REVIEW ARTICLE ISSN (Print) ISSN (Online) X 대한간호학회지제 44 권제 6 호, 2014 년 12 월 J Korean Acad Nurs Vol.44 No.6,

REVIEW ARTICLE ISSN (Print) 2005-3673 ISSN (Online) 2093-758X 대한간호학회지제 44 권제 6 호, 2014 년 12 월 J Korean Acad Nurs Vol.44 No.6, 595-607 박성희 1 박유선 2 1 배재대학교간호학과, 2 고려대학교안산병원적정진료관리팀 Predictive Validity of the Braden Scale for Pressure Ulcer Risk: A Meta-analysis Park, Seong-Hi 1 Park, Yu-Sun 2 1 School of Nursing, Pai Chai University, Daejeon 2 Quality Improvement Team, Korea University Ansan Hospital, Ansan, Korea Purpose: The Braden Scale is one of the most intensively studied risk assessment scales used in identifying the risk of developing pressure sore. However, not all studies show that the predictive validity of this scale is sufficient. The purpose of this study was to evaluate the Braden Scale for predicting pressure ulcer development. Methods: Articles published 1946 and 2013 from periodicals indexed in Ovid Medline, Embase, CINAHL, KoreaMed, NDSL and other databases were selected, using the following keywords: pressure ulcer. The QUADAS-II was applied to assess the internal validity of the diagnostic studies. Selected studies were analyzed using meta-analysis with MetaDisc 1.4. Results: Thirty-eight diagnostic studies with high methodological quality, involving 17,934 patients, were included. Results of the meta-analysis showed that the pooled sensitivity and specificity of the Braden Scale were 0.74 (95% CI: 0.72-0.76), 0.75 (95% CI: 0.74-0.76) respectively. However the predictive validity of the Braden Scale has limitation because there was high heterogeneity between studies. Conclusion: The Braden Scale s predictive validity of risk for pressure ulcer is interpreted as at a moderate level. However there is a limitation to the interpretation of the results, because of high heterogeneity among the studies. Key words: Pressure sore, Sensitivity, Specificity, Meta-analysis 서론 1. 연구의필요성욕창예방은오늘날보건의료가당면한가장중요한도전과제중하나이다 [1]. 욕창은한부위에지속적인압력이가해질때그부위에순환장애가일어나조직이손상된상태로, 환자에게흔한잠재적실제적인간호문제이다 [2]. 욕창은재원기간과사망률및합병증발생률을증가시키며, 삶의질을감소시키고전세계적으로경제적손실 을야기시킨다 [3]. 국제의료기관인증기구인 The Joint Commission[4] 은미국의경우매해 2,500만명의욕창환자가발생되고환자당 4 만달러의의료비용이소요되며, 연간 6만명의환자가욕창관련합병증으로사망한다고보고하였다. 또한, 총 9개국이참여한국제욕창발생률연구에따르면 [5], 1989년 9.2% 였던욕창발생은 1999년 14.8%, 2005년 15.2% 로그증가폭이점점커졌으며, 2005년미국의의료기관별욕창발생률은병원이 14.6%, 장기급성치료시설 (long term acute care) 27.3%, 장기요양시설 14.4% 로분석되었다. 국내욕창발생률이나치료비용등에관한국가통계는체계적으 주요어 : 욕창, 민감도, 특이도, 메타분석 Address reprint requests to : Park, Yu-Sun Quality Improvement Team, Korea University Ansan Hospital, 123 Jeokgeum-ro, Danwon-gu, Ansan 425-707, Korea Tel: +82-31-412-5577 Fax: +82-31-412-6813 E-mail: ssun104@hotmail.com Received: July 28, 2014 Revised: August 21, 2014 Accepted: November 10, 2014 This is an Open Access article distributed under the terms of the Creative Commons Attribution NoDerivs License. (http://creativecommons.org/licenses/by-nd/4.0) If the original work is properly cited and retained without any modification or reproduction, it can be used and re-distributed in any format and medium. 2014 Korean Society of Nursing Science

596 박성희 박유선 로조사되어있지않으나 [6] 몇몇연구에서입원한경우는 11.0%, 중환자실에서는 21.7-45.5%[7], 노인요양시설거주자는 9.8%[8], 요양병원입원환자는 7.0% 로보고되고있다 [9]. 특히, 우리나라의경우노인인구와빠르게확산되는장기요양시설, 보호자개입감소등의사회적촉진요인을갖고있어욕창예방은더욱시급히해결해야할간호문제라고해석된다. 욕창은움직임이자유롭지않은모든환자에서발생될가능성이있으므로병원이나장기요양시설에입원하는경우표준화된욕창위험사정도구를이용하여욕창위험도를평가하고, 이에따라욕창발생이예상되는위험군을선별하여욕창예방중재를시행해야한다. 따라서, 이는주로국가별의료기관인증기준에포함되어그시행이강조되고있으며, 우리나라에서도예외는아니다 [10]. 현재사용되는욕창위험사정도구는 1962년영국에서개발된 Norton 도구, 1973년 Gosnell 도구, 1985년 Waterlow 도구, 1991 년중환자실환자를위한 Cubbin과 Jackson 도구가있지만국제적으로가장많이활용되는것은 1987년미국에서장기요양시설노인을대상으로개발된 Braden 도구이다 [6,11,12]. 도구개발당시 Bergstrom 등 [12] 의연구팀은그동안 Norton 및 Gosnell 도구가사용되어왔지만도구에대한신뢰도와타당도가검증된바가없어욕창발생이예상되는환자가과다추정되고, 욕창이발생된환자도잘못분류하고있다고지적하였다. Braden 도구는욕창발생에가장중요한요소를외적압력 (pressure) 과조직의내성 (tolerance) 으로보고, 외부에서압력을높이는요인으로동작제한, 활동부족, 감각인지장애를들었으며조직내성에관련된요소는외적요인으로는습기, 마찰및전단력 (shear) 을, 내적요인으로는영양상태, 연령, 세동맥압 (arteriolar pressure) 을포함하였다 [11,12]. 그러나이렇게개발된 Braden 도구도이후다른연구자들에의해재검증되는과정에서어떤환자를대상으로하느냐에따라민감도와특이도에차이를보였으며 [13], 경계점수 (cut-off point) 도다양하게보고되었다 [3]. 욕창위험사정도구의예측타당도에대한영향요인으로인종 [3,12], 성별 [7,12], 연령 [6,11], 입원장소 [14], 피부상태 [6,7], 영양 [13], 재원기간 [15,16], 간호사의감별능력 [9] 등을그원인으로언급되고있으나일관적이지는않았으며, 영향요인으로언급되었더라도연구에따라다른결과를보였다. 따라서, 기존의다른도구들과비교하는몇몇연구에서현재사용되는욕창위험사정도구들이어느정도욕창발생을예측할수있지만이도구들을임상에서표준적일상도구로사용하는것은비효율적이라고지적되었다 [14-17]. 이에본연구에서는국내에서가장많이사용되고있는 Braden 도구의진단법평가연구결과를중심으로한메타분석을통해욕창발생여부를감별하는 Braden 도구의예측타당도에대해확인하고, 이에영향을미치는요인을분석함으로간호사정선별도구로서 과학적인근거를제시하고자한다. 연구방법 1. 연구설계본연구는 Braden 욕창위험사정도구의욕창발생예측의진단정확도를보고한연구에대한메타분석연구이다. 2. 문헌검색전략본연구는코크란연합의진단법정확도에대한체계적문헌고찰핸드북 [18] 및 PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 그룹이제시한체계적문헌고찰보고지침 [19] 에따라수행되었다. 1) 문헌검토기준 (criteria for considering studies for this review) (1) 연구유형 (type of studies) Braden 욕창위험사정도구를사용하여욕창발생예측의정확성을평가한연구를포함하였다. 그러나연구결과에서민감도, 특이도만제시하고 2 2 이원분류표를작성하는데충분한진단정확도자료를제공하지않은연구는제외하였다. (2) 연구대상 (type of participants) 병원이나장기요양시설에입원한환자를대상으로하였다. (3) 시험진단법 (index test) Braden 욕창위험사정도구를대상으로하였다. 욕창발생위험사정의판단은각연구에서연구자가제시한경계점수를기준으로하였다. (4) 참조표준검사 (reference standards) 욕창발생은미국및유럽욕창협의회 (National Pressure Ulcer Advisory Panel [NPUAP], European Pressure Ulcer Advisory Panel [EPUAP]) 나 AHCPR (Agency for Health Care Policy and Research [AHCPR]) 의전문가그룹과개별학자 (Bergstrom, Lowthian, Lyder) 에의해개발된피부상태평가척도나 International Classification of Diseases-9 (ICD-9) 에따른진단명을통해확인하였다. (5) 진단정확도결과 (types of outcome measures) 연구결과를토대로진양성 (true positive), 위양성 (false positive), 위

