Author Affiliations: Department of Pediatrics (Drs Hewes, Keenan, McDonnell, Dudley, and Herman) and S. J. Quinney College of Law (Dr McDonnell), University of Utah, Salt Lake City.
Objectives To determine whether convictions and sentencing differ between child abuse homicide cases and adult homicide cases and to identify characteristics of the victim, suspect, or crime that influence conviction and sentencing results.
Design Retrospective case review.
Setting Homicide data abstracted from the National Violent Death Reporting System in Utah.
Participants All deaths classified as homicide in Utah between January 1, 2002, and December 31, 2007.
Main Exposure Judicial processing of homicide cases for conviction and sentencing results.
Main Outcome Measures Conviction rate, level of felony conviction, and severity of sentencing for suspects of child abuse homicide vs adult homicide.
Results Utah had 373 homicide victims during the study period; 52 cases were child abuse homicide. Among 211 homicide cases with an identified suspect, conviction rates for child abuse homicide (88.2%) and adult homicide (83.0%) were similar (risk ratio, 1.0; 95% confidence interval [CI], 0.8-1.4). There were no significant differences in level of felony conviction (adjusted risk ratio, 0.8; 95% CI, 0.4-1.3) or severity of sentencing (adjusted risk ratio, 0.8; 95% CI, 0.5-1.5) for suspects of child abuse homicide vs adult homicide. Among child abuse homicide cases, no demographic factor was significantly associated with felony conviction results.
Conclusion Suspects of child abuse homicide are convicted at a rate similar to that of suspects of adult homicide and receive similar levels of felony conviction and severity of sentencing.
Homicide ranks as one of the top 5 causes of childhood death in the United States.1 The National Child Abuse and Neglect Data System reported 1740 child fatalities in 2008 resulting from child abuse and neglect, representing a fatality rate from abuse and neglect of 2.33 per 100 000 children in the general population.1 This rate likely underestimates the problem, as the number of child deaths caused by abuse or neglect has been documented to be underreported by as much as 50% to 60%.1- 3
Child abuse homicide is defined as an act or a failure to act on the part of a guardian that results in a child's death. The relative rates of conviction and severity of sentencing for suspects of adult and child abuse homicide are unknown. Data vary on the judicial outcomes of criminal child physical and sexual abuse cases. Literature comparing sexual crimes against children vs other adult felonies demonstrates that in some cases the judicial experience is similar, while other studies find that suspects convicted of child sexual abuse receive lesser penalties than suspects of similar adult violent crimes.4- 6 A study7 of documented physical child abuse found that only 40% of cases resulted in criminal prosecution. Similarly, studies8,9 of the prosecution of abusive head trauma cases have found that charges and sentencing vary widely for this type of physical abuse. Sentencing of convicted suspects of abusive head trauma is inconsistent, with sentences ranging from months to several years.9 It is unknown if the judicial system is more consistent when the crime involves a child's death.
To understand whether the judicial system handles child abuse homicide cases differently than adult homicide cases, we undertook a review of the conviction and sentencing results among all suspects of homicide in Utah. We hypothesized that suspects of child abuse homicide are convicted less often than suspects of adult homicide and that suspects receive less severe sentences than those committing adult homicide. We also hypothesized that suspect characteristics would be predictive of sentencing decisions.
We performed a retrospective review of all homicide cases in Utah between January 1, 2002, and December 31, 2007, using the following 3 data sources: the Utah Department of Health, the Utah state information from the National Violent Death Reporting System (NVDRS) of the Centers for Disease Control and Prevention, and the Utah Commission on Criminal and Juvenile Justice. First, we identified homicide cases from death certificates by the Utah Department of Health. Second, we abstracted data concerning each homicide case from the NVDRS to identify suspects, to gain knowledge about the circumstances of the crime, and to obtain demographic information about suspects and victims, including age, sex, race/ethnicity, relationship of victim to suspect, and weapon type used in the crime. The NVDRS database compiles information concerning violent deaths from death certificates, crime laboratories, and other reports, including police, medical examiner, and supplementary homicide data. Third, we obtained conviction and sentencing information of suspects identified from the NVDRS, as well as any known prior suspect felony conviction history in Utah, from the Utah Commission on Criminal and Juvenile Justice. The matching of these 3 sources resulted in a complete data set containing deidentified demographic information about suspects, victims, crimes, and judicial outcomes using unique case numbers for each homicide case. This study was reviewed and approved by the institutional review board of the University of Utah and the Utah Department of Health.
The Utah State Code10 defines a child as a person younger than 18 years. Homicide cases were categorized as child homicide (victim <18 years) or as adult homicide. Child homicide cases were subcategorized as child abuse homicide or as non–child abuse homicide using the definition of child abuse homicide from the Utah State Code. The Utah State Code states that criminal homicide constitutes child abuse homicide if, under circumstances not amounting to aggravated murder, the actor causes the death of a person younger than 18 years and the death results from child abuse. The Utah State Code holds all caretakers, not just legal guardians, accountable for child abuse homicide.
