SCAN for abuse: Electronic health record-based universal child abuse screening

Abstract

Background:

Identification of physical abuse at the point of care without a systematic approach remains inherently subjective and prone to judgement error. This study examines the implementation of an electronic health record (EHR)-based universal child injury screen (CIS) to improve detection rates of child abuse.

Methods:

CIS was implemented in the EHR admission documentation for all patients age 5 or younger at a single medical center, with the following questions:

1. “Is this patient an injured/trauma patient?”

2. “If this is a trauma/injured patient, where did the injury occur?”

A “Yes” response to Question 1 would alert a team of child abuse pediatricians and social workers to determine if a patient required formal child abuse clinical evaluation. Patients who received positive CIS responses, formal child abuse work-up, and/or reports to Child Protective Services (CPS) were reviewed for analysis. CPS rates from historical controls (2017–2018) were compared to post-implementation rates (2019–2021).

Results:

Between 2019–2021, 14,150 patients were screened with CIS. 286 (2.0%) patients screened received positive CIS responses. 166 (58.0%) of these patients with positive CIS responses would not have otherwise been identified for child abuse evaluation by their treating teams. 18 (10.8%) of the patients identified by the CIS and not by the treating team were later reported to CPS. Facility CPS reporting rates for physical abuse were 1.2 per 1000 admitted children age 5 or younger (pre-intervention) versus 4.2 per 1000 (post-intervention).

Conclusions:

Introduction of CIS led to increased detection suspected child abuse among children age 5 or younger.

INTRODUCTION

Although child maltreatment, defined as child abuse or neglect, is among the leading causes of death in American children [1], guidelines for maltreatment screening remain scarce. The adverse effects and mortality of maltreatment are well known [2] and mandatory reporters such as physicians, nurses, and social workers are legally obligated to report suspected maltreatment. Despite this, the Children’s Hospital Association (formerly NACHRI) does not suggest any screening methods or provide guidelines on implementing a screening protocol [3]. Published studies and systematic reviews discuss a variety of hospital screening protocols but provide limited information about their performance. Even fewer studies have discussed the implementation of universal child abuse screening [4–14]. Contributing to the issue is a lack of standardized metrics for measuring and reporting the performance of a hospital in screening for child maltreatment over time, making systematic comparisons difficult.

The challenge of diagnosing child maltreatment lies not only in the difficulty of eliciting clear information from caregivers and identifying non-specific warning signs, but also in the consequences of false positives and false negatives, which incur a high cost [15]. Caregivers falsely accused of maltreatment may face costly legal implications, which could include loss of custody, while missed diagnoses expose children to higher risks of injury and mortality resulting from recurrent maltreatment [16]. Child maltreatment disproportionally affects young children who often lack the capacity to advocate for themselves [17]. Systematic reviews conclude that standardized questionnaires are essential to improving child maltreatment detection [4, 18]. Our objective was to assess and improve child abuse screening at our institution, a 360-bed standalone children’s hospital with a Level 1 Pediatric Trauma center, by developing a standardized, universal framework to screen for physical child abuse.

METHODS

Initial Assessment

We began by assessing existing processes for child abuse assessment and reporting at our institution through staff interviews, qualitative patient flow analysis, and quantitative electronic health record (EHR) analysis. We identified patient intake points to find gaps in screening, examined EHR functionality to find opportunities for EHR-based screening, and used EHR and nationally reported data to compare our institution’s Child Protective Services (CPS) reporting data with those of surrounding counties and the State of California. We noted that CPS reporting rates at our center were significantly lower than those of surrounding counties and the State of California.

Development of new screening protocol

Following assessment of existing screening methods, we considered the workload of nurses, physicians, social workers, and the Suspected Child Abuse and Neglect (SCAN) team (a group of specially trained child abuse pediatricians and social workers) to guide the design of our screening protocol. We developed a prototype EHR-based screening questionnaire with an emphasis on high sensitivity and conducted a feasibility analysis using retrospective chart data to test the capacity of hospital staff to implement the screening protocol.

Implementation of new screening protocol

The new screening protocol was implemented in the nursing admission workflow as a required screening element for all admitted children in January 2019. Counts of patients screened, flagged as at-risk after chart review, and reported to CPS were collected. Although nurses received training on the new screening protocol, treatment teams were not made aware of the new screening protocol to avoid potential confounding of education gained through protocol training that could potentially alter a provider’s routine screening practice. Treatment teams were only alerted when the screening protocol determined that additional testing was necessary and had not been performed by the treatment team. Data was collected on the number of patients who were flagged by the screening protocol who should have been referred for abuse work-up by their treatment teams but were not.

