Screening for mental health symptoms following pediatric traumatic injury: A practice management guideline (from the Pediatric Trauma Society, Society of Trauma Nurses, and Center for Pediatric Traumatic Stress)

Decades of research have documented risk of mental health difficulties, particularly posttraumatic stress symptoms (PTSS), among children with injuries.1,2 A smaller body of literature indicates risk for depression, which may co-occur with PTSS.3,4 A previous systematic review suggested that roughly 22% to 42% of youth will experience PTSS in the first month following injury, with 10% to 19% continuing to experience symptoms after 10 months.1 Importantly, the risk of re-injury, significant impairment in daily functioning, and poor health-related quality of life are associated with symptoms of posttraumatic stress disorder (PTSD), even if full diagnostic criteria are not met.4–10 Given this risk for impairing mental health symptoms, the American College of Surgeons (ACS) updated the standards in the Resources for Optimal Care of the Injured Patients in 2022 to require mental health screening for Level I and Level II Pediatric Trauma Centers.11 Specifically, trauma centers “must meet the mental health needs of trauma patients by having (1) a protocol to screen patients at high risk for psychological sequelae with subsequent referral to a mental health provider (Level I, Level II) or (2) a process for referral to a mental health provider when required (Level III).” The standard does not define which patients are “at high risk for psychological sequalae.” Further guidance on meeting this new requirement was provided in an ACS Best Practices Guidelines document.12 While this document provides a table of potential screening tools, the list is not based on a rigorous meta-analysis or systematic literature review. We aimed to conduct a systematic review to develop evidence-based practice management recommendations regarding mental health screening in pediatric patients with injuries employing a modified Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology. Specifically, guidance is needed regarding which patients, if any, to prioritize for mental health screening and which screening tools accurately and efficiently identify injured pediatric patients at risk for concurrent and/or prospective mental health difficulties. METHOD Using a modified GRADE approach for this systematic review, we assembled a group of pediatric trauma experts (pediatric psychologists, social worker, nurses, advance practice providers, trauma surgeons) from the Pediatric Trauma Society (PTS), Society of Trauma Nurses (STN), and the Center for Pediatric Traumatic Stress (CPTS) and collaborated with a GRADE expert. In a consensus conference and using voting to determine outcomes, these experts developed Population, Intervention, Comparator, and Outcomes (PICO) questions (see Supplemental Digital Content, Supplemental materials for AGREE Reporting Checklist for Clinical Practice Guidelines, https://links.lww.com/TA/E312). Relevant outcomes were identified by the working group and then each member voted independently on outcomes using a scale of 1 to 9. Rounded average scores for each outcome were categorized as critical (7 to 9), important but not critical (4 to 6), or limited importance (1 to 3). Only critically important outcomes were considered in decision making for generating final recommendations. PICO Question 1: Among pediatric patients (<18 years) with injury (P; excluding child abuse, self-harm), should screening for significant psychological symptoms only occur among those perceived to be at risk (I) versus all patients with injuries (C) in order to identify those with current symptoms and/or risk for later psychological symptoms (O)? Lay language: Should only those patients perceived to be at risk for psychological symptoms be screened or should all patients with injury be screened? PICO Question 2a: Among pediatric patients (<18 years) with injury (P; excluding child abuse, self-harm), do screening tools for identification of those with concurrent significant psychological symptoms (I) predict outcomes measured using a reliable and valid assessment of psychological symptoms (e.g., PTSD, depression; C) in an accurate manner (e.g., specific, sensitive; O)? Lay language: Which screening tool(s) should be used to identify those at risk for current psychological symptoms? PICO Question 2b: Among pediatric patients (<18 years) with injury (P; excluding child abuse, self-harm), do screening tools for identification of those with later significant psychological symptoms (I) predict outcomes measured using a reliable and valid assessment of psychological symptoms (e.g., PTSD, depression; C) in an accurate manner (e.g., specific, sensitive; O)? Lay language: Which screening tool(s) should be used to identify those at risk for later psychological symptoms? Data Sources and Search Strategy Following the GRADE approach and led by an expert medical librarian, we conducted a search of PubMed, Scopus, PsychINFO, Emcare, CINAHL, and