Child maltreatment reporting during the initial weeks of COVID-19 in the US: Findings from NCANDS.

BACKGROUND: After the national COVID-19 emergency declaration in the U.S. in March 2020, child welfare agencies observed large reductions in maltreatment reporting.
OBJECTIVE: To quantify the impact of the COVID-19 pandemic on child maltreatment reporting nationally to inform policy for future emergencies. PARTICIPANTS AND
SETTING: Administrative data from the National Child Abuse and Neglect Data System (NCANDS) for 48 states for federal fiscal years (FFYs) 2017 through 2020.
METHODS: Analyses focused on reports to child protective services (CPS) between weeks 12 and 24 of calendar years 2017 through 2020 (mid-March through mid-June). Report sources of screened in and substantiated reports were compared with those during the prior year. Likelihood of a report being substantiated in 2020 compared with 2019 based on report source was calculated using odds ratios.
RESULTS: In 2020, CPS screened in 39 % fewer reports than during the same period in 2019 and the proportion of reports substantiated increased from 18 to 22 %. Reports from all report sources decreased, especially from education personnel (90 % decrease) and child daycare providers (65 % decrease). The odds for substantiation were significantly higher during 2020 than in 2019 for reports from all but three sources.
CONCLUSION: During the initial weeks following the national COVID-19 emergency declaration, the number of reports to CPS declined sharply at the national level and across all states, primarily in association with a large reduction in referrals from education sentinels. Explanations for the increase in percent of substantiation in the context of reduction of reports are considered.

Introduction

Soon after the national emergency was declared in the United States (U.S.) in March 2020 for the coronavirus disease 2019 (COVID-19) pandemic, state and local child welfare agencies observed large reductions in child maltreatment reporting (Brown et al., 2022; Jonson-Reid et al., 2021; Weiner et al., 2020). These decreases in maltreatment reporting were attributed to the lower frequency of in-person contact between children and mandated reporters, such as school personnel (Baron et al., 2020; Boserup et al., 2020; Brown et al., 2022; Jonson-Reid et al., 2021; Rapoport et al., 2021; Weiner et al., 2020). Practitioners, researchers, and policymakers expressed concern about the increased risk to children due to undetected and unreported maltreatment (LeBlanc, 2020; Martins-Filho et al., 2020; Mathematica, 2020; Rodriguez et al., 2021). At the same time, household stressors such as employment disruption and school and child daycare closures during the pandemic created added risk factors for child maltreatment (Connell & Strambler, 2021; Lawson et al., 2020).

The few studies published on the impact of disasters on rates of child maltreatment suggest that family violence does increase in the aftermath of emergencies yet reporting of violence is suppressed due to disruption in reporting mechanisms and infrastructure (Curtis et al., 2000; Seddighi et al., 2021). Puls et al. (2021), using National Child Abuse and Neglect Data System (NCANDS) data, found that child maltreatment reporting levels significantly decline during routine school closures (during years prior to 2020), but without significant increases during two-week intervals immediately following closure periods, suggesting that actual incidence of maltreatment, rather than just reporting, may be reduced when children are out of school. Using data from emergency rooms and police reports, some findings indicate that while reporting to child protective services (CPS) went down at the outset of the pandemic, children were experiencing maltreatment at the same level or at even higher levels (Barboza et al., 2021; Boserup et al., 2020; Kovler et al., 2021; Schmidt & Natanson, 2020; Sharma et al., 2021; Sidpra et al., 2021). However, other studies found that increased physical abuse was not observed as measured by reductions in emergency department visits and stable hospitalizations associated with child maltreatment (Sege & Stephens, 2021; Swedo et al., 2020).

