Raising suspicion of maltreatment from burns: Derivation and validation of the BuRN-Tool

Deliberate Practice as an Effective Method for Reducing Diagnostic Error in Identifying Burn and Bruise Injuries Suspicious for an Abusive Injury

OBJECTIVES

To examine the effectiveness of an education intervention for reducing physician diagnostic error in identifying pediatric burn and bruise injuries suspicious for abuse, and to determine case-specific variables associated with an increased risk of diagnostic error.

STUDY DESIGN

This was a multicenter, prospective, cross-sectional study. A convenience sample of pediatricians and other front-line physicians who treat acutely injured children in the United States and Canada were eligible for participation. Using a web-based education and assessment platform, physicians deliberately practiced with a spectrum of 300 pediatric burn and bruise injury image-based cases. Participants were asked if there was a suspicion for abuse present or absent, were given corrective feedback after every case, and received summative diagnostic performance overall (accuracy), suspicion for abuse present (sensitivity), and absent (specificity).

RESULT

Of the 93/137 (67.9%) physicians who completed all 300 cases, there was a significant reduction in diagnostic error (initial 16.7%, final 1.6%; delta -15.1%; 95% CI 13.5, 16.7), sensitivity error (initial 11.9%, final 0.7%; delta -11.2%; 95% CI 9.8, 12.5), and specificity error (initial 23.3%, final 6.6%; delta -16.7%; 95% CI 14.8, 18.6). Based on 35,627 case interpretations, variables associated with diagnostic error included patient age, sex, skin color, mechanism of injury, and size and pattern of injury.

CONCLUSION

The education intervention substantially reduced diagnostic error in differentiating the presence versus absence of a suspicion for abuse in children with burn and bruise injuries. Several case-based variables were associated with diagnostic error, and these data can be used to close specific skill gaps in this clinical domain.

Identification and Evaluation of Non-Accidental Trauma in the Pediatric Population: A Clinical Review

Non-accidental trauma (NAT) is a major cause of morbidity and mortality for children around the world and most significantly impacts children under one year of age. Prompt and comprehensive treatment of these children relies on a high index of suspicion from any medical provider that treats pediatric patients. This review discusses those most at risk for experiencing NAT, and common initial presentations, to assist providers in the identification of potential victims. In addition, this review provides guidance on the recommended workup for these patients so that the full extent of associated injuries may be identified and the appropriate healthcare team may be assembled.

The impact of a clinical prediction tool (BuRN-Tool) for child maltreatment on social care outcomes for children attending hospital with a burn or scald injury

Burns are common childhood injuries and 10-20% are associated with maltreatment. This prospective before/after study investigated the impact of introducing the BuRN-Tool (a child maltreatment clinical prediction tool), on actions taken by children's social care department (CSC). Before introduction (pre-intervention): we collected standardised data on cause and characteristics of burns, in four regional hospitals. A BuRN-Tool-score was calculated retrospectively pre-intervention and by the attending clinician post-intervention. CSC involvement and actions taken relative to BuRN-Tool-score were compared pre- and post-BuRN-Tool. Data were collected for 1688 children from 17 local authorities. The percentage that received a CSC action decreased post-BuRN-Tool (pre: 58.0%, 51/88; post: 37.5%, 33/88, p = 0.007). A greater percentage of cases with a BuRN-Tool-score of ≥ 3 had a CSC action, than those with a BuRN-Tool-score 3, pre-intervention (≥3 70.0%, 35/50; = 0.04) and post-intervention (≥3 50.0%, 21/42; = 0.01). Children with a BuRN-Tool-score ≥ 3 but no contact/referral recorded by CSC for the burn, and those who had a contact/referral but no action taken, were significantly more likely than those scoring 3 to have new CSC involvement within six months following the burn. The BuRN-Tool-score ≥ 3 has the potential to alert clinicians to maltreatment concerns.

Development and Validation of a Natural Language Processing Tool to Identify Injuries in Infants Associated With Abuse

OBJECTIVES

Medically minor but clinically important findings associated with physical child abuse, such as bruises in pre-mobile infants, may be identified by frontline clinicians yet the association of these injuries with child abuse is often not recognized, potentially allowing the abuse to continue and even to escalate. An accurate natural language processing (NLP) algorithm to identify high-risk injuries in electronic health record notes could improve detection and awareness of abuse. The objectives were to: 1) develop an NLP algorithm that accurately identifies injuries in infants associated with abuse and 2) determine the accuracy of this algorithm.

