The Relationship Between Temperature and Temporal Patterns and Incidence of Abusive Head Trauma in a Midwest Region Hospital

Abstract

Background: Abusive head trauma (AHT) is a leading cause of death and disability in children and one of the most lethal forms of child abuse. Most known risk factors for AHT pertain to the infant’s caregiver and limited research has assessed external influences beyond the familial or caregiver/infant dyad. Objective: Our primary objective was to determine if temperature patterns are associated with AHT events. Secondary outcomes included associations between AHT and specific days of the week, months, or seasons. Methods: This was a retrospective review of 198 patients under 24 months old who were diagnosed with AHT at Saint Louis Children’s Hospital. Demographic information was obtained from the medical record for each patient. For each AHT incident, the date and zip code of the incident were recorded. Temperature on the date of incident was identified using the Midwestern Regional Climate Center (MRCC). Chi square tests were utilized to calculate differences in cases per year as well as temperature and seasonal variation. Results: Temperature was not associated with a statistically significant increase in cases of AHT. There was an increase in cases as temperatures rose, but no statistically significant associations between incidence of AHT and day of the week, month, or season. Conclusion: Our study suggests no significant association between AHT incidence and temperature or temporal patterns in this Midwest hospital.

Keywords

abusive head trauma child abuse infants injury prevention non-accidental trauma trauma

Introduction

Abusive head trauma (AHT) is one of the leading causes of death and disability in young children and one of the most lethal forms of child abuse (Barlow & Minns, 2000). The estimated annual incidence of abusive head trauma lies between approximately 20–40 cases per 100,000 person-years for children < 1 year of age in the US (Shanahan et al., 2013). Although AHT is rare relative to other form of abuse, those affected often experience severe morbidity or mortality (Hennes et al., 2001).

Many aspects of AHT have already been studied and reported in the literature. Certain characteristics of victims of AHT, such as male sex and age < 1 year, have been well replicated throughout a multitude of studies and are thought to be main risk factors of AHT (Hennes et al., 2001; Herman et al., 2011; Kesler et al., 2008; Parks et al., 2012). Studies have also noted that perpetrators are typically of male sex and are often younger, less educated, unmarried, and of lower socioeconomic status than the general population (Coulton et al., 1995; Herman et al., 2011; Kotch et al., 1995; Scribano et al., 2013).

In addition, efforts have been made to understand the external stressors that may contribute to the incidence of AHT, as these are crucial to directing interventions aimed at preventing its occurrence. Caregiver stress is thought to be a significant risk factor for abusive head trauma. Household stressors, most notably a crying infant and sleep-deprived caregiver, are thought to be a trigger for abuse (Barr et al., 2006). Strong associations have also been noted between socioeconomic stressors such as family poverty and housing instability and rates of child abuse (Berger & Brooks-Gunn, 2005; Cancian et al., 2010; Freisthler et al., 2006; Kotch et al., 1995). Economic stress not only within the family, but also to the surrounding community is also a well-documented risk factor for abuse (Berger & Brooks-Gunn, 2005; Cancian et al., 2010; Coulton et al., 1995; Drake & Pandey, 1996; Eckenrode et al., 2014; Freisthler et al., 2006; Kotch et al., 1995). Even macroeconomic factors have been shown to have an effect on physical abuse risk. A recent study examined data for children < 5 years old who were diagnosed with AHT between the years 2004 and 2012 and found that the rate of AHT was higher during and shortly after the 2007–2009 recession compared to prior to the recession (Wood et al., 2016). Environmental factors outside of economic hardship have also been shown to have an effect on the incidence of abuse. For example, increased rates of AHT and other forms of physical abuse have been observed after natural disasters (Curtis et al., 2000; Keenan et al., 2004). However, the effects of temporal, seasonal, and certain weather patterns such as temperature variations have not been studied in large populations.

