A Qualitative Study of Pediatricians’ Adverse Childhood Experiences Screening Workflows

Abstract

Background & Objective:

Adverse childhood experiences (ACEs) are associated with poor health outcomes over the life course. Interest in ACEs screening is growing, but standard ACEs screening workflows have yet to be established. We aimed to describe common workflow processes and variation among pediatricians who have successfully implemented ACEs screening and response protocols.

Methods:

We conducted semi-structured interviews with members of the American Academy of Pediatrics who practiced in clinical pediatric settings that implemented standardized ACEs screening (n = 18 physicians). Interviews were coded and analyzed using thematic content analysis and clinical processes were examined for differences across ACEs screening workflow processes.

Results:

ACEs screening workflows varied considerably, hinging primarily on determination of a positive screen, the type of interventions recommended in response, and protocolization of the workflow. We identified 5 major theme domains related to ACEs screening workflows: 1) degree of protocolization of the workflow, 2) screening tool(s) used, 3) timing of screening, 4) clinic staff involvement, and 5) interventions recommended and/or initiated by the physician. Common workflow processes were identified and grouped based on determination of and thresholds for response to a positive screen. Clinicians used symptoms, ACE score, or a combination of the 2 as criteria for deciding when to intervene and to what degree, though protocolization of this approach varied.

Conclusions:

ACEs screening workflow variability was largely driven by clinical feasibility and availability of ACEs intervention resources. This variability demonstrates that a one-size-fits-all standardized screening protocol may not be universally feasible or appropriate across practices.

Adverse childhood experiences (ACEs) are stressful and/or traumatic experiences that include abuse, neglect, and household dysfunction. ACEs are associated with poor health outcomes over the life course and have significant public health implications.^1,2 While there is mounting interest in implementation of ACEs screening, standard ACEs screening workflows have yet to be established.³

One of the largest initiatives to standardize and expand ACEs screening to date began in January of 2020, when California launched a statewide effort, the “ACEs Aware Initiative,” to train and reimburse health care providers to screen for ACEs. Interest and initiatives like those in California have driven health systems and payers toward developing ACEs screening approaches, often denovo, in the absence of specific information on ACEs screening and response workflow models. More information about screening workflows and implementation guidance is needed as more and more clinical systems adopt ACEs screening. The current literature that explores provider ACEs screening workflows is limited, and data surrounding the determination of and response to a positive screen is needed.^4,5 This lack of information has been cited by providers as a key barrier to ACEs screening implementation.^6,7 And while there are recommended workflows (such as the California state ACEs Aware initiative’s), there is little information on how these workflows can best be adapted to diverse clinical practice settings.

In this study, we describe clinical ACEs screening and response processes used by physicians whose practices have successfully incorporated ACEs screening in a pediatric setting, as well as barriers, facilitators, and variation in key implementation factors and screening process workflows.

Methods

Participants

We conducted 22 semi-structured interviews with individual physician members of the California Chapter-2 of the American Academy of Pediatrics. Participants were recruited via an electronic survey on ACEs screening distributed through the California Chapter-2 of the American Academy of Pediatrics listserv and via snowball sampling and direct referral. Physicians who had previously or were presently conducting standardized ACEs screening via a defined clinical protocol were eligible for this study. Four physicians interviewed did not meet inclusion criteria because they had not conducted or were not conducting ACEs screening using a standardized approach in a pediatric setting, and their interviews were not included in our analysis, leaving 18 interviews in the sample. All but 2 participants were pediatricians (the remaining 2 were a family physician and an emergency medicine physician), and only one pair of participating pediatricians was part of the same clinical practice.

Development of Interview Guide

The key informant interview facilitator guide question domains were developed from items fielded in a survey on barriers and facilitators to ACEs screening among pediatricians in Southern California (manuscript in preparation). Interview question items and probes were adapted from the survey items and from barriers to screening identified in the literature and were revised to consensus by the study authors A.S. and J.R. The facilitator guide structure was adapted from the facilitator guide of a prior key informant, semi-structured interview study exploring barriers and facilitators to clinical social adversity screening and intervention in a large Southern California integrated health system.⁸

The interview guide was divided into domains to explore 1) approaches to screening and intervention, including screening process workflow details; 2) differences in ACEs screening training; and 3) perceptions, attitudes, and beliefs regarding ACEs screening. These domains also asked about key factors in the design of ACEs screening workflows, including screening tool used, population and timing of screening, role of ancillary staff, threshold for intervention, and subsequent intervention (Appendix).