METHODS

An NLP algorithm was designed to identify ten specific injuries known to be associated with physical abuse in infants. Iterative cycles of review identified inaccurate triggers, and coding of the algorithm was adjusted. The optimized NLP algorithm was applied to emergency department (ED) providers' notes on 1344 consecutive sample of infants seen in 9 EDs over 3.5 months. Results were compared with review of the same notes conducted by a trained reviewer blind to the NLP results with discrepancies adjudicated by a child abuse expert.

RESULTS

Among the 1344 encounters, 41 (3.1%) had one of the high-risk injuries. The NLP algorithm had a sensitivity and specificity of 92.7% (95% confidence interval [CI]: 79.0%-98.1%) and 98.1% (95% CI: 97.1%-98.7%), respectively, and positive and negative predictive values were 60.3% and 99.8%, respectively, for identifying high-risk injuries.

CONCLUSIONS

An NLP algorithm to identify infants with high-risk injuries in EDs has good accuracy and may be useful to aid clinicians in the identification of infants with injuries associated with child abuse.

Screening Tools for Child Abuse Used by Healthcare Providers: A Systematic Review

BACKGROUND

The early detection of child abuse is essential for children at risk. Healthcare professionals working at hospitals and in the community are often the first to encounter suspected cases of child abuse. Therefore, an accurate identification of child abuse is critical for intervention. However, there is no consensus on the best method to screen for child abuse.

PURPOSE

This systematic review was designed to evaluate the relevant psychometric properties and critically appraise the methodological quality of child abuse screening tools used by healthcare providers with children less than 18 years old.

METHODS

We searched the Cochrane Library, MEDLINE, Embase, CINAHL, Education Resources Information Center, PubMed, Airiti Library, and OpenGray databases for studies on screening tools used to identify abuse in children published through October 2019 in English or Chinese. Information regarding populations, assessment methods, and accuracy parameters were extracted. Study quality was assessed using the COnsensus-based Standards for the selection of health Measurement INstruments checklist and Grading of Recommendation, Assessment, Development, and Evaluation criteria.

RESULTS

Nine hundred thirty-nine abstracts and 23 full-text articles were reviewed for eligibility, and 15 screening tools for child abuse used by healthcare providers were identified. Screening tools often assess the presence of more than one form of abuse, but no single tool covered all forms. Of these, 10 tools screened for a single, discrete type of abuse, including nine physical abuse screening tools (three abusive head trauma tools) and one sexual abuse tool. Eighty percent (n = 12) of the screening tools had a moderate-to-high quality of evidence based on the Grading of Recommendation, Assessment, Development, and Evaluation criteria. However, none of these screening tools achieved an adequate level of evidence based on the COnsensus-based Standards for the selection of health Measurement INstruments checklist.

CONCLUSIONS/IMPLICATIONS FOR PRACTICE

In this systematic literature review, 15 assessment tools of child abuse used by healthcare providers were identified, of which nine screened for physical abuse. Screening tools must be valid, succinct, user-friendly, and amenable for use with children at every point of care in the healthcare system. Because of the paucity of informative and practical studies in the literature, findings related to the quality of child abuse screening tools were inconclusive. Therefore, future research should focus on the use of screening tools in the healthcare system to identify effective screening interventions that may help healthcare providers identify child abuse cases as early as possible.

Fifteen-minute consultation: Management of paediatric minor burns

Burn injuries are a common presentation to the paediatric emergency department (PED) and are painful, distressing and may have long-term sequelae. In adhering to the first aid principles of burns management, we aim to alleviate pain, prevent bacterial contamination and minimise the extent of injury. First aid involves cooling the burn and covering with an appropriate material while simultaneously providing analgesia. Assessing the severity (depth and total body surface area) of the burn are important for prognostication with implications for management. It is imperative to consider non-accidental injury in burns, which be present in 10% of cases .

The impact of a child abuse guideline on differences between pediatric and community emergency departments in the evaluation of injuries

BACKGROUND

Although child physical abuse is missed more frequently in community (CEDs) vs. pediatric emergency departments (PEDs), little information exists describing how evaluations of high-risk injuries differ between these settings.

