key: cord-0886932-v5icfz82
authors: Hartwell, Micah; Hendrix, Amy D.; Sajjadi, Nicholas B.; Baxter, Michael A.; Chesher, Tessa; Coffey, Sara; Passmore, Sarah J.
title: Child maltreatment during COVID‐19: Deviations from forecasted projections of criminal filings in Oklahoma in 2020
date: 2022-03-01
journal: Child Abuse Rev
DOI: 10.1002/car.2754
sha: 69e82b85f6359c9dbb69f00dd6828c945dbc3cea
doc_id: 886932
cord_uid: v5icfz82

nan

The state of Oklahoma ranks among the nation's highest in child abuse and neglect prevalence, with 15.9 confirmed cases per 1000 children (Oklahoma State Department of Health, 2018) . Moreover, Oklahoma has consistently ranked low in overall child wellbeing compared to other states (Annie E. Casey Foundation, 2020). Oklahoma's poor child wellbeing status is attributable to many factors including high prevalence of poverty, housing instability, food insecurity, lack of health care coverage, and low parent educational attainmentall of which are risk factors for child maltreatment (Annie E. Casey Foundation, 2020; Hunter & Flores, 2021) .

Previously, we found that criminal filings of child abuse and neglect significantly decreased in the state of Oklahoma during the first six months of 2020 during the SARS-CoV-2 (COVID-19) pandemic (Whelan et al., 2020) . The first case of COVID-19 was identified in Oklahoma on March 6, 2020, prompting a state-wide Safer-at-Home order on March 24, 2020 (Oklahoma Executive Order: Safer at Home, 2020). Following these Safer-at-Home orders, many families faced increasing stress and risk of parental substance misuse due to financial hardships from lost wages, thus adding to the risk for abuse (Bradbury-Jones & Isham, 2020; Flory et al., 2009; Liu & Doan, 2020) . Our initial findings showed that criminal filings were significantly lower than a projected model based on 10 years of previous data, culminating in a 63 per cent difference in actual filings compared to forecasted filings in June of 2020 (Whelan et al., 2020) . Beyond the Safer-at-Home order, Oklahoma courts reopened with few interruptions (Oklahoma State Courts Network, 2021) and crime rates remained consistent with pre-pandemic levels (Carter, 2020) . As the pandemic continued to take its toll, the state did little in the way of policy change aimed at alleviating these burdens (Polansky, 2020) . Additionally, most schools did not return to in-person learning in the fall semester of 2020, likely limiting child abuse and neglect surveillance by teachers and school staff. Thus, we hypothesised that the rates of child abuse filings continued to decline during the summer and autumn months of 2020.

Data collected from Oklahoma State Court Network (OSCN) was provided by a third party with identifiable information removed. Further, duplicative entries and filings attributed to the same individual pertaining to the same case number and crime were removed prior to the dataset being provided to us, thus eliminating duplicate counts. Using this data, we compiled monthly counts of criminal cases of physical and sexual child abuse, neglect, domestic violence in the presence of minors, failure to protect a child, and solicitation of a minor, from January 2010 through December 2020. To measure deviations in filings, we used auto-regressive integrative moving average (ARIMA) models (Hyndman & Khandakar, 2008) to forecast the projected criminal filings from February 2020 through December 2020 and assessed differences between the predictions and the actual values. ARIMA modelling has been used to inform and direct policy in the economic (Dhahri & Chabchoub, 2007) and energy sectors (Erdogdu, 2007) , as well as in the field of controlled substance laws and prevention (Cunningham & Liu, 2008; Troelstra et al., 2016; Wyatt et al., 2013) . Corrected Akaike tests (AICc) were used to determine the best model and Ljung-Box test results were reported. Next, we used an integral function to calculate the area between the actual and forecasted curves to approximate the number of criminal filings which may have gone unreported during these months. All analyses were conducted in R. This study followed STROBE reporting guidelines, and an institutional review board determined this study did not constitute human subjects research.