597 음성 (false negative), 진음성 (true negative) 값을기술하고, 이를토대로민감도, 특이도, 양성및음성우도비, 진단교차비를비교하였다. 2) 문헌검색방법 (search methods for identification of studies) (1) 전자데이터베이스검색 (electronic searches) 전자데이터베이스 (DB) 를이용한검색은 2013 년 10월 1일에시행되었다. 국내는 KoreaMed 와국가과학기술정보센터 (NDSL) 및한국교육학술정보원 (KERIS) 을, 국외는 Ovid-Medline (1946년 1월부터 2013 년 10월 ), Embase (1947 년 1월부터 2013 년 10월 ) 와 Cochrane Library 및 CINAHL Plus with Full Text를이용하였다. 된문헌이었다. 원저가아니거나참조표준검사와비교한진단법평가연구가아니거나진단정확도를계산할수없는연구및수지등부분피부궤양에대한연구등은배제하였다. (2) 문헌선택과정 (selection of studies) 검색된문헌에서일차적으로중복된문헌을제거하고, 문헌의제목및초록을검토하였으며, 정확한판단이곤란한경우는원문을찾아선택및배제기준을적용하여문헌을선별하였다. 이과정은 2 명의저자가각자독립적으로검토한후합의를통해결정하였으며, 의견의불일치가있는경우는제3자개입의원칙을정하였다. (2) 기타문헌검색원 (searching other resources) 욕창관련분야의학술지가빠짐없이포함될수있도록한국간호과학회, 기본간호학회, 성인간호학회, 지역사회간호학회, 노인간호학회, 임상간호연구, 한국재활간호학회와대한가정의학회, 대한재활의학회의홈페이지를통해학술지검색을추가하였다. 학술지검색범위는한국간호과학회의대한간호학회지는 1970년, 기본간호학회지 1994년, 성인간호학회지과지역사회간호학회지는 1989년, 노인간호학회지 1999년, 임상간호연구는 1995년, 한국재활간호학회지 1998년, 대한가정의학회지 1980년, 대한재활의학회지는 1977년부터 2013 년 10월 7일검색당일학회홈페이지에게시된전체문헌을대상으로하였다. (3) 검색어 (search terms) 핵심질문의구성요소인연구대상및진단법으로부터주요개념어를도출하여검색전략을구성하였다. 국내 DB는 욕창 과 욕창도구, 욕창사정 및 pressure ulcer risk assessment, decubitus ulcer risk assessment 를이용하였다. 국외 DB는효율적검색을위해욕창관련의학주제용어 (Medical Subject Heading [MeSH]) 및주요개념어를도출하여검색전략을구성하였다. 사용된검색어는 pressure ulcer, decubitus ulcer, skin ulcer, bed sore 와 risk assessment 이었으며, 각단어가사용되는형태를검토하여절단검색, 만능문자및인접연산자를활용하였다. 진단법평가연구에대한검색필터 (search filter) 는 SIGN (Scottish Intercollegiate Guidelines Network) 에서제시한검색전략을활용하였다. 3) 자료수집및분석 (data collection and analysis) (1) 문헌선정기준 (selection and exclusion criteria) 문헌선택은입원환자를대상으로 Braden 욕창위험사정도구를이용하여진양성, 위양성, 진음성, 위음성결과를보고한진단법평가연구로, 동료심사 (peer-reviewed) 를거친한국어와영어로출판 (3) 문헌의질평가문헌의질은 2명의저자에의해 QUADAS-II (Quality Assessment of Diagnostic Accuracy Studies-II) 도구를활용하여독립적으로평가하였다 [20]. QUADAS-II 는진단법평가연구에대한질평가도구로 편중의위험 (risk of bias) 과 적용 (applicability) 을평가하게되며, 평가항목은대상자선택, 시험진단법, 참고표준검사와과정및적시성 (flow and timing) 의 4가지영역 (domain) 으로구성되어있다. (4) 자료추출및분석 (data extraction and management) 자료추출은우선근거표 (evidence table) 기본서식을작성하여시범적으로서식의적절성을검토한후사용되었다. 이역시두명의저자가독립적으로작성한후그결과를서로교차확인하였다. 선택된문헌의연구유형, 연구장소, 연구기관수, 연구대상의성별분포, 연령, 재원기간, Braden 도구의경계점수와아울러욕창위험사정도구의진단결과인진양성, 위양성, 위음성, 진음성값을기술하였다. 이를토대로 2 2 이원분류표로작성하였고, 2-way Contingency Table Analysis 프로그램을이용하여민감도, 특이도, 양성우도비, 음성우도비및진단정확도와각항목의 95% 신뢰구간 (95% confidence interval [95% CI]) 을재계산하였다. 3. 통계적분석과자료통합 MetaDiSc 1.4 프로그램을이용하여메타분석하였다. 진단법메타분석에서통합추정치를분석하고자하는경우, 연구간이질성을반영하기위해랜덤효과모형사용을권고하므로 [18], 통계적모델에대한일반적원칙은임의효과모형 (random effect model) 을토대로하였으며, 통합민감도와특이도, 양성및음성우도비와진단교차비및 sroc 곡선 (symmetric summary receiver-operating characteristic [sroc] curve) 으로분석하였다. sroc 곡선의통계량은곡선아래면적 (Area under the curve [AUC]) 과 index Q * 값을통해검사정확

598 박성희 박유선 도를기술하였다. AUC의수치는 AUC= 0.5인경우비정보적인검사, 0. 5 < AUC 0.7은덜정확한, 0.7 < AUC 0.9는중등도, 0.9<AUC<1 는매우정확한및 AUC=1은완벽한검사로판단하였고 [21], ROC 곡선에서민감도와특이도의동등점을반영하는 index Q * 는 100% 정확할때이를 1 을기준으로판단하였다 [22]. 연구들간이질성 (heterogeneity) 의존재여부는우선숲그림을통해신뢰구간및효과추정치에공통적인부분이있는지시각적으로확인하였다. 또한, 유의수준 5% 미만으로하여 Higgins과 Thompson[23] 의 I 2 동질성검사 (I 2 test) 로평가하였으며, I2의판단기준은 I 2 25% 이면이질성이낮은것으로, 25%<I 2 75% 는중간정도의이질성이있는것으로, I 2 >75% 이상은이질성이있다고판단하였다. 연구결과 1. 문헌선택결과전자데이터베이스를통해총 492개문헌이검색되었다. 국내의경우 KoreaMed 13개, NDSL 8개, KERIS 37개였고, 국외의경우는 Ovid-Medline 230개, Embase 130개, Cochrane Library 12개및 CI- NAHL 62개였다. 관련분야학술지에서는총 27개문헌이검색되었다. 대한간호학회지 8개, 성인간호학회지 9개, 임상간호연구 8개, 지역사회간호학회지와노인간호학회지에서각각 1개씩검색되었고, 기본간호학회지와한국재활간호학회지, 대한가정의학회지및대한재활의학회지에서는욕창관련문헌이검색되지않았다. 우선, 검색된총 519개문헌중중복된문헌은 165개였다. 따라서, 이를제외한 354개문헌을토대로제목및초록을검토하여핵심질문을중심으로구성된선택및배제기준에따라선정하였으며, 62 개문헌은원문을찾아선별하였다. 최종적으로 316 개 (89.3%) 문헌이배제되고, 38개문헌이최종선택되었다. 자세한문헌선택과정은흐름도로제시하였다 (Figure 1). 2. 문헌의질평가결과최종선택된 38편의문헌의질을평가한결과, 전체적으로각영역에대해편중의위험이 높음 으로평가된문헌은하나도없었다. 모든영역과항목에서편중의위험이없다고평가된연구는 9편이었으며 [a1-a9] 일부연구에서평가항목별편중의위험을우려하였으나모두불확실 (unclear) 로판단되었다. 따라서, 선택된문헌은모두질평가영역을충족하는고품질의문헌이었다. 3. 선택된문헌의일반적특성 Braden 욕창위험사정도구의예측타당도를보고한문헌은총 38편으로총대상자는 17,934 명이었다. Braden 도구는개발이후진단정확도를높이기위해일부연구자에의해수정되었는데본연구에수정 Braden 도구를사용한 6편의연구가포함되었다. 수정 Braden 도구는 Pang과 Wong[a10] 에의해보완된 2편 [a11,a12] 의연구와 Song과 Choi[24] 에의해수정개발된 4편의연구 [a13-a16] 에서보고되었다. 선택된문헌들은미국이 14편 (36.8%) 으로가장많았고 [a1-a4, a7, a9,a17-a24], 한국이 10편 (26.3%)[a13-a16,a25-a30], 캐나다가 4편 Identification 492 of records identified through electronic database searching (KoreaMed, NDSL, KERIS, Ovid-Medline, Embase, Cochrane and CINAHL) 27 additional records identified through related Korean journal sources Abstract screened (n=519) in duplicate by two reviewers 165 of record duplicates removed Screening 354 of records screened by abstract based on inclusion and exclusion criteria by two reviewers independently Total of 292 records excluded as follows; Pressure ulcer risk assessment tool was not Index test= 218 Not Braden scale= 24 Not designed diagnostic test accuracy= 24 Non original articles= 10 Irrelevant outcomes= 8 Others= 8 Eligibility Included Figure 1. Flow diagram of article selection. 62 of full-text articles assessed for eligibility by two reviewers independently 38 studies included in quantitative synthesis (meta-analysis) Total 24 of records excluded as follows; Not designed diagnostic test accuracy= 6 Irrelevant outcomes= 14 Others= 4