OBJECTIVES

To determine differences in evaluations of infants for abuse between a PED and CEDs and whether a child abuse guideline reduced these differences.

PARTICIPANTS AND SETTING

Infants presenting to one PED (n = 162) and three CEDs (n = 159) with 3 injury categories: 1) Injuries for which the American Academy of Pediatrics recommends skeletal survey (SS) testing (infants <5-months with an oral injury or bruising, <9-months with a non-skull fracture, and < 12-months with an intracranial hemorrhage); 2) an oral injury or high-risk bruising in older infants; and 3) multiple types of high-risk injuries.

METHODS

We assessed differences in SS testing and child protective services (CPS) reporting between the PED and CEDs before and after implementation of a child abuse guideline.

RESULTS

The median (IQR) age was 4 months (2-7). Before guideline implementation, infants with injuries in categories 1 and 2 had an increased odds of SS testing in the PED vs. the CEDs (Category 1: aOR 2.83, 95% CI: 1.01-8.10; Category 2: aOR 10.1, CI: 1.2-88.0) and CPS reporting (Category 1: aOR 7.96, CI: 2.3-26.7; Category 2: aOR 12.0, CI: 1.4-103.5). After guideline implementation, there were no statistically significant differences in testing and reporting for any injury category.

CONCLUSIONS

Implementation of a child abuse guideline minimized differences between a PED and CEDs in the evaluation of infants with injuries concerning for abuse.

Validation of a Clinical Decision Rule to Predict Abuse in Young Children Based on Bruising Characteristics

IMPORTANCE

Bruising caused by physical abuse is the most common antecedent injury to be overlooked or misdiagnosed as nonabusive before an abuse-related fatality or near-fatality in a young child. Bruising occurs from both nonabuse and abuse, but differences identified by a clinical decision rule may allow improved and earlier recognition of the abused child.

OBJECTIVE

To refine and validate a previously derived bruising clinical decision rule (BCDR), the TEN-4 (bruising to torso, ear, or neck or any bruising on an infant <4.99 months of age), for identifying children at risk of having been physically abused.

DESIGN, SETTING, AND PARTICIPANTS

This prospective cross-sectional study was conducted from December 1, 2011, to March 31, 2016, at emergency departments of 5 urban children's hospitals. Children younger than 4 years with bruising were identified through deliberate examination. Statistical analysis was completed in June 2020.

EXPOSURES

Bruising characteristics in 34 discrete body regions, patterned bruising, cumulative bruise counts, and patient's age. The BCDR was refined and validated based on these variables using binary recursive partitioning analysis.

MAIN OUTCOMES AND MEASURES

Injury from abusive vs nonabusive trauma was determined by the consensus judgment of a multidisciplinary expert panel.

RESULTS

A total of 21 123 children were consecutively screened for bruising, and 2161 patients (mean [SD] age, 2.1 [1.1] years; 1296 [60%] male; 1785 [83%] White; 1484 [69%] non-Hispanic/Latino) were enrolled. The expert panel achieved consensus on 2123 patients (98%), classifying 410 (19%) as abuse and 1713 (79%) as nonabuse. A classification tree was fit to refine the rule and validated via bootstrap resampling. The resulting BCDR was 95.6% (95% CI, 93.0%-97.3%) sensitive and 87.1% (95% CI, 85.4%-88.6%) specific for distinguishing abuse from nonabusive trauma based on body region bruised (torso, ear, neck, frenulum, angle of jaw, cheeks [fleshy], eyelids, and subconjunctivae), bruising anywhere on an infant 4.99 months and younger, or patterned bruising (TEN-4-FACESp).

CONCLUSIONS AND RELEVANCE

In this study, an affirmative finding for any of the 3 BCDR TEN-4-FACESp components in children younger than 4 years indicated a potential risk for abuse; these results warrant further evaluation. Clinical application of this tool has the potential to improve recognition of abuse in young children with bruising.