Examining existing screening process

Interviews with staff and administrators revealed that providers and social workers were expected to leverage their individual expertise to identify cases of possible abuse. There was no mandated child maltreatment identification training for staff. If abuse was suspected, cases could be directly reported to CPS or referred to the SCAN team, but the referral process was neither automated nor standardized.

Identifying gaps in existing screening

We mapped the patient entry points for our center and examined EHR data to quantify the number of patients passing through each entry point (Figure 1). We identified three entry points: scheduled check-in (78.9% of the 19,414 patients admitted in 2016) which includes planned admissions for procedures and outpatient surgery, arrival from the emergency department (ED) (11.8%), or transfer from another facility (9.3%). Social workers were available for consults in the ED, acute care units, and outpatient units, but did not evaluate patients entering via scheduled check-in. Many patients never passed through a department with social workers during their hospitalization; of the 12,037 procedures performed in 2016, 56% were outpatient, meaning patients passed through a procedural area and were discharged from the procedural area or post-anesthesia care unit, neither of which include social workers.

Fig. 1. Patient Flow Diagram.

Identifies pre-intervention gaps in social work coverage where there is increased risk of missed cases of abuse. Three entry points for hospital admission are shown on the left, with receiving departments shown in the box on the right. CVICU = cardiovascular intensive care unit, ICU = intensive care unit, NICU = neonatal intensive care unit, PACU = post-anesthesia care unit, PICU = pediatric intensive care unit.

Development of new screening protocol

After assessing the capacity for nursing and the SCAN team to accommodate a new screening protocol, a system was developed that relied on two steps: 1) an initial, child injury screen (CIS) based on an automated questionnaire in the EHR and 2) a secondary chart review by the SCAN team with follow-up tests, if necessary.

We determined that a brief EHR-based questionnaire for patients in a high-risk age group for physical abuse (age 5 or younger) would limit the number of patients requiring further evaluation by the SCAN team without placing too much initial screening burden on nurses. Children in this age group experience a higher-than-average rate of maltreatment when compared to all children under 18. In 2021, children with less than 1 year of age experienced maltreatment at a rate of 24.5 per 1,000 children. Children ages 1–5 experienced maltreatment at a rate of 9.2 per 1,000 children, versus a national average of 7.9 per 1,000 children under age 18 [17]. Our new screening tool was adapted from the system used at Nationwide Children’s Hospital Emergency Department (Columbus, Ohio) [6]. Our modifications included changing the screening process to take place during admission, rather than on arrival to the ED, as this would miss a large percentage of admitted patients. CIS consists of the following two items and appears automatically in the EHR nursing admission workflow for all children age 5 and under:

1. “Is the patient an injured/trauma patient?” (Response choices: “Yes” or “No”)

2. “If the patient is an injured/trauma patient, where did the injury occur?” (Response choices: “Care center”, “home”, or “other”)

Answering yes to Question 1 automatically flags the patient’s chart for review by the SCAN team to ensure high sensitivity and minimize missed cases of abuse. Question 2 was added based on empirical SCAN data showing that abuse was more likely in injuries occurring in private settings such as homes or daycare centers. Answers are used by the SCAN team to determine the likelihood of abuse during chart review.

Subsequent screening steps rely on the SCAN team’s expertise to remove false positives later in the protocol. The initial silent review of flagged charts by the SCAN team involves no documentation of their involvement in the patient’s chart nor direct patient interaction; rather, they review the chart for any concerning signs of maltreatment within 24 hours of the positive CIS result. The SCAN team examines patients’ histories for vague or changing explanations for significant injuries, delays in seeking care, and previous histories of inflicted injury. Additionally, they look for injury or burn patterns frequently identified in abuse such as torso, ear, and neck injuries or widely spaced bilateral burns in different stages of healing. Existing radiographic studies are examined for findings such as long bone fractures, rib fractures, or subdural hemorrhage. If concerning signs are found, the SCAN team documents their findings and begins a formal workup involving further chart review, family and staff interviews, and orders for additional evaluation as needed. This two-step review process was implemented to reduce the high number of false positives we expected to collect with the CIS questionnaire. If formal workup suggests patients are victims of maltreatment, SCAN files a Suspected Child Abuse Report, which at our institution triggers a report to CPS. The screening workflow is documented in Figure 2.

