Using a Decision Aid in the Emergency Room to Help Parents of Children with Head Trauma Understand Options for Diagnosing Brain Injury

Erik P. Hess; Nilay Shah; Jessica Westphal; Jeph Herrin; Michael Ferrara; Victor M. Montori; Nathan Kuppermann

doi:10.25302/11.2018.CE.12114435

Using a Decision Aid in the Emergency Room to Help Parents of Children with Head Trauma Understand Options for Diagnosing Brain Injury

Authors

Erik P. Hess, MD, MSc,¹ Nilay Shah, PhD,¹ Jessica Westphal,¹ Jeph Herrin, PhD,³ Michael Ferrara,² Victor M. Montori, MD, MSc,¹ and Nathan Kuppermann, MD, MPH².

Affiliations

¹ Mayo Clinic, Rochester, Minnesota

² University of California Davis Medical Center, Sacramento

³ Yale University, New Haven, Connecticut

Washington (DC): Patient-Centered Outcomes Research Institute (PCORI); 2018 Nov.

Structured Abstract

Background:

The Pediatric Emergency Care Applied Research Network (PECARN) prediction rules risk-stratify children who experience blunt head trauma into those at low, intermediate, and high risk for clinically important traumatic brain injury (ciTBI). The rules recommend observation or cranial computed tomography (CT) in children at intermediate (0.9%) risk of ciTBI but provide little guidance regarding how to engage parents in the decision of cranial CT or home observation.

Objectives:

We aimed to refine a shared decision-making intervention, Head CT Choice, and to test if the decision aid (DA) improves validated patient-centered outcome measures and safely decreases health care utilization.

Methods:

We conducted a pragmatic multicenter randomized controlled trial in 7 emergency departments (EDs) across the United States. Clinicians and parents caring for children less than 18 years of age with 1 or 2 PECARN risk factors for ciTBI were eligible. We assigned clinicians randomly (1:1) to DA or to usual care (UC) for the duration of the study. The primary outcome, selected by parent and caregiver representatives, was parent knowledge regarding their child's risk of ciTBI and the available diagnostic options. Secondary outcomes were parent involvement in the decision, satisfaction, decisional conflict related to feeling uninformed regarding the management options, the proportion of patients who had a cranial CT during the ED visit, and health care utilization and safety (missed ciTBI) at 7 days.

Results:

We enrolled 172 emergency clinicians caring for 971 (493 DA, 478 UC) children with head trauma. Compared with the UC arm, parents of children cared for by clinicians randomized to the DA had greater knowledge of their child's risk for ciTBI and the available diagnostic options (questions correct out of 10: 6.20 DA vs 5.26 UC; mean difference [MD] = 0.95; 95% CI, 0.60-1.30), were more involved in the decision (observing patient involvement scores: 25.0 [8.5] DA vs 13.3 [6.5] UC; MD = 11.7, 9.6 to 13.9), were more satisfied with their choice (“strongly satisfied”: 51.5% DA vs 43.9% UC; odds ratio, 1.4, 95% CI, 0.1-1.9), and reported less decisional conflict related to feeling uninformed (decisional conflict scale: 14.8 DA vs 19.2 UC; MD = −4.9; 95% CI, −7.3 to −2.4). Parents cared for by DA clinicians also had greater trust in their physician (trust in physician scale: 91.5 [11.9] DA vs 89.3 [13.7] UC; MD = 2.3; 95% CI, 0.4-4.1). Although there was no difference in the ED rate of cranial CT between study arms (22.1% DA vs 24.3% UC; odds ratio, 0.80; 95% CI, 0.51-1.26), DA patients had a lower rate of imaging (mean number of imaging tests, 0.65 DA vs 0.88 UC; MD = −0.23; 95% CI, −0.35 to −0.11) and blood testing (mean number of tests, 0.41 DA vs 0.70 UC; MD = −0.29; 95% CI, −0.40 to −0.16) within 7 days of ED discharge. There were no detected cases of missed ciTBI.

Conclusions:

Use of a DA in parents of children at moderate risk for ciTBI safely increased parental knowledge, engagement, and satisfaction and decreased decisional conflict. There was no difference in ED CT use but lower health care utilization in the intervention arm at 7 days (a secondary outcome). Overall, the DA safely improved care with less utilization.

Limitations and Subpopulation Considerations:

The DA was tested only in parents of children at moderate risk of ciTBI, and the findings apply only to this population. Use of the DA had similar effects across all parent subgroups.

Background

Blunt head trauma is the leading cause of trauma-related death and disability in children worldwide. Every year in the United States, more than 450 000 children present to emergency departments (EDs) for evaluation of head trauma.¹ Underlying pathology following blunt head trauma can be related to structural brain damage or concussion. Although these pathologies are different and require different treatments, the physical presentation can look similar, which creates a challenge for both clinicians and parents: What is the most effective way to evaluate the underlying cause? With the increased availability and speed of contemporary computed tomography (CT) imaging, clinicians obtain cranial CT imaging in 37% to 50% of children with minor head trauma in the United States.² Fewer than 10% of these CT scans, however, show evidence of traumatic brain injury (TBI) and only 0.2% require neurosurgical intervention.³

To avoid unnecessary CT imaging and limit ionizing radiation exposure,⁴ the Pediatric Emergency Care Applied Research Network (PECARN) developed 2 clinical prediction rules, 1 for children younger than 2 years old and 1 for children 2 to 18 years old.⁵ Each of these 2 prediction rules consists of 6 readily available clinical factors, 2 of which are high risk (Glasgow Coma Scale <15 or other signs of altered mental status and signs of skull fracture) (Table 1). If none of these risk factors are present, CT scanning is not recommended. If the child has a high-risk factor, CT is recommended. If the child has 1 or 2 non–high-risk factors (ie, those at intermediate risk), other considerations such as clinician experience, parental preference, and/or symptom progression guide the decision of whether to obtain CT scanning. However, the PECARN rules provide little guidance regarding how to engage parents of children at intermediate risk of clinically important traumatic brain injury (ciTBI) in CT decision-making.

Table 1

Clinical Factors Included in the PECARN Prediction Rules for TBI.

Decision aids (DAs), designed as patient-centered tools to facilitate shared decision-making,⁶ help clinicians and patients work together to apply both the latest scientific evidence and patients' values and preferences to care decisions. Use of DAs has been shown to increase patient knowledge, increase the accuracy of risk perception, increase patient engagement in decision-making, and appropriately tailor testing to disease risk.^6,7 Given that robustly developed clinical prediction rules guiding the choice of cranial imaging in children with blunt head trauma recommend either cranial CT or observation in those at moderate risk of ciTBI and the demonstrated effectiveness of DAs, the purpose of this study was to (1) give parents a voice and incorporate the perspectives of multiple stakeholders by involving them in refining the Head CT Choice DA, and (2) test the hypothesis that the DA will improve validated patient-centered outcome measures and safely decreases health care utilization in children at moderate risk of ciTBI.

Participation of Patients and Other Stakeholders in the Design and Conduct of Research and Dissemination of Findings

The design, conduct, and dissemination of findings included the following key steps: (1) organizing a steering committee, (2) refining outcomes important to stakeholders, (3) selecting primary and secondary study outcomes, and (4) refining the DA based on the input of parents and other key stakeholders. The key stakeholders involved included a parent, the ED patient advisory council at Saint Marys Hospital in Rochester, Minnesota (which includes 3 additional parent representatives), clinician champions, shared decision-making researchers, and a payer representative.

We invited the ED patient advisory council (EDPAC) at Saint Marys to participate early in the proposal development process. We recruited from the EDPAC a parent advisor (JW) with prior experience of ED care for hockey-related head injuries in her children to join the study as a coinvestigator and investigative steering committee member. We recruited a clinician champion from each participating site to participate in the development and refinement process of the DA and invited them to participate in the study as coinvestigators. We invited the chief medical officer of Mayo Clinic Health Solutions, an insurer of more than 20 000 employees and dependents at Mayo Clinic Rochester, to engage in the study as a consultant and provide input on outcome selection, particularly those related to health care utilization. The parent advisor and the EDPAC at Saint Marys Hospital, the clinician representatives, and the payer representative were involved in the design of the intervention, selection outcomes important to patients, and submission of the application for funding. In the process of refining the DA, we tested it in clinical encounters by both the clinicians and parents and sought feedback immediately after its use from both parties. We used this feedback to inform improvement of the tool to facilitate discussion and to increase parental understanding. Input from the parent advisor and EDPAC influenced the investigative team regarding the importance of addressing parents' concern about concussion early in the clinical encounter. This is the reason the definition and diagnosis of concussion is addressed on the first page of the DA.

The parent advisor, who was a key member of the investigative steering committee, was additionally involved in monitoring the study conduct for progress; presenting her perspectives at a regional PCORI meeting in Minneapolis, Minnesota; interpreting the data; reviewing abstracts and manuscripts for important intellectual content; and approving final manuscripts for publication. In the process of outcome selection, we discussed potential outcomes that could be validly measured with parent representatives and the EDPAC in face-to-face meetings. We prioritized parental perspectives over outcomes of interest to other key stakeholders, which were classified as secondary outcomes.

Methods

Study Design

Aim 1: Give Parents a Voice and Incorporate the Perspectives of Multiple Stakeholders by Refining the Head CT Choice DA

A multidisciplinary research team including clinicians, health services researchers, a graphic designer, a radiation physicist, and parent representatives collaboratively developed the initial Head CT Choice DA. Using methods we have employed to develop other DAs,^8-10 we designed a prototype and subsequently refined the DA through an iterative process that involved revision of the prototype based on the feedback received, use of the prototype in simulated parent/patient encounters in the ED setting, and use in actual ED clinical encounters in the flow of patient care. After each use of the DA in the ED and prior to conducting the trial, perspectives of the parent and clinician on the design and ability of the DA to prompt an evidence-based conversation were elicited, recorded by study personnel, and reported to the investigative steering committee. This process resulted in several common themes emerging, particularly the need to limit the topics included in the DA to only those that would facilitate conversation between the clinician and the parent and the need to address parental anxiety regarding whether the child had a concussion up front. Once thematic saturation had been achieved (ie, key themes repetitively emerged with use of the DA in clinical encounters and no substantial new information was gleaned with additional use), we considered the DA ready for testing.

