Relationships between psychological trauma, resilience, and professional quality of life in personnel of an HIV Service Organization in the Southeast United States

Abstract

Introduction:

High rates of burnout have been observed among HIV care personnel, which has been connected to reductions in professional quality of life and patient care quality. However, limited literature explores how personnel mental wellbeing might influence professional quality of life, including burnout.

Methods:

We assess the relationship between trauma and adversity, psychological resilience, and Professional Quality of Life (PQoL), including burnout, compassion fatigue, and compassion satisfaction, with 114 personnel in a community-based HIV Service Organization (HSO) in urban Tennessee between October 2017 and November 2019. Firth’s Logistic regression models were used to estimate the association between childhood adversity, adult trauma effects, and professional quality of life with resilience as a moderator, adjusting for gender, race, age, professional role, and adversity scores.

Results:

Higher resilience and trauma scores and older age were associated with lower odds of burnout and compassion fatigue, with some differences by gender. While the odds of compassion satisfaction increased as resilience increased, there was insufficient evidence to conclude an interaction effect between trauma and resilience for each of the PQoL outcomes.

Discussion:

Future research should explore resilience as a multi-dimensional construct to assess if factors such as organizational trauma resilience influence professional quality of life among HSO personnel.

Introduction

Professional quality of life (PQoL) is an umbrella term generally comprising compassion satisfaction, compassion fatigue, and burnout among individuals working in human service care environments. Compassion satisfaction is fulfillment derived by work from care personnel; compassion fatigue is the expected cost of caring for patients, in which fatigue ebbs and flows; burnout, sometimes referred to as Secondary Traumatic Stress (STS), is emotional or physical exhaustion resulting from work, generally worsening over time, producing reductions in occupational satisfaction and performance (Figley & McCubbin, 2016). Healthcare personnel who work directly with patients are at risk for both compassion fatigue and burnout, with the latter associated with chronic emotional and interpersonal stressors (Friganović et al., 2019). Consequences of burnout can be severe, affecting personnel’s psychological states (e.g., depression, suicidality) (Sales et al., 2019; Shanafelt et al., 2011) and behavioral practices (e.g., aggression toward patients, mistakes in treatment, avoiding diagnostic tests to save time) as well as patient experiences (e.g., suboptimal care satisfaction and quality, increased patient safety incidents) (Salyers et al., 2017).

A robust literature indicates contributors to burnout include occupational environmental factors like intensity of role expectations and psychological demands. This includes occupations where providers endure greater psychological demands such as working with patients with high rates of Post-Traumatic Stress Disorder (PTSD) (van Dernoot Lipsky, 2010). Given that people with HIV (PWH) endure disproportionately high rates of PTSD (29-34%), compared with the public (3.9% rate), there are higher rates of burnout among HIV care personnel versus other care personnel (Tang et al., 2020; van Dernoot Lipsky, 2010) – up to 77% of HIV care personnel report experiencing burnout (Kim et al., 2018; Qiao et al., 2016). Other factors associated with increased risk for burnout include environmental cultures where work autonomy and decision latitude are limited, especially when coupled with moral distress, role conflict, poor social climate/team relationships (Dall’Ora et al., 2020; Friganović et al., 2019; LeClaire et al., 2022), and psychological components among providers. Notably, primary trauma, including childhood abuse, has been connected with adult burnout; compared with personnel in non-direct care roles, frontline social service workers reported higher rates of primary trauma (e.g., nearly a third) (Buchanan et al., 2006). Provider resilience and emotion regulation coping have been shown to moderate the relationship between primary trauma and burnout (Ben-Zur & Michael, 2007); however, limited literature explores how psychological wellbeing influences burnout among HIV care personnel, and if so then what methods might enhance resilience among this population.

As HIV care providers routinely operate in high trauma burden care spaces (Brown et al., 2024; Tang et al., 2020), more interventions are needed to improve wellness outcomes among patients and providers in these environments. In the current study, we explore the relationship between the three sub-dimensions of PQoL among HIV care providers in an HIV Service Organization (HSO) in the urban southeastern United States, as this area is disproportionately burdened by the HIV epidemic (Centers for Disease Control and Prevention (CDC), 2019). Specifically, we sought to explore associations between trauma exposures and effects, psychological resilience, and PQoL, and to determine if psychological resilience acts as a moderator. We hypothesized there would be a positive association between trauma and PQoL, with resilience moderating the relationship.