599 [a5,a8,a31,a32], 홍콩 [a10,a11] 과브라질 [a33,a34] 이각각 2편이었으며, 네덜란드 [a6], 영국 [a35], 벨기에 [a36], 인도네시아 [a37], 이란 [a38], 중국 [a12] 이각각 1편으로전세계적으로활발한연구경향을나타냈다. 연구대상은대부분성인이었으며, 평균연령이 30대 [a1] 와 40대 [a37] 인연구는각각 1편이었고 50대가 8편 [a3,a8,a12-a14,a25,a26,a29] 60대가 12편 [a6,a9,a15,a16,a21,a22,a28,a30-a33,a38], 70대가각각 11편 [a2,a5,a7,a11,a17-a20,a24,a27,a34] 이었으며 80대도 2편 [a35,a36] 있었다. 100명이상의대규모연구가 29편 (76.3%) 이었고, 이중천명이넘는경우도 5편 [a19,a20,a25,a29,a36] 이었다. 참조표준기준으로는 NPUAP 가 15편의연구 [a2,a4,a7,a8,a11,a12,a21,a22,a26-a28,a30,a32,a33,a37] 에서가장많이이용되었고, AHCPR 도구는 7편 [a14-a16,a25,a30,a32,a38] 에서, 개별학자들에의해개발된피부사정도구는주로 1900년대초발표된연구 [12,25] 에서활용되었다. 또한, ICD-9 코드를참조표준기준으로비교한연구는모두후향적연구 [a17,a19,a20] 이었다. Braden 도구의경계점수는연구자가가장효과적이라고제시한경우를사용하였는데 16점이 13편 [a3,a7,a8,a11,a13,a18,a23,a24,a26,a30,a31,a35,a 38] 으로가장많았고 18점이 11편 [a2,a4,a10,a17,a19,a21,a22,a25,a27,a2 9,a36], 19점이상이 4편 [a5,a6,a20,a32] 이었으며 15점이하는 5편 [a1,a14,a28,a33,a37] 에서보고되었다 (Table 1). 4. Braden 욕창위험사정도구의예측타당도평가 Braden 도구의예측타당도평가결과는선택된문헌을대상으로통합민감도와특이도, 양성및음성우도비와진단교차비, sroc 통계량을종합적으로기술하였다. 문헌들간이질성의원인을찾아내기위한세부분석은충분한문헌이선택된 Braden 도구만을대상으로연구장소, 성별, 평균연령, 입원장소및참조표준기준등의특성에따라시행하였다 (Table 2). 1) Braden 도구별예측타당도 Braden 도구의예측타당도는총 36편 ( 총대상자 17,611 명 ) 의문헌에서보고되었다. 메타분석결과, 통합민감도는 0.74 (95% CI: 0.72-0.76) 였으며 (Figure 2-A), 통합특이도는 0.75 (95% CI: 0.74-0.76) 로분석되었다 (Figure 2-B). 문헌들간이질성은각각 77.8% (χ 2 =157.38, p<.001), 98.7% (χ 2 =2716.10, p<.001) 로모두높은수준을보였다. 통합양성우도비는 3.13 (95% CI: 2.47-3.97), 통합음성우도비는 0.36 (95% CI: 0.30-0.43), 통합진단교차비는 10.23 (95% CI: 6.94-15.07) 이었고 sroc AUC 는 0.83 (SE= 0.02), Q * 값은 0.77 (SE= 0.02) 이었다. Song과 Choi의수정 Braden 도구는 4편 ( 총대상자 688명 ) 의연구에서보고되었다 [a13-16]. 메타분석결과, 통합민감도는 0.97 (95% CI: 0.92-0.99) 였으며, 문헌들간이질성은 59.8% (χ 2 =7.47, p=.06) 로중 간수준이었고 (Figure 2-C), 통합특이도는 0.70 (95% CI: 0.66-0.73), 문헌들간이질성은각각 97.7% (χ 2 =128.15, p<.001) 로높은수준을보였다 (Figure 2-D). 통합양성우도비는 3.47 (95% CI: 1.33-9.06), 통합음성우도비는 0.08 (95% CI: 0.04-0.20), 통합진단교차비는 56.56 (95% CI: 21.88-146.1) 이었고 sroc AUC는 0.95 (SE=0.02), Q * 값은 0.90 (SE= ) 이었다. Pang과 Wong 의수정 Braden 도구는 2편의연구 ( 총대상자 626명 ) 에서보고되었다 [a11,a12]. 메타분석결과, 통합민감도는 0.89 (95% CI: 0.71-0.98) 였으며, 문헌들간이질성은없었다 (0.0%, χ 2 = 0.00, p =1.000) (Figure 2-E). 통합특이도는 0.71 (95% CI: 0.67-0.75) 이었고, 문헌들간이질성은 90.0% (χ 2 =1, p<.001) 로높은수준을보였다 (Figure 2-F). 통합양성우도비는 2.87 (95% CI: 1.88-4.38), 통합음성우도비는 0.17 (95% CI: 0.06-0.49), 통합진단교차비는 16.06 (95% CI: 4.75-54.35) 이었다. 2) Braden 도구의예측타당도에영향하는요인분석 (1) 연구장소별미국에서연구된 14편 [a1-a4,a7,a9,a17-a24] 문헌의메타분석결과, 통합민감도는 0.69 (95% CI: 0.65-0.73), 통합특이도는 0.67 (95% CI: 0.66-0.68) 이었고, 문헌들간이질성은각각 77.8% (χ 2 =58.60, p<.001), 96.0% (χ 2 =58.60, p<.001) 로높았으며, sroc AUC는 0.80 (SE= 0.04), Q * 값은 0.74 (SE= ) 였다. 한국에서연구된 8편 [a13,a16,a25-a30] 문헌의메타분석결과, 통합민감도는 0.87 (95% CI: 0.83-0.91), 문헌들간이질성은 39.7% (χ 2 =11.60, p =.114) 였다. 통합특이도는 0.87 (95% CI: 0.86-0.88), 문헌들간이질성은 99.4% (χ 2 =1107.05, p<.001) 였으며, sroc AUC는 0.92 (SE= ), Q * 값은 0.85 (SE= ) 였다. 캐나다문헌 4편 [a5,a8,a31,a32] 의메타분석결과, 통합민감도는 0.69 (95% CI: 0.64-0.74) 였고문헌들간이질성은 0.0% (χ 2 =2.42, p=.490) 로전혀없었으나통합특이도는 0.66 (95% CI: 0.63-0.69) 이었으며, 문헌들간이질성은 92.8% (χ 2 = 41.70, p <.001) 로높았다. sroc AUC는 0.80 (SE= 0.04), Q * 값은 0.73 (SE= 0.04) 이었다. (2) 성비연구대상자의성비특성에따라세부분석하였다. 연구대상자의남성비가높은문헌은총 11편 [a12-a14,a17,a18,a21,a25,a26,a28,a33,a37] 이었다. 메타분석결과, 통합민감도는 0.83 (95% CI: 0.80-0.87) 이었고문헌들간이질성은 58.8% (χ 2 = 24.28, p =.006), 통합특이도는 0.84 (95% CI: 0.83-0.85), 문헌들간이질성은 98.9% (χ 2 = 933.55, p <.001) 로높았다. sroc AUC는 0.89 (SE = ), Q* 값은 0.82 (SE = ) 였다. 여성의성비가높았던연구는총 9편 [a2,a11,a19- a21,a24,a27,a35,a36] 으로, 통합민감도는 0.73 (95% CI: 0.69-0.76), 통합

600 박성희 박유선 Table 1. Characteristics of Selected Studies Year* Authors Location Braden Scale Types of studies Setting 2012 Cowan et al. USA R ICU & at adm. (±) 2011 Kim et al. Korea P LTCF (±) 2011 Serpa et al. Brazil P ICU 2010 Cho & Noh Korea R ICU & 2010 de Souza et al. Brazil P LTCF 2009 Chan et al. Hong Kong P OS 2009 Kim et al. Korea P SICU 2007 Oh et al. Korea P ICU & 2006 Suriadi et al. Indonesia 2005 Defloor & Grypdonck P ICU Belgium P LTCF 2005 Jalali & Rezaie Iran P ICU & 2005 Kwong et al. China P Wards 2004 Cho et al. Korea P Wards (±) 2004 Jun et al. Korea P ICU (±) Participants Reference standards Age (year) (M±SD) M:F Total (year) 71.0±10.6 206:7 213 ICD-9 codes 71.9±8.2 30:73 103 NPUAP (2007) 60.9±16.5 48:24 72 NPUAP (2007) 59.4-62.3 433:282 715 NPUAP (2007) 76.6±9.2 104:129 233 NPUAP (2007) 79.4±10.9 30:167 197 NPUAP (2007) 58.1±1.2 145:74 219 AHCPR Cut off point 2 2 Table Value (95% confidence interval) TP FP FN TN SN SP PLR NLR DOR < 18 65 34 5 79 0.93 (0.85-0.97) < 18 24 30 2 47 0.92 (0.76-0.99) < 13 6 11 2 53 0.75 (0.38-0.95) < 13 32 355 10 318 0.76 (0.61-0.87) < 17 27 48 10 148 0.73 (0.57-0.85) < 16 12 64 6 115 0.67 (0.42-0.85) < 14 37 54 3 125 0.93 (0.80-0.98) 51.1 997:885 1,882 - < 18 4 114 0 1,764 1.00 (0.40-1.00) 47.5-50.9 72:33 105 NPUAP 84.6±7.9 369:1,403 1,772 