Data Collection

One author (J.R.) conducted individual phone interviews in a semi-structured format between February and May 2021, which were audio recorded and transcribed. Interview length was 51 minutes on average (range 25–81 minutes). Interviews were concluded once thematic saturation was reached.

We asked all participants to report key characteristics of their practice, including practice or work setting, description of work or practice, years in practice, proportion of their practice’s patients covered by Medi-Cal or Medicaid, proportion of time seeing patients clinically, California county and geography, participant gender, race and ethnicity, and personal experience with ACEs.

Data Analysis

Audio recordings were de-identified, transcribed, and coded based on thematic content using Atlas.ti, version 8. Transcripts were independently reviewed by 2 authors (J.R. and A.S.). An initial codebook was developed (J.R.), refined to consensus (J.R. and A.S.), then definitively coded, and discrepancies were harmonized by consensus between authors (J.R. and A.S.). We used thematic content analysis and drafted process flow diagrams for each pediatrician’s ACEs screening workflow according to guidelines from the Institute for Healthcare Improvement, with swim lanes representing the role of the parent/caregiver and/or patient, clinical staff, and the pediatric provider.⁹ Common process flow archetypes were identified according to similarities and differences found in ACEs screening and response protocols reported across participants.

The University of California Los Angeles Institutional Review Board approved the study.

Results

The majority of the 18 participants identified as female and white were part of an ambulatory large group practice, and were affiliated with a Federally Qualified Health Center or a community health center. Practices were distributed across suburban, urban inner-city, and urban non-inner-city geographies. Participants had been in practice an average of 19 (standard deviation 12) years. Nearly all reported that patients seen in their practice were predominantly covered by Medicaid. Of those reporting, half of the participants had experienced ACEs themselves (Table 1).

Table 1.

Participant Characteristics

Variable	% (n = 18)
Description of work or practice*
Small group practice	22 (4)
Large group or Health Maintenance Organization practice	39 (7)
Federally Qualified Health Center or Community Health Center	50 (9)
Hospital- or medical center-affiliated	22 (4)
Academically affiliated (with medical school or residency)	6 (1)
Practice or work setting*
Ambulatory—primary care	94 (17)
Urgent care	6 (1)
Emergency room	6 (1)
Academic	11 (2)
Administrative	17 (3)
Other	11 (2)
Years in practice
1–5 years	17 (3)
6–10 years	11 (2)
11–20 years	33 (6)
21–30 years	17 (3)
31–40 years	11 (2)
41–50 years	11 (2)
Proportion of patients covered by Medi-Cal or Medicaid
Some	28 (5)
Most	67 (12)
All	6 (1)
California county of clinic location
San Bernardino	17 (3)
Los Angeles	61 (11)
Santa Barbara	11 (2)
Riverside	6 (1)
Kern	6 (1)
Geography of clinic location
Suburban	33 (6)
Urban, inner city	28 (5)
Urban, not inner city	39 (7)
Gender
Female	67 (12)
Male	33 (6)
Other	0
Race/ethnicity
Asian	27 (5)
Black/African American	17 (3)
White	44 (8)
Hispanic/Latinx	6 (1)
Multiracial	6 (1)
% time seeing patients clinically
0%–25%	11 (2)
26%–50%	22 (4)
51%–75%	22 (4)
76%–100%	39 (7)
Missing	6 (1)
Personal experience with ACEs
Yes	44 (8)
No	44 (8)
Prefer not to answer	11 (2)

ACEs indicates adverse childhood experiences.

^*Providers had the option to select more than one answer.

We identified 5 major themes related to the ACEs screening workflow (Table 2). These included the degree of workflow protocolization, screening tool used, timing of screening, clinic staff involvement, and interventions recommended and/or initiated by the physician.

Table 2.