Studies regarding CPS reporting during the pandemic published to date have relied on data collected at the state or local levels (Baron et al., 2020; Boserup et al., 2020; Brown et al., 2022; Rapoport et al., 2021; Whelan et al., 2021) or in other countries (Katz & Cohen, 2021; Katz, Katz, et al., 2021; Katz, Priolo-Filho, et al., 2021; Martins-Filho et al., 2020; Wong et al., 2021). The annual Child Maltreatment report from the U.S. Department of Health & Human Services Children's Bureau compared federal fiscal year (FFY) 2020 quarterly data from NCANDS with the same quarters from FFY 2019 (U.S. Department of Health & Human Services, 2022). Assignment to quarters was based on disposition date (i.e., the date a determination is made by the CPS agency about whether the maltreatment occurred) as are all analyses for the Child Maltreatment report. During the third quarter of FFY 2020 (April–June) the number of screened-in reports that reached a disposition was 22.8 % lower than during the same period in FFY 2019 (U.S. Department of Health & Human Services, 2022). Unlike the Child Maltreatment Report, this study focuses on reports with report dates (rather than disposition dates) during the initial lockdown and provides additional analytic support for the changes during this period.

The goal of the present study is to use NCANDS data to document the impact of the COVID-19 pandemic on child welfare systems at the national level by comparing the number of CPS responses resulting from reports during the weeks of the initial COVID-19 lockdown period with the data during the same weeks in previous years and to consider potential drivers of these changes. Specifically, we examined how the total number of screened-in referrals (reports) changed during the initial COVID-19 lockdown period, as well as reports from each report source. We also assessed the impact on the percent of substantiated reports during the initial COVID-19 lockdown period overall and from each report source compared with the same period in prior years.

This study adds to the literature documenting the COVID outbreak in three unique ways; 1) it incorporates a nearly complete census of reported maltreatment in the U.S. during the initial outbreak, 2) it describes reported maltreatment confined to the specific time period of the initial COVID outbreak (calendar weeks 12–24 of 2020, corresponding to mid-March through mid-June), and 3) in contrast to previously published national data, it includes reports initiated during the lockdown period rather than those that reached a disposition.

Method

Data source

This study examined administrative data reported to NCANDS, a federally sponsored effort to collect nationally standardized data about children who were known to CPS agencies. NCANDS data are collected annually on an FFY schedule (October 1 through September 30) from all 50 states, the District of Columbia, and the Commonwealth of Puerto Rico. States construct an electronic file called the Child File containing child-specific records for each report of alleged maltreatment for which a determination about the CPS response was reached during the reporting year (U.S. Department of Health & Human Services, 2022).

When a CPS agency receives an initial notification—called a referral—alleging child abuse or neglect, agency hotlines or intake units conduct a screening to determine whether the referral is appropriate for further action. Referrals that meet CPS agency criteria are screened in (reports) and receive a CPS response. More than one child can be included in a report, and children can be included in multiple reports. While this study primarily analyzed data at the report level, we also observed the changes in the number of children who received a CPS response during this period. In most states, the majority of reports receive an investigation, which results in a determination (disposition) about the alleged maltreatment. In some states, reports may be assigned to an alternative response track, which focuses on the service needs of the child and family. A report is substantiated when an investigation concludes that the allegation of maltreatment, or risk of maltreatment, is supported or founded by state law or policy. A report is unsubstantiated when an investigation concludes with insufficient evidence under state law that the child was maltreated or is at risk of being maltreated. Some states (eight during FFY 2020) also use the disposition of indicated, when maltreatment could not be substantiated under state law or policy, but there is a reason to suspect that at least one child may have been maltreated or is at risk of maltreatment For this study, the term substantiated reports will refer to reports that were either substantiated or indicated, among all reports The annual Child Maltreatment report contains a glossary defining all NCANDS terms (U.S. Department of Health & Human Services, 2022).

Human subject protection

NCANDS data do not include identifiable private information, so this study is not considered research involving human subjects; U.S. Department of Health & Human Services Policy for the Protection of Human Subjects does not apply (Protection of Human Subjects, 2018).