EPUAP (1999) 60.0 100:130 230 AHCPR 54.1±16.9 253:176 429 NPUAP 55.3±15.8 2339:1555 3,894 AHCPR 62.0 64:48 112 AHCPR < 11 23 15 12 55 0.66 (0.51-0.78) < 18 160 724 27 861 0.86 (0.80-0.90) < 16 39 0 35 156 0.53 (0.47-0.53) < 15 8 118 1 302 0.89 (0.52-0.99) < 18 94 248 15 3524 0.86 (0.78-0.92) < 16 34 57 1 20 0.97 (0.86-1.00) 0.70 (0.65-0.73) 0.61 (0.55-0.63) 0.83 (0.78-0.85) 0.47 (0.46-0.48) 0.76 (0.73-0.78) 0.64 (0.62-0.66) 0.70 (0.67-0.71) 0.94 (0.94-0.94) 0.79 (0.71-0.85) 0.54 (0.54-0.55) 1.00 (0.97-1.00) 0.72 (0.71-0.72) 0.93 (0.93-0.94) 0.26 (0.21-0.27) 3.09 (2.43-3.55) 2.37 (1.70-2.68) 4.36 (1.73-6.52) 1.44 (1.13-1.68) 2.98 (2.09-3.83) 1.87 (1.10-2.52) 3.07 (2.41-3.39) 16.47 (6.41-16.47) 3.07 (1.79-5.06) 1.87 (1.72-2.00) 0.10 (0.04-0.23) 0.13 (0.02-0.44) 0.30 (0.05-0.80) 0.50 (0.27-0.85) 0.36 (0.19-0.59) 0.52 (0.22-0.94) 0.11 (-0.30) 0.00 (0.00-6.42) 0.44 (0.26-0.68) 0.27 (0.18-0.38) - 0.47 (0.47-0.54) 3.16 (1.77-3.57) 13.12 (11.51-14.31) 1.31 (1.08-1.37) 0.16 (0.01-0.69) 0.15 (0.09-0.23) 0.11 (0.01-0.69) 30.21 (10.46-93.91) 18.8 (3.86-124.44) 14.46 (2.17-121.64) 2.87 (1.33-6.34) 8.33 (3.54-19.98) 3.59 (1.18-11.37) 28.55 (7.96-121.67) - 7.03 (2.62-19.30) 7.05 (4.55-10.97) - 20.48 (2.57-441.31) 89.05 (49.51-162.59) 11.93 (1.57-249.20) *Year of publication; =Pressure ulcer; TP=True positive; FP=False positive; FN=False negative; TN=True negative; SN=Sensitivity; SP=Specificity; PLR=Positive likelihood ratio; NLR=Negative likelihood ratio; DOR=Diagnosis odds ratio; R=Retrospective; P=Prospective; ICU=Intensive care unit; OS=Orthopedic surgery; SICU=Surgical intensive care unit; ICD-9=International Classification of Diseases-9; LTCF=Long term care facilities; NPUAP=National Pressure Ulcer Advisory Panel; AHCPR=Agency for Health Care Policy and Research; TDCPS=Torrence Developmental Classification of Pressure Sore.

601 Table 1. Characteristics of Selected Studies (Continued) Year* Authors Location Types of studies Setting at adm. 2003 Lee Korea P NSICU Participants Reference standards Age (year) (M±SD) M:F Total (year) 54.1 48:18 66 NPUAP 2003 Marrie et al. Canada R Wards 61.0±18.0 98:90 188 NPUAP (2002) 2002 Bergstrom & Braden USA P Wards LTCF 2001 Bergquist USA R LTCF 2001 Bergquist & Frantz 2000 Halfens et al. Netherlands 1998 Baldwin & Ziegler 1998 Goodridge et al. USA R LTCF P Wards USA P Trauma center Canada P GW & LTCF 1998 Lyder et al. USA P Wards 1998 Olson et al. Canada P Wards 1998 Pang & Wong Hong Kong 1998 Schue & Langemo P OS USA R RM 1997 Watkinson UK P Wards 1996 Capobianco & McDonald USA P Wards 1996 Harrison et al. Canada P Wards 58.1-63.2 825 NPUAP (1998) 76.4±8.6 633:1,051 1,684 ICD-9 codes 78.8-76.3 641:1,070 1,696 ICD-9 codes Cut off point 2 2 Table Value (95% confidence interval) TP FP FN TN SN SP PLR NLR DOR <16 26 14 4 22 0.87 (0.73-0.95) 16 35 33 11 109 0.76 (0.63-0.86) <18 76 166 32 551 0.70 (0.62-0.78) <18 64 725 43 852 0.60 (0.50-0.69) <19 66 508 42 1080 0.61 (0.52-0.70) 60.9±18.3 167:153 320 - <20 34 82 13 191 0.72 (0.58-0.84) 31.7±10.9 16:10 36 - <10 10 1 1 24 0.91 (0.65-0.99) 78.6±8.5-330 - <19 22 134 10 164 0.69 (0.51-0.83) 71.0±6.5 15:21 36 NPUAP 16 5 0 9 22 0.36 (0.18-0.36) 54.8-62.4-128 NPUAP <16 9 19 2 98 0.82 (0.50-0.97) 45-92 52:54 106 TDCPS (1983) 69.2±10.9 170:0 170 NPUAP 82.7 24:68 92 Lowthian (1987) 66.9±19.3 18:32 50 NPUAP 60.0±19.0 376:362 738 AHCPR (1992) <18 19 32 2 53 0.91 (0.70-0.98) <18 33 50 13 74 0.72 (0.59-0.83) <16 14 18 1 59 0.93 (0.69-1.00) 18 10 6 4 30 0.71 (0.47-0.89) <19 147 176 72 343 0.67 (0.62-0.72) 0.61 (0.49-0.68) 0.76 (0.73-0.80) 0.77 (0.76-0.78) 0.54 (0.53-0.55) 0.68 (0.67-0.69) 0.70 (0.68-0.72) 0.96 (0.85-1.00) 0.55 (0.53-0.57) 1.00 (0.86-1.00) 0.84 (0.81-0.85) 0.62 (0.57-0.64) 0.60 (0.55-0.64) 0.77 (0.72-0.78) 0.83 (0.74-0.90) 0.66 (0.64-0.68) 2.23 (1.44-3.01) 3.27 (2.30-4.31) 3.04 (2.51-3.55) 1.30 (1.08-1.52) 1.91 (1.58-2.23) 2.41 (1.80-2.97) 22.73 (4.26-279.82) 1.53 (1.08-1.91) 0.22 (0.07-0.55) 0.31 (0.17-0.51) 0.39 (0.28-0.51) 0.74 (0.57-0.93) 0.57 (0.44-0.72) 0.40 (0.23-0.62) 0.10 (0.01-0.41) 0.57 (0.30-0.93) - 0.64 (0.64-0.93) 5.04 (2.58-6.52) 2.40 (1.65-2.75) 1.78 (1.29-2.28) 3.99 (2.44-4.50) 4.29 (1.77-8.96) 1.98 (1.70-2.28) 0.22 (0.04-0.62) 0.15 (-0.52) 0.47 (0.27-0.76) 0.09 (0.00-0.44) 0.34 (0.12-0.73) 0.50 (0.40-0.60) 10.21 (2.60-43.90) 10.51 (4.53-24.87) 7.88 (4.93-12.66) 1.75 (1.15-2.66) 3.34 (2.20-5.09) 6.09 (2.92-12.89) 240.00 (10.37-24609.34) 2.69 (1.17-6.34) - 23.21 (4.15-170.28) 15.73 (3.19-105.01) 3.76 (1.70-8.39) 45.89 (5.57-1000.60) 12.50 (2.43-72.06) 3.98 (2.81-5.65) *Year of publication; =Pressure ulcer; TP=True positive; FP=False positive; FN=False negative; TN=True negative; SN=Sensitivity; SP=Specificity; PLR=Positive likelihood ratio; NLR=Negative likelihood ratio; DOR=Diagnosis odds ratio; R=Retrospective; P=Prospective; ICU=Intensive care unit; OS=Orthopedic surgery; SICU=Surgical intensive care unit; ICD-9=International Classification of Diseases-9; LTCF=Long term care facilities; NPUAP=National Pressure Ulcer Advisory Panel; AHCPR=Agency for Health Care Policy and Research; TDCPS=Torrence Developmental Classification of Pressure Sore.

