研究背景：估計兒虐死亡數是探討兒虐問題與擬定兒虐防制計畫的第一步。但是，過去研究已經指出，使用官方死因統計資料估計兒虐死亡數會低估。只能呈現冰山浮出海面的部份，無法掌握兒虐冰山的整體輪廓。

研究目的：連結不同全人口行政資料庫，估計台灣五歲以下兒虐死亡人數冰山全貌。

研究材料與方法: 本研究連結2011-2015年死因統計、全民健保申報與兒童保護司通報等三個資料庫，探討不同層級資料庫分別可以發現多少五歲以下兒虐死亡數。第一層級死亡數估計的操作定義是死因統計中原死因出現國際疾病分類第十版(ICD-10)編碼T74.X(確認兒虐)與T76.X(疑似兒虐)與X85 -Y09(侵害攻擊)的個案數。第二層級是死因統計中多重死因出現ICD-10編碼T74.X、T76.X、X85-Y09的個案數。第三層級健保急診或住院申報診斷出現國際疾病分類臨床修訂第九版編碼(ICD-9-CM)有995.5X或E960.X-E969.X且就醫日與死亡日同一天或是一週內的個案數。第四層級是承第三層級定義且就醫日是死亡日一週後至前六個月內的個案數。第五層級是死亡日同年保護司通報資料有紀錄者的個案數。第六層級是死亡日前一年保護司通報資料有紀錄者的個案數。統計分析首先呈現整體估計兒虐死亡個案的不同特徵，接著比較不同資料庫找出個案的特徵差異。最後再分析連結資料庫內容，以死因分類做區分，各項死因選取一位個案進行分析，以時間的概念描繪動態就醫歷程，質性文字敘述個案死亡前在就醫歷程與相關診斷。

研究結果：符合第一層級定義兒虐死亡數有57人，第二層級2人，第三層級35人，第四層級3人，第五層級331人，第六層級14人，總計442人。若將第一第二層級合併，第三第四層級合併，第五第六層合併，冰山輪廓由頂端至底層比約為1:1.6:7.5。五層級合計，2011年至2015年人數分別為86, 112, 126, 84, 34人。男孩(248人)略高於女孩(194人)。死因診斷分類，190人(43%)歸因於自然病死，128人(29%)歸因於意外。相較於死因統計定義的兒虐死亡個案特徵分布，由健保申報資料定義的兒虐死亡個案，小於一歲人數比例較高(50% vs. 27%)。由通報資料定義的兒虐死亡個案，一樣是小於一歲人數比例較高(64% vs. 27%)。由四位個案分析可以發現：一位死因統計是他殺個案，在死前就有多次外傷診斷。三位死因統計是自然病死、意外死或不詳者，在死前就醫也出現過兒虐診斷，硬腦膜外出血，股骨幹骨折，嬰兒搖晃症候群等前哨傷害診斷，有強烈證據證明這些個案是兒虐死亡個案。

研究結論：本資料庫連結研究顯示，如果採取較嚴格定義(只考慮健保申報資料庫有提及兒虐診斷且就醫日在死亡日一週內)，台灣2011-2015年五歲以下兒虐死亡數約94人，比只用死因統計資料估計得59人多出近兩倍。如果採取較寬鬆定義(考慮死亡同年有兒童保護通報紀錄)，兒虐死亡數約442人，比只用死因統計資料估計得59人多出近八倍。本研究也發現，許多低估個案在健保申報就醫資料還是有出現前哨傷害診斷，宜多多教育訓練醫護人員針對前哨傷害進行進一步的評估，確定是否有兒虐傷害。