In the process of DA development and refinement the following themes emerged: (1) clarifying the difference between concussion and structural brain injury, given recent social media attention to concussion; (2) communicating individual patient risk using natural frequencies (eg, “x out of 100 patients with these characteristics will have a ciTBI and y will not”), a consistent denominator, a prose description of the natural frequencies, and pictographs; (3) limiting the amount of information to that which is relevant to the decision at hand and best facilitates conversation between the parent and clinician without drawing undue focus on the DA itself; and (4) clearly outlining the choice of a head CT scan during the ED visit, home observation if no CT scan is obtained, or having the clinician make the decision on the patient's behalf if that is the parent's preference.

Head CT Choice educates caregivers regarding the definitions of and difference between a concussion and other forms of TBI, how the clinician determined the severity of their child's head injury, their child's risk of ciTBI, what signs and symptoms should prompt a return visit to the ED should the parent opt to forgo cranial CT imaging during the ED visit, and the pros and cons of cranial CT imaging compared with active observation (Figure 1).

Figure 1

Head CT Choice.

Aim 2: Test if the DA Improves Validated Patient-Centered Outcome Measures and Safely Decreases Health Care Utilization in Children at Moderate Risk of ciTBI

A study protocol describing the methods of this trial has been published,¹¹ and the IRB-approved protocol is available upon request from the lead investigator (EPH). This pragmatic clinician-level multicenter randomized trial¹² was designed to compare an intervention group participating in shared decision-making facilitated by a DA (Head CT Choice) against usual care in parents of children at moderate risk for ciTBI. We intentionally chose a pragmatic comparator—usual care—and did not seek to control or standardize the control arm in order to most closely reflect the conditions under which cranial CT imaging decisions are made in routine emergency practice in the United States. Additionally, we opted to randomize at the clinician level. We initially considered randomizing at the patient level, but in the process of DA design it became clear that the PECARN prediction rules had already been widely disseminated and were frequently used in routine emergency pediatric care at the participating sites. To limit the risk of contamination related to randomizing at the patient level (which would involve clinicians using the DA in some cases and not others), we opted to randomize at the clinician level. Although randomizing at the site level would have optimally limited the risk of contamination, this would substantially increase the number of sites required to measure the impact of the intervention, would limit the precision with which we could detect differences related to use of the DA, and would not be feasible given the amount of funding available. For these reasons, randomizing at the clinician level was the best balance of methodological rigor and financial and logistical feasibility.

Forming the Study Cohort

To decrease the risk of contamination related to the sharing of information between clinicians, we emphasized to participating clinicians the importance of not sharing information with nonintervention clinicians about the DA during the study period. Study coordinators also restricted DA access to clinicians randomized to the intervention and monitored for contamination by closely observing the clinician during each clinical encounter and throughout the trial. After informed consent was obtained from participating clinicians, a statistician at a centralized location performed randomization to conceal allocation. Clinicians were dynamically randomized and stratified^13,14 by site and whether their primary clinical training was in a pediatric specialty (pediatrics or pediatric emergency medicine) or another clinical specialty (general emergency medicine, family practice, internal medicine).

Eligible clinicians were attending-level physicians, pediatric fellows, and midlevel providers caring for children with minor head trauma. Parents seeking care for a child or adolescent less than 18 years of age within 24 hours of blunt head trauma (above the eyebrows and not isolated to the face or eyes) and with 1 or 2 non–high-risk PECARN factors for ciTBI were eligible for enrollment. Children with any of the following characteristics were excluded: signs of skull fracture, Glasgow Coma Scale <15 or other signs of altered mental status, brain tumor, penetrating head trauma, bleeding disorder or coagulopathy, ventricular shunt, preexisting neurological disorder complicating assessment, syncope or seizure preceding the head trauma, transferal to the ED with imaging already obtained, known pregnancy, >2 PECARN risk factors, or parents with communication barriers such as visual or hearing impairment precluding use of the DA.

Study coordinators identified potentially eligible parent–patient dyads based on a chief complaint of head trauma recorded at the time of patient registration and through communicating with practicing clinicians in real time. Prior to the encounter, the study coordinator instructed the clinician to obtain the PECARN risk factors during the initial history and physical examination but to defer discussing the diagnostic options with the parent. If the clinician confirmed that the patient had at least 1 PECARN non–high-risk factor (Table 1), the study coordinator then reviewed the remaining eligibility criteria and, if the patient met all other criteria, obtained written informed consent from the parent(s) and assent from the patient to participate in the study and to video record the diagnostic discussion with his or her clinician.

Study Setting

We conducted the study at 7 diverse sites across the United States, including 1 academic ED serving a largely rural population (Mayo Clinic, Rochester, Minnesota), 4 academic centers serving urban populations (University of California Davis Children's Hospital, Sacramento; University of Minnesota Masonic Children's Hospital, Minneapolis; Nationwide Children's Hospital, Columbus, Ohio; and Boston Children's Hospital, Boston, Massachusetts), and 2 community EDs serving an urban/suburban population (Children's Hospitals and Clinics of Minnesota, Minneapolis and Saint Paul). We chose these sites based on their regional diversity, track record in enrolling patients in prospective trials, and representativeness of both community and academic practices, settings representative of those in which we anticipated the DA to be used on completion of the trial.

Interventions

The design of the DA is included in the description of the methods for Aim 1 above. A description of how the DA was delivered in the context of the trial is included below.

Delivery of the DA

Prior to beginning enrollment, we obtained the public access database from the original PECARN head injury prediction rules study and calculated risk estimates for ciTBI based on the presence or absence of individual clinical predictors or combinations of predictors.¹⁵ We provided to the study coordinator at each site a table summarizing these data. (Go to http://shareddecisions.mayoclinic.org/decision-aid-information/head-ct-choice-decision-aid/ to view or download a table that displays individual patient risk based on the PECARN risk factors present.) We educated clinicians randomized to the intervention arm regarding the information on and use of the DA with parents by means of a video demonstration of its use in practice and a standardized document providing answers to frequently asked questions. For patients whose clinician was randomized to the DA arm, study coordinators selected the precise PECARN risk estimate for that patient and provided the clinician with an 11 × 17 color copy of the DA corresponding to that level of risk prior to discussing the options of cranial CT vs observation with the parent. (Go to http://shareddecisions.mayoclinic.org/decision-aid-information/head-ct-choice-decision-aid/ to download a DA specific to a patient's level of risk.) Study coordinators also provided brief 3- to 5-minute just-in-time refreshers of DA content and its use with parents to clinicians on an as-needed basis throughout the trial. Intervention clinicians used the DA to educate parents regarding differences between concussion and other forms of TBI, communicate to parents a quantitative risk estimate for ciTBI, review with parents the pros and cons of cranial CT scanning vs active observation and the signs and symptoms that should prompt reevaluation in the ED. The clinician subsequently engaged the caregiver in shared decision-making regarding the choice of cranial CT in the ED or active observation.

Usual Care

For parents whose clinician was randomized to usual care, study coordinators instructed the clinician to discuss the management options with the parent in his or her usual manner.

Clinicians in the usual care arm did not have access to the precise risk estimates for ciTBI calculated from the PECARN public access database or to the DA. As the trial was intentionally pragmatic in design, the usual care arm was not standardized.¹²

Follow-up

Study coordinators contacted parents starting at 7 days after enrollment to assess health care utilization and safety. (As the intervention was delivered during the clinical encounter in the ED, follow-up occurred after exposure to the intervention.) Study coordinators made at least 5 attempts to contact parents by phone for follow-up during different times of the day and on different days of the week. If parents were unable to be reached by phone or email and no subsequent visits were documented in the medical record, we reviewed continuous quality improvement reports at each participating hospital and the morgue records at each participating county for any missed injuries or fatalities not identified by other methods.

Study Outcomes

In the process of selecting the study outcomes, the investigative team and parent representatives collaboratively reviewed and considered outcomes of potential importance to each of the stakeholders (parent representatives, clinicians, shared decision-making researchers, and a payer representative). Parent representatives participated in this process, and the outcome they considered to be most important was classified as the primary outcome, with the remaining outcomes classified as secondary outcomes.

Primary Outcome

We selected parental knowledge as the primary outcome, and this outcome was assessed by the immediate postvisit survey. In meetings with parent representatives, it became clear that knowledge of whether their child had experienced a concussion, their child's risk for more serious brain injury, and the available management options were of greatest importance.

Because a primary aim of patient-centered outcomes research is to assist patients and caregivers in making decisions that reflect their desired health outcomes,¹⁶ we selected parental knowledge as the primary outcome. Parental knowledge was assessed by immediate postvisit survey.