602 박성희 박유선 Table 1. Characteristics of Selected Studies (Continued) Year* Authors Location Types of studies Setting at adm. Participants Reference standards Age (year) (M± SD) M:F Total (year) Cut off point 2 2 Table Value (95% confidence interval) TP FP FN TN SN SP PLR NLR DOR 1996 VandenBosch et al. 1994 Braden & Bergstrom 1993 Barnes & Payton 1992 Salvadalena et al. USA P ICU & USA P LTCF USA P Wards USA P Wards 1991 Choi & Song Korea P Wards 1987 Bergstrom et al. 1987 Bergstrom et al. USA P LTCF USA P ICU Modified Braden Scale by Song & Choi 2009 Kim et al. Korea P SICU 2004 Kim & Choi Korea P NSICU & ward 2003 Lee at al. Korea P ICU 1991 Choi & Song Korea P Wards Modified Braden Scale by Pang & Wong 2009 Chan et al. Hong Kong P OS 2005 Kwong et al. China P Wards 62.4-67.0 49:54 103 Bergstrom 76.1± 9.7 29:73 102 NPUAP 50-90 183:178 361 Lyder (1991) 72±13 34:63 99 Bergstrom (1987) 54.1 89:57 146 Bergstrom (1987) <17 17 30 12 44 0.59 (0.42-0.74) <18 22 19 6 55 0.79 (0.62-0.90) <16 16 32 6 307 0.73 (0.51-0.88) <16 8 24 12 55 0.40 (0.21-0.61) <16 13 8 3 122 0.81 (0.57-0.95) 75.0±20.0 85:14 99 - <16 7 9 0 83 1.00 (0.59-1.00) 58.5±14.5 28:32 60 Bergstrom (1987) 58.1±1.2 145:74 219 AHCPR 60.7-211 AHCPR (1996) 62.0 64:48 112 AHCPR 54.1 89:57 146 Bergstrom (1987) 79.4±10.9 30:167 197 NPUAP (2007) 54.1±16.9 253:176 429 NPUAP <16 20 13 4 23 0.83 (0.66-0.94) <21 38 55 2 124 0.95 (0.83-0.99) <23 34 42 0 135 1.00 (0.88-1.00) <24 35 63 0 14 1.00 (0.90-1.00) <24 14 11 2 119 0.88 (0.64-0.98) <19 16 68 2 111 0.89 (0.65-0.98) <14 8 105 1 315 0.89 (0.51-0.99) 0.60 (0.53-0.66) 0.74 (0.68-0.79) 0.91 (0.89-0.92) 0.70 (0.65-0.75) 0.94 (0.91-0.96) 0.90 (0.87-0.90) 0.64 (0.53-0.71) 0.69 (0.67-0.70) 0.76 (0.74-0.76) 0.18 (0.14-0.18) 0.92 (0.89-0.93) 0.62 (0.60-0.63) 0.75 (0.74-0.75) 1.44 (0.88-2.15) 3.06 (1.92-4.26) 7.71 (4.72-10.42) 1.32 (0.60-2.44) 13.20 (6.31-21.03) 10.22 (4.58-10.22) 2.31 (1.40-3.25) 3.09 (2.48-3.33) 4.21 (3.39-4.21) 1.22 (1.04-1.22) 10.34 (5.57-13.58) 2.34 (1.61-2.65) 3.56 (1.98-4.01) 0.70 (0.39-1.11) 0.29 (0.12-0.57) 0.30 (0.13-0.55) 0.86 (0.52-1.22) 0.20 (0.06-0.47) 0.00 (0.00-0.47) 0.26 (0.08-0.64) 0.07 (0.01-0.26) 0.00 (0.00-0.16) 0.00 (0.00-0.73) 0.14 (0.02-0.41) 0.18 (-0.59) 0.15 (0.01-0.66) 2.08 (0.80-5.46) 10.61 (2.40-34.77) 25.58 (8.60-79.55) 1.53 (0.49-4.70) 66.08 (13.46-376.46) - 8.85 (2.18-39.15) 42.84 (9.59-266.55) - - 75.73 (13.54-559.83) 13.06 (2.75-84.99) 24.00 (3.01-517.73) *Year of publication; =Pressure ulcer; TP=True positive; FP=False positive; FN=False negative; TN=True negative; SN=Sensitivity; SP=Specificity; PLR=Positive likelihood ratio; NLR=Negative likelihood ratio; DOR=Diagnosis odds ratio; R=Retrospective; P=Prospective; ICU=Intensive care unit; OS=Orthopedic surgery; SICU=Surgical intensive care unit; ICD-9=International Classification of Diseases-9; LTCF=Long term care facilities; NPUAP=National Pressure Ulcer Advisory Panel; AHCPR=Agency for Health Care Policy and Research; TDCPS=Torrence Developmental Classification of Pressure Sore.

603 Table 2. Summary Results of Meta Analysis Scales Studies Sensitivity Specificity Results of diagnostic test accuracy Positive likelihood ratio Negative likelihood ratio Diagnostic odds ratio Results of summary ROC curve Braden 36 0.74 (0.72-0.76) 0.75 (0.74-0.76) 3.13 (2.47-3.97) 0.36 (0.30-0.43) 10.23 (76.94-15.07) 0.83 0.02 0.77 0.02 Modified Song & Choi braden Pang & Wong 4 2 Sub-group analysis by Braden Nations USA 14 Korea 8 Canada 4 Gender ratio M>F M<F Mean age 50 60 70 Setting Wards ICU LTCF Reference standard NPUAP AHCPR ICD-9 11 9 7 10 11 15 6 7 15 5 3 0.97 (0.92-0.99) 0.89 (0.71-0.98) 0.69 (0.65-0.73) 0.87 (0.83-0.91) 0.69 (0.64-0.74) 0.83 (0.80-0.87) 0.73 (0.69-0.76) 0.87 (0.82-0.91) 0.69 (0.65-0.73) 0.69 (0.65-0.73) 0.73 (0.69-0.76) 0.85 (0.79-0.90) 0.74 (0.70-0.78) 0.74 (0.70-0.78) 0.74 (0.69-0.78) 0.68 (0.63-0.74) 0.70 (0.66-0.73) 0.71 (0.67-0.75) 0.67 (0.66-0.68) 0.87 (0.86-0.88) 0.66 (0.63-0.69) 0.84 (0.83-0.85) 0.60 (0.58-0.61) 0.91 (0.90-0.92) 0.63 (0.60-0.65) 0.63 (0.61-0.64) 0.86 (0.85-0.86) 0.65 (0.60-0.69) 0.60 (0.59-0.62) 0.68 (0.66-0.70) 0.89 (0.88-0.90) 0.61 (0.60-0.63) 3.47 (1.33-9.06) 2.87 (1.88-4.38) 2.80 (2.08-3.78) 4.04 (1.80-9.07) 2.53 (1.69-3.79) 3.89 (2.13-7.12) 2.09 (1.72-2.54) 5.33 (2.57-11.07) 2.18 (1.64-2.89) 2.33 (1.78-3.04) 3.69 (2.26-6.02) 2.48 (1.54-4.02) 2.37 (1.84-3.04) 2.73 (2.22-3.35) 4.56 (1.57-13.23) 1.94 (1.26-2.99) 0.08 (0.04-0.20) 0.17 (0.06-0.49) 0.44 (0.33-0.58) 0.19 (0.12-0.32) 0.44 (0.33-0.60) 0.23 (0.15-0.37) 0.43 (0.29-0.63) 0.17 (0.12-0.23) 0.47 (0.41-0.54) 0.46 (0.33-0.63) 0.35 (0.26-0.49) 0.25 (0.15-0.43) 0.37 (0.24-0.58) 0.40 (0.33-0.48) 0.26 (0.14-0.48) 0.42 (0.22-0.80) 56.56 (21.88-146.21) 16.06 (4.75-54.35) 7.35 (4.13-13.08) 23.09 (6.82-78.12) 5.81 (2.86-11.81) 17.52 (6.94-44.24) 5.20 (2.90-9.31) 31.05 (12.57-76.70) 5.69 (3.54-9.14) 5.89 (3.19-10.86) 12.23 (5.96-25.10) 11.06 (6.59-18.55) 6.34 (3.25-12.36) 7.49 (5.51-10.18) 29.28 (4.59-186.64) 5.02 (1.62-15.58) AUC SE (AUC) Q* SE (Q*) 0.95 0.02 0.90 ROC curve=receiver operating characteristic curve; AUC=Area under the curve; SE=Standard error; M=Male; F=Female; ICU=Intensive care unit; LTCF=Long term care facilities; NPUAP=National pressure ulcer advisory panel; AHCPR=Agency for health care policy and research; ICD-9=International Classification of Diseases-9. 0.80 0.92 0.80 0.89 0.73 0.93 0.77 0.77 0.84 0.82 0.75 0.80 0.91 0.67 0.04 0.04 0.04 0.01 0.04 0.05 0.02 0.04 0.10 0.74 0.85 0.73 0.82 0.67 0.87 0.71 0.71 0.77 0.76 0.69 0.73 0.84 0.63 0.04 0.01 0.02 0.05 0.02 0.05 0.08 특이도는 0.60 (95% CI: 0.58-0.61) 이었다. 문헌들간이질성은각각 85.1% (χ 2 = 53.56, p<.001), 93.2% (χ 2 =117.68, p<.001) 였고, sroc AUC 는 0.73 (SE= 0.04), Q * 값은 0.67 (SE= ) 이었다. (3) 평균연령대별평균연령을 50대로보고한 7편 [a3,a8,a12,a13,a25,a26,a29] 의문헌의통합민감도는 0.87 (95% CI: 0.82-0.91) 이었고, 문헌들간이질성은 0.0% 로전혀없었다 (χ 2 =3.12, p =.793). 통합특이도는 0.91 (95% CI: 0.90-0.92) 이었으며, 문헌들간이질성은 97.9% (χ 2 =286.97, p<.001) 로높았다. sroc AUC 는 0.93 (SE= 0.01), Q * 값은 0.87 (SE= 0.01) 이었다. 평균연령 60대인 10편 [a6,a9,a21,a22,a28,a30-a33,a38] 연구에서통합민감도는 0.69 (95% CI: 0.65-0.73) 통합특이도는 0.63 (95% CI: 0.60-0.65) 이었으며, 문헌들간이질성은각각 71.5% (χ 2 =31.59, p =.001), 97.0% (χ 2 =297.13, p <.001), sroc AUC는 0.77 (SE = ), Q* 값은 0.71 (SE = 0.02) 이었다. 