Focus period

The focus of this research was a 13-week period from weeks 12 through 24 of calendar year 2020, corresponding to Sunday, March 15 through Saturday, June 13, with comparable 13-week periods defined for the prior three years. While stay-at-home orders and school closures varied across the country, the national health emergency for COVID-19 was declared on Friday, March 13, 2020 (Centers for Medicare & Medicaid Services, 2020). The beginning of the following week (March 15th) corresponding to week 12 is the beginning of the defined comparison period. Two factors were considered in defining the end of the comparison period. First, child maltreatment reports typically decline substantially during the summer months when children are not in school. Because the goal of this research was to identify the impact of the COVID-19 pandemic on maltreatment reporting, we wanted to limit the seasonal variability during the comparison periods, so we ended the focus period near the start of summer. Second, the NCANDS Child File is submitted by states on an FFY basis and includes all reports that reach a disposition between October 1 and September 30. Any reports that come to CPS toward the end of the FFY but have not yet reached disposition prior to September 30 are included in the Child File for the subsequent FFY.

Exclusions of reports still in process

State NCANDS FFY submissions only include completed reports that have reached a disposition (or finding) as an outcome of the CPS response during the reporting year, therefore reports that occur later that have not yet reached a disposition in the FFY are not included until the following year. Consequently, the FFY 2020 Child File used for analysis in this study included reports that reached a disposition prior to September 30, 2020. To maintain similarity between the comparison groups, this study excluded the small portion of reports that came in during weeks 12 to 24 for 2017, 2018, or 2019 but did not reach disposition within the same FFY.

Exclusions of states

All analyses included 48 states; while this study is intended to reflect national trends, four states were excluded due to data submission issues unrelated to COVID-19 (U.S. Department of Health & Human Services, 2022). Iowa, Maryland, and Oregon were excluded due to insufficient data on report source, and North Carolina was excluded due to a change in definition of substantiation between 2019 and 2020. These four states did not differ significantly from the included states in the number of reports, percent of reports substantiated, or the percent change in number of reports from 2019 to 2020.

Analyses

After constructing the dataset, we produced descriptive statistics to assess data quality. Descriptive analyses were also used to quantify the size and directionality of changes between reporting and substantiation of reports during the initial COVID-19 lockdown period and the same months in previous years. For the percent of reports substantiated by report source, we calculated odds ratios to estimate the likelihood that a report would be substantiated in 2020 compared with in 2019 based on the source of the report (Tenny & Hoffman, 2021). IBM SPSS Statistics for Windows, Version 24.0 and Tableau, Version 2020.2.17 were used to conduct data screening and statistical analysis (IBM SPSS Statistics, 2016; Tableau, 2021).

Results

Number of reports

During the COVID-19 comparison period in 2020, CPS screened in 344,509 reports; 39 % lower than the number during the same period in 2019. The number of reports that were screened in was consistent during the same period in 2017 through 2019, slightly >560,000 reports each year (see Table 1 , Fig. 1 ). During the comparison period, the number of reports decreased in all 48 states from 2019 to 2020, with decreases ranging from 2 % to 66 %. In 41 states, the number of reports decreased by at least 25 %, and in 11 of these states, the number of reports decreased by >50 %.

Table 1

Number of substantiated and unsubstantiated reports and percent of reports with substantiated dispositions,a 2017–2020b (n = 48 states).

	2017	2018	2019	2020
Substantiated reports	107,369	105,489	102,959	76,858
Unsubstantiated reports	452,649	460,534	458,928	267,651
Total reports	560,018	566,023	561,887	344,509
% reports substantiated	19 %	19 %	18 %	22 %

Reports with disposition of “indicated” are included in substantiated reports.

Data include reports accepted between weeks 12–24 of each calendar year.

Fig. 1

Number of substantiated and unsubstantiated reports and percent of reports with substantiated dispositionsa 2017–2020b (n = 48 states).

aReports with disposition of “indicated” are included in substantiated reports.

bData include reports accepted between weeks 12–24 of each calendar year.