Secondary Outcomes

In addition to the primary outcome, we assessed 8 additional secondary outcomes in both the usual care and DA arms:

We assessed the degree to which the clinician engaged parents in the decision-making process using the validated Observing Patient Involvement (OPTION) scale.¹⁷ Five trained raters independently viewed videos of the parent–clinician discussion in assessing the OPTION scale. The scale contains 12 items with a value of 0 to 4, which are summed, divided by 48, and then multiplied by 100. Scores range from 0 to 100, where higher scores indicate greater parental engagement. The mean OPTION score in the usual care arm of our prior shared decision-making (SDM) multicenter trial in adults with low-risk chest pain was 7.9 (SD: 5.4). Although a clinically meaningful difference in OPTION score has not been defined, we anticipated that, if effective, use of the Head CT Choice DA would increase OPTION scores by more than twofold as observed in prior work.⁷
We assessed the degree of uncertainty parents experienced related to choosing between management options with which they were unfamiliar using the validated decisional conflict scale.¹⁸ The decisional conflict scale includes 16 items that are scored from 0 to 4. The items are summed, divided by 16, and then multiplied by 25. The scale is from 0 to 100, where higher scores indicate increased parent uncertainty about the choice. A prior study found that for every unit increase in decisional conflict scale scores, patients were 19% more likely to blame their doctor for bad outcomes.¹⁹ For this reason, we considered a 1-unit difference in decisional conflict scale score to be clinically meaningful.
We measured parents' trust in their clinician using the validated trust in physician scale.²⁰ The trust in physician scale includes 9 items scored from 1 to 5. The items are subtracted by 1, summed, divided by 9, and then multiplied by 25. The scale ranges from 0 to 100, where higher values indicate higher levels of physician trust. To the best of our knowledge, a clinically meaningful difference in trust in physician scale scores has not been published.
We assessed parental satisfaction with the decision made by asking whether they were satisfied with their choice using a 5-point Likert scale and with the amount, clarity, and helpfulness of the information shared about whether their child should receive a head CT during the ED visit using a 7-point Likert scale. For the analysis, we classified responses into satisfied/very satisfied vs other responses.
We measured the proportion of children who underwent head CT during the ED visit, the most immediate utilization decision.
We assessed health care utilization for the subsequent 7 days after the ED visit, including measures such as hospitalization, rehospitalization, primary and specialty visits, diagnostic imaging including CT and magnetic resonance imaging, and blood testing by review of itemized hospital charges on the UB-92 and UB-04 forms (summary billing statements) and parental report at the time of the 7-day phone follow-up.
We assessed the safety of the intervention by comparing the rate of ciTBI in each arm of the study. We defined ciTBI as we did in the original PECARN study: death from TBI, intubation for more than 24 hours for TBI, neurosurgical procedure, or hospital admission of 2 nights or more associated with TBI on CT.⁵ We measured fidelity—the degree to which the intervention was implemented as intended—in both arms of the study using a checklist of items when reviewing video recordings of clinical encounters.

Modification of the Original Study Protocol

Although we planned on measuring health care utilization at 30 days from the index ED visit as described in the initial study protocol, upon further discussion among the investigative team prior to beginning enrollment for the trial, we decided to measure health care utilization at 7 days as documented in our published protocol.¹¹ We did this for 2 reasons: (1) Safety outcomes (rate of ciTBI) were assessed at 7 days in the original PECARN prediction rule investigation⁵; and (2) we thought it likely that measuring utilization up to 30 days increased the risk of having medical issues other than head trauma impact utilization measures.

Data Collection and Sources

We collected data documenting the process of screening and identifying potentially eligible participants and the progress of all participants throughout the trial in compliance with CONSORT guidelines.^21,22 We accomplished data collection via administration of a pre-encounter parent survey (Appendix 1), a postencounter parent survey (Appendix 2), a post encounter clinician survey (Appendix 3), health record review, video recordings of the clinical encounter, and phone follow-up starting 7 days after enrollment. On the pre-encounter survey we assessed parent literacy using the subjective literacy scale,²³ parent numeracy using the subjective numeracy scale,²⁴ and annual household income.

The subjective literacy scale consists of 3 items, each with a 5-point Likert scale, that were summed into a total score from 3 to 15 after reverse coding 1 item: The higher the summed score, the higher the patient's subjective assessment of his or her general health literacy skills. For the subjective numeracy scale, parents were asked to respond to an 8-item questionnaire, each with a 6-point Likert scale. Numeracy responses to all 8 questions were reversed and averaged, creating an overall score ranging from 6 to 48, where higher scores are indicative of higher levels of numeracy. The postencounter parent survey collected data assessing parents' knowledge regarding their child's risk for ciTBI and the available management options; parents' decisional conflict related to feeling uninformed regarding the management options; parents' trust in their clinician; and the amount, clarity, and helpfulness of the information provided by the clinician. Study coordinators collected the following data during the visit and confirmed accuracy via health record review: the date and time of ED registration, the date and time of ED discharge, whether the patient was observed in the ED after the initial evaluation, the specific PECARN criteria qualifying the patient for enrollment in the study, whether a cranial CT was obtained, any positive findings on CT, and any return visits to the ED or ciTBI criteria within 7 days of the visit. The clinician–parent discussion was audio and video recorded when consent for recording was obtained. Video and audio recordings were time stamped, and we determined the duration of the clinician–parent discussion from these recordings. The recordings were uploaded to a secure server and deleted from the portable devices after upload. Access to the server was protected by a 2-step access: password-protected access to all of Mayo's computers, and password-protected access to the secure server. Audio and video recordings from facilities outside of Mayo Clinic were downloaded onto a password-protected flash drive, sent to the prime site by secure courier, and uploaded to a secure server on receipt.

Analytical and Statistical Approaches

We estimated that 950 patients would provide 99% power to detect a 16% difference in parent knowledge between DA and usual care arms and 95% power to determine a 15% difference in the rate of head CT. We assumed an intraclinician correlation of ρ = 0.01 and adjusted sample sizes using a variance inflation factor of VIF = [1 + (m − 1)ρ], where the average number of patients per clinician is assumed to be 15 clinicians per site. This number of patients provided 82.5% power to detect a difference from a baseline ciTBI rate of 0.9% between study arms, using a 1-sided noninferiority test with an α of .05 and a minimally important difference of 2%.¹¹ We conducted the study according to the intention-to-treat principle, including all caregiver–child dyads in the arm to which they were randomized, regardless of whether they received the intervention assigned. We compared patient characteristics between study arms using t tests for continuous outcomes and χ² tests for dichotomous outcomes, adjusted for clustering by clinician and stratified by study site.²⁵ To test for differences in outcomes, we estimated a series of mixed effects regression models, each of which included indicators for study group. For continuous outcomes we used linear models, and for categorical outcomes we used multinomial (polytomous) logistic models; these models included a random effect for clinician. For the health care utilization analysis, we used the negative binomial model—a count data model—to measure differences in utilization. To account for nonindependence of outcomes by site, we included indicators for study site in each model. In each model the only independent variable was the study group.

Heterogeneity of DA Effect Analyses

We also performed descriptive analyses to describe any potential heterogeneity of treatment effect and to facilitate synthesis of subgroup results in future meta analyses. We dichotomized each of the following variables to evaluate the differential effect of the DA across parent characteristics: gender, race, level of education, annual income, health literacy, numeracy, age and number of PECARN risk factors for the child, and whether the father only or both parents were present. We prespecified the following hypotheses: (1) Use of the DA would increase parent knowledge, decrease decisional conflict, and increase parent engagement similarly across all subgroups; and (2) parents with lower numeracy would have lower knowledge scores on the post visit survey. We based the hypothesis that parents with lower numeracy would have lower knowledge scores based on an interaction between numeracy and knowledge that was observed in a prior heterogeneity of DA effect analysis of patients with low-risk chest pain.²⁶ We chose to dichotomize for 2 reasons: (1) Given the overall enrollment of 971 patients, more than 2 classifications would have resulted in many subgroups that were too small to analyze; and (2) binary classifications made the heterogeneity of effect simpler to analyze and interpret by way of interactions.²⁷ These classifications were based on both the distribution of the values, which we report in full, and conceptual considerations regarding the mostly likely contrasts to show heterogeneity of effect. Classifications were as follows: gender as female vs male; race as White vs non-White; education as less than or equal to high school/General Education Development (GED) vs greater than high school/GED; annual income as less than $40 000 vs greater than or equal to $40 000; literacy as typical (≥12) vs low (<12); and numeracy as typical (≥34) vs low (<34).²⁸ We excluded patients missing a subgroup variable from the analysis for that subgroup. For race, we included the “other” group with non-White. For education, we excluded the “other” category from the dichotomous groups, as it could not be assumed to indicate either of the 2 categories. We also tested for interactions among each of these patient characteristics and trial outcomes with each variable classified as a continuous or ordinal variable, as appropriate for the data, and observed similar findings. For these reasons, we present the results of interaction testing between each dichotomized patient characteristic and trial outcomes.

Missing Data

A study analyst conducted frequency reports every 2 weeks to identify and obtain missing data. Patients with missing outcome data were not included in the assessment for that outcome.

Rates of missing data are reported in the tables where appropriate. Rates of missing data as well as known reasons for missing data are described in the results section. The outcome with the greatest proportion of missing data was parent participation as measured by the OPTION scale, as this required video recordings of the clinical encounter (461 out of 971, or 47% missing). Despite the degree of missingness in this outcome, we had sufficient power (>99%) to detect meaningful differences with only 221 videos, and we video recorded 510 encounters.

We were also unable to contact 8% of parents by phone within 7 days to confirm the absence of a head injury at 7 days (a 92% phone follow-up rate), and these data were also not imputed.

Finally, the greatest degree of missingness in parent survey data was 4% (96% data completion rate), and survey data were not imputed.

Conduct of the Study

We added 2 additional study sites—Boston Children's Hospital in Boston, Massachusetts, and Nationwide Hospital in Columbus, Ohio—after trial enrollment had begun. We added these 2 sites in order to increase recruitment rates to meet the sample size target of 950 patients. We made no other additional modifications to the original study protocol. The IRB at each participating site approved the study protocol without additional modifications.

Results

Aim 1: Give Parents a Voice and Incorporate the Perspectives of Multiple Stakeholders by Refining the Head CT Choice DA

The Head CT Choice DA (Figure 1) is a tool designed by an investigative team that included a variety of stakeholders and using a validated iterative process to ensure that patient, clinician, and other stakeholder voices were represented. The completed aid is a 3-page paper tool that takes into account 3 themes in numerically ordered sections to facilitate dialogue between the clinician and the parent. The DA first graphically highlights key differences between concussions and other types of significant brain injuries and communicates the child's risk for ciTBI. The second page clearly outlines the decision that is to be made (head CT scan in the ED vs home observation) and educates the parent on the concerning signs and symptoms of TBI and the indications for a return visit to the ED should they opt for home observation without a CT scan. Next, a table compares the key advantages and disadvantages of each management option and invites the parent to further discuss the issues that are most important to him or her. Third, the DA outlines the 3 management options: head CT scan in the ED, further observation at home with no CT scan, or having the ED clinician decide on the parent's behalf. It includes check boxes to document the decision along with a reminder that the parent will have the opportunity to revisit the decision with his or her clinician while in the ED. The third and final page includes a space to provide a contact phone number should the parent wish to contact a health care provider after leaving the ED.