It is widely known that weather can have a substantial impact on mood. This is partially thought to be mediated through one of the main parameters of weather—temperature (Denissen et al., 2008). Therefore, in areas of the Midwest that harbor large fluctuations in temperatures, weather has the potential to be a significant stressor. Several studies have attempted to correlate a temporal pattern such as weather, seasonality, or time of year to the incidence of AHT, but have only studied this phenomenon in a small cohort of subjects or have not taken into account other confounders that may also be associated (Daly & Connor, 2001; Kesler et al., 2008; Parks et al., 2012). The purpose of this study was to determine whether the incidence of abusive head trauma is associated with certain temperature patterns or with other temporal factors including day of the week, month of the year, or the presence of holidays.

Methods

This was a retrospective chart review, approved by the Institutional Review Board (IRB ID #:201609020) and carried out in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki).

Medical records of 2428 children were obtained through the Saint Louis Children’s Hospital Trauma Registry and Coding Records from Saint Louis Children’s Hospital from January 2006 to December 2015. Specific charts extracted included those of children under the age of 24 months with a discharge diagnosis of “trauma” and/or any International Classification of Disease, 9^th Edition (ICD-9) (Appendix), child abuse-related diagnosis. These charts were then reviewed by two individual researchers (LB and NM) who separately identified patients from these records who met the inclusion criteria. The specific inclusion criteria consisted of patients age 0–24 months with a diagnosis of abusive head trauma. This diagnosis was made by: 1. A confession by or witness of a perpetrator violently shaking the child or inflicting blunt impact to the head; 2. An expired patient with an autopsy report declaring death by homicide secondary to abusive head trauma; and/or 3. A child with multiple injuries suspicious for abusive head trauma (such as retinal hemorrhages or subdural hematomas) with a confirmation of the diagnosis by a child abuse pediatrician. These criteria were selected to include all patients who met the diagnosis of abusive head trauma per the Center for Disease Control (CDC), which defines abusive head trauma as “an injury to the skull or intracranial contents of an infant or young child due to inflicted blunt impact and/or violent shaking” (Parks et al., 2012). Exclusion criteria in the study consisted of patients > 24 months of age at the time of incident; patients who presented and were later diagnosed with a non-acute presentation of abusive head trauma; inability by the researchers to determine the date of the incident when the AHT took place; patients who presented with non-abusive etiologies of head trauma; and/or patients who presented with abuse but not including head trauma. Physician, nursing, and social work documentation as well as radiology reports within each chart were reviewed by both researchers to determine whether a patient met inclusion criteria. A third reviewer who is a child abuse pediatrician (JSK) then cross-referenced both reviewers’ selections with a database of all known abusive head trauma cases seen at Saint Louis Children’s Hospital during the study time period to identify the final patient cohort used in the study.

Once the final patient cohort was determined, data from each chart was extracted. This included: age of patient at diagnosis, sex, race or ethnicity, approximate date abusive head trauma occurred, zip code of location of incident, perpetrator’s relation to patient (if known), type of injuries sustained by the victim, if a consult was obtained by a child abuse pediatrician, and whether the patient died as a result of his or her injuries. Approximate date the abusive head trauma occurred was determined by either: 1. Perpetrator confession which mentioned a specific date; 2. Date given by a witness to the abuse; and/or 3. Approximate date extracted from chart information based on when patient first became symptomatic. Zip code of location of incident was similarly determined by perpetrator confession or witness statement of where incident took place or approximation based on patient’s history and known addresses listed in chart. Approximation of date and zip code was agreed upon by all three reviewers. Once the approximate date of incident was obtained, the day of the week was determined based on that year’s calendar. In addition, whether the date was within 2 days before, after, or on a holiday was noted. Days that were considered holidays included New Year’s Day (January 1), Martin Luther King Day (third Monday of January), Saint Patrick’s Day (March 17), Memorial Day (last Monday of May), Independence Day (July 4), Labor Day (first Monday of September), Thanksgiving (fourth Thursday in November), and Christmas (December 25).

