Structured Topic Modeling of Child Fatal Injury Narratives Using Machine Learning with County-Level Demographic Indicators to Strengthen Prevention Strategies
Restricted accessResearch articleFirst published online 2025
Structured Topic Modeling of Child Fatal Injury Narratives Using Machine Learning with County-Level Demographic Indicators to Strengthen Prevention Strategies
Interventions for children at risk for fatal injury have been limited by the lack of systematic methods to detect and classify the circumstances surrounding deaths recorded in unstructured text narratives. To address this gap, we analyzed N = 453 child fatality reports from the Pennsylvania Department of Human Services (2016–2023). Each report was merged with the 2016–2021 American Community Survey (ACS) five-year estimates by county of death, incorporating county-level indicators of poverty and racial composition to provide sociodemographic context, as well as with report-level metadata on year and county of death. We then applied Natural Language Processing and Structured Topic Modeling, a machine learning algorithm that incorporated both ACS data and report metadata, to identify common themes in the narratives. The model revealed 11 distinct categories that included Severe Traumatic Injury (12%); Homicide due to Parental Neglect (10%); Medical History or Illness (11%); Institutional Negligence (5.3%); Sleep-Related Deaths (12%); Aggravated Assault (11%); Substance Misuse (13.4%); Failure to Supervise (6%); Supervisory Neglect (6%); Drowning (8%); and Firearm-Related Injury (6%). Analysis of temporal and sociodemographic covariates showed that some themes (e.g., Homicide due to Parental Neglect) declined over time, while others (e.g., Substance Misuse) became more prevalent. Counties with higher poverty levels and larger non-White populations showed a greater prevalence of Firearm-Related Injury and Severe Traumatic Injury, while counties with larger non-White populations had lower rates of Drowning and a marginal association with fewer Sleep-Related Deaths. Integrating ACS data and report-level metadata with computational text analysis and machine learning models can generate actionable insights into the circumstances of child deaths, strengthen fatality surveillance, and inform targeted prevention strategies.
ArbogastK. B.CurryA. E.PfeifferM. R.ZonfrilloM. R.Haarbauer-KrupaJ.BreidingM. J.CoronadoV. G.MasterC. L. (2016). Point of health care entry for youth with concussion within a large pediatric care network. JAMA Pediatrics, 170(7), e160294. https://doi.org/10.1001/jamapediatrics.2016.0294
2.
BarbozaG. E.SchiambergL. B.PachlL. (2021). A spatiotemporal analysis of the impact of COVID-19 on child abuse and neglect in the city of Los Angeles, California. Child Abuse & Neglect, 116(Pt 2), Article 104740. https://doi.org/10.1016/j.chiabu.2020.104740
3.
Barboza-SalernoG. E. (2020). Variability and stability in child maltreatment risk across time and space and its association with neighborhood social & housing vulnerability in New Mexico: A bayesian space-time model. Child Abuse & Neglect, 104(1), Article 104472. https://doi.org/10.1016/j.chiabu.2020.104472
4.
Barboza-SalernoG. E.DuhaneyS.LiebhardB.Schockley-McCarthyK. (2025). Pushing the boundary of child well-being: A spatial examination of child mortality in transition zones of extreme economic inequality and material hardship. PLoS One, 20(8), Article e0330449. https://doi.org/10.1371/journal.pone.0330449
5.
Barboza-SalernoG. E.SteinkeH.MeshelemiahJ. C. A.StanekC.DuhanyS.CashS. (2024). A multilevel analysis of individual and community factors associated with case dispositions following child maltreatment investigations. Child Maltreatment, 30(1), Article 10775595241258452. https://doi.org/10.1177/10775595241258452
6.
BatraE. K.PalusciV. J.BergA. (2022). Factors associated with child maltreatment fatality among young children with an open child protective services case at death. Child Abuse Review, 31(2), car.2734. https://doi.org/10.1002/car.2734
7.
BatraE. K.QuinlanK.PalusciV. J.NeedelmanH.CollierA.AAP SECTION ON CHILD DEATH REVIEW AND PREVENTIONCOUNCIL ON INJURY, VIOLENCE, AND POISON PREVENTIONCOUNCIL ON CHILD ABUSE AND NEGLECT. (2024). Child fatality review. Pediatrics, 153(3), Article e2023065481. https://doi.org/10.1542/peds.2023-065481
8.