Aim 2. Test if the DA Improves Validated Patient-Centered Outcome Measures and Safely Decreases Health Care Utilization in Children at Moderate Risk of ciTBI

The patient population by gender and race/ethnicity is included in Table 2. Additional PICOTS descriptors (interventions, comparators, and timing of follow-up) are included in Figure 2. Study outcomes are described in “6. Study Outcomes” and included in Tables 4 to 6, and the study settings are described in the “3. Study setting” section of this report. The 4 tables required by ClinicalTrials.gov (patient flow, baseline characteristics, outcomes and statistical analyses, and adverse events) are included as a supplementary file to this research report. The remainder of the results below adhere to the CONSORT guidelines for the reporting of randomized trials.²¹

Table 2

Baseline Characteristics (N = 971).

Figure 2

Head CT CONSORT.

Table 4

Effect of Decision Aid on Parent Knowledge, Decisional Conflict, Trust in the Physician, Parent/Caregiver Involvement in the Decision, Patient Acceptability and Satisfaction, and Clinician Acceptability.

Table 6

Health Care Utilization Data Obtained From Hospital-Level Billing Data, Including the ED Visit and Utilization Within the Subsequent 7 Days.

We assessed 8055 patients for eligibility from April 2014 to September 2016 (Figure 2). Overall, 172 clinicians participated in the study. We randomized and included in the final analysis 971 patients (493 in the DA arm, 478 in the usual care arm). We audio or video recorded the parent–clinician discussion in 516 (53%) encounters. Clinician or patient refusal (n = 293) and technical difficulties with recording equipment (n = 10) were the main reasons recordings were not obtained. We contacted 890 (91.6%; n = 449 DA) parents by telephone or email for follow-up. Of the remaining 81 patients, 31 had subsequent visits to a health care site, the electronic medical record of which documented no serious head injury within 7 days, and 50 had no adverse outcomes documented in continuous quality improvement reports or county morgue records at each participating site.

Table 2 summarizes baseline characteristics of the enrolled participants. The mean (SD) age of the children was 6.7 (6.0) years, with the majority (739; 76.1%) being 2 to 18 years. Of the patients, 575 (59.2%) were male. Most patients were White (718; 73.9%) or Black (115; 11.8%). There were 159 (16.4%) parents whose highest level of education was high school, General Educational Diploma, or less. There were no significant differences in baseline characteristics between study arms.

Table 3 summarizes the PECARN risk factors for head injury for children <2 years of age and those 2 to 18 years. For children <2 years of age, severe mechanism of injury was the most common predictor present, and for those 2 to 18 years of age, any vomiting since the injury was the most common predictor. Most (780; 80.3%) enrolled patients had only 1 PECARN predictor.

Table 3

PECARN Risk Factors for the Enrolled Patients (N = 971).

Parent Knowledge, Decisional Conflict, and Trust

Parents of children cared for by clinicians randomized to the DA had greater knowledge (questions correct out of 10: 6.20 DA vs 5.26; mean difference = 0.95; 95% CI, 0.60-1.30; Table 4). Parents in the DA arm also reported less decisional conflict (decisional conflict scale: 14.8 vs 19.2; mean difference = −4.9; 95% CI, −7.3 to −2.4) and greater trust in their physician (91.5 vs 89.3; mean difference = 2.3; 95% CI, 0.4-4.1).

Parent Participation, Acceptability, and Satisfaction

Clinicians randomized to the DA involved patients in decision-making more than in the usual care arm as indicated by higher OPTION scores (25.0 vs 13.3; mean difference = 11.7; 95% CI, 9.6-13.9, on a scale from 0 to 100). The correlation coefficient between raters for OPTION scale assessments was 0.72 (95% CI, 0.67-0.76). Parents randomized to the DA found the information communicated by their physician to be of greater clarity (77.7% vs 71.5%; odds ratio, 1.5; 95% CI, 1.1-2.1). Although most (828; 85.3%), parents agreed or strongly agreed with the statement, “I am satisfied with my choice,” a greater proportion of parents in the DA arm strongly agreed with the statement (51.5% vs 43.9%; odds ratio, 1.4; 95% CI, 1.1-1.9).

Clinician Acceptability

A greater proportion of clinicians in the DA arm would recommend the way they presented information to parents and caregivers to other providers (61.9% vs 41.4%; odds ratio, 2.9; 95% CI, 1.2-6.8). Many clinicians in the DA arm also found the information they shared with parents to be helpful, though this was not significantly different from usual care (50.1% vs 40.2%; odds ratio, 1.8; 95% CI, 1.0-3.3).

Management and 7-Day Health Outcomes

There were 558 (57.5%) children observed in the ED after being evaluated by their clinician for worsening clinical symptoms to determine the need for cranial CT (Table 5). Overall, 225 (23.2%) children underwent cranial CT, and there was no significant difference in the frequency of cranial CT between study arms (22.1% vs 24.3%; odds ratio, 0.81; 95% CI, 0.51-1.27). A greater proportion of children who had 2 PECARN risk factors underwent cranial CT (39.3% vs 19.2%; odds ratio, 2.71; 95% CI, 1.90-3.87). The ED length of stay was significantly shorter in the intervention arm (minutes: 176.0 vs 198.8; mean difference = −22.8; 95% CI, −41.6 to −4.0). One child in the usual care arm had a ciTBI. This child had an extra-axial hematoma diagnosed on cranial CT, and it was confirmed after hospital admission that the patient was a victim of nonaccidental trauma.

Table 5

Effect of Decision Aid on Management and 7-Day Outcomes.

There was no significant difference in the frequency of hospital admission between the DA and usual care (9% vs 9%; odds ratio, 0.97; 95% CI, 0.38-2.45) or the frequency of ED return visits (10% vs 18%, respectively; odds ratio, 0.54; 95% CI, 0.24-1.24) within 7 days between study arms.

Fidelity of DA Use

Of the 516 (53.1%) parents for whom video recordings of the clinician–patient interaction were obtained, 267 (51.7%) were in the DA arm of the study. In 259 (97%), the clinician brought the DA into the room, and in 251 (94%) the clinician used the DA during the decision-making process with the parent. In 266 (99.6%) of these cases, the clinician discussed options of care with the parent. The option of getting a head CT during the initial ED visit was discussed in 256 (96.2%) cases, active observation at home in 259 (97.4%) cases and letting the ED clinician decide in 14 (5.3%) cases.

Health Care Utilization Within 7 Days

Table 6 displays total health care utilization, including the ED visit and utilization within the subsequent 7 days, for the enrolled patients obtained from hospital-level billing data. There was no difference in provider evaluation and management coding between the DA and usual care arms (1.84 vs 1.88, respectively; mean difference = −0.04; 95% CI, −0.12 to 0.04). Patients of clinicians randomized to the DA, however, had a lower rate of imaging (mean number of imaging tests, 0.65 vs 0.88; mean difference = −0.23; 95% CI, −0.35 to −0.11) and blood testing (mean number of tests, 0.41 vs 0.70; mean difference = −0.29; 95% CI, −0.40 to −0.16) within 7 days of ED discharge than patients of clinicians randomized to usual care.

Heterogeneity of DA Effect Analyses

Table 7 shows parent and patient baseline sociodemographic characteristics by subgroup. There were no significant differences in baseline characteristics between the 2 study arms. We observed no significant interactions between the DA and parent sociodemographic characteristics on the outcomes of parent knowledge, decisional conflict, physician trust, and parent involvement in the decision (Table 8). The DA increased parent knowledge, decreased decisional conflict, increased physician trust, and increased parent involvement in the decision similarly across all subgroups. (Figures 3 and 4; Table 8). However, we observed significant interactions between parent health literacy, patient age, when the father was the only parent present during the encounter, and the rate of cranial CT (Figure 5; Table 9). Parents with low health literacy were less likely to have their child undergo CT scanning (odds ratio, 0.36; 95% CI, 0.16-0.83). Children <2 years of age were more likely to undergo CT scanning (odds ratio, 1.94; 95% CI, 0.95-3.97), and when the father was the only parent present during the encounter, there was a lower odds ratio of CT imaging (odds ratio, 0.26; 95% CI, 0.07-0.97).

Table 7

Parent and Patient Sociodemographic Characteristics by Subgroup.

Table 8

Differential Effect of the Decision Aid on Parent Knowledge, Decisional Conflict, Physician Trust, and Parent Involvement in the Decision (OPTION Score) Based on Parent Sociodemographic Characteristics.

Figure 3

Knowledge (%) Subgroup Effects.

Figure 4

Trust in Physician Scale Subgroup Effects.

Figure 5

Cranial CT Rate Subgroup Effects.

Table 9

Differential Effect of the Decision Aid on the Emergency Department Cranial CT Rate Based on Parent Sociodemographic Characteristics.

Discussion

Decisional Context

The decision addressed in this study—whether to obtain a cranial CT or to actively observe the child at home for the development of signs or symptoms that require further evaluation—typically takes place in an ED setting. Thus, this decision occurs at the level of the ED (or the hospital). In this clinical setting there is typically no previously existing doctor–patient relationship and no future follow-up appointments in which the decision can be reconsidered. Parents are often concerned with their child's risk for concussion and are frequently unaware of their child's risk for ciTBI. With the increased availability and speed of cranial CT imaging and parental concern and anxiety, clinicians opt for imaging in 37% to 50%²² of children presenting with minor head trauma, while fewer than 10% show evidence of TBI² and only 0.2% require neurosurgical intervention.⁵ This exposes children unnecessarily to ionizing radiation and increases health care utilization.

Study Results in Context

In children with minor head trauma who were at moderate risk for clinically important traumatic brain injury according to the PECARN prediction rules,³ shared decision-making facilitated by a DA increased parent knowledge, increased parent engagement in the decision-making process, decreased decisional conflict, and increased parents' trust in their child's clinician. Both parents and clinicians found the DA to be acceptable, and a greater proportion of parents in the DA arm reported being highly satisfied with their choice. Although there was no difference in the frequency of ED cranial CT imaging between study arms, DA patients less frequently underwent outpatient imaging and blood testing after ED discharge. No cases of clinically important traumatic brain injury were missed in either arm of the trial.