Clinical workflow Themes From Interviews With Physicians Who Screened for ACEs With Quotes Illustrating Each Theme

Themes	Representative Quotes
ACEs Screening Workflows without a Protocolized Intervention	“There’s quite a limitation on services like child therapists and social workers…So it is not like you can just refer everybody that has an ACE greater than something for an assessment. I think that’s contributing to that case-by-case basis kind of situation.”
ACEs Screening Workflows with a Protocolized Intervention Threshold	“Whether [their ACE score is] a zero through three, our protocol is to always reflect that each additional ACE or ongoing adversity can increase the risk to physical and behavioral health, and therefore we want to try to connect the dots for the patient.”
Screening Tool(s) Used	“…The formal anonymous screening is easier for parents to admit that they are struggling, as opposed to me just saying, “Hey, like are you able to feed your kid every day?”
Timing of Screening	“We decided we would screen …patients between the ages of zero and five…hoping that we’d be likely to encounter fewer positive screens.”
Clinic Staff Involvement	“I don’t have to think about [screening], honestly. The receptionist just gives it. The patient fills it out. The nursing puts it into the computer. So, all I have to do is see the score. And then my only burden, of course, is to discuss it with the patient but at that point, all the screening is done already.”
Intervention	“The real opportunity is for that pediatrician to make a connection with [the] parent or whoever the caregiver is.”

ACEs indicates adverse childhood experiences.

Variation in ACEs Screening Workflow Protocolization

We found substantial variation in the degree to which participants’ ACEs screening and response workflows were protocolized. This variation hinged primarily on the interventions recommended by clinicians in response to the ACEs they identified (ie, referral to mental health providers, parent supports, social work, and/or other resources). Other variability was attributed to factors related to feasibility of administration, beliefs about patient risk, availability of support staff, and/or intervention resource accessibility.

The minority of participants reported that their ACEs screening and response workflows did not have a protocolized approach to intervention (Figure, Category 1). These workflows lacked both a set threshold for intervention and a method to modulate intervention intensity in response to identified ACEs or their consequences. Intervention decisions were left fully up to the discretion of the primary clinician without guidance from a protocol.

Figure.

Three categories of ACEs screening workflows were identified and are depicted. Categories were based on the degree of protocolization of ACEs screening. Category 1 depicts a workflow without protocolization. Category 2 depicts a protocolized approach to screening with the use of a predetermined threshold to determine which patients received an intervention to address their identified ACEs. Category 3 goes one step further, where physicians also protocolized modulation of the intensity of their interventions based on assessed patient risk levels.

The majority of participants reported that their clinical workflow processes had a protocolized approach based on a predetermined threshold to determine which patients received an intervention to address their identified ACEs (Figure, Category 2). Some participants’ ACEs screening and response process workflows went further to protocolize modulation of the intensity of their interventions based on assessed patient risk levels using information gathered during ACEs screening (Figure, Category 3).

A key difference across providers who used a protocolized approach was the criteria used in the clinical decision-making processes for risk assessment and referral for intervention services. While many providers used an ACE score threshold, there was no consensus regarding which ACE score level should trigger or modify the decision to intervene by the clinician. One provider felt that “any young child that has an ACE score of two would probably benefit from at least being offered the services.” Others used a threshold based on the presence of symptoms thought to be related to ACEs by participants, and some used a combination of ACE score and symptoms. Examples of those symptoms thought to be related to ACEs included recurrent viral infections, abnormal weight, difficult to control asthma or eczema behavioral health concerns, or somatic symptoms with no other identifiable etiology.

Three different types of workflow types based on their protocolization are described in more detail in the 3 sections below.

ACEs Screening Workflows Without a Protocolized Intervention

Less than a quarter of participants did not utilize a protocolized intervention threshold or a protocolized approach to modulating the intensity of the intervention, making intervention recommendations and resource referrals on a case-by-case basis. Many of these providers felt that a case-by-case basis model was the most effective approach for managing patients and their families. One physician explained, “…If a child is perfectly healthy… [and] their ACEs score is two, I’m not necessarily going to do anything different.” However, a few felt they were reduced to using an ad hoc approach despite wanting greater consistency due to limited access to intervention resources.

ACEs Screening Workflows With a Protocolized Intervention Threshold

More than three quarters of participants utilized a protocolized threshold for intervention. These thresholds were based on ACE score, the presence of symptoms thought to be related to ACE exposure, or both. About two thirds of that group of participants reported that they also modified the intensity of their clinical ACEs interventions based on their patient risk assessment.

More than half of providers who modified their intervention intensity in a protocolized way based on patient risk assessment used a combination of score and symptoms to do so. The remaining providers considered the presence of symptoms related to ACEs alone as sufficient criteria to escalate intervention intensity. For example, some classified patients as “low risk” when they had an ACE score of 1 to 3 without related symptoms, whereas a comparable number considered patients who fell within this score range “low risk” regardless of the presence of related symptoms. By comparison, the majority classified patients as “high risk” if symptoms were present, while the minority required at least a score of 4 or more for this classification. Notably, almost all providers in this category provided at least some intervention to any patient with an ACE score of 1 or more, and the presence of symptoms was used as a key step in the workflow protocol.