BengtssonH. (2021). A unifying framework for parallel and distributed processing in R using futures. The R Journal, 13(2), 208–227. https://doi.org/10.32614/RJ-2021-048
9.
BenoitK.WatanabeK.WangH.NultyP.ObengA.MüllerS.MatsuoA. (2018). Quanteda: An R package for the quantitative analysis of textual data. Journal of Open Source Software, 3(30), 774. https://doi.org/10.21105/joss.00774
10.
BironD.ReynaldD. (2016). Developing a revised typology of child homicide. Children Australia, 41(1), 8–15. https://doi.org/10.1017/cha.2015.18
11.
BlumenthalA.MartinK. A. (2023). Boys at risk beyond Male fragility: The gender gap in child injury deaths in the United States, 2000–2020. Center for Open Science. https://osf.io/download/655d626b874c2e2ec14e80f1/
12.
BourgetD.BradfordJ. M. W. (1990). Homicidal parents. Canadian Journal of Psychiatry. Revue canadienne de psychiatrie, 35(3), 233–238. https://doi.org/10.1177/070674379003500306
CunninghamR. M.WaltonM. A.CarterP. M. (2018). The Major causes of death in children and adolescents in the United States. New England Journal of Medicine, 379(25), 2468–2475. https://doi.org/10.1056/NEJMsr1804754
16.
DeanJ. R.KaczorK.LorenzD.MasonM.SimontonK. (2024). Characteristics of child abuse fatalities: Insights from a statewide violent death reporting system. Child Abuse & Neglect, 149, Article 106649. https://doi.org/10.1016/j.chiabu.2024.106649
17.
DietterichT. G.KongE. B. (1995). Machine learning bias, statistical bias, and statistical variance of decision tree algorithms.
18.
D’OrbánP. T. (1979). Women who kill their children. The British Journal of Psychiatry, 134(6), 560–571.
19.
DouglasE. M. (2017). Child death review teams. In DouglasE. M. (Ed.), Child maltreatment fatalities in the United States (pp. 71–89): Springer. https://doi.org/10.1007/978-94-017-7583-0_5
20.
DouglasE. M.LeeK. A. (2020). Challenges in determining child maltreatment fatalities: What do we really know?Child Abuse Review, 29(6), 505–517. https://doi.org/10.1002/car.2642
21.
DouglasE. M.MohnB. L. (2014). Fatal and non-fatal child maltreatment in the US: An analysis of child, caregiver, and service utilization with the National Child Abuse and Neglect Data Set. Child Abuse & Neglect, 38(1), 42–51. https://doi.org/10.1016/j.chiabu.2013.10.022
22.
EhrenthalD. B.KuoH.-H. D.KirbyR. S. (2020). Infant mortality in rural and nonrural counties in the United States. Pediatrics, 146(5), Article e20200464. https://doi.org/10.1542/peds.2020-0464
23.
FeinererI.HornikK.MeyerD. (2008). Text mining infrastructure in R. Journal of Statistical Software, 25(5), 1–54. https://doi.org/10.18637/jss.v025.i05
24.
FowlerP. J.HenryD. B.SchoenyM.LandsverkJ.ChaviraD.TaylorJ. J. (2013). Inadequate housing among families under investigation for child abuse and neglect: Prevalence from a national probability sample. American Journal of Community Psychology, 52(1–2), 106–114. https://doi.org/10.1007/s10464-013-9580-8
25.
GoldstickJ. E.CunninghamR. M.CarterP. M. (2022). Current causes of death in children and adolescents in the United States. New England Journal of Medicine, 386(20), 1955–1956. https://doi.org/10.1056/NEJMc2201761
JennyC.IsaacR. (2006). The relation between child death and child maltreatment. Archives of Disease in Childhood, 91(3), 265–269. https://doi.org/10.1136/adc.2004.066696
28.
JungK.KimH.LeeE.ChoiI.LimH.LeeB.ChoiB.KimJ.KimH.HongH.-G. (2020). Cluster analysis of child homicide in South Korea. Child Abuse & Neglect, 101, Article 104322. https://doi.org/10.1016/j.chiabu.2019.104322
KimB.MerloA. V.KimY. (2024). Intrafamilial child homicide: An umbrella review of systematic reviews. Journal of Family Violence, 39(6), 995–1012. https://doi.org/10.1007/s10896-023-00565-z
31.