Findings from this trial indicate that parents of children with minor head trauma can be feasibly informed of their child's risk for clinically important traumatic brain injury and engaged in the decision of whether to obtain cranial CT imaging in the busy emergency department setting. The significant increase in knowledge in the intervention arm indicates that clinicians' usual approach to educating parents about their child's risks and options for care can be improved on with use of a standardized discussion tool such as the Head CT Choice DA. Moreover, in the process of designing the DA, we learned, to the surprise of the investigative team, that parents are often most concerned about whether their child has suffered a concussion, and that clinicians often must address the diagnosis of concussion prior to discussing the risk of structural brain injury. The heightened awareness of concussion in the lay public may be one reason that parents often express this concern when seeking emergency care for their child.⁵ Use of the DA equips clinicians to directly elicit and address parents' concerns in the diagnostic decision-making process. In addition, making parents aware of the availability of cranial CT imaging for diagnosis and the advantages and disadvantages of CT imaging did not increase imaging rates. Rather, it likely had the opposite effect; parents, once informed of their child's risk for structural brain injury and the options for diagnosis, opted to have their child undergo less extensive evaluation after ED discharge. In addition, a greater proportion of parents who were engaged in SDM using the DA were strongly satisfied with their choice between imaging or observation for the child. This is consistent with prior research in which more than 90% of parents expressed a desire to be involved in health care decisions regarding their child.^29,30 Patients of clinicians randomized to the DA had a shorter ED length of stay. The rationale for this is not entirely clear, as there was no difference in the proportion of patients who were further observed in the ED to determine the need for CT between study arms. It is possible that parents who were explicitly engaged in the decision of imaging vs observation were more involved their child's care and in closer communication with their care team, facilitating a timelier discharge, but this is speculation. There were no patients in either arm of the study who experienced a missed clinically important traumatic brain injury, suggesting the intervention is as safe as usual care. Prior research guiding the choice of cranial CT imaging in pediatric minor head trauma has largely focused on developing optimal prediction rules to guide clinician decision-making.^31,32 These prediction rules have largely focused on the clinician as the primary decision maker; none have intentionally sought to educate and engage parents in the decision of whether to obtain cranial CT. This is also the first and largest trial of a shared decision-making intervention in children with minor head trauma.^5,33,34

Implementation of Study Results

The Head CT Choice DA is designed to be used simultaneously by both the clinician and the parent at the patient's bedside, with the clinician sitting down with the parent to discuss his or her child's care. The DA assists the clinician in addressing parents' concerns regarding the diagnosis of concussion and facilitates transitioning the conversation to the child's risk for structural brain injury (described and depicted pictorially as “bleeding in or around the brain”) and clearly frames the decision for parents of whether to obtain cranial CT imaging in the ED or to further observe their child. The tool then displays the child's risk of ciTBI in ways parents can understand, using prose, natural frequencies with a common denominator, and a pictogram. With their child's risk for ciTBI clearly in mind, parents can then consider the advantages and disadvantages of the available diagnostic options and participate in a shared decision of whether to have their child undergo cranial CT imaging in the ED or to further observe their child. The results of this trial invite clinicians to consider whether their current approach to interacting with parents of children at moderate risk of ciTBI educates, engages, and empowers parents in the decision-making process to this degree, facilitating both heartfelt trust in the clinician and cognitive understanding and agreement of the diagnostic plan. Use of the tool requires only 2 additional minutes of clinicians' time.

We recommend clinicians consider using the DA for parents of children at moderate risk for ciTBI. The clinician can obtain an estimate of the child's risk for ciTBI, download the DA corresponding to the patient's level of risk at http://shareddecisions.mayoclinic.org and print it out, and engage the parent in a shared decision-making discussion. The DA should be given to the parent to take home for further use and review as needed, and the shared decision-making discussions should be documented in the medical record to facilitate communication with downstream providers and continuity of care.

In the emergency care setting, the PECARN prediction rules were routinely used at the sites in which the trial was conducted. When seeking to implement the DA as part of routine practice, the DA will need to be integrated with existing pathways for evaluation of children with minor head trauma, and clinical practice protocols at the participating sites will require modification. A standardized approach to educating and training clinicians to use the DA with parents must be developed and implemented, particularly given the role of study coordinators in the trial and their provision of just-in-time training. The fidelity of implementation will also need to be tracked. Although clinicians may be concerned about additional time required for communication when using the DA with parents, the trial data show that the ED length of stay was actually shorter in intervention patients. These potential barriers must be explored and solutions to these and other potentially unintended consequences of implementation will need to be developed in the process of implementation. Without further implementation studies, the DA may not have the same effect on patient outcomes that was observed in this trial.

Generalizability

We enrolled only children with 1 to 2 PECARN risk factors who were at moderate risk of ciTBI. We did not test the DA in parents of children at either low or high risk for ciTBI and did not measure the impact of the tool in these risk populations. For these reasons, the study findings can be confidently applied only to parents of children at moderate risk of ciTBI. However, we enrolled patients from both community and academic centers.

We conducted the study at 7 US EDs. Two of the 7 EDs were community practice settings in which clinicians provided direct patient care, and 5 were academic medical centers in which trainees frequently interacted with parents and patients and attending physicians supervised the care delivered. Two of the 7 EDs were general emergency departments providing care for both adults and children, and the remaining 5 were dedicated pediatric emergency departments. The geographic diversity and different clinical environments in which the DA was tested suggest that the effect estimates observed in the trial would likely be observed in the settings in which the DA would be used in practice. Given this diversity of practice contexts, the parents who were engaged in the decision-making process in the trial likely represent the variety of the types of parents who will be encountered by clinicians in the course of usual clinical practice. As described below, there was a lower likelihood of imaging in parents with low health literacy and a lower likelihood of imaging if only the father was present during the clinical encounter. Clinicians and systems should keep these observations in mind as they seek to incorporate use of the DA in their clinical practice workflow.

Subpopulation Considerations

Use of the DA increased parent knowledge, decreased decisional conflict, increased physician trust, and increased parent involvement in the decision similarly across all subgroups. There was a higher rate of CT scanning in patients with 2 PECARN risk factors compared with those with only 1 risk factor, which is what one would expect given the higher level of patient risk for ciTBI in those with 2 risk factors. However, there was also a higher likelihood of imaging in children less than 2 years of age, a lower likelihood of CT imaging in parents with low health literacy, and a lower likelihood of imaging if only the father was present during the clinical encounter. None of these subgroup effects were prespecified, and we were likely underpowered to detect other potentially important subgroup effects. As such, these findings should be considered hypothesis generating. Children less than 2 years of age are typically preverbal and use of the DA in this subgroup may have unexpectedly increased parental anxiety and the rate of CT scanning in this subgroup. Parents with lower health literacy may lack the confidence or skill needed to clearly communicate their values and preferences, potentially resulting in a lower rate of CT imaging in this subgroup. However, for the reasons mentioned above, these hypotheses require further testing.

Study Limitations

Several limitations of this trial should be taken into consideration. We randomized at the clinician level. Although this method decreased the risk of contamination compared with randomizing at the patient level, there is still some risk that clinicians in the intervention and control arms of the study communicated their experiences in participating in the trial with one another. To decrease this risk, the principal investigator emphasized to participating clinicians the importance of not speaking to one another about their trial experiences prior to completion of the trial, and site investigators reinforced this message and monitored for evidence of contamination throughout the trial. Furthermore, even if contamination were to occur, this would likely decrease the magnitude of any differences observed between the intervention and control arms, biasing the effect of the intervention toward the null. We were unable to contact 81 (8.3%) of the patients by phone for assessment of a secondary outcome.

Of these patients, 31 had visits documented in the electronic medical record after 7 days, and the remaining 50 did not have any adverse events identified on review of continuous quality improvement reports at each participating hospital or the morgue records at each participating county. The 92% phone follow-up rate supplemented by mortality review of quality improvement and morgue records is robust and comparable to the original PECARN investigation³⁵ and a prior multicenter shared decision-making trial conducted in the ED setting.⁵ Despite these rigorous measures to ensure close follow-up, it is possible that some outcomes were missed. We were unable to obtain video recordings in 455 (46.9%) of the clinical encounters. However, the 516 recordings that were obtained exceeded the required sample size of 221 needed to meet power estimates.⁷ Given the relatively low rate of missingness of data in this trial, we decided not to impute missing data, which increases the degree of confidence we can place in the trial results.

The primary limitations of the heterogeneity of DA effect analyses relate to issues of multiple testing and imprecision around estimates of subgroup effects. To limit the risk of bias associated with multiple testing, we prespecified hypotheses based on prior observations in SDM trials.^11,22 We also followed guideline recommendations for reporting subgroup analyses in clinical trials³⁶ by presenting only those subgroup analyses that were prespecified or based on a primary study outcome in the abstract, basing analyses of the heterogeneity of DA effect on tests for interaction, and exercising caution in interpreting subgroup differences. Our analyses often yielded imprecise results of potentially important subgroup effects. However, this limitation is inherent in subgroup analyses of clinical trials, and, to the best of the investigators' knowledge, the current investigation represents the largest cohort of patients enrolled in an SDM trial to date and has potential to reveal important insights related to the effect of a DA in potentially vulnerable patients.

Finally, some of the effect sizes of the impact of the DA were small. However, the trial was well designed to test the effects of the DA, and the findings inform clinicians and other stakeholders regarding the potential effects of SDM in this population.

Future Research

While the findings from this multicenter trial suggest that the DA is effective across a variety of practice settings, further work is needed to determine how to embed and routinize the intervention in the flow of emergency care and the degree of uptake that can be achieved through targeted implementation of the intervention. Development of an interactive electronic version of the DA as well as incorporation of the aid within the electronic medical record and/or other web-based platform may also aid dissemination and implementation of the intervention. Consideration of inclusion of the DA in clinical guidelines for the care of children with traumatic brain injury also requires further exploration. Finally, the effect of the DA in specific parent subgroups requires additional testing in a larger population of parents of children with head trauma to determine if parent characteristics impact the direction and magnitude of the DA's effect on cranial CT imaging.