Cochrane Trials Register databases on 5/8/2023, and repeated this search on 5/8/2024, leading to the addition of one article (see Supplemental Digital Content, Supplemental materials for search terms and strategy, https://links.lww.com/TA/E313). Identified articles were uploaded into Covidence software (Melbourne, Australia). At each review phase (abstract/title, full-text), two reviewers independently applied inclusion/exclusion criteria to each article. All conflicts were discussed among the lead author and at least one other reviewer and consensus was reached. Study Eligibility To be included, articles had to meet the following criteria: (1) describe results of a quantitative empirical study in a medical setting (e.g., inpatient, emergency department) of pediatric patients (<18 years) who experienced an unintentional injury; (2) describe use of a screening tool, which does not require administration by a trained mental health professional, to categorize pediatric patients into groups based on risk for current and/or future mental health symptoms and/or report outcomes regarding accuracy of the tool (e.g., sensitivity, specificity); (3) report relations between screening tool results and outcomes of interest, namely child mental health symptoms (acute stress disorder [ASD], PTSD, anxiety, depression, behavioral problems) or child behavioral health service utilization (e.g., psychology consultations). We excluded non-English articles, abstracts, systematic or narrative reviews, meta-analyses, qualitative studies, intervention studies, policy articles, editorials, protocol articles, and consensus documents. We also excluded studies that involved only patients with traumatic brain injuries or focused on child abuse or self-harm, as these are unique populations with distinct outcomes (e.g., changes in cognitive functioning) or correlates (e.g., foster care placement) that complicate prediction of mental health outcomes. Given the focus on child mental health outcomes that are related to exposure to unintentional injury, we also excluded studies that measured exclusively the following mental health outcomes: Attention Deficit/Hyperactivity Disorder, autism, schizophrenia, psychosis, cognitive disorders, intellectual disability, procedural anxiety, and reported solely on parent or caregiver mental health outcomes. For PICO 1, included articles had to explicitly a) compare outcomes of interest (e.g., utilization of mental health services) across patients for whom a screening tool was and was not used, b) examine outcomes of interest among a targeted subsample of patients for whom a screening tool was used, or c) examine outcomes of interest among those for whom a universal screening approach was taken. For PICO 2, manuscripts had to include measures of accuracy (i.e., sensitivity, specificity) for concurrent prediction of mental health symptoms (PICO 2a) or for prediction of later mental health symptoms (PICO 2b). Data Extraction and Methodology One reviewer (JP) extracted data points from each manuscript and a different reviewer (CY) checked accuracy and completeness of extraction. The two reviewers met to resolve conflicts and consulted with other reviewers as needed. For PICO 2a and PICO 2b, sensitivity, specificity, positive predictive values (PPV), negative predictive values (NPV), and the area under the receiver operating curve (AUC) were extracted when available.13 In evaluating screening tool performance to identify children at-risk for mental health symptoms following injury, higher sensitivity and higher NPV are desirable. Tools that optimize the odds of correctly identifying those with mental health symptoms (sensitivity) and optimize the probability that those who screen negatively are truly not at risk for mental health symptoms (NPV) will ensure patients with any risk are identified and mental health supports are utilized by those truly in need.14,15 In this context, the consequences of lower specificity or PPV, i.e., increased odds of incorrectly capturing those not truly at-risk, are minimal in terms of harms to patients and families and cost. Thus, screening tools that show promising sensitivity (≥70%) and NPV (≥70%) and modest specificity (≥50%) and PPV (≥50%) are considered effective. Recommendation Creation PICO 1 All reviewers voted independently on the strength of practice management recommendation for PICO 1, considering the quality and depth of the evidence, as well as weighing risks and potential benefits. PICO 2a and 2b All authors reviewed an in-depth qualitative synthesis of results for PICO 2a and 2b. Given the lack of a gold standard screening tool for comparison and significant variability across studies, all authors agreed that meta-analyses for PICO 2a and 2b were not feasible. All authors agreed to a qualitative process to evaluate which screening tools met specific criteria to be recommended for clinical use. These criteria were developed collaboratively by all authors and revised based on individual feedback and group discussion. The Essential Screening Tool