Official records of abused child deaths are underestimated. The purpose of this study was to estimate the number of child abuse fatalities in Taiwan under the age of five by linking data from multiple population-based administrative records. Data were obtained from three databases through 2011 to 2015, including Cause-of-Death Statistics, National Insurance Claimed Data, and Reported Data of Protection of Child and Youths. Cases were defined if they had ICD-10 code T74.X, T76.X or X85-Y09 in cause of death, ICD-9-CM code 995.5X or E960.X-E969.X within six months before the date of death by medical claim data, or reports in Reported Data of Protection of Child and Youths within one year before death. We analyzed characteristics of all deaths from child maltreatment. Four hundred and forty-two cases were found from the three databases. The number of child abused death from the first level to sixth level was 57, 2, 35, 3, 331, 14 respectively. The number of death from maltreatment we estimated was two to eight times higher than Cause-of-Death Statistics. Compared with Cause-of-Death Statistics, the proportion of cases who were younger than age one from the medical claim was higher (50% vs. 27%), and cases from the Reported Data of Protection of Child and Youths was also higher (65% vs. 27%). Four cases we reviewed had several diagnoses of trauma and sentinel injuries before death. Therefore, we suggest medical staff have more education and training about further assessments of sentinel injuries to determine whether cases are abused children.

呂宗學、蕭安芝、徐俊強、彭花春、陳麗華(2013)，死因診斷與死因統計品質，台灣醫學, 17(5), 551-557。
邱鈺鸞、白璐、鍾其祥、蔡沛然、簡戊鑑(2010)，台灣 1997-2008 年受虐住院兒童傷害特性及趨勢，亞洲家庭暴力與性侵害期刊, 6(2), 65-81。
吳敏髣(2011)，建立中文版幼兒傷害區辨意圖之判斷標準，成功大學護理學研究所學位論文。
衛生福利部統計處(2016)，105年死因統計年報，取自 https://dep.mohw.gov.tw/DOS/mp-113.html
衛生福利部保護司(2016)，兒童及少年保護工作指南，取自https://dep.mohw.gov.tw/DOPS/cp-1240-6870-105.html
衛生福利部統計處(2018)，兒童少年保護通報處理情事，取自https://dep.mohw.gov.tw/DOS/lp-2985-113.html
Centers for Disease Control and Prevention. (2016). Leading Causes of Death Reports, 1981 - 2016. from https://webappa.cdc.gov/sasweb/ncipc/leadcause.html
Chang, D. C., Knight, V., Ziegfeld, S., Haider, A., Warfield, D., & Paidas, C. (2004). The tip of the iceberg for child abuse: the critical roles of the pediatric trauma service and its registry. Journal of Trauma and Acute Care Surgery, 57(6), 1189-1198.
Crume, T. L., DiGuiseppi, C., Byers, T., Sirotnak, A. P., & Garrett, C. J. (2002). Underascertainment of child maltreatment fatalities by death certificates, 1990–1998. Pediatrics, 110(2), e18-e18.
Euser, S., Alink, L. R., Pannebakker, F., Vogels, T., Bakermans-Kranenburg, M. J., & Van IJzendoorn, M. H. (2013). The prevalence of child maltreatment in the Netherlands across a 5-year period. Child abuse & neglect, 37(10), 841-851.
Ewigman, B., Kivlahan, C., & Land, G. (1993). The Missouri child fatality study: underreporting of maltreatment fatalities among children younger than five years of age, 1983 through 1986. Pediatrics, 91(2), 330-337.
Gallagher, S. S., Finison, K., Guyer, B., & Goodenough, S. (1984). The incidence of injuries among 87,000 Massachusetts children and adolescents: results of the 1980-81 Statewide Childhood Injury Prevention Program Surveillance System. American journal of public health, 74(12), 1340-1347.
Heinrich, H. W. (1941). Industrial Accident Prevention. A Scientific Approach. Industrial Accident Prevention. A Scientific Approach.(Second Edition).
Herman-Giddens, M. E., Brown, G., Verbiest, S., Carlson, P. J., Hooten, E. G., Howell, E., & Butts, J. D. (1999). Underascertainment of child abuse mortality in the United States. JAMA, 282(5), 463-467.
Hooft, A., Ronda, J., Schaeffer, P., Asnes, A. G., & Leventhal, J. M. (2013). Identification of physical abuse cases in hospitalized children: accuracy of International Classification of Diseases codes. The Journal of pediatrics, 162(1), 80-85.