Conclusions

Use of a DA during the clinical encounter for parents of children at moderate risk of clinically important traumatic brain injury increased parent knowledge and satisfaction, decreased decisional conflict, and was perceived to be acceptable to both parents and clinicians. Use of the DA had no effect on the rate of cranial CT imaging in the ED but decreased patient length of stay and the rate of CT imaging and testing at 7 days with no increase in adverse events. We believe that the direction and weight of evidence generated by this comparative effectiveness research trial warrant additional implementation studies. The findings suggest use of the DA will benefit parents, patients, clinicians, and the efficiency of emergency care delivery, and will safely decrease downstream health care utilization after ED discharge.

References

1.: Faul M, Xu L, Wald MM, Coronado VG. Traumatic Brain Injury in the United States: Emergency Department Visits, Hospitalizations and Deaths 2002–2006. Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. March 2010. https://www.cdc.gov/traumaticbraininjury/pdf/blue_book.pdf

2.: Quayle KS, Powell EC, Mahajan P, et al. Epidemiology of blunt head trauma in children in U.S. emergency departments. N Engl J Med. 2014;371(20):1945-1947. [PubMed: 25390756]

3.: Atabaki SM, Hoyle JD Jr, Schunk JE, et al. Comparison of prediction rules and clinician suspicion for identifying children with clinically important brain injuries after blunt head trauma. Acad Emerg Med. 2016;23(5):566-575. doi:10.1111/acem.12923 [PubMed: 26825755] [CrossRef]

4.: Miglioretti DL, Johnson E, Williams A, et al. The use of computed tomography in pediatrics and the associated radiation exposure and estimated cancer risk. JAMA Pediatr. 2013;167(8):700-707. doi:10.1001/jamapediatrics.2013.311 [PMC free article: PMC3936795] [PubMed: 23754213] [CrossRef]

5.: Kuppermann N, Holmes JF, Dayan PS, et al. Identification of children at very low risk of clinically-important brain injuries after head trauma: a prospective cohort study. Lancet. 2009;374(9696):1160-1170. [PubMed: 19758692]

6.: Stacey D, Legare F, Lewis K, et al. Decision aids for people facing health treatment or screening decisions. Cochrane Database Syst Rev. 2017;4(4):CD001431. doi:10.1002/14651858.CD001431.pub5 [PMC free article: PMC6478132] [PubMed: 28402085] [CrossRef]

7.: Hess EP, Hollander JE, Schaffer JT, et al. Shared decision making in patients with low risk chest pain: prospective randomized pragmatic trial. BMJ. 2016;355:i6165.doi:10.1136/bmj.i6165 [PMC free article: PMC5152707] [PubMed: 27919865] [CrossRef]

8.: Montori VM, Breslin M, Maleska M, Weymiller AJ. Creating a conversation: insights from the development of a decision aid. PLoS Med. 2007;4(8):e233. doi:10.1371/journal.pmed.0040233 [PMC free article: PMC1939861] [PubMed: 17683195] [CrossRef]

9.: Montori VM, Shah ND, Pencille LJ, et al. Use of a decision aid to improve treatment decisions in osteoporosis: the osteoporosis choice randomized trial. Am J Med. 2011;124(6):549-556. [PubMed: 21605732]

10.: Hess EP, Knoedler MA, Shah ND, et al. The chest pain choice decision aid: a randomized trial. Circ Cardiovasc Qual Outcomes. 2012;5(3):251-259. [PubMed: 22496116]

11.: Hess EP, Wyatt KD, Kharbanda AB, et al. Effectiveness of the head CT choice decision aid in parents of children with minor head trauma: study protocol for a multicenter randomized trial. Trials. 2014;15:253. doi:10.1186/1745-6215-15-253 [PMC free article: PMC4081461] [PubMed: 24965659] [CrossRef]

12.: Karanicolas PJ, Montori VM, Devereaux PJ, Schunemann H, Guyatt GH. A new ‘mechanistic-practical” framework for designing and interpreting randomized trials. J Clin Epidemiol. 2009;62(5):479-484. doi:10.1016/j.jclinepi.2008.02.009 [PubMed: 18468856] [CrossRef]

13.: Pocock SJ, Simon R. Sequential treatment assignment with balancing for prognostic factors in the controlled clinical trial. Biometrics. 1975;31(1):103-115. [PubMed: 1100130]

14.: Signorini DF, Leung O, Simes RJ, Beller E, Gebski VJ, Callaghan T. Dynamic balanced randomization for clinical trials. Stat Med. 1993;12(24):2343-2350. [PubMed: 8134737]

15.: Study datasets. Pediatric Emergency Care Applied Research Network. Accessed January 26, 2017. http://www.pecarn.org/studyDatasets/Default

16.: About us. Patient-Centered Outcomes Research Institute. Accessed February 2, 2017. http://www.pcori.org/about-us

17.: Elwyn G, Hutchings H, Edwards A, et al. The OPTION scale: measuring the extent that clinicians involve patients in decision-making tasks. Health Expect. 2005;8(1):34-42. [PMC free article: PMC5060272] [PubMed: 15713169]

18.: O'Connor AM. Validation of a decisional conflict scale. Med Decis Making. 1995;15(1):25-30. [PubMed: 7898294]

19.: Gattellari M, Ward JE. Will men attribute fault to their GP for adverse effects arising from controversial screening tests? An Australian study using scenarios about PSA screening. J Med Screen. 2004;11(4):165-169. [PubMed: 15563771]

20.: Thom DH, Ribisl KM, Stewart AL, Luke DA. Further validation and reliability testing of the Trust in Physician Scale. The Stanford Trust Study Physicians. Med Care. 1999;37(5):510-517. [PubMed: 10335753]

21.: Schulz KF, Altman DG, Moher D; the CONSORT Group. CONSORT 2010 statement: updated guidelines for reporting parallel group randomized trials. Ann Intern Med. 2010;152(11):726-732. [PubMed: 20335313]

22.: Anderson RT, Montori VM, Shah ND, et al. Effectiveness of the Chest Pain Choice decision aid in emergency department patients with low-risk chest pain: study protocol for a multicenter randomized trial. Trials. 2014;15:166. [PMC free article: PMC4031497] [PubMed: 24884807]

23.: McNaughton, C., et al., Short, subjective measures of numeracy and general health literacy in an adult emergency department. Acad Emerg Med., 2011. 18(11):1148-55. doi:10.1111/j.1553-2712.2011.01210.x. [PMC free article: PMC3886121] [PubMed: 22092896] [CrossRef]

24.: Fagerlin A, Zikmund-Fisher BJ, Ubel PA, Jankovic A, Derry HA, Smith DM. Measuring numeracy without a math test: development of the Subjective Numeracy Scale. Med Decis Making. 2007;27(5):672-680. [PubMed: 17641137]

25.: Donner A, Klar N. Design and Analysis of Cluster Randomized Trials in Health Research. Arnold Publishers; 2000.

26.: Rising KL, Hollander JE, Schaffer JT, et al. Effectiveness of a decision aid in potentially vulnerable patients: a secondary analysis of the chest pain choice multicenter randomized trial. Med Decis Making. 2018;38(1):69-78. [PubMed: 28525723]

27.: Wang R, Lagakos SW, Ware JH, Hunter DJ, Drazen JM. Statistics in medicine--reporting of subgroup analyses in clinical trials. N Engl J Med. 2007;357(21):2189-2194. [PubMed: 18032770]

28.: McNaughton CD, Collins SP, Kripalani S, et al. Low numeracy is associated with increased odds of 30-day emergency department or hospital recidivism for patients with acute heart failure. Circ Heart Fail. 2013;6(1):40-46. [PMC free article: PMC3663074] [PubMed: 23230305]

29.: Pinker S. The perilous aftermath of a simple concussion. The Wall Street Journal. April 6, 2016. http://www.wsj.com/articles/the-perilous-aftermath-of-a-simple-concussion-1459963724

30.: Reddy, S., New Help for Concussions, in The Wall Street Journal. October 8, 2016. http://online.wsj.com/public/resources/documents/print/WSJ_-D001-20161018.pdf [link no longer works]

31.: Yamamoto LG. Application of informed consent principles in the emergency department evaluation of febrile children at risk for occult bacteremia. Hawaii Med J. 1997;56(11):313-317, 320-322. [PubMed: 9420350]

32.: Yamamoto LG, Young LL, Roberts JL. Informed consent and parental choice of anesthesia and sedation for the repair of small lacerations in children. Am J Emerg Med. 1997;15(3):285-289. [PubMed: 9148989]

33.: Osmond MH, Klassen TP, Wells GA, et al. CATCH: a clinical decision rule for the use of computed tomography in children with minor head injury. CMAJ. 2010;182(4):341-348. [PMC free article: PMC2831681] [PubMed: 20142371]

34.: Babl FE, Borland ML, Phillips N, et al. Accuracy of PECARN, CATCH, and CHALICE head injury decision rules in children: a prospective cohort study. Lancet. 2017;389(10087):2393-2402. doi:10.1016/S0140-6736(17)30555-X [PubMed: 28410792] [CrossRef]

35.: Flynn D, Knoedler MA, Hess EP, et al. Engaging patients in health care decisions in the emergency department through shared decision-making: a systematic review. Acad Emerg Med. 2012;19(8):959-967. [PubMed: 22853804]

36.: Coylewright M, Branda M, Inselman JW, et al. Impact of sociodemographic patient characteristics on the efficacy of decision AIDS: a patient-level meta-analysis of 7 randomized trials. Circ Cardiovasc Qual Outcomes. 2014;7(3):360-367. doi:10.1161/HCQ.0000000000000006 [PubMed: 24823953] [CrossRef]

Acknowledgment

Research reported in this report was [partially] funded through a Patient-Centered Outcomes Research Institute® (PCORI®) Award CE-12-11-4435). Further information available at: https://www.pcori.org/research-results/2013/using-decision-aid-emergency-room-help-parents-children-head-trauma-understand

Appendix

PCORI ID: CE-12-11-4435

ClinicalTrials.gov ID: NCT02063087

Suggested citation:

Hess EP, Shah N, Westphal J, et al. (2018). Using a Decision Aid in the Emergency Room to Help Parents of Children with Head Trauma Understand Options for Diagnosing Brain Injury. Patient-Centered Outcomes Research Institute (PCORI). https://doi.org/10.25302/11.2018.CE.12114435

Disclaimer

The [views, statements, opinions] presented in this report are solely the responsibility of the author(s) and do not necessarily represent the views of the Patient-Centered Outcomes Research Institute® (PCORI®), its Board of Governors or Methodology Committee.