Criteria required for recommending tool use and the additional Desirable Criteria are listed in Table 1. TABLE 1 - Essential and Desirable Criteria to Recommend a Screening Tool for PICO Questions 2a and 2b Essential Screening Tool Criteria 1. Screening tool performance has been evaluated with injured children in a medical setting 2. At least one study provides evidence that screening tool detects concurrent psychological distress and/or predicts persistent, ongoing distress 3. Tool is feasible for use in the emergency department or inpatient setting (i.e., time to administer, ease of scoring)* 4. Evidence-based scoring guidelines are provided and include guidance on interpreting the meaning of a positive screen so that healthcare team can take appropriate action (e.g. monitor, assess further, refer for mental health services) 5. No rigorous study/substantial evidence showing that the tool is not effective at detecting psychological distress and/or predicts persistent, ongoing distress Additional Desirable Screening Tool Criteria 1. Multiple studies provide evidence for prediction of concurrent or future/ongoing psychological distress 2. At least one study offers evidence for prediction of broader spectrum of distress—e.g., posttraumatic stress, depression, anxiety, behavioral difficulties, and/or impact on functioning 3. Demonstrated feasibility and utility based on real world use in multiple settings 4. Tool has been evaluated in, and has evidence for good performance with, injured patients similar to those in the target setting (e.g., age, culture, language, injury types and severity) Note. Criteria developed through discussion and revisions among all authors.* Based on authors' judgment when formal feasibility data (ie time to administer) was not reported. External Review The PTS Guidelines Committee and the Boards of STN and CPTS conducted an external review of the methods, recommendations, and full manuscript. All endorsed this guideline. RESULTS In the initial search 6,228 studies were identified, 377 reached full-text review, and 25 were included in analyses (see Fig. 1).16–40Figure 1: PRISMA Flow Diagram.PICO 1: Screen Only Those Perceived to Be at Risk vs. Screen All Patients With Injury Among the final 25 papers, five offered data regarding who should be screened for mental health difficulties. Qualitative Synthesis Two studies compared utilization of psychology consultations for admitted pediatric injury patients who were and were not part of a PTSD screening program.29,37 One study reported a 40% increase in appropriate referrals to psychology when comparing referrals preimplementation and postimplementation of a universal PTSD screening program.31 Of those referred, 95% to 96% were diagnosed with ASD and 90% to 92% of these patients were then diagnosed with PTSD. of patients who should have been to based on the screening tool, were of these referrals among children years) and children with In a of inpatient psychology consultations were higher among those for whom a screening admitted youth years) compared with those who were not for this i.e., on medical provider clinical to refer admitted youth Data from studies that universal use of a screening tool, patients who from mental health referrals are articles evaluated of positive PTSD screening data regarding which pediatric injury populations may be at high risk for PTSD and should be for across studies on which a positive PTSD For of and high Injury high of positive PTSD screen for but not other One study that those who screen show higher of later positive data are limited and on which predict positive PTSD screening In terms of one study in screening tool administration across and vs. among those for a screening screening may ensure are to needs and to Synthesis While populations were studies reported data (e.g., one not report age, two not report not report or years) were and of these had health and had or health In the one study of children that reported this was and and only had health health All five studies required that patients were admitted to Only one study reported injury studies reported data on injury which included injury, and were and studies reported one reported and two provided One study reported average studies conducted screening the and two While all studies screened for PTSD, tools One study an two the Screening Tool for of PTSD one used the Trauma Screening years) and the PTSD Screen and used only the Grading the Only two studies compared a screening tool protocol to clinical as of for a psychology or other mental health from significant risk of a universal screening studies used data from clinical two prospective studies of positive for PTSD. While data may populations to prioritize in a screening these on the that the screening tool provides a valid and reliable of PTSD or PTSD Two studies used the which has not well in prospective prediction of One study used a screening tool developed by an team at the which has not been All studies screening for PTSD only and not other mental health symptoms, as or All