The season and the temperature at the approximate date of the incident were then determined. The mean, minimum, and maximum temperature of the date of incident for each case was identified using the Midwestern Regional Climate Center (MRCC) using the data from the collection center closest to the zip code of the location of the incident. The maximum and minimum temperatures were classified into five groups: cold (≤ 32°F), cool (33–49°F), mild (50–64°F), moderate (65–79°F), or hot (≥ 80°F). These temperature categories were selected to allow for standardization amongst related published literature in this field (Daly & Connor, 2001). Temperatures were then further classified into 1. Endpoint temperatures, which consisted of the maximum and minimum daily temperature categories “cold” and “hot” together to indicate the upper and lower limit at either end; and 2. Non-endpoint temperature, which included the cool, mild, and moderate temperatures. Seasons were then classified into two groups: April through September, and October through March. The number of days over the full study period for each of the five temperature groups in this region was then obtained through the MRCC database.

We tested for differences within the population of children with AHT by categorizing all cases by their year and examining potential differences in age, race, sex, and season. We categorized age (in months) into four categories: 1) 0–6; 2) $\geq$ 6 & < 12; 3) $\geq$ 12 & < 18; and 4) $\geq$ 18 & < 24. We used two seasons instead of four because of small cell size issues—the number of cases each year categorized by four seasons resulted in no cases for some of the cells. We used chi square tests to explore differences in cases per year by race, sex, age, endpoint classification, and our two-season classification. Alpha levels were set at < 0.05 for significance testing. To investigate the relationship between cases and temperature, we computed the following rate for the study period: the number of cases in each temperature group over the number of days in each temperature group.

Data management and analysis was conducted using Stata 14 (StataCorp. 2015. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP).

Results

From a total of 2428 trauma cases, 2197 did not meet the criteria of head trauma concerning for abuse and 17 who met criteria were > 24 months of age at the time of hospitalization (Figure 1). Of the remaining 214 head trauma cases, 16 cases were described as “chronic” and date of injury could not be determined; therefore, 198 cases remained that met inclusion criteria. Of these patients, 63% were male and 62% white, with a median age of 4.3 months (Table 1). Thirty of the cases (15%) died as a result of injuries. While 77 cases did not have a confirmed perpetrator, among those that did (n = 121) 54% were perpetrated by the father and 24% by the mother’s male partner.

Figure 1.

Consort diagram demonstrating patients who were included in and excluded from the study. SLCH = St Louis Children’s Hospital.

Table 1.

Demographic Information.

Characteristic	N (%)
Sex (male)	125 (65)
Race/Ethnicity
White	123 (62)
Black	64 (32)
Hispanic	5 (3)
All other	6 (3)
Age
< 6 months	130 (65)
7–12 months	29 (15)
13–18 months	25 (13)
19–23 months	14 (7)
Died	30 (15)
Perpetrator
Father	65 (33)
Mother’s male partner	29 (14)
Mother	10 (5)
Babysitter	9 (5)
Relative	4 (2)
Foster/Other parent	4 (2)
Unknown	77 (39)

There was an average of 19.8 cases (SD = 2.9) per year over the ten-year study period. The year 2006 saw the greatest number of cases at 25, and 2013 saw the fewest at 16. The month of August across all years saw the highest number of cases (n = 23), whereas September and October across all years shared the lowest counts (n = 11 each; Figure 2). There were no statistically significant differences by season, as cases were as likely to occur in April–September (n = 106) as they were in October–March (n = 92), year by year (p = .38). Upon visual inspection, there was also a fairly even distribution of cases by day of the week (Figure 3). A total of 13 cases (7%) occurred 2 days before, after, or on a holiday. Cases were as likely to occur in endpoint temperatures than in non-endpoint temperatures, year over year (p = .67). The number of AHT cases increased steadily when categorized by their daily max temperature. When determining absolute number of cases in each temperature category over the ten-year period, there was a steady increase in number of cases as the temperature rose (Figure 4). The total number of abusive head trauma cases in each temperature category was also compared to the overall total number of days with average temperatures in each category in St Louis region between 2006–2015 (Figure 4). Taking the total days into account, however, the rates of cases in each temperature category were similar. The largest rate was 5.91 for hot cases compared to the smallest rate of 5.80 for cool cases (Table 2).