평균연령 70대이상인 11편 [a2,a5,a7,a11,a17- a20,a24,a27,a34] 의연구에서통합민감도는 0.69 (95% CI: 0.65-0.73), 통합특이도는 0.63 (95% CI: 0.61-0.64) 이었으며, 문헌들간이질성은각각 83.3% (χ 2 =59.84, p<.001), 93.6% (χ 2 =156.82, p<.001) 로모두높았다. sroc AUC 는 0.77 (SE= 0.04), Q * 값은 0.71 (SE= ) 이었다. (4) 입원장소별병동환자를대상으로한 15편 [a6-a8,a10-a13,a21-a25,a31,a32,a35] 의연구에서통합민감도는 0.73 (95% CI: 0.69-0.76), 통합특이도는 0.86 (95% CI: 0.85-0.86) 이었으며, 문헌들간이질성은각각 68.1% (χ 2 = 43.95, p=.001), 97.8% (χ 2 = 624.56, p<.001) 였고 sroc AUC는 0.84 (SE= ), Q * 값은 0.77 (SE= ) 이었다. 중환자실환자를대상으로한총 6편 [a3,a14,a26,a30,a33,a37] 연구의통합민감도는 0.85 (95% CI: 0.79-0.90), 통합특이도는 0.65 (95% CI: 0.60-0.69) 였으며, 문헌들간이질성은각각 69.7% (χ 2 =16.52, p=.005), 92.5% (χ 2 = 66.95, p<.001) 였다. sroc AUC는 0.82 (SE= ), Q * 값은 0.76 (SE= ) 이었다. 장기요양시설거주자를대상으로한총 7편 [a2,a18-a20,a27,a33,a36] 연구에서의통합민감도는 0.74 (95% CI: 0.70-0.78), 통합특이도는 0.60 (95% CI: 0.59-0.62) 이었으며, 문헌들간이질성은각각 86.3% (χ 2 = 43.75, p<.001), 96.2% (χ 2 =157.70, p<.001) 였다. sroc AUC는 0.75 (SE= 0.05), Q * 값은 0.69 (SE= 0.05) 였다. (5) 참조표준기준 NPUAP를참조표준기준으로한 15편 [a2,a4,a7,a8,a11,a12,a21,a22, a26-a28,a30,a32,a33,a37] 연구의통합민감도는 0.74 (95% CI: 0.70-0.78), 통합특이도는 0.68 (95% CI: 0.66-0.70) 이었으며문헌들간이질성은각각 37.2% (χ 2 = 22.31, p =.073), 93.7 (χ 2 = 220.56, p<.001) 이었다. sroc AUC 는 0.80 (SE= 0.02), Q* 값은 0.73 (SE= 0.02) 이었다. AHCPR 을참조기준으로한 5편 [a14,a25,a30,a32,a38] 의연구에대한메타분석결과, 통합민감도는 0.74 (95% CI: 0.69-0.78), 통합특이도는 0.89 (95% CI: 0.88-0.90) 였으며문헌들간이질성은각각 92.5% (χ 2 = 53.31, p<.001), 99.2% (χ 2 = 530.95, p<.001) 였다. sroc AUC는 0.91 (SE= 0.04),

604 박성희 박유선 A. Sensitivity in Braden scale. B. Specificity in Braden scale. C. Sensitivity in modified Braden scale by Song & Choi. D. Specificity in modified Braden scale by Song & Choi. E. Sensitivity in modified Braden scale by Pang & Wong. Figure 2. Diagnosis test accuracy of Braden scale in total selected studies. F. Specificity in modified Braden scale by Pang & Wong. Q * 값은 0.84 (SE= 0.05) 였다. ICD-9 을참조기준으로한 3편 [a17,a19,a20] 의연구에서통합민감도는 0.68 (95% CI: 0.63-0.74) 통합특이도는 0.61 (95% CI: 0.60-0.63) 이었으며, 문헌들간이질성은각각 93.5% (χ 2 =30.93, p<.001), 97.1% (χ 2 = 69.06, p<.001) 로높았다. sroc AUC는 0.67 (SE= 0.10), Q* 값은 0.63 (SE= 0.08) 이었다. 논의예측타당도는검사나평가도구에서얻은점수와미래의어떤결과와의관계로추정되는준거타당도이다. 욕창위험사정도구는욕창발생을확진하는진단검사가아니라욕창위험을예측하는선별검사이다. 선별검사는잠재적위험요인은있지만무증상인일반인을대상으로시행되고, 대개쉽고간단한비침습적인도구가권장 되기때문에확진검사에비해제한적조건을가지게된다. 또한, 선별검사의결과는환자에게부담감이나해를줄수있기때문에해당검사를사전에시행하는것이해보다득이많음을증명하여야하고, 잠재적질병가능성을놓치지않아야하므로민감도가높아야한다 [26]. 2002년 Schoonhoven 등 [27] 은현재사용되는욕창사정도구들의진단정확도가낮아욕창예방중재의사용이비효율적이므로좀더예측성을높인도구개발이필요하다고제시하였다. 그러나그간연구들은욕창위험사정을통한실제욕창발생률에대한차이에만초점을두고있어 [28] 본연구에서는욕창위험사정도구로가장널리사용되는 Braden 도구의예측타당도를규명하기위해총 38편, 총 17,934 명의진단법평가결과를토대로체계적으로고찰한후메타분석을시행하였다. 분석결과, Braden 도구의욕창이발생될대상자를욕창고위험

605 군으로검출하는능력인통합민감도는 0.74, 욕창이발생되지않을대상자를욕창저위험군으로판별하는통합특이도는 0.75였다. 또한, Braden 도구의진단적가치를반영하는통합양성우도비는 3.13, 음성우도비는 0.36이었고, 실제욕창발생유무에따라대상자를고위험군또는저위험군으로구분하는통합진단교차비는 10.23, 검사의진단정확도를나타내는 sroc AUC는 0.83이었다. 민감도와특이도는검사성능의일반측정값으로해석이용이하지만숫자자체로는검사성능의적절성을결정하는데바람직하지않으므로우도비나 AUC 값을활용하는것이바람직하다 [26]. 본연구에서는 sroc AUC 값을기준으로욕창위험사정도구의예측타당도를판별하였다. 연구방법에서언급된바와같이 Braden 도구의 sroc AUC는 0.83으로분석되어 0.7< AUC 0.9 사이의값을가지므로중등도의검사정확도를갖는다고설명할수있다. 그러나문헌들간이질성은 75% 를넘는높은수준을나타내충분한근거를제시하기에는한계를보였다. 수정 Braden 도구는선별검사로서의조건을보다향상시키기위해민감도를높이는수준으로보완되었다. Song과 Choi의수정 Braden 도구는기존 Braden 도구에체온과약품사용을추가하였고 [a13-a16], Pang과 Wong 의수정 Braden 도구는피부상태 (skin type) 와체형 (body build for height) 을추가하였다 [a11,a12]. 수정 Braden 도구는모두민감도가크게높아지면서동시에문헌들간이질성도크게감소되는양상을보였으나통합특이도는오히려더낮아졌으며, 문헌들간이질성도여전히 90% 이상의높은수준을보였다. 뿐만아니라 Braden 도구의세부분석에서도민감도는연구대상의특성에따라문헌들간이질성이감소되는경향을보였으나특이도는차이를보이지않았다. 비록욕창예방요법이위험도가높은간호중재가아니므로민감도와특이도가모두높은검사법을요구하지는않지만어떤상황에서도문헌들간이질성이높게나타남은선별도구로서의적용에한계를보인다고해석된다. 일반적으로진단검사메타분석에서문헌들간이질성은연구대상이나검사방법등각연구자료의특성과관련되며, 해결되지않는문제점중하나이다 [26,28]. 따라서, 연구대상자의특성에따라세부분석하였다. 민감도의이질성은연구장소, 성비, 평균연령, 입원장소, 참조표준기준모두에서변화가있었다. 각특성에따라세분하였을때모든부분에서이질성이낮아진것은아니었으나전체와비교했을때상대적으로크게줄어드는경향을보였으며, 특히참조표준기준에대해서는 ICD-9 코드를제외하고는문헌들간이질성이 37.2% 로낮아졌다. ICD-9 코드만으로는그원인을설명하기곤란하나 ICD-9 코드를참조표준기준으로활용한연구의공통적특징은모두후향적연구라는점이었다. 따라서, 문헌들간이질성에변화를보이지않은이유는간호기록에쓰여진내용만을토대로욕창여부를판별함에따라야기된정보의제한성때문에야기 된문제로추정해볼수있다. 본연구결과에별도제시하진않았지만전향적연구와후향적연구간의 Braden 욕창사정도구의예측타당도는큰차이를보이지않는것으로분석되었다. 특히, Braden 도구의예측타당도에영향하는요인으로알려진연령, 성별및입원장소등의공통점만을선별하여재분석한결과, 남성의성비가높은경우모든연령대에서문헌들간이질성이전혀없는것으로나타났으며, 평균연령과입원장소별분석에서문헌들간의이질성이없어진점은, 연구대상의특성에따라 Braden 도구가다르게적용되어야함을간접적으로시사하는결과로추정된다. 본연구에서는각연구자들이가장최선으로정했던경계점수를그대로사용하였으며, 이에따른세부분석은진행하지않았다. 그러나여기에서주목할점은욕창위험여부를결정짓는경계점수는임상에서대개일률적으로적용한다는점이다. 총 38편의문헌들에서연령이나입원장소등연구대상특성에따라경계점수는어떤일관성을보이지않았다. 다만, 일부연구 [a1,a36,a37] 에서연령이적은경우경계점수가낮은경향을보이긴하였다. 이결과로만해석하면 Braden 도구를임상에서일상적으로누구에게나적용되는표준화된도구로활용하기에는적절하지않으며, 경계점수는특정임계치보다는범주로서적용되어야한다는점이다. 좋은판별력을가진검사가반드시유용한검사가되는것은아니지만 2008년코크란연합의체계적문헌고찰연구 [29] 에서지적된바와같이욕창위험사정도구를이용한사전욕창발생위험환자의선별결과가이후욕창발생률을감소시키지못하며, 기존 Braden 도구의판별력을보완한수정 Braden 도구에서도비록민감도를 0.74 에서 0.97로향상시켰으나특이도는오히려 0.75에서 0.70로낮아지는문제가야기되어 [a11-a16] Braden 도구로만욕창발생의위험을판별하기에는충분하지못하다고해석된다. 따라서, 보다욕창위험사정도구의예측타당도를높이기위해서는욕창발생위험이높은환자의경우 Braden 도구만을단독으로적용하기보다는 Norton 이나 Waterlow 도구등의기존욕창위험사정도구들을중복적 (add-on) 으로적용하여현재수준에서욕창위험사정의정확도를높이는전략이요구되며, 그간욕창발생요인으로검증된위험요인들에대한체계적고찰을통해욕창위험사정도구의타당도를보다높이기위한추후연구가필요하다고생각된다. 결론본연구에서는 38편의잘설계된진단법평가문헌을토대로한메타분석을통해 Braden 욕창위험사정도구의예측타당도를확인하였다. Braden 도구는 sroc AUC가 0.7 이상이므로중등도의예측타당도를보이는것으로분석되었지만문헌들간이질성이높아해

606 박성희 박유선 석에한계를보였다. 비록본연구결과, Braden 도구를표준화된기 준하에모든환자를대상으로일률적으로활용하긴곤란하다고 판단되었으나연구대상의특성에따라 Braden 도구를구별하여적 용한다면예측타당도를높일수있다는가능성을보여주었다. 욕 창예방간호는움직임이자유롭지못한환자에게적용되어야하는 기본간호중하나이다. 따라서, 보다효율적인간호제공을위해서 는 Braden 도구뿐아니라다른욕창위험사정도구의진단정확도를 분석하는연구가요구되며, 기존도구들의장단점을보완함으로보 다예측타당도를높인수정형욕창위험사정도구의개발을위한노 력이필요할것으로보인다. REFERENCES 1. Armstrong DG, Ayello EA, Capitulo KL, Fowler E, Krasner DL, Levine JM, et al. New opportunities to improve pressure ulcer prevention and treatment: Implications of the CMS inpatient hospital care present on admission indicators/hospital-acquired conditions policy: A consensus paper from the international expert wound care advisory panel. Advances in Skin & Wound Care. 2008;21(10):469-478. http://dx.doi.org/10.1097/01.asw.0000323562.52261.40 2. National Pressure Ulcer Advisory Panel. NPUAP pressure ulcer stages/ categories [Internet]. Washington, DC: Author; 2007 [cited 2010 July 12]. Available from: http://www.npuap.org/. 