One physician who used the presence of symptoms alone as a threshold said, “…A lot of people have ACE scores, but they get through life…I’m a child of a holocaust survivor, and my mother was an alcoholic. But I don’t need counseling…Some people have a lot of resilience and…are doing well in school…[it’s] just…a problem to keep watching.”

Of those physicians who considered score alone as part of their threshold, equal proportions used a score of 1 or more, a score of 3 or more, or a more complicated algorithm for their protocolized intervention thresholds. One physician said, “…If [the score is] four or more, but the patient is not exhibiting current mental health symptoms, it’s highly likely they will at some point.”

Screening Tools Used

Roughly equal proportions of participants used either the Adverse Childhood Experiences Questionnaire or the Pediatric Adverse Childhood Experiences and Related Life-Events Screener, while a minority used either the Whole Child Assessment or a site-specific screening tool developed by that practice. Many participants cited the reimbursement incentive under the California ACEs Initiative as a primary reason for choosing to use Pediatric Adverse Childhood Experiences and Related Life-Events Screener. The majority of those who chose to use either the Whole Child Assessment or a site-specific tool did so primarily to decrease the time and paperwork burden of screening for patients and providers. Some of these physicians acknowledged the added benefit of avoiding singling out ACEs screening questions, which was believed to help to normalize these questions for patients and their families. One provider said, “When you have a completely separate questionnaire about ACEs…it says, this is a separate topic…and it continues that stigma [that] mental health is different and separate than physical health.”

Almost all providers used either a parent- or child-completed paper or electronic version of their ACEs questionnaire, rather than asking questions directly to parents, children, and/or caregivers.

Half of the participants used a de-identified screening tool, which refers to a tool that asks patients to report a total score without the need to endorse specific experiences. The most stated benefit for using a de-identified tool was that this might increase disclosure rates, while the most common drawback mentioned was that recommending interventions can be challenging without knowing specifics. One physician shared, “If I don’t…find out which [experiences] are the problems, I can’t actually help them very much.” Some physicians felt that this knowledge about specifics was not as important as knowing the overall ACE score. One shared, “Studies have shown that it’s really the number of ACEs that are more problematic.” Many physicians felt use of a de-identified tool was more time efficient, with one physician sharing, “I don’t have time to sit with everyone and discuss [their answers] every time.”

Timing of Screening

Almost all participants conducted screening at preventive care visits, and over a third added ACEs screening to any visit if relevant ACE-related disease symptoms were reported or clinical surveillance indicated that a complete ACEs assessment was needed.

Most participants conducted regular ACEs screening on patients at or after specific ages or within specific age ranges. Of those, more than half screened starting at a predetermined age and continued until the end of childhood. Most chose 6 months, 4 or 5 years, or 11 or 12 years, with a range of 4 months to 12 years. A third of participants only screened at discrete ages or within specific age range (s), with an average of 4 time points selected for a given pediatric patient. One common reason for selecting certain ages or age ranges to screen was the increased feasibility of incorporating ACEs screening into their practice at those visits, which allows for fewer patients overall that require screening and/or the ability to select particular ages with less to cover in the well-child visit. Others chose to screen certain age ranges because they felt patients might be less likely to screen positively. One provider shared that, in doing so, they could conduct ACEs screening “…without adding a huge level of work to [their] infrastructure, to [their] providers, and to [their] care teams.”

Few providers who included adolescents as part of their screening protocol made an effort to separate adolescents from caregivers during the information collection process.

Only a minority of providers also screened for parent/caregiver ACEs, although many physicians noted that screening for parent/caregiver ACEs was a future goal. One provider shared, “[Parental ACEs are] a very important predictor of child health problems. But it’s just one more thing.” Of those that did screen for parent/caregiver ACEs, most did so only if the patient met certain criteria (ie, age, ACE score).

Clinic Staff Involvement

Almost all providers involved ancillary staff (including front desk staff, medical assistants, and/or nurses) at some point during the screening workflow. The most common role for ancillary staff was in delivery of ACEs screening questions, almost all involving distribution of a paper screening tool. Nearly two thirds utilized the front desk staff to distribute their screening tool, while 28% (n = 5) had either the medical assistant or nurse do so. Only one practice had staff (a medical assistant) deliver the ACEs screening questions verbally to the parent/caregiver and/or patient. This practice first provided training to their staff such that they could respond to endorsed ACEs with a trauma-informed care approach. Almost half of physicians indicated that their practice workflow asked the medical assistant or nurse to enter the responses or scores into the electronic medical record.