KimD. (2019). Social determinants of health in relation to firearm-related homicides in the United States: A nationwide multilevel cross-sectional study. PLoS Medicine, 16(12), Article e1002978. https://doi.org/10.1371/journal.pmed.1002978
32.
KimM.-H.SubramanianS. V.KawachiI.KimC.-Y. (2007). Association between childhood fatal injuries and socioeconomic position at individual and area levels: A multilevel study. Journal of Epidemiology & Community Health, 61(2), 135–140. https://doi.org/10.1136/jech.2006.047738
33.
KoballA. M.RasmussenC.Olson-DorffD.KlevanJ.RamirezL.DomoffS. E. (2019). The relationship between adverse childhood experiences, healthcare utilization, cost of care and medical comorbidities. Child Abuse & Neglect, 90, 120–126. https://doi.org/10.1016/j.chiabu.2019.01.021
34.
KrugmanS. D.LaneW. G. (2022). Fatal child abuse. In KrugmanR. D.KorbinJ. E. (Eds.), Handbook of child maltreatment (pp. 165–180). Springer International Publishing. https://doi.org/10.1007/978-3-030-82479-2_7
35.
KulshresthaS.DligachD.JoyceC.BakerM. S.GonzalezR.O’RourkeA. P.GlazerJ. M.SteyA.KruserJ. M.ChurpekM. M.AfsharM. (2021). Prediction of severe chest injury using natural language processing from the electronic health record. Injury, 52(2), 205–212. https://doi.org/10.1016/j.injury.2020.10.094
36.
LambertA. B. E.Shapiro-MendozaC. K.ParksS. E.CottengimC.FaulknerM.HauckF. R. (2024). Characteristics of sudden unexpected infant deaths on shared and nonshared sleep surfaces. Pediatrics, 153(3), Article e2023061984. https://doi.org/10.1542/peds.2023-061984
37.
LawrenceR. (2004). Understanding fatal assault of children: A typology and explanatory theory. Children and Youth Services Review, 26(9), 837–852. https://doi.org/10.1016/j.childyouth.2004.02.024
38.
LeeS. H.LeeC. H.HwangS. H.KangD. H. (2022). A machine learning–based prognostic model for the prediction of early death after traumatic brain injury: Comparison with the corticosteroid randomization after significant head injury (CRASH) model. World Neurosurgery, 166, e125–e134. https://doi.org/10.1016/j.wneu.2022.06.130
39.
LittletonT.FreisthlerB.BoydR.SmithA. M.Barboza-SalernoG. (2024). Historical redlining, neighborhood disadvantage, and reports of child maltreatment in a large urban county. Child Abuse & Neglect, 156, Article 107011. https://doi.org/10.1016/j.chiabu.2024.107011
Maguire-JackK.MarçalK. E. (2022). The mediating role of housing insecurity in the relationship between neighborhood social cohesion and child maltreatment among low-income urban families. Child Abuse & Neglect, 132, Article 105792. https://doi.org/10.1016/j.chiabu.2022.105792
42.
MantellD. M.ScharlachK. M.IyerS. K.ChergouH. M. (2021). State-by-State analysis of child fatality reporting: A 21-Year review. International Journal on Child Maltreatment: Research, Policy and Practice, 3(4), 409–419. https://doi.org/10.1007/s42448-020-00059-1
43.
MatsuoK.AiharaH.NakaiT.MorishitaA.TohmaY.KohmuraE. (2020). Machine learning to predict In-Hospital morbidity and mortality after traumatic brain injury. Journal of Neurotrauma, 37(1), 202–210. https://doi.org/10.1089/neu.2018.6276
44.
McClureR.KeglerS.DaveyT.ClayF. (2015). Contextual determinants of childhood injury: A systematic review of studies with multilevel analytic methods. American Journal of Public Health, 105(12), e37–e43. https://doi.org/10.2105/AJPH.2015.302883
MillarJ. A.DaoH. D. N.StefopulosM. E.EstevamC. G.Fagan-GarciaK.TaftD. H.ParkC.AlruwailyA.DesaiA. N.MajumderM. S. (2021). Risk factors for increased COVID-19 case-fatality in the United States: A county-level analysis during the first wave. PLoS One, 16(10), Article e0258308. https://doi.org/10.1371/journal.pone.0258308
47.