Total criminal filings of child abuse were lower than projected models for February and April through November, before returning to normal expected values in December ( Figure 1 ). Statistically significant decreases began in May with 80 filings compared to an expected 109.2 (95%CI: 80.6-137.7), which continued to decline through September, during which only eight filings occurred compared to the projected 102.0 (95%CI: 71.3, 132.7) -92.2 per cent fewer than expected (Table 1) . All types of criminal filings were lowest during August and September. From the integral calculation, the cumulative difference between forecasted values and actual filings from February through November was 457.3 (442.2-472.4). Findings for crimes were within predicted confidence intervals in November, while filings of neglect were significantly higher than predicted in December. The parameters for the ARIMA model for p, d and q were 0, 1 and 1, respectively, with seasonality parameters P, D and Q set to 1, 0 and 1. The model fit AICc was 976.95, and Ljung-Box Q was 35.28 (df = 21, P = 0.03).

Our findings suggest that Oklahoma criminal filings for child abuse and neglect continued to decline for most of the latter half of 2020 before rising to expected levels near the end of the year. Child abuse and neglect in Oklahoma likely F I G U R E 1 Total and predicted criminal cases of child abuse (physical and sexual), neglect, domestic violence in the presence of a minor, failure to prevent injury or neglect, and solicitation of minors Non-seasonal and seasonal ARIMA parameters (p,d, q)(P,D,Q) for models were as follows: Total: (0,1,1)(1,0,1), Neglect: (1,1,1)(0,0,2), DV in presence of a minor: (1,0,1)(1,0,1), Physical abuse: (1,1,2) (1,0,0); Sexual abuse: (1,1,1) (1,0,0). continued and may have even increased to levels higher than that seen prior to the COVID-19 pandemic during this time, especially in the presence of heightened risk factors (Storz, 2020; World Health Organization, 03/, 2020) . Though our findings are correlational in naturea limitation of our studywe posit our findings of lower criminal filings are likely due to interrupted child abuse and neglect surveillance and may only represent a small fraction of actual child abuse.

Aside from law enforcement and social service workers, which were also in reduced capacity and services during the pandemic, schools, hospitals and relatives are the top reporting sources of child abuse in Oklahoma (Robinson, 2020) . A potential reason for the largest deviation occurring in August and September is that caregivers plan for summer childcare and rely on children returning to school in order to return to work at full capacityespecially families who are less socially advantaged (Xu et al., 2021) . As schools were provided more instructional guidance for virtual learning (Tiffany & Levi, 2020) and many districts returned to in-person learning in late September (Martinez-Keel, 2020), reports of abuse and neglect started increasing. The filing trends normalising, or exceeding estimated projections in the case of neglect, near the end of the year may be the result of increased family interactions during the holidays or of some schools reopening under the advice of the American Academy of Paediatrics (American Academy of Pediatrics, 2021) and the Centers for Disease Control and Prevention's recommendations (Honein et al., 2021) . The relative increase in neglect over projected filings in December of 2020 also likely points to the further erosion of financial safety nets during a time of extreme stress and economic hardship as America endured the worst wave of the pandemic.

Interrupted surveillance during the COVID-19 pandemic could leave sentinel injuries from abuse untreatedlikely leading to further complications or more severe subsequent injuries from ongoing, inescapable abuse. Further, emotional trauma resulting from abuse and neglect, particularly without access to support, is likely to have long-lasting impacts, as adverse childhood experiences can detrimentally affect individuals well into adulthood (Anda et al., 2006) . The inadvertent cost of school closures and lockdown measures on children may bear significant psychosocial consequences (Ghosh et al., 2020; in a state that is already burdened with a high prevalence of child abuse and neglect. While the COVID-19 pandemic has highlighted the importance of telehealth and its vital role during disasters in Oklahoma (Canady, 2020) , its use for identifying abuse or neglect among children has limitations (Racine et al., 2020) . Therefore, supporting organisations that provide social services during natural disasters, national emergencies and pandemics is essential for mitigating child abuse risk factors.