One of the most commonly cited reasons for involving nonphysician staff was the improved efficiency of screening, which was thought to be a major factor in establishing and incorporating a successful screening workflow. One said, “…I would credit [our nurses and administrators] so that… [screening is] done for us so we can do our jobs.” The minority felt that involvement of non-physician staff helped to improve disclosure rates, which some attributed to the delivery of an introductory script by the staff prior to distributing the screening tool.

Intervention

All providers utilized referral to either social workers or mental health providers as part of their ACEs screening and response workflow. Thirty-nine percent (n = 7) of providers had social workers, mental health providers, or other behavioral health staff readily accessible to patients in clinic, and many of those providers felt that this accessibility was important in better connecting patients to resources. Most providers incorporated referrals to community resources into their workflow model interventions, and one third of providers involved a case worker, care coordinator, or community health worker to facilitate connections to these resources. Others were only able to provide patients and/or families with contact information rather than direct referrals. Thirty-nine percent (n = 7) utilized a handout detailing information about local resources as part of their intervention practices, though one provider shared, “It’s just disheartening to feel that you’re saying, ‘You have increased risk for depression and suicide, and here’s a packet.’ That doesn’t feel right.” A few of the providers felt that access to mental health and community resources for their patients was limited. Those providers were more often directly involved in the intervention process, relying heavily on counseling and close follow-up.

Though we did not include counseling as an intervention when categorizing workflow models, most providers regularly provided counseling regarding ACEs. Counseling involved education surrounding ACEs screening but could also involve discussion of protective factors and resilience strategies. While some physicians only provided counseling if the patient screened “positive,” others provided at least some form of education to any patient screened. One shared, “I’m not professionally trained to do the counseling…but people still depend a lot on what the doctor says…You can say small things like this is not your fault…Otherwise, why are you doing this?” Of those providers who provided little to no counseling, limited time available was cited as a barrier.

Discussion

In this first study to examine commonalities and variation across clinical ACEs screening and response workflow processes used by physicians in multiple practices, as well as barriers, facilitators, and key implementation factors, we found substantial variation in implemented clinical process workflows. This variation included screening timing and tool used, involvement of different clinic staff, types of interventions used, and degree of protocolization for intervention and modulation in response to identified ACEs. Three basic types of ACEs screening and response process workflows were identified based on these differences.

Despite the many similarities across participants, there was a lack of consensus regarding best screening and response practices. Additionally, while providers could identify when their practices were using protocols for some or all of the ACEs screening and response workflows, there was still considerable clinical discretion applied by providers when protocols were not in place or to adapt to situations in which the protocols were poorly-suited. This lack of consensus and high degree of individual- and system-level practice variation poses a major challenge in developing a standard of practice for screening, but also in identifying individual best practices that are most suitable for a particular clinical environment. This study may serve not only to provide some examples of successful screening protocols, but to demonstrate that there may not be a single best practice standard without additional evidence on outcomes.

Policies like reimbursement and medical professional society position statements may not drive ACEs screening if the mechanics of screening workflows or the resource landscape remain as uneven as we found in this study. Even if providers are incentivized to perform ACEs screening, guidance for screening and infrastructure to facilitate successful integration of ACEs screening into a clinical practice will need to balance best practices for patients with feasibility of implementation. More evidence is needed on screening workflow effectiveness to guide clinicians toward ACEs screening and response best practices.

Limitations

Though the sample was recruited from a large and diverse area and chapter in Southern California, the state where ACEs screening is most supported through reimbursement, generalizability of our study may be limited by sample size, geography, and overrepresentation of urban and suburban pediatricians. Our study was not designed to assess which workflows were most efficient or effective, or the best fit between workflows and practice characteristics.

Conclusion

ACEs screening workflow variation in practice was largely driven by differences in intervention thresholds and access to ACEs intervention resources. This variability demonstrates that workflows can be tailored to the needs and limitations of an individual practice. However, even if greater uniformity in screening is achieved, a single standardized screening protocol may not be universally feasible or effective across practices. Rather, there is a need for evidence on how ACEs screening can be adapted to different types of practices and resource landscapes.