3. Cowan LJ, Stechmiller JK, Rowe M, Kairalla JA. Enhancing Braden pressure ulcer risk assessment in acutely ill adult veterans. Wound Repair and Regeneration. 2012;20(2):137-148. http://dx.doi.org/10.1111/j.1524-475x.2011.00761.x 4. The Joint Commission. Strategies for preventing pressure ulcers. Joint Commission Perspectives on Patient Safety. 2008;8(1):5-7. 5. Vangilder C, Macfarlane GD, Meyer S. Results of nine international pressure ulcer prevalence surveys: 1989 to 2005. Ostomy/Wound Management. 2008;54(2):40-54. 6. Cho MS, Park IS, Kim GH, Woo KS, Joo YH, Jung EH, et al. Evaluation of predictive validity for the pressure ulcer risk assessment tool in a medical ward inpatient: Using Braden scale. The Korean Nurse. 2004; 43(2):68-82. 7. Lee JK. The relationship of risk assessment using Braden scale and development of pressure sore in neurologic intensive care unit. Journal of Korean Academy of Adult Nursing. 2003;15(2):267-277. 8. Song HJ, Kim SM, Kim NC. A study of voiding patterns and pressure ulcer for the residents of long term care facilities. Journal of the Korean Continence Society. 2003;7(2):91-97. 9. Lee EJ, Yang SO. Clinical knowledge and actual performance of pressure ulcer care by hospital nurses. Journal of Korean Clinical Nursing Research. 2011;17(2):251-261. 10. Jeong IS, Kim S, Jeong JS, Hong EY, Lim EY, Seo HJ, et al. Development of pressure ulcer management guideline by adaptation process. Journal of Korean Clinical Nursing Research. 2014;20(1):40-52. 11. Choi KS, Song MS. Test of predictive validity for the new pressure risk assessment scale. Journal of Korean Academy of Adult Nursing. 1991;3 (1):19-28. 12. Bergstrom N, Braden BJ, Laguzza A, Holman V. The Braden scale for predicting pressure sore risk. Nursing Research. 1987;36(4):205-210. 13. Chou R, Dana T, Bougatsos C, Blazina I, Starmer AJ, Reitel K, et al. Pressure ulcer risk assessment and prevention: A systematic comparative effectiveness review. Annals of Internal Medicine. 2013;159(1):28-38. http://dx.doi.org/10.7326/0003-4819-159-1-201307020-00006 14. Defloor T, Grypdonck MF. Pressure ulcers: Validation of two risk assessment scales. Journal of Clinical Nursing. 2005;14(3):373-382. http://dx.doi.org/10.1111/j.1365-2702.2004.01058.x 15. Bergquist S. Subscales, subscores, or summative score: Evaluating the contribution of Braden scale items for predicting pressure ulcer risk in older adults receiving home health care. Journal of Wound, Ostomy, and Continence Nursing. 2001;28(6):279-289. 16. Suriadi, Sanada H, Sugama J, Thigpen B, Kitagawa A, Kinosita S, et al. A new instrument for predicting pressure ulcer risk in an intensive care unit. Journal of Tissue Viability. 2006;16(3):21-26. 17. Jalali R, Rezaie M. Predicting pressure ulcer risk: Comparing the predictive validity of 4 scales. Advances in Skin & Wound Care. 2005;18(2): 92-97. 18. Deeks JJ, Bossuyt PM, Gatsonis C, editors. Cochrane handbook for systematic reviews of diagnostic test accuracy version 1.0.0. Oxford, UK: The Cochrane Collaboration 2010. 19. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. Annals of Internal Medicine. 2009;151(4):264-269. 20. Whiting PF, Rutjes AW, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, et al. QUADAS-2: A revised tool for the quality assessment of diagnostic accuracy studies. Annals of Internal Medicine. 2011;155(8):529-536. http://dx.doi.org/10.7326/0003-4819-155-8-201110180-00009 21. Greiner M, Pfeiffer D, Smith RD. Principles and practical application of the receiver-operating characteristic analysis for diagnostic tests. Preventive Veterinary Medicine. 2000;45(1-2):23-41. 22. Walter SD. Properties of the summary receiver operating characteristic (SROC) curve for diagnostic test data. Statistics in Medicine. 2002;21 (9):1237-1256. http://dx.doi.org/10.1002/sim.1099 23. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Statistics in Medicine. 2002;21(11):1539-1558. http://dx.doi.org/10.1002/sim.1186 24. Song M, Choi KS. Factors predicting development of decubitus ulcers among patients admitted for neurological problems. The Journal of Nurses Academic Society. 1991;21(1):16-26. 25. Lowthian P. The classification and grading of pressure sores. Care: Science and Practice. 1987;5:5-9. 26. Knottnerus JA. The evidence base of clinical diagnosis. Park SH, Kang CB, translator. Seoul: E-Public; 2008. 27. Schoonhoven L, Haalboom JR, Bousema MT, Algra A, Grobbee DE, Grypdonck MH, et al. Prospective cohort study of routine use of risk assessment scales for prediction of pressure ulcers. BMJ: British Medical Journal. 2002;325(7368):797. 28. Sousa MR, Ribeiro AL. Systematic review and meta-analysis of diagnostic and prognostic studies: A tutorial. Arquivos Brasileiros de Cardiologia. 2009;92(3):229-238, 235-245. 29. Moore ZE, Cowman S. Risk assessment tools for the prevention of pressure ulcers. The Cochrane Database of Systematic Reviews. 2008;16(3):CD006471. http://dx.doi.org/10.1002/14651858.cd006471.pub2

607 APPENDIX: 38 selected studies in meta-analysis a1. Baldwin KM, Ziegler SM. Pressure ulcer risk following critical traumatic injury. Advances in Wound Care. 1998;11(4):168-173. a2. Braden BJ, Bergstrom N. Predictive validity of the Braden scale for pressure sore risk in a nursing home population. Research in Nursing and Health. 1994;17(6):459-470. a3. Bergstrom N, Demuth PJ, Braden BJ. A clinical trial of the Braden scale for predicting pressure sore risk. The Nursing Clinics of North America. 1987;22(2):417-428. a4. Bergstrom N, Braden BJ. Predictive validity of the Braden scale among black and white subjects. Nursing Research. 2002;51(6):398-403. a5. Goodridge DM, Sloan JA, LeDoyen YM, McKenzie JA, Knight WE, Gayari M. Risk-assessment scores, prevention strategies, and the incidence of pressure ulcers among the elderly in four Canadian health-care facilities. The Canadian Journal of Nursing Research. 1998;30(2): 23-44. a6. Halfens RJ, Van Achterberg T, Bal RM. Validity and reliability of the Braden Scale and the influence of other risk factors: A multi-centre prospective study. International Journal of Nursing Studies. 