Percent of reports substantiated

While the total number of substantiated reports dropped during the COVID-19 comparison period of 2020, the proportion of reports that were substantiated increased to 22 % during 2020 compared with 18 % in 2019 (see Table 1, Fig. 1). During the comparison period, the percent of reports substantiated increased in 43 states, with 14 states increasing >5 percentage points and three states increasing >10 percentage points. In five states the percent of reports substantiated decreased.

Weekly reports and substantiation

Fig. 2 shows the weekly number of reports (as shown in the bars) and the weekly percent of reports substantiated (as shown in the lines) for the entire period from week 2 through week 30 of 2020. Consistent with the findings for the COVID-19 comparison period described above, the weekly number of reports declined, and the weekly percent of reports substantiated increased between 2019 and 2020. Thus, overall findings indicate that reports in 2020 were more likely to be substantiated than reports in 2019 during the comparison period. This divergence occurred during the COVID-19 comparison period, but numbers and percentages were similar during the weeks before and after.

Fig. 2

Number of reports and percent of reports substantiated 2019–2020, weeks 2–30 (n = 48 states).

Number of children and victims

Since more than one child can be included in a report, and children can be included in multiple reports, the 344,509 reports during the COVID-19 comparison period in 2020 included 589,021 unique children who received a CPS response, representing a 38 % decrease from 943,150 during the same period in 2019. Of these children, 17 % were found to be victims in 2019 and 20 % in 2020, also consistent with the changes in the percent of reports substantiated.

Report source

Reports from each report source

The report source indicated in each record in the Child File is the community role of the person who notified a CPS agency of the alleged child abuse or neglect. As the total number of reports decreased during the COVID-19 comparison period, so did the number of reports from all report sources. The largest decreases were in the number of reports from education personnel and child daycare providers (see Table 2 ). During the COVID-19 comparison period of 2020, 12,526 reports came to CPS from education personnel, a 90 % decrease from the same period in 2019. Reports from legal, law enforcement, or criminal justice personnel showed the smallest reduction of all report sources. Law enforcement sources initiated 92,782 reports during the COVID-19 comparison period, a 12 % decrease from the same period in 2019. Table 2 also shows the proportion of reports received by each source during the COVID-19 comparison period of 2020, compared with 2019. Reports from education personnel accounted for 22 % of all reports during this period of the year in 2019, but in 2020 they accounted for only 4 %. This was the largest decrease for any category. Reports from legal, law enforcement, or criminal justice personnel accounted for 27 % of reports during 2020. During 2019, reports from law enforcement comprised 19 % of all reports.

Table 2

Number and percent of reports by report source, 2019–2020a (n = 48 states).

	2019		2020		Percent difference
Child daycare provider	3877	1 %	1359	0.4 %	−65 %
Education personnel	126,307	22 %	12,526	4 %	−90 %
Foster care provider	2336	0 %	1525	0 %	−35 %
Legal, law enforcement, or criminal justice person	105,949	19 %	92,782	27 %	−12 %
Medical personnel	60,446	11 %	45,326	13 %	−25 %
Mental health personnel	34,231	6 %	20,339	6 %	−41 %
Social services personnel	55,607	10 %	35,534	10 %	−36 %
Nonprofessionalb	85,335	15 %	72,928	21 %	−15 %
Unclassifiedc	87,799	16 %	62,190	18 %	−29 %
Total	561,887	100 %	344,509	100 %

Data include reports accepted between weeks 12–24 of each calendar year.

Nonprofessional report sources include: alleged perpetrator, alleged victim, parent, other relative, and friends/neighbors.

Unclassified report sources include: anonymous reporter, other, and unknown or missing.