Fig. 2. Screening Protocol.

CPS = child protective services.

This study was approved under the Stanford Institutional Review Board (IRB-59963).

RESULTS

Feasibility assessment using historical controls

By applying a screening tool that included only patients who were under 5 and injured in the home to retrospective chart review data from 2016, we determined the protocol would have referred 131 patients per year for review by the SCAN team. The screening method would have captured 48 of 54 total inpatients reported to CPS and flagged an additional 83 for chart review. Of the approximately 8,000 inpatients under the age of 5, 440 patients were admitted with an injury. We reviewed this anticipated case load with the SCAN team prior to rollout and confirmed their ability to handle these additional reviews. After additional discussions, it was decided to refer all patients age 5 or younger with any type of injury or suspected abuse to the SCAN team for chart review, regardless of where the injury took place.

Implementation of new screening workflow

Counts of patients, ages, and genders through each step of the screening workflow from January 1, 2019 to December 31, 2021 are summarized in Table 1. 14,150 patients were screened using CIS. Of those screened, 286 (2.0%) patients received a positive result on the CIS (indicating they were a patient with injury).

Table 1:

CIS Demographics (January 1, 2019 – December 31, 2021)

	Screened using CIS	Positive CIS	Formal Workup	CPS Report
Number of patients	14,150	286	106	78
Age, mean (SD)	2.2 (1.8)	2.3 (2.0)	1.3 (1.4)	1.5 (1.6)
Sex, n (%)
Male	8,069 (57)	172 (60)	61 (58)	43 (55)
Female	6,081 (43)	114 (40)	45 (42)	35 (45)

Following silent review of the 286 patients with a positive CIS, 106 (37.1%) of these patients progressed to a formal workup of possible abuse. 78 (73.6%) of the patients who received a formal workup were ultimately reported to CPS.

Missed cases caught by CIS

Of the 286 patients with positive CIS responses, 166 (58%) of these patients did not have a child abuse consultation ordered by their medical team. Of these patients who were identified as at risk of physical abuse by CIS but not by their medical teams, 18 (10.8%) were reported to CPS after a SCAN team consultation; 14 of these cases occurred in 2019, 3 in 2020, and 1 in 2021.

CPS reporting rates following rollout

The CPS reporting rate for physical abuse in patients age 5 and younger at our institution rose from 1.2 per 1000 admitted patients in the two years prior to rollout of the screening tool (2017–2018) to 4.2 per 1000 admitted patients after rollout (2019–2021). CPS reporting rates demonstrated even greater difference among the subpopulation of patients admitted from the emergency department; CPS reports rose from 3.6 per 1000 admits to 13.6 per 1000 admits (Table 2). CPS reporting rates for physical abuse only in surrounding counties were not available; however, overall CPS reporting rates for all forms of maltreatment in surrounding counties modestly fell from the pre-intervention period to the post-intervention period. CPS reporting rates for children 5 and under in San Mateo County fell from 22.3 per 1000 in 2017–2018 to 20.6 in 2020–2021 and in Santa Clara County fell from 29.4 per 1000 in 2017–2018 to 26.9 in 2020–2021 [19].

Table 2:

CPS report rates before and after CIS Implementation

Time period	All admissions ≤5y/o	Admits ≤5y/o except newborns	Urgent/emergent admissions ≤5y/o	Number of total admissions
2017–2018	1.2 per 1000	2.49 per 1000	3.6 per 1000	17,312
2019–2021	4.2 per 1000	9.4 per 1000	13.6 per 1000	24,641

Urgent/emergent admissions indicate admissions from the emergency department

DISCUSSION

Child abuse disproportionately affects children who are often too young to advocate for themselves or to benefit from monitoring by schoolteachers [1]. Given that mortality rates more than double in victims who experience repeated abuse compared to the first incidence of abuse, it is incumbent upon healthcare providers to develop standard methods for early identification of cases of abuse and minimizing missed cases [16]. The development process of CIS at our institution included a screening tool design with appropriate sensitivity/specificity, assessment of feasibility among stakeholders, and reporting outcomes using standardized metrics. Following implementation of CIS, we observed increased detection of CPS-reportable cases of physical abuse and increased utilization of the SCAN team by medical teams. Our standardized screening protocol identified 18 children who were reported to CPS who were not identified by their medical teams as being under suspicion for abuse. Interestingly, the number of cases caught by the screening tool but missed by medical teams has declined each year, which we attribute to increased awareness of child abuse resources and increased direct case referral to the SCAN team by providers. The combined effect of increased referrals and increased detection with the novel screening process can be seen in the increased post-intervention CPS reporting rate.