2000;37 (4):313-319. a7. Lyder CH, Yu C, Stevenson D, Mangat R, Empleo-Frazier O, Emerling J, et al. Validating the Braden scale for the prediction of pressure ulcer risk in blacks and Latino/Hispanic elders: A pilot study. Ostomy/ Wound Management. 1998;44(3A Suppl):42S-49S. a8. Olson K, Tkachuk L, Hanson J. Preventing pressure sores in oncology patients. Clinical Nursing Research. 1998;7(2):207-224. a9. VandenBosch T, Montoye C, Satwicz M, Durkee-Leonard K, Boylan- Lewis B. Predictive validity of the Braden scale and nurse perception in identifying pressure ulcer risk. Applied Nursing Research. 1996;9 (2):80-86. a10. Pang SM, Wong TK. Predicting pressure sore risk with the Norton, Braden, and Waterlow scales in a Hong Kong rehabilitation hospital. Nursing Research. 1998;47(3):147-153. a11. Chan WS, Pang SM, Kwong EW. Assessing predictive validity of the modified Braden scale for prediction of pressure ulcer risk of orthopaedic patients in an acute care setting. Journal of Clinical Nursing. 2009;18(11):1565-1573. http://dx.doi.org/10.1111/j.1365-2702.2008.02757.x a12. Kwong E, Pang S, Wong T, Ho J, Shao-ling X, Li-jun T. Predicting pressure ulcer risk with the modified Braden, Braden, and Norton scales in acute care hospitals in Mainland China. Applied Nursing Research. 2005;18(2):122-128. http://dx.doi.org/10.1016/j.apnr.2005.01.001 a13. Choi KS, Song MS. Test of predictive validity for the new pressure risk assessment scale. Journal of Korean Academy of Adult Nursing. 1991; 3(1):19-28. a14. Kim EK, Lee SM, Lee E, Eom MR. Comparison of the predictive validity among pressure ulcer risk assessment scales for surgical ICU patients. The Australian Journal of Advanced Nursing. 2009;26(4):87-94. a15. Kim SS, Choi KS. Evaluating the predictive validity for the new pressure sores risk assessment scale. Journal of Korean Academy of Adult Nursing. 2004;16(2):183-190. a16. Lee YH, Jeong IS, Jeon SS. A comparative study on the predictive validity among pressure ulcer risk assessment scales. Journal of Korean Academy of Nursing. 2003;33(2):162-169. a17. Cowan LJ, Stechmiller JK, Rowe M, Kairalla JA. Enhancing Braden pressure ulcer risk assessment in acutely ill adult veterans. Wound Repair and Regeneration. 2012;20(2):137-148. http://dx.doi.org/10.1111/j.1524-475x.2011.00761.x a18. Bergstrom N, Braden BJ, Laguzza A, Holman V. The Braden scale for predicting pressure sore risk. Nursing Research. 1987;36(4):205-210. a19. Bergquist S. Subscales, subscores, or summative score: Evaluating the contribution of Braden scale items for predicting pressure ulcer risk in older adults receiving home health care. Journal of Wound, Ostomy, and Continence Nursing. 2001;28(6):279-289. a20. Bergquist S, Frantz R. Braden scale: Validity in community-based older adults receiving home health care. Applied Nursing Research. 2001;14(1):36-43. http://dx.doi.org/10.1053/apnr.2001.21079 a21. Schue RM, Langemo DK. Pressure ulcer prevalence and incidence and a modification of the Braden scale for a rehabilitation unit. Journal of Wound, Ostomy, and Continence Nursing. 1998;25(1):36-43. a22. Capobianco ML, McDonald DD. Factors affecting the predictive validity of the Braden scale. Advances in Wound Care. 1996;9(6):32-36. a23. Barnes D, Payton RG. Clinical application of the Braden scale in the acute-care setting. Dermatology Nursing. 1993;5(5):386-388. a24. Salvadalena GD, Snyder ML, Brogdon KE. Clinical trial of the Braden scale on an acute care medical unit. Journal of ET Nursing. 1992;19(5): 160-165. a25. Cho MS, Park IS, Kim GH, Woo KS, Joo YH, Jung EH, et al. Evaluation of predictive validity for the pressure ulcer risk assessment tool in a medical ward inpatient: Using Braden scale. The Korean Nurse. 2004;43(2):68-82. a26. Lee JK. The relationship of risk assessment using Braden scale and development of pressure sore in neurologic intensive care unit. Journal of Korean Academy of Adult Nursing. 2003;15(2):267-277. a27. Kim DH, Jeong HS, Lee DW. Evaluation of risk for pressure ulcers using the Braden scale in elderly patients receiving long-term care. Journal of the Korean Geriatrics Society. 2011;15(4):191-199. http://dx.doi.org/10.4235/jkgs.2011.15.4.191 a28. Cho I, Noh M. Braden scale: Evaluation of clinical usefulness in an intensive care unit. Journal of Advanced Nursing. 2010;66(2):293-302. http://dx.doi.org/10.1111/j.1365-2648.2009.05153.x a29. Oh DY, Kim JH, Lee PK, Ahn ST, Rhie JW. Prevention of pressure ulcer using the pressure ulcer risk assessment based on Braden scale. Journal of the Korean Society of Plastic and Reconstructive Surgeons. 2007;34(4):465-469. a30. Jun S, Jeong I, Lee Y. Validity of pressure ulcer risk assessment scales; Cubbin and Jackson, Braden, and Douglas scale. International Journal of Nursing Studies. 2004;41(2):199-204. a31. Marrie RA, Ross JB, Rockwood K. Pressure ulcers: Prevalence, staging, and assessment of risk. The Canadian Geriatrics Journal. 2003;6 (3):134-140. a32. Harrison MB, Wells G, Fisher A, Prince M. Practice guidelines for the prediction and prevention of pressure ulcers: Evaluating the evidence. Applied Nursing Research. 1996;9(1):9-17. a33. Serpa LF, Santos VL, Campanili TC, Queiroz M. Predictive validity of the Braden scale for pressure ulcer risk in critical care patients. Revista Latino-Americana de Enfermagem. 2011;19(1):50-57. a34. de Souza DM, Santos VL, Iri HK, Sadasue Oguri MY. Predictive validity of the Braden scale for pressure ulcer risk in elderly residents of long-term care facilities. Geriatric Nursing. 2010;31(2):95-104. http://dx.doi.org/10.1016/j.gerinurse.2009.11.010 a35. Watkinson C. Developing a pressure sore risk assessment scale. Professional Nurse. 1997;12(5):341-346. a36. Defloor T, Grypdonck MF. Pressure ulcers: Validation of two risk assessment scales. Journal of Clinical Nursing. 2005;14(3):373-382. http://dx.doi.org/10.1111/j.1365-2702.2004.01058.x a37. Suriadi, Sanada H, Sugama J, Thigpen B, Kitagawa A, Kinosita S, et al. A new instrument for predicting pressure ulcer risk in an intensive care unit. Journal of Tissue Viability. 2006;16(3):21-26. a38. Jalali R, Rezaie M. Predicting pressure ulcer risk: Comparing the predictive validity of 4 scales. Advances in Skin & Wound Care. 2005;18 (2):92-97.