Percent of reports substantiated from each report source

Table 3 shows the number and percent of reports from each report source that were substantiated and unsubstantiated during the COVID-19 comparison period of 2020 compared with 2019. The percent of reports substantiated was highest during both years for reports from legal and law enforcement personnel (34 % during 2020, an increase from 33 % in 2019), followed by medical personnel (29 %, an increase from 27 % in 2019) and social services personnel (25 %, and increase from 23 % in 2019). The percent of reports substantiated was lowest for reports from education personnel; 11 % in 2019 and 10 % in 2020. While percent substantiated varied from source to source, the percent of reports that were substantiated increased in 2020 over the same period in 2019 for all report source categories except education personnel. Table 3 and Fig. 3 show the odds ratios for the substantiation of reports from each source during 2020 relative to 2019. The odds for substantiation were significantly higher during 2020 compared with 2019 for reports from all but three sources—child daycare provider, education personnel, and foster care providers.

Table 3

Substantiation of reports in 2020 by report source, compared with 2019a (48 states).

	2019			2020			Odds ratio and 95 % CI
	Unsub.	Sub.	% Sub.	Unsub.	Sub.	% Sub.
Child daycare provider	3411	466	12 %	1178	181	13 %	1.12 (0.94, 1.35)
Education personnel	112,855	13,452	11 %	11,250	1276	10 %	0.95 (0.9, 1.01)
Foster care provider	1966	370	16 %	1251	274	18 %	1.16 (0.98, 1.38)
Legal, law enforcement, or criminal justice personnel⁎⁎⁎	71,408	34,541	33 %	61,023	31,759	34 %	1.08 (1.06, 1.1)
Medical personnel⁎⁎⁎	44,055	16,391	27 %	31,997	13,329	29 %	1.12 (1.09, 1.15)
Mental health personnel⁎⁎⁎	30,353	3878	11 %	17,736	2603	13 %	1.15 (1.09, 1.21)
Social services personnel⁎⁎⁎	42,943	12,664	23 %	26,530	9004	25 %	1.15 (1.12, 1.19)
Nonprofessionalb, ⁎⁎⁎	75,103	10,232	12 %	63,075	9853	14 %	1.15 (1.11, 1.18)
Unclassifiedc, ⁎⁎⁎	76,834	10,965	12 %	3611	8579	14 %	1.12 (1.09, 1.16)

Abbreviations:

Sub., Substantiated

Unsub., Unsubstantiated

CI., Confidence Interval

p < .001.

Data include reports accepted between weeks 12–24 of each calendar year.

Nonprofessional report sources include: alleged perpetrator, alleged victim, parent, other relative, and friends/neighbors.

Unclassified report sources include: anonymous reporter, other, and unknown or missing.

Fig. 3

Odds of substantiation of reports in 2020 by report source, compared with 2019a (48 states).

aData include reports accepted between weeks 12–24 of each calendar year.

***p < .001.

Discussion

As the COVID-19 pandemic has disrupted child welfare system operations, it has created an opportunity to observe and learn from the impact of this disruption (Weiner et al., 2020). This is the first study, to our knowledge, that examines the volume of maltreatment reporting and substantiation nationally during the early weeks of the COVID-19 pandemic compared with a similar timeframe in 2019. Our findings confirm that reductions in reporting documented at state and local levels (Barboza et al., 2021; Brown et al., 2022; Rapoport et al., 2021) were national in scope and occurred in all states to varying degrees. Additionally, we found that a higher proportion of reports were substantiated during the initial COVID-19 lockdown, an important contribution to the literature. Documenting and understanding the details of what occurred during this period can inform the development of policies for protecting children and providing services during future national emergencies.

Reporting declined

Analysis of NCANDS data revealed that during the initial weeks following the declaration of national emergency due to the COVID-19 pandemic, the number of reports to CPS declined sharply at the national level, similar to declines in maltreatment reports during routine school closures (Puls et al., 2021). Our findings support the conclusions of other researchers that these declines resulted from decreased opportunities for surveillance of maltreatment by community sentinels, particularly educators, normally in a position to observe children. The duration and scope of the school closures during the COVID-19 response far exceeded routine closures, as well as closures during prior emergencies in recent U.S. history (Rapoport et al., 2021). The sudden, dramatic reduction in reports during the initial weeks of COVID-19 yields a new perspective on mandatory reporters and their critical role in bringing reports to CPS.