studies evaluated screening tools among patients who were admitted for injury, but only studies the screening tool the the of evidence for targeted compared with universal screening was to among study and study with Recommendation We universal mental health screening among pediatric patients with injury to identify those at-risk for mental health difficulties. screening (e.g., and accurate referrals for potential for functioning and lower and high risk of in pediatric patients current or future mental health (e.g., or distress and functioning) the minimal risk of in universal screening (e.g., potential in PICO 2a: Screening Tools for Outcomes vs. Among the included articles, manuscripts offered data regarding accuracy of screening tools for concurrent mental health One study from each PICO 2a and 2b have Qualitative Synthesis All studies involved screening for concurrent PTSD or each a different screening Screening tool administration studies evaluated only one tool and a of tools and evaluated the performance of of the tools in concurrent PTSD. All studies used clinical the gold to evaluate PTSD outcomes. Synthesis the studies, of patients from All studies reported only two included data Only one study included a of One study reported types of injuries All studies of injury, were and injuries No study reported or average of the performance of a of screening tools, the of and the good sensitivity and specificity PPV a of multiple is not feasible. The in concurrent PTSD specificity The in concurrent PTSD specificity or of of the of with scores to good performance sensitivity specificity PPV NPV in concurrent PTSD or PTSD The PTSD good prediction of concurrent PTSD across multiple sensitivity specificity Stress Checklist for of the Stress Checklist for the first month of injury to concurrent prediction of ASD sensitivity specificity PPV NPV Grading the While sensitivity and specificity data are promising and PTSD outcomes were measured using gold standard studies data to this PICO Only one study concurrent screening tool performance the first of the of evidence is and to lack of and of screening tool Recommendation to of recommendation voting not be the group developed the Essential and Desirable Criteria for a screening The following screening tools were at identifying concurrent risk of PTSD or one month of of to Table TABLE - Screening Tools for of (PICO 2a child symptoms child PTS symptoms child PTS symptoms child symptoms in child PTS symptoms risk concurrent vs. later distress or 10 to of can to for to No Essential Criteria Screening tool performance has been evaluated with injured children in a medical setting At least one study provides evidence that screening tool detects concurrent psychological distress and/or predicts persistent, ongoing distress Tool is feasible for use in the emergency department or inpatient setting (i.e., time to administer, ease of Evidence-based scoring guidelines are provided and include guidance on interpreting the meaning of a positive screen so that healthcare team can take appropriate action (e.g., monitor, assess further, refer for mental health services) No rigorous study/substantial evidence showing that the tool is not effective at detecting psychological distress and/or predicts persistent, ongoing distress Desirable Criteria Multiple studies provide evidence for prediction of concurrent or future/ongoing psychological distress for prediction of concurrent and future psychological At least one study offers evidence for prediction of broader spectrum of distress—e.g., posttraumatic stress, depression, anxiety, behavioral difficulties, and/or impact on functioning Demonstrated feasibility and utility based on real world use in multiple settings Tool has been evaluated in, and has evidence for good performance with, injured patients similar to those in the target setting (e.g., age, culture, language, injury types and severity) Note. Criteria developed through discussion and revisions among all 1 or for and posttraumatic to but to use and report results if used in Based on judgment not clinical feasibility PICO 2b: Screening Tools for Outcomes vs. Among the included articles, manuscripts offered data regarding accuracy of screening tools in later mental health of studies have data for PICO 2b. Two are in PICO and the other included one study each from PICO 2b and Qualitative Synthesis was significant variability across studies in the screening tools outcome and of when the screening tool and outcome measures were (see TABLE - Screening Tools in PICO 2b Screening Tools with in screening tools were evaluated across the included studies, with screening tools child PTSS or risk for one child depression, and one multiple risk for (see Table 3). studies the studies each evaluated the and Of the studies, compared two or screening the 10 one screening studies the screening tool of the injury (see Table TABLE - of Screening Tool in PICO Screening 1 1 with in of groups studies that the screening tool in the emergency at or of injury or group studies that the screening