MiyamotoS.RomanoP. S.Putnam-HornsteinE.ThurstonH.DharmarM.JosephJ. G. (2017). Risk factors for fatal and non-fatal child maltreatment in families previously investigated by CPS: A case-control study. Child Abuse & Neglect, 63, 222–232. https://doi.org/10.1016/j.chiabu.2016.11.003
48.
MokdadA. A.WolfL. L.PandyaS.RyanM.QureshiF. G. (2020). Road traffic accidents and disparities in child mortality. Pediatrics, 146(5), Article e20193009. https://doi.org/10.1542/peds.2019-3009
49.
MorrongielloB. A.HoggK. (2004). Mothers’ reactions to children misbehaving in ways that can lead to injury: Implications for gender differences in children’s risk taking and injuries. Sex Roles, 50(1–2), 103–118. https://doi.org/10.1023/B:SERS.0000011076.43831.a6
50.
NasejjeJ. B.MwambiH. G.AchiaT. N. O. (2015). Understanding the determinants of under-five child mortality in Uganda including the estimation of unobserved household and community effects using both frequentist and Bayesian survival analysis approaches. BMC Public Health, 15(1), 1003. https://doi.org/10.1186/s12889-015-2332-y
OruçM.DündarA. S.OkumuşH.GörmezM.ŞamdancıE. T.CelbişO. (2021). Shaken baby syndrome resulting in death: A case series. Turkish Journal of Pediatrics, 63(1), 31–36. https://doi.org/10.24953/turkjped.2021.01.004
PalusciV. J.CovingtonT. M. (2014). Child maltreatment deaths in the U.S. National Child Death Review Case Reporting System. Child Abuse & Neglect, 38(1), 25–36. https://doi.org/10.1016/j.chiabu.2013.08.014
58.
PalusciV. J.SchnitzerP. G.CollierA. (2023). Social and demographic characteristics of child maltreatment fatalities among children ages 5–17 years. Child Abuse & Neglect, 136(1), Article 106002. https://doi.org/10.1016/j.chiabu.2022.106002
59.
ParkerS. T. (2020). Estimating nonfatal gunshot injury locations with natural language processing and machine learning models. JAMA Network Open, 3(10), e2020664. https://doi.org/10.1001/jamanetworkopen.2020.20664
PhillipsJ. A.DavidsonT. R.Baffoe-BonnieM. S. (2023). Identifying latent themes in suicide among black and white adolescents and young adults using the National Violent Death Reporting System, 2013–2019. Social Science & Medicine, 334(1), Article 116144. https://doi.org/10.1016/j.socscimed.2023.116144
62.
Putnam-HornsteinE.NeedellB.KingB.Johnson-MotoyamaM. (2013). Racial and ethnic disparities: A population-based examination of risk factors for involvement with child protective services. Child Abuse & Neglect, 37(1), 33–46. https://doi.org/10.1016/j.chiabu.2012.08.005
ResnickP. J. (1969). Child murder by parents: A psychiatric review of filicide. American Journal of Psychiatry, 126(3), 325–334. https://doi.org/10.1176/ajp.126.3.325
65.
RichardsC. E. (2000). Introduction. In The loss of innocents: Child killers and their victims (pp. xi–xxvi). Scholarly Resources.
66.
RoachJ.BryantR. (2015). Child homicide: Generating victim and suspect risk profiles. Journal of Criminal Psychology, 5(3), 201–215. https://doi.org/10.1108/jcp-04-2015-0013
67.
RobertsD. (2009). Shattered bonds: The color of child welfare. Basic Books.
68.
RobertsM. E.StewartB. M.TingleyD. (2019). stm: An R package for structural topic models. Journal of Statistical Software, 91(2), 1–40. https://doi.org/10.18637/jss.v091.i02
69.
RobertsM. E.StewartB. M.TingleyD.LucasC.Leder‐LuisJ.GadarianS. K.AlbertsonB.RandD. G. (2014). Structural topic models for open‐ended survey responses. American Journal of Political Science, 58(4), 1064–1082. https://doi.org/10.1111/ajps.12103
70.
SchnitzerP. G.CovingtonT. M.DykstraH. K. (2012). Sudden unexpected infant deaths: Sleep environment and circumstances. American Journal of Public Health, 102(6), 1204–1212. https://doi.org/10.2105/AJPH.2011.300613
71.