This book is distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License which permits noncommercial use and distribution provided the original author(s) and source are credited. (See https://creativecommons.org/licenses/by-nc-nd/4.0/

Bookshelf ID: NBK591248PMID: 37104637DOI: 10.25302/11.2018.CE.12114435

Children <2 y	Children 2-18 y
High-risk factors
GCS ≤14 or other signs of altered mental status^a	GCS ≤14 or other signs of altered mental status^a
Palpable skull fracture	Signs of basilar skull fracture
Non–high-risk factors
Occipital, parietal, or temporal scalp hematoma History of LOC ≥5 s	History of vomiting History of LOC of any duration
Not acting normally per parent	Severe headache
Severe mechanism of injury^b	Severe mechanism of injury^b

	Usual care (n = 478)	Decision aid (n = 493)
	No. (%)	No. (%)
Age, mean (SD), y	6.8 (5.3)	6.6 (6.7)
Age group
<2 y	109 (22.8)	123 (24.9)
2-18 y	369 (77.2)	370 (75.1)
Male	285 (59.6)	290 (58.8)
Race
White	347 (72.6)	371 (75.3)
Black	54 (11.3)	61 (12.4)
AmI/Asian/Pac/other	19 (4.0)	24 (4.9)
Hispanic	62 (13.0)	54 (11.0)
Insurance
Government	121 (25.3)	133 (27.0)
Private FFS	278 (58.2)	280 (56.8)
HMO	67 (14.0)	65 (13.2)
None	12 (2.5)	15 (3.0)
Annual income ($)
Less than 20 000	66 (13.8)	63 (12.8)
20 000-29 999	26 (5.4)	30 (6.1)
30 000-39 999	34 (7.1)	45 (9.1)
40 000-59 999	55 (11.5)	60 (12.2)
60 000-79 999	51 (10.7)	44 (8.9)
80 000-99 999	60 (12.6)	47 (9.5)
100 000 or more	168 (35.1)	185 (37.5)
Missing	18 (3.8)	19 (3.9)
Respondent education
High school or less	31 (6.5)	27 (5.5)
High school or graduate education diploma	55 (11.5)	46 (9.3)
College or vocational school	126 (26.4)	152 (30.8)
College graduate (4 y)	147 (30.8)	151 (30.6)
Graduate degree	101 (21.1)	93 (18.9)
Missing data	11 (2.3)	9 (1.8)
Parent present during the ED visit
Both parents	170 (35.8)	156 (31.4)
1 parent	303 (63.4)	334 (67.7)
Other guardian	4 (0.8)	4 (0.8)
Literacy scale^a
Mean (SD)	13.57 (1.96)	13.43 (2.00)
Median (IQR)	14 (13-15)	14 (12-15)
Low (<12), No. (%)	66 (14.1)	76 (15.8)
Numeracy scale^a
Mean (SD)	35.69 (9.71)	36.09 (8.61)
Median (IQR)	38 (30-43)	38 (32-42)
Low (<34), No. (%)	158 (33.0)	158 (32.0)

PECARN risk factors^a	Usual care (n = 478)	Decision aid (n = 493)
PECARN risk factors^a	No. (%)	No. (%)
<2 y^b (n = 232)
Presence of occipital, temporal, or parietal scalp hematoma
No	85 (17.8)	92 (18.6)
Yes	23 (4.8)	32 (6.5)
Loss of consciousness ≥5 s
No	94 (19.7)	116 (23.5)
Yes	13 (2.7)	7 (1.4)
Unknown	1 (0.2)	1 (0.2)
Severe mechanism (PECARN definition)^a
No	51 (10.7)	39 (7.9)
Yes	57 (11.9)	83 (16.8)
Unknown	0 (0.0)	2 (0.4)
Acting abnormally per parent
No	74 (15.5)	90 (18.2)
Yes	34 (7.1)	34 (6.9)
2-18 y^b (n = 739)
Any loss of consciousness
No	266 (55.8)	260 (52.6)
Yes	88 (18.4)	93 (18.8)
Unknown	15 (3.1)	17 (3.4)
Any vomiting since injury
No	195 (40.9)	199 (40.3)
Yes	174 (36.5)	171 (34.6)
Severe mechanism (PECARN definition)^a
No	270 (56.6)	272 (55.1)
Yes	98 (20.5)	96 (19.4)
Unknown	1 (0.2)	2 (0.4)
Severe headache in ED
No	279 (58.5)	295 (59.7)
Yes	88 (18.4)	71 (14.4)
Unknown	2 (0.4)	4 (0.8)
No. of PECARN variables present
1	380 (79.5)	400 (81.1)
2	98 (20.5)	93 (18.9)

Outcome	Decision aid (n = 493)	Usual care (n = 478)	Effect, MD or OR (95% CI)	P value
Parent/caregiver knowledge
Mean (SD) knowledge, No. of questions correct out of 10)	6.20 (2.0)	5.26 (2.0)	MD 0.95 (0.60-1.30)	<.001
Decisional conflict and trust
Decisional conflict scale (range 0-100)	14.8 (15.5)	19.2 (16.6)	MD −4.9 (−7.3 to −2.4)	<.001
Trust in physician scale (range 0-100)	91.5 (11.9)	89.3 (13.7)	MD 2.3 (0.4-4.1)	.015
Parent/caregiver involvement in the decision
OPTION scale (n = 510) (range 0-100)	25.0 (8.5)	13.3 (6.5)	MD 11.7 (9.6-13.9)	<.001
Parent/caregiver acceptability
Amount of information			OR 1.4 (0.8-2.5)	.287
Satisfied	455 (92.3)	441 (92.3)
Unsatisfied	21 (4.3)	28 (5.9)
Clarity of information satisfied
Satisfied	382 (77.5)	342 (71.5)	OR 1.5 (1.1-2.1)	.016
Unsatisfied	94 (19.1)	124 (25.9)
Helpfulness of information satisfied
Satisfied	377 (76.5)	344 (72.0)	OR 1.4 (1.0-2.0)	.054
Unsatisfied	101 (20.5)	124 (25.9)
Would recommend to others			OR 0.7 (0.4-1.2)	.202
Yes	460 (93.3)	442 (92.5)
Not sure/not at all	19 (3.9)	27 (5.6)
Would want to use for other decisions			OR 0.9 (0.6-1.4)	.549
Yes	460 (93.3)	442 (92.5)
Not sure/not at all	47 (9.5)	52 (10.9)
Parent/caregiver satisfaction
Satisfied with the choice			OR 1.4 (1.1-1.9)	.015
Strongly agree	254 (51.5)	210 (43.9)
Agree	165 (33.5)	199 (41.6)
Not sure/not at all	46 (9.3)	56 (11.7)
Clinician acceptability
Helpfulness of the information			OR 1.8 (1.0-3.3)	.05
Yes	247 (50.1)	192 (40.2)
Not sure/not at all	237 (48.1)	278 (58.2)
Present information on other diagnostic choices in the same way?			OR 1.5 (0.7-3.4)	.343
Yes	281 (57.0)	216 (45.2)
Not sure/not at all	207 (42.0)	259 (54.2)
Recommend the way you presented information to other providers?			OR 2.9 (1.2-6.8)	.014
Yes	305 (61.9)	198 (41.4)
Not sure/not at all	183 (37.1)	275 (57.5)

Characteristic	Decision aid (n = 493), No. (%)	Usual care (n = 478), No. (%)	Odds Ratio, 95% CI	P value
Observed in the ED to determine the need for CT after being evaluated by the clinician	271 (54.0)	287 (60.0)	0.81 (0.63-1.05)	.11
Cranial CT obtained in the ED	109 (22.1)	116 (24.3)	0.81 (0.51-1.27)	.35
ED length of stay^a	176.0 (135.2)	198.8 (162.1)	−22.8 (−41.6 to −4.0)	.02
Admitted to the hospital	9 (1.8)	9 (1.9)	0.97 (0.38-2.45)	.94
Clinically important traumatic brain injury at 7 d	0 (0.0)	1 (0.2)^b	1.0 (1.0-1.0)	1.0
ED return visit within 7 d	10 (2.0)	18 (3.8)	0.54 (0.24-1.24)	.147

Utilization category	Decision aid (n = 493), mean (SD)	Usual care (n = 478), mean (SD)	Mean difference (95% CI)	P value^a
Evaluation and management^b	1.84 (0.23)	1.88 (0.2)	−0.04 (−0.12 to 0.04)	.727
Imaging	0.65 (0.41)	0.88 (0.56)	−0.23 (−0.35 to −0.11)	.045
Tests	0.41 (0.32)	0.7 (0.55)	−0.29 (−0.40 to −0.16)	.046
Procedures	0.17 (0.13)	0.26 (0.19)	−0.09 (−0.17 to −0.03)	.082
Other	0.23 (0.28)	0.38 (0.46)	−0.15 (−0.26 to −0.04)	.009
Unclassified	0.02 (0.02)	0.02 (0.03)	0.00 (−0.03 to 0.03)	.802

