Study population.

All 51 intervention arm families were included in the analysis (Table 1). Mean child age was 46.5 months; 78.4% of families identified as Latine; 74.5% of parents reported speaking English “Very well.” Over 70% of families reported yearly income less than $40,000 in 2021; 98% of families were insured through Medicaid, with a small percentage being uninsured.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of sampled children and parents (n = 51).

https://doi.org/10.1371/journal.pone.0356066.t001

Mean ACE scores.

The mean ACE score from PEARLS Part 1 was 1.35 (95% CI 1.20–1.51) in clinic screening and 2.08 (95% CI 1.64–2.51) with 211LA screening; the mean score from PEARLS Part 2 was 0.49 (95% CI 0.27–0.71) in clinic screening and 1.12 (95% CI 0.75–1.48) with 211LA screening (Fig 1). In a paired t-test, PEARLS PART 1 ACE scores in 211LA screening were significantly higher than in clinic screening (difference = 0.73, 95% CI: 0.29–1.16, p = 0.002).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Paired t-test testing agreement in PEARLS scores within individuals between clinic and 211LA.

Mean PEARLS Part I and Part 2 scores are shown for the clinic administered ACEs screening in orange and the 211LA ACEs screening in blue with 95% confidence intervals. P-values result from paired t-tests testing whether scores differed within individuals across these settings.

https://doi.org/10.1371/journal.pone.0356066.g001

Qualitative interviews.

Thematic analysis of parents’ answers on perceived differences in ACE screening between 211LA and clinic revealed several key themes. Parents attributed the difference in ACE reporting to greater convenience, anonymity, privacy, focus, and the prospect of getting help from 211LA (Table 2).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Themes and illustrative quotes of perceived differences between 211LA and clinic screening.

https://doi.org/10.1371/journal.pone.0356066.t002

Parents felt screening in clinic was less comfortable for several reasons: needing to provide sensitive information in a non-private setting, being distracted by their children, and needing to answer other questionnaires given by the clinic. Meanwhile, screening over the phone afforded the convenience of flexibility, as parents could postpone screening and ask to be called back later. It also allowed greater anonymity and privacy, as parents could take the call from a private location and discuss sensitive issues without seeing someone face-to-face. Completing it over the phone also allowed better focus, as they didn’t have to simultaneously keep their children entertained in the clinic. Parents also felt more likely to receive help through referrals from 211LA. Some parents felt “more embracing of [211LA]” because they saw 211LA as a “follow-up,” going “hand-in-hand” with the clinic visit.

Although most parents found screening tolerable, helpful, and worthwhile regardless of modality, several parents cited fear of being reported to Child Protective Services or having their children taken away as reasons not to report to either 211LA or the clinic.

Factors associated with a greater difference in 211LA ACE scores versus clinic scores.

In linear regressions, greater trust in providers (β = 0.64, 95% CI: 0.01–1.27, p = 0.049), more parental stress (β = 0.06, 95% CI: 0.00–0.12, p = 0.028), and speaking English “Well” versus “Very well” (β = 1.85, 95% CI: 0.11–3.59, p = 0.043) were associated with higher 211LA scores relative to clinic PEARLS Part I ACE scores (Table 3). Other factors that were tested but showed no significant relationship included financial wellbeing, income in 2021, at least one parent born outside the US, at least one parent with a high school education or greater, Latine ethnicity, child age in months, and child gender.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Unadjusted association of family contextual factors with difference between PEARLS Part I ACE scores via 211LA versus clinic screening (n = 51)*.

https://doi.org/10.1371/journal.pone.0356066.t003

Parents reported more ACEs to 211LA than to clinics.

The a priori hypothesis for our primary aim was that parents would be more likely to report ACEs (Part 1 of the PEARLS) and social determinants of health (Part 2 of the PEARLS) to their clinicians than to 211LA care coordinators, based on assumed trust between parent and provider; thus, the finding that ACE scores were significantly higher in 211LA screening relative to clinic screening was somewhat surprising. This result suggests that screening context may influence ACE reporting or that some screening modalities may result in more positive screens. With respect to modality, a study of population-based surveys suggested that telephone ACE screening may increase ACE reporting compared to self-completed web-based screening, which the authors attributed to trust developed through verbal communication [12]. Another study, however, showed more positive PHQ-2 screens with self-completed paper screening compared to verbal screening [20]. Most studies on screening of sensitive topics show no clear difference between modalities [18,19,21].

The finding that phone screening results in higher ACE scores compared to self-completed paper surveys in clinics is novel to the literature. Existing literature suggests that de-identified in-person ACE screening leads to more positive responses on both Part 1 and Part 2 of the PEARLs screener than identified in-person ACE screening [17], but prior studies have not examined how phone screening affects reporting. Evaluations of screening of sensitive topics over the phone are scant: multiple studies have found smartphone apps effective at screening for depression, but no studies exist on verbal depression screening over the phone [29–31]. Multiple studies have also examined telemedicine interventions for patients with known depression or intimate partner violence (IPV), but not telemedicine screening [32–34].

There were many differences between 211LA and clinic screening, including modality (telephone vs. in-person, verbal vs. written, administered vs. self-completed, identified vs. de-identified), location (at-home vs. in-clinic), agenda (screening for and responding only to ACEs vs. addressing multiple clinical issues), and timing (211LA screening happened after clinic screening). It is unclear which aspects of 211LA screening contributed to the difference in reporting, but parents reported clear differences between 211LA and clinic screening in qualitative interviews.

