Health and Welfare Konyang University; Korea, 2013. [8] Jung K, Yun Y, Baek S, Jee S, Kim I. Smokingattributable mortality among Korean adults. Journal of the Korea Society of Health Informatics and Statistics 2013; 38(2): 36-48. (Korean) [9] Kim S. A study on how to improve the quality of causes of death statistics : focused on the death reporting system [dissertation]. Korea University Graduate School of Public Administration; Korea, 2013. [10] Park C. Order selection based on scaled lift. Journal of the Korean Data & Information Science Society 2011; 22(2): 227-234. (Korean) [11] Kim J. Social-environment factors by region for cause of death of elderly people in Korea. Journal of the Korea Gerontological Society 2005; 25(2): 1-14. (Korean) [12] Shin W. Logistic regression analysis of disease cause of death in 2008 [dissertation]. Ewha Womans University Graduate School; Korea, 2009. [13] Lee S. Social contextual effects on regional mortality and self-rated health status [dissertation]. Department of Public Health the Graduate School Yonsei University; Korea, 2002. [14] Jung J, Kim I, Lee D, Shin J, Kim B. Study on the vulnerability regarding high erature related mortality in Korea. Korean Geographical Society 2014; 49(2): 245-263. (Korean) Journal of Health Informatics and Statistics (JHIS), Vol. 39, No. 2, 2014