Figure 2.

Total number of cases of abusive head trauma by month of year at Saint Louis Children’s Hospital between 2006–2015.

Figure 3.

Total number of abusive head trauma cases by day of week at Saint Louis Children’s Hospital between 2006–2015.

Figure 4.

Comparison of total abusive head trauma cases by temperature to overall temperature pattern in Saint Louis region in 2006–2015. Bar graph demonstrates the number of total days St Louis experienced in each temperature category, categorized using maximum temperature noted on each day between the years 2006–2015. Definitions of each temperature category: cold (≤ 32°F), cool (33–49°F), mild (50–64°F), moderate (65–79°F), or hot (≥ 80°F). Line graph demonstrates total number of abusive head trauma cases that occurred in each temperature category at Saint Louis Children’s Hospital between 2006–2015.

Table 2.

Comparison of Rates of Abusive Head Trauma Cases Over Total Days in Each Temperature Category Between 2006 and 2015.

	Number of Days During Study Period	Number of Cases (Max Temperature)	Rate of Cases
Cold (< 32)	153	9	5.9
Cool (33–49)	586	34	5.8
Mild (50–64)	649	38	5.9
Moderate (65–79)	765	45	5.9
Hot (> 80)	1218	72	5.9

The rates of abusive head trauma for each temperature category were obtained by dividing the total cases by the number of days that fell into each temperature category over the specified time period.

There were no significant differences in sex, age, or race/ethnicity by year or by season.

Discussion

This was a retrospective study evaluating for the presence of any seasonal or temporal patterns to the incidence of AHT to determine whether these may be external stressors contributing to the occurrence of AHT.

Cohort Demographics

The demographics in this study overall matched those of previous studies. 63% of patients in our cohort were male, which is consistent with the known male predominance between 54% and 66% (Hennes et al., 2001; Kelly et al., 2015; Scribano et al., 2013; Starling et al., 2004). Racial demographics were also largely similar, with a majority of patients identified as Caucasian (Scribano et al., 2013; Wood et al., 2016). The city of St. Louis has a mixture of African American and Caucasian predominance, and the surrounding areas are of predominantly Caucasian population. 75% of patients with AHT in this cohort were less than 1 year of age and there was a 15% mortality rate amongst all patients, also similar to previously documented observations (Scribano et al., 2013; Wood et al., 2016).

Seasonality

Previous studies have examined the possibility of a seasonal association with deadly violence against children and have thus far not found a correlation. A study investigating the incidence of homicide deaths in children < 5 years old in Indiana, Ohio, Missouri, Oklahoma, and Washington between the years 1999 and 2006 identified no seasonality by month of year or season (Laskey et al., 2010). Similarly, a recent study found no seasonal variation in deaths due to homicide in children < 19 years old in Japan between the years 2000–2010 (Shinsugi et al., 2015). With regard to fatal and non-fatal AHT specifically, previous studies attempting to correlate seasonality with its incidence have had mixed results. A 2008 study examining the demographics of AHT in Pennsylvania found a seasonal effect with significantly more cases occurring during the winter holiday months of October through December (Kesler et al., 2008). Alternatively, a study by the CDC found that non-fatal AHT rates among children < 5 years old showed no variation across seasons in the US between the years 2003–2008 (Parks et al., 2012). Additionally, a study examining 29 cases of shaken baby syndrome presenting to Rainbow Babies and Children’s Hospital in Cleveland, Ohio between the years 1995 and 1999 found an increased incidence of AHT in winter months, but this was not statistically significant (Daly & Connor, 2001).

In this current study, the month of August saw the greatest number of cases (n = 23 across the 10 years) while September and October had the fewest cases (n = 11 across the 10 years), however, there was overall no statistically significant difference between the incidence of AHT by month. There was also no difference between the incidence of AHT during the months of April–September compared to October–May. Overall, this study suggests that unlike the common myth amongst healthcare providers that winter months tend to harbor “shaken baby season,” as referenced by Daly & Connor, AHT is in fact prevalent and should be highly suspected in all months and seasons of the year in this Midwest region and regions with similar temperature patterns.