The Use of Experts to Evaluate a Child Abuse Guideline in Community Emergency Departments

OBJECTIVE

Guidelines and pathways exist to help frontline providers evaluate injured children for suspected child abuse. Little, however, is known about whether the decision-making resulting from these interventions is correct. Therefore, in the absence of an available gold-standard test, we used experts' judgments to examine the appropriateness of these clinical decisions. We evaluated community emergency department (ED) providers' adherence to a guideline recommending a child protection team (CPT) consultation for infants with injuries associated with abuse. We then compared providers' decision-making to experts' recommendations before and after guideline implementation.

METHODS

Two experts conducted a blinded, retrospective review of injured infants from 3 community EDs (N = 175). Experts rated the likelihood that an injury was abusive, indeterminate, or accidental, and made recommendations that were compared with skeletal survey (SS) testing and child protective services (CPS) reporting by providers before and after guideline implementation.

RESULTS

Postguideline implementation, there was a significant increase in CPT consultations in indeterminate cases (14.3% vs 72.2%, P < .001) and in SS testing when experts recommended SS (20.6% vs 56.8%, P = .002). In contrast, a higher percentage of cases for whom the experts did not recommend reporting were reported to CPS (1.8% vs 14.6%, P = .02).

CONCLUSIONS

Providers consulted the CPT most often for indeterminate cases. SS testing was in line with expert recommendations, but CPS reporting diverged from expert recommendations. Interventions linking community ED providers with a CPT may improve the evaluation of infants with injuries concerning for abuse.

A natural language processing and deep learning approach to identify child abuse from pediatric electronic medical records

Child physical abuse is a leading cause of traumatic injury and death in children. In 2017, child abuse was responsible for 1688 fatalities in the United States, of 3.5 million children referred to Child Protection Services and 674,000 substantiated victims. While large referral hospitals maintain teams trained in Child Abuse Pediatrics, smaller community hospitals often do not have such dedicated resources to evaluate patients for potential abuse. Moreover, identification of abuse has a low margin of error, as false positive identifications lead to unwarranted separations, while false negatives allow dangerous situations to continue. This context makes the consistent detection of and response to abuse difficult, particularly given subtle signs in young, non-verbal patients. Here, we describe the development of artificial intelligence algorithms that use unstructured free-text in the electronic medical record-including notes from physicians, nurses, and social workers-to identify children who are suspected victims of physical abuse. Importantly, only the notes from time of first encounter (e.g.: birth, routine visit, sickness) to the last record before child protection team involvement were used. This allowed us to develop an algorithm using only information available prior to referral to the specialized child protection team. The study was performed in a multi-center referral pediatric hospital on patients screened for abuse within five different locations between 2015 and 2019. Of 1123 patients, 867 records were available after data cleaning and processing, and 55% were abuse-positive as determined by a multi-disciplinary team of clinical professionals. These electronic medical records were encoded with three natural language processing (NLP) algorithms-Bag of Words (BOW), Word Embeddings (WE), and Rules-Based (RB)-and used to train multiple neural network architectures. The BOW and WE encodings utilize the full free-text, while RB selects crucial phrases as identified by physicians. The best architecture was selected by average classification accuracy for the best performing model from each train-test split of a cross-validation experiment. Natural language processing coupled with neural networks detected cases of likely child abuse using only information available to clinicians prior to child protection team referral with average accuracy of 0.90±0.02 and average area under the receiver operator characteristic curve (ROC-AUC) 0.93±0.02 for the best performing Bag of Words models. The best performing rules-based models achieved average accuracy of 0.77±0.04 and average ROC-AUC 0.81±0.05, while a Word Embeddings strategy was severely limited by lack of representative embeddings. Importantly, the best performing model had a false positive rate of 8%, as compared to rates of 20% or higher in previously reported studies. This artificial intelligence approach can help screen patients for whom an abuse concern exists and streamline the identification of patients who may benefit from referral to a child protection team. Furthermore, this approach could be applied to develop computer-aided-diagnosis platforms for the challenging and often intractable problem of reliably identifying pediatric patients suffering from physical abuse.

Evaluation of the efficacy and impact of a clinical prediction tool to identify maltreatment associated with children's burns

BACKGROUND

An estimated 10%-24% of children attending emergency departments with a burn are maltreated.

OBJECTIVE

To test whether a clinical prediction tool (Burns Risk assessment for Neglect or abuse Tool; BuRN-Tool) improved the recognition of maltreatment and increased the referral of high-risk children to safeguarding services for assessment.