Homicide is an umbrella term for several types of inflicted death. The most severe felony charge for homicide is a capital crime, which carries the possibility of a death sentence. The next most severe charge is a felony 1. A felony 1 conviction usually results in a penalty with the possibility of life in prison. Capital crimes and felony 1 crimes are frequently associated with acts of aggression and premeditation. Felony 2 and felony 3 charges are progressively less severe than felony 1. Child abuse homicide also falls under the umbrella of criminal homicide. It can be considered a capital crime or a felony 1 if the child abuse homicide meets those specific criteria or can be classified as a felony 2 or a felony 3. Felony convictions for homicide in Utah carry prison sentencing guidelines but no minimum prison sentences; therefore, judicial discretion is allowed in sentencing decisions (eFigure 1).
Descriptive statistics were used to characterize the case flow of all homicide victims and demographics of the victims and suspects. Medians and interquartile ranges were calculated for continuous variables and were compared using the Wilcoxon test of nonnormally distributed values. Categorical data were compared, and risk ratios (RRs) and 95% confidence intervals (CIs) were calculated. Two Poisson regression models were built. The first model calculated the adjusted risk of a felony 1 or higher conviction dependent on whether the case was a child abuse homicide or an adult homicide. The second model calculated the adjusted risk of a high sentence (a sentence carrying the possibility of life in prison) dependent on whether the case was a child abuse homicide or an adult homicide. All covariates, including victims' sex and race/ethnicity and suspects' sex, race/ethnicity, and prior felony conviction, were entered into the models. Covariates that did not alter the estimate by at least 10% were removed from the model in a backward stepwise selection.
Utah had 373 homicide victims during the study period. Thirty-nine (10.5%) were ruled as justifiable homicide because the decedent was killed by an active-duty police officer. These cases were excluded from further analysis. The remaining 334 homicide victims comprised the study cohort.
Sixty-six of 334 cases (19.8%) were child homicide. Fifty-two cases were considered child abuse homicide according to Utah statutes, and the remaining 14 cases were primarily gang-related deaths of children perpetrated by children. Therefore, 15.6% of all nonjustifiable homicides in Utah were child abuse homicide. Eleven of 52 child abuse homicide cases were not under Utah state jurisdiction, so felony type and sentencing information were unavailable. Of the remaining 41 child abuse homicide cases, 39 had a suspect identified through the NVDRS. Five of these suspects committed suicide before legal proceedings, leaving 34 victim-suspect pairs for analysis. Conviction occurred in 30 of 34 pairs (88.2%).
There were 268 adult homicide cases. Of these, 172 (64.2%) had a suspect identified in the NVDRS. Thirty-seven suspects committed suicide before legal proceedings, leaving 135 victim-suspect pairs for analysis. Among these, 112 convictions (83.0%) occurred (eFigure 2).
Child abuse homicide suspects and adult homicide suspects were compared for conviction rate, level of felony conviction, and severity of sentencing. Among 211 homicide cases with an identified suspect, conviction rates for child abuse homicide (88.2%) and adult homicide (83.0%) were similar (RR, 1.0; 95% CI, 0.8-1.4). Among child abuse homicide cases, no demographic factor of the victim or suspect was significantly associated with level of felony conviction. Among adult homicide cases, sex of the victim was the only covariate significantly associated with level of felony conviction. A suspect who killed an adult female victim was significantly more likely to receive a felony 1 or higher conviction (RR, 1.5; 95% CI, 1.2-2.0). The risk of a child abuse homicide suspect's receiving a capital crime or felony 1 conviction was approximately half that of an adult homicide suspect (RR, 0.6; 95% CI, 0.3-1.1) (Table 1). After adjusting for sex and race/ethnicity of the victim and sex of the suspect, the risk remained similar (adjusted RR, 0.8; 95% CI, 0.4-1.3). Prior felony conviction and race/ethnicity of the suspect did not change the model estimate by 10%. Similarly, the risk of a child abuse homicide suspect's receiving a sentence that included the possibility of life in prison was less than that of an adult homicide suspect (RR, 0.7; 95% CI, 0.4-1.3). After adjusting for sex of the victim, sex and race/ethnicity of the suspect, and prior felony conviction, the adjusted RR remained similar (adjusted RR, 0.8; 95% CI, 0.5-1.5). Race/ethnicity of the victim did not change the model estimate by 10%.
In cases with a conviction (30 child abuse homicides and 112 adult homicides), victims of child abuse homicide tended to be young, male, and of white race/ethnicity. The mechanism of injury was most frequently (36.7%) abusive head trauma (Table 2). Suspects of child abuse homicide were most frequently male, the victim's parent, and of white race/ethnicity.
Our study found that suspects of child abuse homicide are convicted and sentenced at a rate similar to that of suspects of adult homicide in Utah. These results represent one of the first comparisons of suspects of child abuse homicide vs adult homicide. The findings add a new perspective to the literature, in which much of the prior work has focused on sexual abuse or mixed abuse judicial outcomes.