Variable	Usual care	DA	P value
Variable	No. (%)	No. (%)	P value
1	478 (100.0)	493 (100.0)
Gender of child			.8
Male	285 (59.6)	290 (58.8)
Female	193 (40.4)	203 (41.2)
Race			.345
White	347 (72.6)	371 (75.3)
Non-White	131 (27.4)	122 (24.7)
Respondent education			.261
<High school	31 (6.5)	27 (5.5)
High school or graduate educational diploma	55 (11.5)	46 (9.3)
Some college/Associate's degree	126 (26.4)	152 (30.8)
College graduate	147 (30.8)	151 (30.6)
Graduate/professional	101 (21.1)	93 (18.9)
Other	7 (1.5)	15 (3.0)
Missing	11 (2.3)	9 (1.8)
Education			.205
≤High school	86 (18.0)	73 (14.8)
>High school	374 (78.2)	396 (80.3)
Missing	18 (3.8)	24 (4.9)
Family income			.568
Less than $20 000	66 (13.8)	63 (12.8)
$20 000 to $29 999	26 (5.4)	30 (6.1)
$30 000 to $39 999	34 (7.1)	45 (9.1)
$40 000 to $59 999	55 (11.5)	60 (12.2)
$60 000 to $79 999	51 (10.7)	44 (8.9)
$80 000 to $99 999	60 (12.6)	47 (9.5)
$100 000 or more	168 (35.1)	185 (37.5)
Missing	18 (3.8)	19 (3.9)
Literacy			.48
Typical (≥12)	401 (83.9)	406 (82.4)
Low (<12)	66 (13.8)	76 (15.4)
Missing	11 (2.3)	11 (2.2)
Numeracy			.738
Typical (≥34)	320 (66.9)	335 (68.0)
Low (<34)	158 (33.1)	158 (32.0)
Adult			.357
Both parents	171 (35.8)	155 (31.4)
1 parent	303 (63.4)	334 (67.7)
Other	4 (0.8)	4 (0.8)
No. of PECARN risk factors^a		0.521
1	380 (79.5)	400 (81.1)
2	98 (20.5)	93 (18.9)
Literacy scale		0.268
Mean (SD)	13.573876 (2.0)	13.431535 (2.0)
Numeracy		0.494
Mean (SD)	35.686192 (9.7)	36.089249 (8.6)

	Knowledge (%)	P value	Decisional conflict	P value	Trust in physician	P value	OPTION score	P value
Gender of child		.774		.583		.192		.439
Male	9.47 (6.30-12.65)		−4.91 (−7.46 to −2.37)		3.31 (1.21-5.41)		12.23 (10.55-13.91)
Female	8.88 (4.74-13.03)		−3.98 (−7.50 to −0.46)		1.02 (−1.60 to 3.63)		11.08 (8.98-13.18)
Race		.199		.05		.293		.17
White	9.70 (6.91-12.48)		−3.07 (−5.45 to −0.68)		1.55 (−0.21 to 3.30)		12.27 (10.75-13.79)
Non-White	5.91 (0.84-10.97)		−8.14 (−12.33 to −3.95)		4.29 (0.51-8.06)		9.98 (7.29-12.67)
Education		.619		.375		.893		.762
≤HS	6.49 (0.55-13.53)		−6.16 (−12.18 to −0.14)		2.23 (−2.45 to 6.92)		12.55 (9.70-15.40)
>HS	9.46 (6.85-12.07)		−4.02 (−6.26 to −1.78)		2.11 (0.41-3.82)		11.81 (10.30-13.33)
Income		.999		.294		.576		.367
<$40 000	9.53 (4.73-14.33)		−6.18 (−10.13 to −2.23)		3.23 (−0.55 to 7.00)		12.90 (10.50-15.30)
$40 000 or more	9.65 (6.85-12.45)		−3.55 (−6.00 to −1.11)		2.26 (0.49-4.03)		11.35 (9.75-12.95)
Literacy		.063		.402		.343		.802
Typical	8.25 (5.60-10.90)		−4.23 (−6.42 to −2.04)		2.30 (0.61-3.99)		11.85 (10.40-13.31)
Low	14.61 (7.36-21.85)		−5.87 (−12.03 to 0.29)		4.28 (−0.99 to 9.56)		11.78 (8.57-14.99)
Numeracy		.142		.741		.251		.756
Typical	10.22 (7.36-13.08)		−4.25 (−6.62 to −1.88)		1.52 (−0.23 to 3.28)		11.51 (9.89-13.13)
Low	6.50 (1.89-11.12)		−4.86 (−8.84 to −0.88)		3.82 (0.44-7.21)		11.72 (9.51-13.93)
Age^a		.764		.308		.746		.681
<2	8.89 (3.49-14.30)		−2.02 (−6.45 to 2.42)		1.96 (−1.37 to 5.29)		12.18 (9.57-14.79)
≥2	9.42 (6.58-12.27)		−5.15 (−7.49 to −2.80)		2.61 (0.72-4.49)	*	11.48 (9.95-13.01)
No. of risk factors^a		.662		.346		.815		.132
1	8.88 (6.03-11.73)		−4.93 (−7.19 to −2.67)		2.24 (0.37-4.11)	*	11.34 (9.88-12.80)
2	10.15 (4.82-15.47)		−2.38 (−7.47 to 2.71)		2.64 (−0.70 to 5.97)		14.07 (10.96-17.18)
Father only		.277		.859		.135		.49
No	8.69 (6.05-11.33)		−4.47 (−6.68 to −2.26)		1.83 (0.11-3.55)	*	11.49 (10.11-12.86)
Yes	14.40 (5.43-23.37)		−3.90 (−10.09 to 2.28)		6.59 (0.92-12.27)	*	12.75 (8.11-17.39)
Both parents		.197		.663		.057		.553
No	10.59 (7.52-13.66)		−4.25 (−6.82 to −1.69)		3.49 (1.50-5.47)	*	12.03 (10.42-13.63)
Yes	6.34 (2.06-10.62)		−4.53 (−8.06 to −1.00)		−0.37 (−3.23 to 2.50)		11.36 (9.02-13.70)

	Cranial CT	Overall P
	DA effect	Overall P
Gender of child		.916
Male	0.96 (0.64-1.42)
Female	0.93 (0.57-1.54)
Race		.423
White	0.99 (0.70-1.41)
Non-White	0.81 (0.42-1.56)
Education		.815
≤HS	0.80 (0.36-1.79)
>HS	0.94 (0.67-1.33)
Income		.912
<$40 000	0.97 (0.53-1.75)
$40 000 or more	0.89 (0.61-1.29)
Literacy		.01
Typical	1.09 (0.77-1.53)
Low	0.36 (0.16-0.83)
Numeracy		.816
Typical	0.93 (0.63-1.36)
Low	0.96 (0.56-1.65)
Age^a		.019
<2	1.94 (0.95-3.97)
≥2	0.78 (0.55-1.11)
No. of PECARN risk factors^a		.193
1	1.09 (0.75-1.58)
2	0.67 (0.37-1.22)
Father only		.041
No	1.02 (0.74-1.41)
Yes	0.26 (0.07-0.97)
Both parents		.432
No	1.03 (0.71-1.51)
Yes	0.75 (0.43-1.29)

Using a Decision Aid in the Emergency Room to Help Parents of Children with Head Trauma Understand Options for Diagnosing Brain Injury

Authors

Affiliations

Structured Abstract

Background:

Objectives:

Methods:

Results:

Conclusions:

Limitations and Subpopulation Considerations:

Background

Participation of Patients and Other Stakeholders in the Design and Conduct of Research and Dissemination of Findings

Methods

Study Design

Aim 1: Give Parents a Voice and Incorporate the Perspectives of Multiple Stakeholders by Refining the Head CT Choice DA

Aim 2: Test if the DA Improves Validated Patient-Centered Outcome Measures and Safely Decreases Health Care Utilization in Children at Moderate Risk of ciTBI

Forming the Study Cohort

Study Setting

Interventions

Delivery of the DA

Usual Care

Follow-up

Study Outcomes

Primary Outcome

Secondary Outcomes

Modification of the Original Study Protocol

Data Collection and Sources

Analytical and Statistical Approaches

Heterogeneity of DA Effect Analyses

Missing Data

Conduct of the Study

Results

Aim 1: Give Parents a Voice and Incorporate the Perspectives of Multiple Stakeholders by Refining the Head CT Choice DA

Aim 2. Test if the DA Improves Validated Patient-Centered Outcome Measures and Safely Decreases Health Care Utilization in Children at Moderate Risk of ciTBI

Parent Knowledge, Decisional Conflict, and Trust

Parent Participation, Acceptability, and Satisfaction

Clinician Acceptability

Management and 7-Day Health Outcomes

Fidelity of DA Use

Health Care Utilization Within 7 Days

Heterogeneity of DA Effect Analyses

Discussion

Decisional Context

Study Results in Context

Implementation of Study Results

Generalizability

Subpopulation Considerations

Study Limitations

Future Research

Conclusions

References

Acknowledgment

Appendix

Supplementary file

Pre Encounter Survey

Post Encounter Survey

Clinician Survey

Suggested citation:

Disclaimer

Table 1Clinical Factors Included in the PECARN Prediction Rules for TBI

Figure 1Head CT Choice

Table 2Baseline Characteristics (N = 971)

Figure 2Head CT CONSORT

Table 3PECARN Risk Factors for the Enrolled Patients (N = 971)

Table 4Effect of Decision Aid on Parent Knowledge, Decisional Conflict, Trust in the Physician, Parent/Caregiver Involvement in the Decision, Patient Acceptability and Satisfaction, and Clinician Acceptability

Table 5Effect of Decision Aid on Management and 7-Day Outcomes

Table 6Health Care Utilization Data Obtained From Hospital-Level Billing Data, Including the ED Visit and Utilization Within the Subsequent 7 Daysa

Table 7Parent and Patient Sociodemographic Characteristics by Subgroup

Table 8Differential Effect of the Decision Aid on Parent Knowledge, Decisional Conflict, Physician Trust, and Parent Involvement in the Decision (OPTION Score) Based on Parent Sociodemographic Characteristics

Figure 3Knowledge (%) Subgroup Effects

Figure 4Trust in Physician Scale Subgroup Effects

Figure 5Cranial CT Rate Subgroup Effects

Table 9Differential Effect of the Decision Aid on the Emergency Department Cranial CT Rate Based on Parent Sociodemographic Characteristics

Table 6Health Care Utilization Data Obtained From Hospital-Level Billing Data, Including the ED Visit and Utilization Within the Subsequent 7 Days^a