METHODS

A prospective study of children presenting with burns to four UK hospitals (2015-2018), each centre providing a minimum of 200 cases before and after the introduction of the BuRN-Tool. The proportions of children referred to safeguarding services were compared preintervention and postintervention, and the relationship between referral and the recommended cut-off for concern (BuRN-Tool score (BT-score) ≥3) was explored.

RESULTS

The sample was 2443 children (median age 2 years). Nurses and junior doctors mainly completed the BuRN-Tool, and a BT-score was available for 90.8% of cases. After intervention, 28.4% (334/1174) had a BT-score ≥3 and were nearly five times more likely to be discussed with a senior clinician than those with a BT-score <3 (65.3% vs 13.4%, p<0.001). There was no overall difference in the proportion of safeguarding referrals preintervention and postintervention. After intervention, the proportion of referrals for safeguarding concerns was greater when the BT-score was ≥3 (p=0.05) but not for scores <3 (p=0.60). A BT-score of 3 as a cut-off for referral had a sensitivity of 72.1, a specificity of 82.7 and a positive likelihood ratio of 4.2.

CONCLUSIONS

A BT-score ≥3 encouraged discussion of cases of concern with senior colleagues and increased the referral of <5 year-olds with safeguarding concerns to children's social care.

Child Protection Team Consultation for Injuries Potentially Due to Child Abuse in Community Emergency Departments

OBJECTIVES

Emergency care for children is provided predominantly in community emergency departments (CEDs), where abusive injuries frequently go unrecognized. Increasing access to regional child abuse experts may improve detection of abuse in CEDs. In three CEDs, we intervened to increase involvement of a regional hospital child protection team (CPT) for injuries associated with abuse in children < 12 months old. We aimed to increase CPT consultations about these infants from the 3% baseline to an average of 50% over 12 months.

METHODS

We interviewed CED providers to identify barriers and facilitators to recognizing and reporting abuse. Providers described difficulties differentiating abusive from nonabusive injuries and felt that a second opinion would help. Using a plan-do-study-act approach, beginning in April 2018, we tested, refined, and implemented interventions to increase the frequency of CPT consultation, including leadership and champion engagement, provider training, clinical pathway implementation, and an audit and feedback process. Data were collected for 15 months before and 17 months after initiation of interventions. We used a statistical process control chart to track CPT consultations about children < 1 year old with high-risk injuries, use of skeletal surveys (SSs), and reports to child protective services (CPS).

RESULTS

Evidence of special cause was identified beginning in June 2018, with a shift of 8 points to one side of the center line. For the subsequent 8-month period, the CPT was consulted for a mean of 47.5% of children with high-risk injuries; this was sustained for an additional 7 months. The average percentage of infants with high-risk injuries who received a SS increased from 6.7% to 18.9% and who were reported to CPS increased from 10.7% to 32.6%.

CONCLUSION

Targeted interventions in CEDs increased the frequency of CPT consultation, SS use, and reports to CPS for infants with high-risk injuries. Such interventions may improve recognition of physical abuse.

Burns and Scalds Assessment Template: standardising clinical assessment of childhood burns in the emergency department

Objectives

The Burns and Scalds Assessment Template (BaSAT) is an evidence-based proforma coproduced by researchers and ED staff with the aim of (1) standardising the assessment of children attending ED with a burn, (2) improving documentation and (3) screening for child maltreatment. This study aimed to test whether the BaSAT improved documentation of clinical, contributory and causal factors of children’s burns.

Methods

A retrospective before-and-after study compared the extent to which information was recorded for 37 data fields after the BaSAT was introduced in one paediatric ED. Pre-BaSAT, a convenience sample of 50 patient records of children who had a burn was obtained from the hospital electronic database of 2007. The post-BaSAT sample included 50 randomly selected case notes from 2016/2017 that were part of another research project. Fisher's exact test and Mann-Whitney U tests were conducted to test for statistical significance.

Results

Pre-BaSAT, documentation of key data fields was poor. Post-BaSAT, this varied less between patients, and median completeness significantly (p<0.001) increased from 44% (IQR 4%–94%) to 96% (IQR 94%–100%). Information on ‘screening for maltreatment, referrals to social care and outcome’ was poorly recorded pre-BaSAT (median of 4% completed fields) and showed the greatest overall improvement (to 95%, p<0.001). Documentation of domestic violence at home and child’s ethnicity improved significantly (p<0.001) post-BaSAT; however, these were still not recorded in 36% and 56% of cases, respectively.