Better convenience, anonymity, privacy, and focus are all reasonable explanations for a difference in reporting between 211LA and clinic screening. However, these factors may be more inherent to phone screening generally than to 211LA. Thus, it is plausible that clinics could offer phone screenings to increase access to ACE screening for underserved patients without 211LA involvement. On the other hand, using phone screening without involving 211LA would not provide the final parent-endorsed advantage of 211LA care navigation for referrals and follow-up, which in initial analyses resulted in greater receipt of services about four months after baseline. [22]

It is not immediately clear why some parents felt the prospect of getting help motivated them to report more through 211LA than in clinic, given that both 211LA and FQHCs can refer families to appropriate services. One possible explanation is that the purpose of screening was clearer with 211LA than with the clinic. Ideally, any ACE screening starts by explaining that a core purpose of screening is to provide help, including via referral to services. It is unclear how often this was explained during busy FQHC visits with multiple objectives. In contrast, the process of enrollment and warm hand-off from the study team to 211LA care coordinators was standardized and presented to parents as focused on getting appropriate help.

Factors associated with larger differences between 211LA and clinic ACE scores.

Higher parent stress was associated with greater reporting to 211LA. Parent stress may reflect a greater need for supportive interventions, perhaps making them more willing to report ACEs in general. However, it is unclear why greater parent stress might motivate parents to report specifically to 211LA over their clinic. Speculatively, perhaps the potentiating effect of higher parent stress on reporting to 211LA is mediated by the convenience, anonymity, and other benefits of 211LA screening. Surprisingly, other factors that we considered proxies for “level of need” such as income and financial wellbeing were not associated with higher reporting to 211LA. That parent stress was associated with reporting while income was not might indicate that a family’s perceived need for services is less dependent on available resources than on parents’ ability to cope with stress related to ACEs.

It is also unclear why parents with greater trust in their healthcare providers would report more ACEs to 211LA than in clinic. We had theorized that trust in providers would increase reporting in clinics, rather than to 211LA. Of note, clinic screening was generally done via self-completed survey and not necessarily face-to-face with a clinician, so rapport and trust may have played less of a role. Speculatively, one possibility is that parents saw 211LA as being part of the clinic follow-up, so their trust toward providers was also transferred to 211LA care coordinators. Several parents noted that 211LA felt like a follow-up to the in-clinic survey. Thus, parents with greater trust in their clinicians may have reported more during “follow-up” with 211LA. The effect size was large enough to be clinically significant – a one-unit increase of a parent’s trust in providers on any one of three questions was associated with a 0.64-point difference in reported ACE score between 211LA and clinic. For example, a parent who answered “Strongly Agree” instead of “Agree” to a statement like “I feel like staff here understand me” could report one additional ACE, which might alter the clinical response and eligibility for other services.

English proficiency was not associated with any difference in ACE reporting except for between speaking English “Very well” and “Well”. One explanation is that 211LA offered Spanish-speaking coordinators to all patients who spoke it, while in-clinic interpretation may often be reserved only for patients who speak little English, leaving patients who speak English “well” but not “very well” in a language doughnut hole. Communicating in one’s preferred language may be particularly important when discussing sensitive topics such as ACEs.

Together, these associations suggest that screening modality and context might be important for ACEs identification, and that the “best” screening approach may vary by patient characteristics.

One major limitation of this study was that all participants received 211LA screening just two weeks after clinic screening and that only those reporting some ACE exposure during the clinic screening participated in this study. Indeed, referral to 211LA was intended to support mitigation of toxic stress that is known to occur in the setting of a positive ACE screen and is thought to be essential for potentially achieving clinical benefit from ACE screening. Reporting could have been higher on the second screening due to greater recall when prompted again or greater familiarity or comfort with the questionnaire items. The latter seems especially plausible, as invasive questions on sensitive issues may be more palatable after repeated exposure. This is in line with a study on postpartum depression that found an increase in positive screens when repeating screening 6 months later, although this may be attributed to the natural progression of postpartum depression itself [35].

Another limitation is the small sample size of this pilot (n = 51). We were unable to perform multivariable analysis, making the results subject to confounders. Thus, it is best to consider these results hypothesis-generating only rather than confirmatory, warranting further study of the effect of telephone screening and screening context on ACE reporting. Finally, participants were drawn from 3 Federally Qualified Health Centers in Los Angeles. Although the demographic characteristics are consistent with low-income families impacted by ACEs in Southern California, results may not generalize to other populations.

Significance.

Our main exploratory finding suggests that screening for ACEs over the phone outside of a clinic visit may be effective at identifying ACE exposure—with the ultimate goal of prompting an effective response to a positive screening result. If confirmed in larger trials, this finding opens potential opportunities to build and test systems for ACE identification and management that only minimally burden busy primary care clinics. Telephone screening and response activities may require fewer resources and personnel and be easier to scale up than a clinic-based screening program.

Such systems may also have the advantage of reaching more vulnerable patient populations, which may be less likely to present for preventative healthcare [3,36] Telephone screening may be especially helpful for families with difficulty accessing care due to cost, insurance, scheduling, or transportation issues. This is an important consideration for those who continue to investigate optimal strategies for prevention of, identification of, and responses to ACEs and toxic stress.