NCANDS findings cannot resolve whether the reduction in maltreatment reporting during the lockdown reflects reduced levels of actual maltreatment, as during routine school closures (Puls et al., 2021) or whether it reflects only the absence of detected maltreatment while undetected maltreatment remains at similar or potentially higher levels, as experts predicted. Data from FY2021, which will include more complete data on reports during the summer of 2020 immediately following the lockdown period we studied, will provide some insight into any rebound effect, in which reports increase to compensate for the earlier reduction, indicating that only reporting, but not actual maltreatment, was reduced.

Higher percent of reports were substantiated

A higher proportion of reports were found to be substantiated during the initial lockdown period. Whether the increase in the percent of substantiated reports and the corresponding proportion of children found to be victims reflects greater severity of maltreatment events cannot be readily assessed using NCANDS data. The increased proportion of substantiated reports could have resulted from efficiencies in the investigation process due to the lower volume of reports and reduction in workload, allowing caseworkers to thoroughly investigate and substantiate cases that would have been overlooked in previous years. Additional analyses showed that on average, during the initial COVID-19 lockdown, reports reached a disposition within 38 days, which was significantly less than the average time to disposition during the same period in 2019 (44 days) (Shusterman et al., 2022). Future research can explore whether an infusion of investigation resources to child protective services agencies could have similar results during non-pandemic circumstances.

Since the increase in substantiation varied across reports from different sources, we considered whether the higher substantiation rate may simply reflect the absence of reports that would have been unsubstantiated anyway, i.e., that obstacles to reporting introduced by the lockdown essentially raised the bar for reportable maltreatment. If this were the case, we would have expected that most report sources would still have identified situations associated with substantiated maltreatment at levels similar to those they reported in prior years, and the number of substantiated reports would have remained constant. Instead, the number of substantiated reports declined, just not as steeply (a 25 % decrease) as the number of unsubstantiated reports (42 % decrease).

During pre-pandemic years, prior researchers have documented the lower substantiation rates for reports originating from school personnel compared with other mandated reporters such as law enforcement, medical professionals, and social workers, (Kesner, 2008; Kesner & Robinson, 2002; King et al., 2013; Weiner et al., 2020). While education personnel typically have frequent opportunities to observe children, school settings may be less conducive for collecting detailed evidence on suspected maltreatment than medical, social services or law enforcement settings, resulting in lower substantiation rates for their reports (King et al., 2013). Lower substantiation of reports by education personnel has also been attributed to teachers' lack of training about child maltreatment (Kesner & Robinson, 2002). During the COVID-19 lockdown period when most classes were conducted remotely, any existing barriers to teachers' thorough documentation of suspected maltreatment could have been reinforced. However, reports from educational sources have also consistently represented the largest percentage of reports since FFY2015 (U.S. Department of Health & Human Services, 2016; U.S. Department of Health & Human Services, 2017), and given that the 2017–2019 numbers of substantiated reports were consistent, we theorized that the 2020 numbers would have been sustained in the absence of a pandemic. We estimated the missed substantiated reports as the percent change in the number of substantiated reports between 2019 and 2020. Even with their low substantiation rate, reports from education personnel accounted for nearly half (47 %) of all missed substantiated reports. The implication of these potentially missed substantiated reports as well as increased efficiencies in investigations will require further study.