Temperature

This study also aimed to identify specific factors that could influence rates of AHT and account for any location-specific patterns to its incidence. One environmental factor that differs based on location is weather and temperature. Previous reports have hypothesized that extreme cold weather could cause the caregiver and infant to be confined indoors for long periods of time and potentially cut them off from outside sources of support, therefore increasing the risk of shaken baby syndrome (Daly & Connor, 2001). In this cohort, there was no significant difference between the number of cases that occurred in endpoint temperatures compared to in non-endpoint temperatures, year over year. The number of AHT cases did increase as the daily maximum temperature increased, and the majority of cases that occurred in the spring and summer months occurred on days with a maximum temperature of > 80°F. However, the calculated rates of cases in each temperature category were similar, indicating that these findings likely reflect this region’s tendency toward a warmer climate. Saint Louis, Eastern Missouri, and Western Illinois are areas of the country that harbor more heat in the summer than cold in the winter. As depicted in Figure 4 and Table 2, the higher number of cases with higher temperature follows the trend of a higher overall number of days with higher temperature seen in this region.

Days of the Week and Holidays

Previous studies have examined day of the week as a factor for abuse in older children. A recent study from 2018 identified higher rates of verified reports of child abuse in children 5–11 years old in the state of Florida on Saturdays immediately following a Friday report card release compared to Saturdays that did not follow a Friday report card release (Bright et al., 2019). Unlike the study performed by Bright et al., however, there was a fairly even distribution of AHT cases in this study among the days of the week with no increased rate of cases on weekends in comparison to weekdays (Figure 3). It is possible that there is a pattern of higher proportions of abuse in older children based on day of the week that does not translate to infants, because the differences between weekends and weekdays plays a larger role in a school-aged child’s life than in an infant’s life. An infant’s crying, which is a known trigger for abuse(Barr et al., 2006), is not expected to differ in frequency depending on the day of the week. Given major holidays can also lead to added stress in a household, we hypothesized that there may be an increased incidence of AHT on or within 2 days of a major holiday; however, this study found no such association.

Identifying temporal patterns, environmental risk factors, and age-related risk factors is important in order to develop preventative measures. Some studies have described a decreased incidence of non-accidental trauma during child abuse prevention month (April), suggesting that interventions and awareness may have a preventative effect (Thomas et al., 2007). For example, report card release presents an opportunity for prevention in school age children, much like education about crying and colic can be a preventative measure against AHT in infants. Although this study did not find a seasonal or temporal pattern in our Midwest region hospital, it is possible that unique community factors may influence local patterns of other regions. Understanding these unique risk factors would likely be helpful to the providers and inform public health initiatives aimed at decreasing the incidence of AHT in that area.

Limitations

This study was limited by its retrospective, single institution nature. There were several factors in the data extraction that were not possible to clearly identify, such as exact date of the incidence of abuse when no perpetrator or witness was able to confirm this information. In addition, the hospital records, and not death records, were accessed; therefore, children with abusive head trauma who died at home or who were treated at other hospitals were not included in our study. Furthermore, temperature at the exact location of the event was also extrapolated, as the zip code listed in patients’ charts indicated their home address which may not necessarily be the address where the abuse took place. However, all of the zip codes were in Eastern Missouri or Western Illinois so variations in temperature between zip codes in this study are likely minimal. Additionally, other stressors that could theoretically increase the amount of stress in the household and increase the risk for non-accidental trauma (socioeconomic status, other children in the house, number of caregivers involved) could not be controlled in the study. Lastly, the study was based in Saint Louis, MO and its surrounding regions, which experience very hot and humid weather, but not as extreme cold or extreme heat as other parts of the country.