SchnitzerP. G.CovingtonT. M.WirtzS. J.Verhoek-OftedahlW.PalusciV. J. (2008). Public health surveillance of fatal child maltreatment: Analysis of 3 state programs. American Journal of Public Health, 98(2), 296–303. https://doi.org/10.2105/AJPH.2006.087783
72.
SchnitzerP. G.EwigmanB. G. (2005). Child deaths resulting from inflicted injuries: Household risk factors and perpetrator characteristics. Pediatrics, 116(5), e687–e693. https://doi.org/10.1542/peds.2005-0296
73.
SharonC.RousseauS. (2024). A machine learning approach to identifying non-parental caregivers’ risk for harsh caregiving towards infants in daycare centers. Early Childhood Research Quarterly, 67(1), 128–138. https://doi.org/10.1016/j.ecresq.2023.12.006
SilvaJ. A.LeongG. B.DassoriA.FerrariM. M.WeinstockR.YamamotoJ. (1998). A comprehensive typology for the biopsychosociocultural evaluation of child-killing behavior. Journal of Forensic Sciences, 43(6), 1112–1118.
StickleyA.BaburinA.JasilionisD.KruminsJ.MartikainenP.KondoN.ShinJ. I.InoueY.LeinsaluM. (2023). Sociodemographic inequalities in mortality from drowning in the Baltic countries and Finland in 2000–2015: A register-based study. BMC Public Health, 23(1), 1103. https://doi.org/10.1186/s12889-023-15999-9
78.
StiffmanM. N.SchnitzerP. G.AdamP.KruseR. L.EwigmanB. G. (2002). Household composition and risk of fatal child maltreatment. Pediatrics, 109(4), 615–621. https://doi.org/10.1542/peds.109.4.615
79.
TirunagariS. (2015). Data mining of causal relations from text: Analysing maritime accident investigation reports(No. arXiv:1507.02447). arXiv. https://arxiv.org/abs/1507.02447
80.
TixierA. J.-P.HallowellM. R.RajagopalanB.BowmanD. (2016). Automated content analysis for construction safety: A natural language processing system to extract precursors and outcomes from unstructured injury reports. Automation in Construction, 62(1), 45–56. https://doi.org/10.1016/j.autcon.2015.11.001
81.
United States Senate Committee on Finance. (2024). Warehouses of neglect: How taxpayers are funding systemic abuse in youth residential treatment facilities.
VictorB. G.PerronB. E.SokolR. L.FedinaL.RyanJ. P. (2021). Automated identification of domestic violence in written child welfare records: Leveraging text mining and machine learning to enhance social work research and evaluation. Journal of the Society for Social Work and Research, 12(4), 631–655. https://doi.org/10.1086/712734
85.
WebsterR. A.SchnitzerP. G.JennyC.EwigmanB. G.AlarioA. J. (2003). Child death review: The state of the nation. American Journal of Preventive Medicine, 25(1), 58–64. https://doi.org/10.1016/S0749-3797(03)00091-6
86.
Weigend VargasE.StallworthP.CarterP. M.GoldstickJ. E. (2024). State-to-State variation in rates and causes of child and adolescent mortality in the US. JAMA Pediatrics, 178(11), e242894. https://doi.org/10.1001/jamapediatrics.2024.2894
87.
WestB. A.RuddR. A.Sauber-SchatzE. K.BallesterosM. F. (2021). Unintentional injury deaths in children and youth, 2010–2019. Journal of Safety Research, 78(1), 322–330. https://doi.org/10.1016/j.jsr.2021.07.001
88.
WestonS. J.ShryockI.LightR.FisherP. A. (2023). Selecting the number and labels of topics in topic modeling: A tutorial. Advances in Methods and Practices in Psychological Science, 6(2), Article 25152459231160105. https://doi.org/10.1177/25152459231160105
WilsonR. F.MintzS.BlairJ. M.BetzC. J.CollierA.FowlerK. A. (2023). Unintentional firearm injury deaths among children and adolescents aged 0–17 years—National Violent Death Reporting System, United States, 2003–2021. MMWR. Morbidity and Mortality Weekly Report, 72(50), 1338–1345. https://doi.org/10.15585/mmwr.mm7250a1
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