Regarding future pandemics and child abuse, prioritising support and resources, including vaccinations and personal protective equipment, to educational facilities to allow for minimal interruption to in-person learning may serve to protect children from prolonged abuse by preventing impaired surveillance. In the event of a future pandemic, targeted economic relief and childcare resources should be administered to families swiftly and consistently throughout the course of the pandemic. Future research should be done evaluating the impact of the COVID-19 pandemic on families to better prepare social service systems for future pandemics. Beyond these targeted approaches, supporting a strong vaccination effort in Oklahoma will play an important role in reopening schools safely, limiting the untoward effects of lockdown measures and bringing us closer to ending the pandemic.

COVID-19 Guidance for Safe Schools

The enduring effects of abuse and related adverse experiences in childhood. A convergence of evidence from neurobiology and epidemiology

Casey Foundation. (2020) 2020 KIDS COUNT Data Book

The pandemic paradox: The consequences of COVID-19 on domestic violence

COVID-19 outbreak represents a new way of mental health service delivery

Crime may be unaffected by COVID-19, while lawsuits proliferate. Oklahoma Council of Public Affairs

Impact of methamphetamine precursor chemical legislation, a suppression policy, on the demand for drug treatment

Nonlinear goal programming models quantifying the bullwhip effect in supply chain based on ARIMA parameters

Electricity demand analysis using cointegration and ARIMA modelling: A case study of Turkey

Alcohol and cigarette use and misuse among Hurricane Katrina survivors: psychosocial risk and protective factors

Impact of COVID -19 on children: special focus on the psychosocial aspect

Data and Policy to Guide Opening Schools Safely to Limit the Spread of SARS-CoV-2 Infection

Social determinants of health and child maltreatment: a systematic review

Automatic time series forecasting: the forecast package for R

Psychosocial Stress Contagion in Children and Families During the COVID-19 Pandemic

Mental health considerations for children quarantined because of COVID-19. The Lancet Child & Adolescent Health

OKC schools plan dates for in-person return to classes

Oklahoma Executive Order: Safer at Home. (2020) 2020-07 Fourth Amended Executive Order

Oklahoma State Courts Network Notices

Oklahoma State Plan for the Prevention of Child Abuse and Neglect

Stitt administration choosing to no longer receive weekly White House Coronavirus task force reports

Telemental health for child trauma treatment during and post-COVID-19: Limitations and considerations

Oklahoma Human Services: Child Welfare Services and Office of Performance Outcomes and Accountability

Child abuse: A hidden crisis during covid-19 quarantine

Instruction Support. Oklahoma State Department of Education

Effect of Tobacco Control Policies on Information Seeking for Smoking Cessation in the Netherlands: A Google Trends Study

Deviations in criminal filings of child abuse and neglect during COVID-19 from forecasted models: An analysis of the state of Oklahoma

WHO Director-General's opening remarks at the media briefing on COVID-19. WHO. Available at: https://www. who.int/director-general/speeches/detail/who-director-general-s-opening-remarks-at-the-media-briefing-on-covid

Population-level administration of AlcoholEdu for college: an ARIMA time-series analysis

Material hardship and child neglect risk amidst COVID-19 in grandparentheaded kinship families: The role of financial assistance

Child maltreatment during COVID-19: Deviations from forecasted projections of criminal filings in Oklahoma in 2020

Micah Hartwell had access to data received for the study and takes responsibility for the integrity and accuracy of the data analysis.

We submitted the methodology used herein for this study to an institutional review board which determined this study did not constitute human subjects research.ORCID Micah Hartwell https://orcid.org/0000-0001-6810-6571 Amy D. Hendrix https://orcid.org/0000-0002-9376-1524