Conclusions

From this study, we detected no association between temperature, seasonality, month, day of the week, or presence of holiday and the incidence of abusive head trauma in this Midwest children’s hospital. Other locations may have unique community or environmental characteristics that influence local patterns and so this study may not be generalizable across multiple locations. For example, communities with extreme cold temperatures, heavy snow fall, or other inclement weather may have a different pattern of AHT incidence than the one presented in this study. More research is needed to identify location-specific risk factors and incidence trends in order to prepare health care providers and inform public health initiatives, as knowledge of risk factors remains the cornerstone of prevention in AHT. This study concludes that a high index of suspicion for non-accidental trauma should be maintained when evaluating patients with concerning exam findings regardless of the temperature, time of year, day of the week, or presence of a holiday. Future directions may include a multicenter study with a control group that would involve several institutions around the country to account for different variations in temperature extremes in order to determine whether seasonal trends exist in all forms of non-accidental trauma in comparison to accidental pediatric trauma.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Laura Bliss

References

Barlow

K. M.

Minns

R. A.

(2000). Annual incidence of shaken impact syndrome in young children. The Lancet, 356(9241), 1571–1572. https://doi.org/10.1016/S0140-6736(00)03130-5.

Barr

R. G.

Trent

R. B.

Cross

(2006). Age-related incidence curve of hospitalized shaken baby syndrome cases: Convergent evidence for crying as a trigger to shaking. Child Abuse and Neglect, 30(1), 7–16. https://doi.org/10.1016/j.chiabu.2005.06.009

Berger

L. M.

Brooks-Gunn

(2005). Socioeconomic status, parenting knowledge and behaviors, and perceived maltreatment of young low-birth-weight children. Social Science Research, 79(2), 237–267. https://doi.org/10.1086/428957

Bright

M. A.

Lynne

S. D.

Masyn

K. E.

Waldman

M. R.

Graber

Alexander

(2019). Association of friday school report card release with saturday incidence rates of agency-verified physical child abuse. JAMA Pediatrics, 173(2), 176–182. https://doi.org/10.1001/jamapediatrics.2018.4346

Cancian

Slack

K. S.

Yang

Y. M.

(2010). The effect of family income on risk of child maltreatment. Institute for Research on Poverty.

Coulton

Korbin

Chow

(1995). Community level factors and child maltreatment rates. Child Development, 66(5), 1262–1276. https://doi.org/10.1111/j.1467-8624.1995.tb00934.x

Curtis

Miller

B. C.

Berry

E. H.

(2000). Changes in reports and incidence of child abuse following natural disasters. Child Abuse and Neglect, 24(9), 1151–1162. https://doi.org/10.1016/S0145-2134(00)00176-9

Daly

S. E.

Connor

S. M.

(2001). Seasonal variations in the incidence of suspected shaken baby syndrome. International Journal of Trauma Nursing, 7(4), 124–128. https://doi.org/10.1067/mtn.2001.119010

Denissen

J. J. A.

Butalid

Penke

van Aken

M. A. G.

(2008). The effects of weather on daily mood: A multilevel approach. Emotion, 8(5), 662–667. https://doi.org/10.1037/a0013497

10.

Drake

Pandey

(1996). Understanding the relationship between neighborhood poverty and specific types of child maltreatment. Child Abuse and Neglect, 20(11), 1003–1018. https://doi.org/10.1016/0145-2134(96)00091-9

11.

Eckenrode

Smith

E. G.

McCarthy

M. E.

Dineen

(2014). Income inequality and child maltreatment in the United States. Pediatrics, 133(3), 454–461. https://doi.org/10.1542/peds.2013-1707

12.

Freisthler

Merritt

D. H.

Lascala

E. A.

(2006). Understanding the ecology of child maltreatment: A review of the literature and directions for future research. Child Maltreatment, 11(3), 263–280. https://doi.org/10.1177/1077559506289524

13.

Hennes

Kini

Palusci

V. J.

(2001). The epidemiology, clinical characteristics and public health implications of shaken baby syndrome. Journal of Aggression Maltreatment & Trauma, 5(1), 19–40. https:/doi.org/10.1300/j146v05n01_03

14.