Systemic impact

Every incidence of child maltreatment recorded in NCANDS is the result of a complex series of events with multiple actors: a child was maltreated (actually or allegedly); another individual witnessed or was made aware of the maltreatment; that individual, either as mandated or not, reported the maltreatment to CPS; CPS workers accepted, investigated, and potentially substantiated the report. The changes introduced by the COVID-19 lockdown could have had an impact on any of these actors at any point along this process—the perpetrator, the reporter, and the CPS workers—so the data reflect the sum of this impact. Researchers found that following natural disasters, communities and CPS systems are affected as much as children and families (Curtis et al., 2000; Seddighi et al., 2021), and during the initial months of COVID-19, social workers were experiencing elevated levels of posttraumatic stress, grief, burnout, and secondary trauma (Holmes et al., 2021). These experiences, along with a myriad of co-occurring changes could have influenced decision-making within CPS systems in screening in or substantiating reports, as well as reporting as reflected in the NCANDS data. Child welfare agencies adapted policies and practices to address the evolving health and safety needs of families as well as their own workers (Schwab-Reese et al., 2020). Remote work opportunities increased, enabling staff to conduct some visits from their homes (Goldhaber, 2020; U.S. Department of Health & Human Services, 2022). Furthermore, during the same period, other historic events unfolded in the U.S. with nationwide protests and civil unrest against police brutality and racism, which also raised attention to racial disparities including those in the child welfare system (Agosti et al., 2021; Crowley, 2020). Thus, changes in system practices in response to these events as well as reporter experiences and attitudes during this time may have contributed to the reductions in reporting reflected in the NCANDS data. Planned analyses will explore the changes in reporting during COVID-19 across age and race groups, as well as among different types of maltreatment which may shed light on these dynamics.

Limitations and conclusions

NCANDS data have important limitations. First, the Child File includes only records of maltreatment that are not only observed, but also reported. Second, substantiation is far from a gold standard as an indicator of maltreatment because state- and local-level policies vary for screening in, investigating, and substantiating maltreatment. Third, while reports that were accepted by CPS but did not reach a disposition prior to September 30 were excluded from all comparison periods, the actual proportion of reports reaching a disposition may not be consistent with that of prior years. Thus, the end of FFY reporting trend observed in the later weeks of 2020 (August–September) may not reflect actual reporting in those months.

This analysis allows us to understand some aspects of what resulted from the multiple and abrupt changes and adaptations affecting the CPS system due to the COVID-19 pandemic at a national level. While the magnitude of the changes varied, nationally observed patterns were largely consistent from state to state, especially for overall reporting. Researchers, the media, and child advocates have speculated whether the drop in reporting obscured an increased danger to children experiencing unreported maltreatment (Ingram, 2020; LeBlanc, 2020; Martins-Filho et al., 2020; Mathematica, 2020; Rodriguez et al., 2021). As schools and child daycare centers were closed, and in some cases transitioned to virtual learning, the highest drops were among school-based reporters, who typically comprise nearly a quarter of reports. (U.S. Department of Health & Human Services, 2021; U.S. Department of Health & Human Services, 2022). Reports also declined but to a far lesser degree from law enforcement, medical, mental health, and social services providers, as well as nonprofessional sources, as families had fewer opportunities to be observed in the community. Because all community sentinels initiated fewer reports that were substantiated, it is likely that some children who experienced maltreatment were missed during the comparison period. This may indicate that children remained unprotected or that families did not receive critical services. However, the wide-ranging impact of COVID-19 and coinciding events during the spring of 2020 limit the conclusions we can draw from NCANDS data about the actual maltreatment of children during this time.

Finally, data from NCANDS suggest that following the initial COVID-19 lockdown period, reporting and substantiation during the summer months of 2020 was consistent with the seasonal pattern of previous years, as shown in Fig. 2. COVID-19 has continued to impact educational and health care settings which in turn may have affected maltreatment reporting. Because of the wide variation in COVID-19 safety regulations and practices around the country, we expect that these patterns may vary as well. As some remote service delivery may remain prevalent event after the pandemic, a reduction in in-person contact between children and service providers could result in a long-term decline in opportunities for surveillance of maltreatment by mandated reporters. Only after NCANDS data for FFY 2021 are available will we know whether reporting during the school year of 2020–2021 remained suppressed, returned to a pattern similar to other years, or if potential spikes in maltreatment reporting are observed that reflect latent maltreatment, and to what extent any national pattern was consistent across states.

Declaration of competing interest

None.