Hunt, R., & Barr, P. (2000). Errors in the certification of neonatal death. Journal of paediatrics and child health, 36(5), 498-501.
Jütte, S., Bentley, H., Miller, P., & Jetha, N. (2014). How safe are our children London: NSPCC.
Kajese, T. M., Nguyen, L. T., Pham, G. Q., Pham, V. K., Melhorn, K., & Kallail, K. J. (2011). Characteristics of child abuse homicides in the state of Kansas from 1994 to 2007. Child abuse & neglect, 35(2), 147-154.
Klevens, J., & Leeb, R. T. (2010). Child maltreatment fatalities in children under 5: Findings from the National Violence Death Reporting System. Child abuse & neglect, 34(4), 262-266.
Lapidus, G. D., Gregorio, D. I., & Hansen, H. (1990). Misclassification of childhood homicide on death certificates. American journal of public health, 80(2), 213-214.
Last, J. M. (2013). The iceberg:‘completing the clinical picture’in general practice. International journal of epidemiology, 42(6), 1608-1613.
Lee, L. K., Fleegler, E. W., Forbes, P. W., Olson, K. L., & Mooney, D. P. (2010). The modern paediatric injury pyramid: injuries in Massachusetts children and adolescents. Injury prevention, 16(2), 123-126.
Lindberg, D. M., Beaty, B., Juarez-Colunga, E., Wood, J. N., & Runyan, D. K. (2015). Testing for abuse in children with sentinel injuries. Pediatrics, 136(5), 831-838.
Lu, T.-H., Shaw, K.-P., Hsu, P.-Y., Chen, L.-H., & Huang, S.-M. (2008). Non-referral of unnatural deaths to coroners and non-reporting of unnatural deaths on death certificates in Taiwan: implications of using mortality data to monitor quality and safety in healthcare. International Journal for Quality in Health Care, 20(3), 200-205.
Meinck, F., Steinert, J., Sethi, D., Gilbert, R., Bellis, M., Alink, L., & Baban, A. (2016). Measuring and monitoring national prevalence of child maltreatment: a practical handbook.
Olsen, S. J., & Durkin, M. S. (1996). Validity of hospital discharge data regarding intentionality of fatal pediatric injuries. Epidemiology, 644-647.
Rimsza, M. E., Schackner, R. A., Bowen, K. A., & Marshall, W. (2002). Can child deaths be prevented? The Arizona child fatality review program experience. Pediatrics, 110(1), e11-e11.
Sahai, V. S., Ward, M. S., Zmijowskyj, T., & Rowe, B. H. (2005). Quantifying the iceberg effect for injury: using comprehensive community health data. Canadian Journal of Public Health/Revue Canadienne de Sante'e Publique, 328-332.
Schnitzer, P. G., Covington, T. M., Wirtz, S. J., Verhoek-Oftedahl, W., & Palusci, V. J. (2008). Public health surveillance of fatal child maltreatment: analysis of 3 state programs. American journal of public health, 98(2), 296-303.
Schnitzer, P. G., Gulino, S. P., & Yuan, Y.-Y. T. (2013). Advancing public health surveillance to estimate child maltreatment fatalities: Review and recommendations. Child welfare, 92(2), 77-98.
Tursz, A., Crost, M., Gerbouin-Rérolle, P., & Cook, J. M. (2010). Underascertainment of child abuse fatalities in France: retrospective analysis of judicial data to assess underreporting of infant homicides in mortality statistics. Child abuse & neglect, 34(7), 534-544.
U.S. Department of Health and Human Services. (2016). Child Maltreatment 2016.
Wadman, M. C., Muelleman, R. L., Coto, J. A., & Kellermann, A. L. (2003). The pyramid of injury: using ecodes to accurately describe the burden of injury. Annals of emergency medicine, 42(4), 468-478.
Winn, D. G., Agran, P. F., & Anderson, C. L. (1995). Sensitivity of hospitals' E-coded data in identifying causes of children's violence-related injuries. Public health reports, 110(3), 277.
Wu, M. F., Lu, T. H., Lin, C. J.,& Feng, J. Y. (2015). Risk factors and physical signs of child abuse in hospitalized children in Taiwan. Children and Youth Services Review, 58,137-111.
World Health Organization (2016)，Child maltreatment，from http://www.who.int/violence_injury_prevention/violence/child/en/