Assessing the accuracy of reporting remains a challenge due to a lack of consistent outcomes data shared between CPS and the reporting institution. Successful tracking of CPS substantiation rates could help inform the accuracy of screening protocols and mitigate the risk of increasing CPS reporting rates only through increasing false positives. A nationally standardized mechanism for hospitals and CPS to share data could improve assessment accuracy for screening programs. Furthermore, the development of standardized metrics for reporting child abuse reporting rates, substantiation rates, and a proxy for identifying patients who may have previously been missed by the screening protocol (i.e., patients reported to CPS who had previously presented with sentinel injuries) would allow for better hospital-to-hospital comparisons of performance.

Given the differences in patient populations, social work resources, and SCAN specialist availability from center to center, the screening approach described here may not be universally replicable. However, the following design principles we identified in developing this screening process should be broadly relevant.

Integrated data collection

Data collection is crucial to gain information on past cases of child maltreatment and to evaluate performance before and after a policy is implemented. Integration of the screening tool into an EHR system allows for improved collection as well as opportunities for automated referral of cases to the proper groups through each step of the screening process.

Appropriate balance of sensitivity vs. specificity and feasibility

Our screening approach maximizes sensitivity and objectivity in the CIS questionnaire and relies on the SCAN team to remove false positives via a chart review that did not involve documentation of suspected abuse if no signs of abuse were found. This also allowed the experts to have the final judgement instead of relying on an automated algorithm to make decisions on diagnosis and reporting.

Feasibility

Our protocol was feasible because of the size of our injured/trauma population, the limited time requirements for nurses answering the questions in the EHR, and the capacity that the SCAN team had to accommodate additional reviews. Other centers may need to adjust questionnaire items, age cutoffs, or review steps to strike a balance between sensitivity, specificity, and feasibility.

Completeness of coverage

Mapping the flow of patients through the hospital can help identify points of entry and gaps in detection of abuse. EHR data can be used to quantify the volume associated with these flows.

Staff education

Standardization of any screening workflow should proceed in tandem with hospital-wide education that discusses the steps in screening and reporting programs to facilitate standardized practice.

Limitations

This study is limited by a lack of county CPS substantiation data, preventing assessment of reporting accuracy. The counties surrounding our institution do not routinely report substantiation or case resolution data back to reporting institutions. However, the SCAN team continues to follow reported patients and captures outcome data whenever possible. Additionally, the screening tool was only focused on child physical abuse detection instead of capturing all forms of child maltreatment, including neglect. Further investigation is needed to expand this simple screen to encompass all maltreatment, a much more heterogeneous and complex entity.

CONCLUSIONS

Following rollout of this EHR-based child abuse screening program, our center experienced increased detection of CPS-reportable cases of children who might have otherwise been missed, as well as an increase in overall CPS reporting rates of physical abuse. Furthermore, these results were accomplished without placing undue burden on nurses, social workers, or providers. As an unintended benefit, health care teams have become more aware of the child abuse resources available at the hospital through learning about the screening process and are now more likely to involve the SCAN team in the evaluation of patients. Although we have identified a need for improved standardization of reporting rates and hospital CPS data exchange on a national level, our framework for developing a standardized screening protocol can be used immediately by hospitals to design a screening approach appropriate for their system.

Highlights.

• Although the adverse effects of child maltreatment are well known, guidelines for maltreatment screening remain scarce.

• Application of a standardized framework for physical child abuse screening allowed for improved detection of physical abuse. A lack of standardized metrics for abuse reporting prevents effective comparisons of screening performance between institutions.

Abbreviations

CIS

Child Injury Screen

CPS

Child Protective Services

CVICU

Cardiovascular Intensive Care Unit

ED

Emergency Department

EHR

Electronic Health Record

ICU

Intensive Care Unit

NACHRI

National Association of Children’s Hospitals and Related Institutions

NICU

Neonatal Intensive Care Unit

PACU

Post-Anesthesia Care Unit

PICU

Pediatric Intensive Care Unit

SCAN

Suspected Child Abuse and Neglect