Conclusion

Introduction of the BaSAT significantly improved and standardised the key clinical data routinely recorded for children attending ED with a burn.

Identifying children exposed to maltreatment: a systematic review update

Background

Child maltreatment affects a significant number of children globally. Strategies have been developed to identify children suspected of having been exposed to maltreatment with the aim of reducing further maltreatment and impairment. This systematic review evaluates the accuracy of strategies for identifying children exposed to maltreatment.

Methods

We conducted a systematic search of seven databases: Medline, Embase, PsycINFO, Cumulative Index to Nursing and Allied Health Literature, Cochrane Libraries, Sociological Abstracts and the Education Resources Information Center. We included studies published from 1961 to July 2, 2019 estimating the accuracy of instruments for identifying potential maltreatment of children, including neglect, physical abuse, emotional abuse, and sexual abuse. We extracted data about accuracy and narratively synthesised the evidence. For five studies—where the population and setting matched known prevalence estimates in an emergency department setting—we calculated false positives and negatives. We assessed risk of bias using QUADAS-2.

Results

We included 32 articles (representing 31 studies) that evaluated various identification strategies, including three screening tools (SPUTOVAMO checklist, Escape instrument, and a 6-item screening questionnaire for child sex trafficking). No studies evaluated the effects of identification strategies on important outcomes for children. All studies were rated as having serious risk of bias (often because of verification bias). The findings suggest that use of the SPUTOVAMO and Escape screening tools at the population level (per 100,000) would result in hundreds of children being missed and thousands of children being over identified.

Conclusions

There is low to very low certainty evidence that the use of screening tools may result in high numbers of children being falsely suspected or missed. These harms may outweigh the potential benefits of using such tools in practice (PROSPERO 2016:CRD42016039659).

[Burns by immersion in children and abuse]

AIM OF THE STUDY

Ten percent of childhood burns could arise from maltreatment. While describing 6 severe cases of inflicted scalds by immersion in children, we expose our systematic diagnostic approach of abuse and confirm the serious nature of burn when they are inflicted.

PATIENTS AND METHOD

The retrospective study concerned children hospitalized for scalds by immersion between 2013 and 2016 and for whom child abuse has been confirmed. Sex, age, burns description, needs of surgery, length of stay at hospital and protection plan set up were collected.

RESULTS

Six cases of burns by immersion due to maltreatment were identified (5 boys, 1 girl) with a median age of 12 months. The average total burn surface area was 19%. Burns were of deep second and third degree and always symmetric. Every child underwent surgery at least once. Concern information was transferred for all of them.

CONCLUSION

Teams taking care of children with burns must be trained to the difficult diagnostic of abuse or neglect so that early social interventions can be set up in case of maltreatment.

First aid for children's burns in the US and UK: An urgent call to establish and promote international standards

INTRODUCTION

Appropriate first aid can reduce the morbidity of burns, however, there are considerable variations between international first aid recommendations. We aim to identify, and compare first aid practices in children who present to Emergency Departments (ED) with a burn.

METHODS

A prospective cross-sectional study of 500 children (0-16 completed years) presenting with a burn to a paediatric ED in the UK (Cardiff) and the USA (Denver, Colorado), during 2015-2017. The proportion of children who had received some form of first aid and the quality of first aid were compared between cities.

RESULTS

Children attending hospital with a burn in Cardiff were 1.47 times more likely (RR 1.47; CI 1.36, 1.58), to have had some form of first aid than those in Denver. Denver patients were 4.7 time more likely to use a dressing and twice as likely to apply ointment/gel/aloe vera than the Cardiff cohort. First aid consistent with local recommendations was only administered to 26% (128/500) of children in Cardiff and 6% (31/500) in Denver. Potentially harmful first aid e.g. application of food, oil, toothpaste, shampoo or ice was applied to 5% of children in Cardiff and 10% in Denver.

CONCLUSION

A low number of children received optimal burns first aid, with potentially harmful methods applied in a considerable proportion of cases. There is an urgent need for internationally agreed, evidence-based burn first aid recommendations.