The law recognizes all child abuse homicide as criminal. However, some believe that suspects of child abuse homicide lack intentionality, in contrast to suspects of adult homicide.11- 13 Because intentionality is an important legal distinction in determining how a suspect is charged, one would hypothesize that suspects of child abuse homicide would face less severe penalties. Contrary to this, we found that sentencing and conviction results that reflect intentionality were similar in child abuse homicide and adult homicide cases. This may reflect a societal view that values a child's life equally to an adult's life.
Violent crimes against children have been reported to be less likely to incur felony charges and convictions compared with other types of adult violent crimes.6,12 In a meta-analysis of criminal justice decisions in 21 studies of prosecutions of sexual or mixed sexual and physical abuse, Cross et al12 demonstrated that child abuse cases were less likely to lead to charges being filed and incarceration compared with most other violent felonies. Once charges were brought forward, conviction rates were similar to those of other violent crimes, and the mean conviction rate (94%) was higher than our conviction rate of 88.2%. Cross et al found large variation in incarceration rates, suggesting no consistent standard in incarceration of suspects of child abuse. In contrast, we found no significant difference in level of felony conviction for suspects of child abuse homicide vs adult homicide, and we found similar severity of sentencing. Our data may be unique in that the severity of injury renders the adult and child cases more similar compared with other studies looking at diverse crimes with resultant varied severity of injury.
We found no demographic characteristics of the victim or suspect that predicted severity of sentencing in the child abuse homicide group, including race/ethnicity of the victim and prior felony conviction of the suspect. Unlike the North Carolina study by Keenan et al,8 we did not find that race/ethnicity influenced severity of sentencing. Our study may be limited by the homogeneity of the Utah population, which is primarily of white race/ethnicity, with a 19% minority population that is 12% Hispanic.14 Similarly, the Utah population may differ from that of other states in that most of our child homicide deaths were young children who were victims of child abuse homicide rather than older adolescents involved in gang violence. Other states with higher urban densities may have a different demographic composition of homicide cases.
The limitations and strengths of our analysis are several. This study reports data from a single state, limiting sample size and generalizability of the results. The use of data from a single state is also a strength, as it allowed us to combine multiple data sets for more complete information concerning the victims and suspects. In addition, because Utah holds all caretakers responsible for a child and not just the parents, we have a complete accounting of child abuse homicide suspects.
Like all retrospective studies, our results have limitations based on the existing database. Our findings were limited by the information available to the Utah Department of Health to compile the NVDRS database and by the data provided by the Utah Commission on Juvenile Justice for the criminal information. This study was restricted to complete analysis only of homicide cases that had a convicted suspect. In many cases, the suspect was not convicted because a suspect was not identified, because formal charges were not filed, or because the suspect committed suicide. This limited our ability to analyze and compare all homicide cases. While our cohort represents an entire state's experience for 6 years, the study size limits the precision of the estimates.
It is possible that some child abuse homicide cases may have been misclassified as accidental deaths and may not have been brought to the attention of law enforcement. Such misclassification would have led us to underestimate the number of child abuse homicide cases. However, Utah is a state with a strong child fatality review team, which minimizes the number of misclassified cases.
In conclusion, child abuse homicide accounted for most child homicide cases in Utah between 2002 and 2007. Suspects in child abuse homicide cases were convicted as often as suspects in adult homicide cases, and no statistically significant differences were noted in the level of felony conviction or severity of sentencing between the 2 groups. Under the Utah criminal justice system, suspects of child abuse homicide are not treated differently than suspects of adult homicide. Much could be learned by replicating this study in other states and by comparing data across state legislations for conviction and sentencing outcomes.
Correspondence: Hilary A. Hewes, MD, Department of Pediatrics, University of Utah, 295 Chipeta Way, PO Box 581289, Salt Lake City, UT 84158 (Hilary.firstname.lastname@example.org).
Accepted for Publication: April 21, 2011.
Author Contributions:Study concept and design: Hewes, McDonnell, and Herman. Acquisition of data: Hewes. Analysis and interpretation of data: Hewes, Keenan, McDonnell, and Dudley. Drafting of the manuscript: Hewes, Keenan, and Dudley. Critical revision of the manuscript for important intellectual content: Hewes, Keenan, McDonnell, Dudley, and Herman. Statistical analysis: Hewes and Keenan. Administrative, technical, and material support: Hewes, Keenan, Dudley, and Herman. Study supervision: Keenan, McDonnell, Dudley, and Herman.
Financial Disclosure: None reported.
Additional Contributions: Anna Fondario, MPH; Teresa Brechlin, Cristy Sneddon, Gayla Whitaker, and Katie McMinn supported this project. Christine Mitchell and Jennifer Hemenway provided criminal data and legal advice.
Hewes HA, Keenan HT, McDonnell WM, Dudley NC, Herman BE. Judicial Outcomes of Child Abuse Homicide. Arch Pediatr Adolesc Med. 2011;165(10):918–921. doi:10.1001/archpediatrics.2011.151
* * SCHEDULED MAINTENANCE * *
The JAMA Network Sites will be conducting routine maintenance from 10/20/2017 through 10/21/2017. During this window access to content and authentication may be intermittently available. The JAMA Store will be completely unavailable during the maintenance window.