Herman

B. E.

Makoroff

K. L.

Corneli

H. M.

(2011). Abusive head trauma. Pediatr Emer Care, 27, 65–69. https://doi.org/10.1016/B978-0-323-37101-8.00103-X

15.

Keenan

H. T.

Marshall

S. W.

Nocera

M. A.

Runyan

D. K.

(2004). Increased incidence of inflicted traumatic brain injury in children after a natural disaster. American Journal of Preventive Medicine, 26(3), 189–193. https://doi.org/10.1016/j.amepre.2003.10.023

16.

Kelly

John

Vincent

A. L.

Reed

(2015). Abusive head trauma and accidental head injury: A 20-year comparative study of referrals to a hospital child protection team. Archives of Disease in Childhood, 100(12), 1123–1130. https://doi.org/10.1136/archdischild-2014-306960

17.

Kesler

Dias

Shaffer

Rottmund

Cappos

Thomas

(2008). Demographics of abusive head trauma in the commonwealth of Pennsylvania. Journal of Neurosurgery. Pediatrics, 1(5), 351–356. https://doi/org/10.3171/PED/2008/1/5/351

18.

Kotch

J. B.

Browne

D. C.

Ringwalt

C. L.

Stewart

P. W.

Ruina

Holt

Lowman

Jung

J. W.

(1995). Risk of child abuse or neglect in a cohort of low-income children. Child Abuse and Neglect, 19(9), 1115–1130. https://doi.org/10.1016/0145-2134(95)00072-G

19.

Laskey

A. L.

Thackeray

J. D.

Grant

S. R.

Schnitzer

P. G.

(2010). Seasonality of child homicide. Journal of Pediatrics, 157(1), 144–147. https://doi.org/10.1016/j.jpeds.2010.01.013

20.

Parks

Sugerman

Coronado

(2012). Characteristics of non-fatal abusive head trauma among children in the USA, 2003-2008: Application of the CDC operational case definition to national hospital inpatient data. Injury Prevention, 18(6), 392–398. https://doi.org/10.1136/injuryprev-2011-040234

21.

Scribano

P. V.

Makoroff

K. L.

Feldman

K. W.

Berger

R. P.

(2013). Association of perpetrator relationship to abusive head trauma clinical outcomes. Child Abuse and Neglect, 37(10), 771–777. https://doi.org/10.1016/j.chiabu.2013.04.011

22.

Shanahan

M. E.

Zolotor

A. J.

Parrish

J. W.

Barr

R. G.

Runyan

D. K.

(2013). National, regional, and state abusive head trauma: Application of the CDC algorithm. Pediatrics, 132(6), e1546–1543. https://doi.org/10.1542/peds.2013-2049

23.

Shinsugi

Stickley

Konishi

C. F. S.

Watanabe

(2015). Seasonality of child and adolescent injury mortality in Japan, 2000–2010. Environmental Health and Preventive Medicine, 20(1), 36–43. https://doi.org/10.1007/s12199-014-0421-7

24.

Starling

S. P.

Patel

Burke

B. L.

Sirotnak

A. P.

Stronks

Rosquist

(2004). Analysis of perpetrator admissions to inflicted traumatic brain injury in children. Archives of Pediatrics and Adolescent Medicine, 158(5), 454–458. https://doi.org/10.1001/archpedi.158.5.454

25.

Thomas

N. J.

Shaffer

M. L.

Rzucidlo

Shirk

B. J.

Dias

M. S.

(2007). Temporal factors and the incidence of physical abuse in young children: Decreased nonaccidental trauma during child abuse prevention month. Journal of Pediatric Surgery, 42(10), 1735–1739. https://doi.org/10.1016/j.jpedsurg.2007.05.032

26.

Wood

J. N.

French

Fromkin

Fakeye

Scribano

P. V.

Letson

M. M.

Makoroff

K. L.

Feldman

K. W.

Fabio

Berger

(2016). Association of pediatric abusive head trauma rates with macroeconomic indicators. Academic Pediatrics, 16(3), 224–232. https://doi.org/10.1016/j.acap.2015.05.008