tool of injury or 1 groups studies a screening tool to 1 month and all studies measured PTSS or PTSD as an outcome studies measured multiple outcomes. all studies, different outcome measures were used to assess PTSD and quality of life and psychological distress Table of outcome of studies measured outcomes between one and (see Supplemental Digital Content, Table 1, TABLE - in PICO 2b 1 10 with in (1) mental health in disorders, disorders, disorders, disorders, disorders, disorders, disorders, disorders, disorders, disorders, disorders, use disorders, and other disorders, (2) in and (3) of PTSD for and and Trauma PTSD, for for PTSD, PTSD for PTSD symptoms of the for and and PTSD and of the PTSD and for and of the of PTSD Center for Pediatric of PTSD Synthesis of included patients or to or Two studies included children under all studies, patients were and of the studies not report and only one reported studies reported of patients one study reported on One study included and/or and studies involved patients admitted for injury and excluded patients who had abuse or injuries 6), were not to in had cognitive or had traumatic brain injuries studies excluded patients who from the were involved with child or were not a studies not report injury and one included only patients with studies reported injury were for across of these studies, and were included in of these studies injury included and Of the studies reported a of two reported a of and five reported One study reported categorized as or The injury study reported body area Of the articles admitted patients, reported of studies reported of at and reported with a and with to or Of the studies comparing the to other only one study that an of the in later PTSD sensitivity specificity the and the had poor screening tools, namely the and a on tools Center for the in the two comparison all studies of the only one good prediction of later PTSD sensitivity PPV NPV The five studies poor performance of the one study in which the was at identifying later PTSD The the in later PTSD sensitivity specificity PPV NPV when considering sensitivity and the sensitivity NPV the with (1 for and in the emergency department or to predict later In the two studies, the was by other screening tools of of screening in later these two studies used an but not of the and the sensitivity specificity NPV and PPV of the in those studies In a study that screening tools of the injury, the the and the Stress in later PTSD specificity NPV PPV Pediatric In a prospective study the Pediatric specificity PPV NPV the of risk in later in the emergency department or of All studies employing performance in PTSS as to good specificity PPV NPV The specificity PPV NPV the The was by the in later PTSS specificity PPV NPV The when with the and the any one tool in later PTSD and the well in PTSD specificity PPV NPV but not specificity PPV NPV The when with the and the any one tool in later PTSD and the poor to performance in PTSD and sensitivity specificity PPV NPV sensitivity specificity PPV NPV Stress The Stress performance in PTSD specificity NPV the of this tool One of two studies The had specificity and sensitivity in later PTSD the authors that this should not be used as a screening tool for PTSD. The Tool for children with injuries well specificity PPV NPV in child psychological outcomes symptoms, conduct Grading the from these studies of sensitivity, specificity, PPV, and NPV data in PTSD for of youth to studies prospective and gold standard for outcomes (e.g., clinical for studies those outcomes to the results of the screening screening tools are measures that can be by health studies offered data on accuracy for screening tools for with the and showing for the of evidence is to risk of and study Recommendation with PICO recommendation voting was not to the Based on the developed Essential and the authors identified the following screening tools as accurate for prospective risk of PTSD or and (see Table Pediatric trauma centers to meet the ACS for mental health screening should screening all injured pediatric trauma centers report screening pediatric patients with injury, particularly for PTSD. describe patients for screening based on with evidence of increased risk (e.g., Pediatric trauma centers should screening tools at a meet Essential Criteria for accuracy and feasibility (see 1 and (e.g., nurses, social who and from mental health can these tools, as by individual and are of this review. The the for additional rigorous research to evaluate and compare the performance of specific screening tools, across a of pediatric injury All future studies should report of those to trauma centers in results to the populations an of this guideline. research should also mental health screening for other (e.g., A reported feasibility of screening into particularly one study at a Level 1 pediatric trauma documented screening and psychology to feasibility of universal screening can trauma centers in and screening to (e.g., should include with mental health of an identifying a and data on screening that