Materials & methods

Study population

We collected cross-sectional data between 2017 and 2019, as part of an on-going longitudinal study with HSO personnel in an urban area of Tennessee. All employees, regular volunteers, and board of directors were invited to participate via an email, which included a link to a REDCap survey and a self-administered informed consent document, requiring all individuals had signed before proceeding to the survey (Harris et al., 2009). All participants completed the self-administered informed consent document by providing a signature prior to initiating the survey. Participants then completed a self-administered survey and were compensated for participation in accordance with the participating organization’s policy, in which employees could receive time off work as compensation. Ethical approval was obtained for this study from Vanderbilt University Review Board (IRB, 171596) in accordance with the principles stated in the Declaration of Helsinki.

Study definitions

Psychological trauma exposure and effects and resilience

Psychological trauma was the primary independent variable and was measured using the Adverse Childhood Experiences (ACE) questionnaire (Felitti et al., 2019). For descriptive statistics ACE scores were dichotomized based on participants’ responses to each question, categorized as “Never” / “No” as “No” (0) versus “Yes”/ “Once” / “More than once” were combined as “Yes” (1). ACE scores are based on the sum of all 11 questions ranging from 0 to 11, with higher scores indicating higher levels of childhood adversity. Responses of “Decline to answer” and “Don’t know/ Not sure” for the ACE and all other questionnaires were treated as missing and singly imputed (McElroy & Hevey, 2014). For analyses, ACE scores were applied as continuous scores. Trauma effects were determined by computing the total of 10 items from the PTSD Checklist 5 (PCL-5) (Weathers et al., 2013) using Likert scoring: “not at all” = 1, “a little bit” = 2, “somewhat” = 3, “quite a bit” = 4, and “totally” = 5, with 10–50 as the potential scoring range, and higher scores indicating higher levels of trauma effects. Resilience was a composite score based on 10 items from the Multidimensional Trauma Recovery and Resiliency Scale (Harvey et al., 2003) using Likert scoring described above, with 10–50 as the potential score range and higher scores indicating higher levels of resilience.

Professional quality of life

We used the 30-item Professional Quality of Life (PQoL) IV scale (Stamm, 2009) to estimate levels of compassion satisfaction, compassion fatigue, and burnout. Total score was computed for each sub-scale using Likert-scoring described above, and the responses were dichotomized based on medians for analysis for ease of interpretation. The outcome variables compassion satisfaction, compassion fatigue, and burnout were dichotomized as greater than vs. less than or equal to their medians and contextually compared with average scores: compassion satisfaction was 37 (interquartile range = 33, 42); compassion fatigue was 13 (interquartile range = 8, 17); and Burnout was 22 (interquartile range = 18, 27) (Stamm, 2009). Analyses were repeated with outcomes kept as continuous.

Demographic items assessed included race (White, Black, Asian, Multi-racial, Other); Age; Gender (cisgender female, cisgender male, transgender female, transgender male, Other); and Role in organization (regular volunteers, paid employees). Missing demographic variables were singly imputed in analyses.

Statistical analysis

To explore predictors and moderators of compassion satisfaction, compassion fatigue, and burnout among HIV care providers in an HSO, we had two primary research questions and models (Figure 1 presents the proposed pathway for the moderation analysis):

Figure 1.

Conceptual diagram.

1. Research Question 1 (RQ1): Are there associations between childhood trauma exposures, adult trauma effects, psychological resilience, and PQoL dimensions (including compassion satisfaction, compassion fatigue, and burnout), notwithstanding the influences of gender, race, age, and work role?

2. Research Question 2 (RQ2): In what ways does psychological resilience moderate the relationship between adult trauma effects on PQoL dimensions, including compassion satisfaction, compassion fatigue, and burnout?

We hypothesized there would be a positive association between trauma (combined exposures + effects) and compassion fatigue and burnout; a positive association between compassion satisfaction and resilience; and resilience would moderate the relationship between trauma and each sub-dimension of PQoL.

For RQ1, Firth’s Logistic regression models were used to estimate the association between trauma (ACE and PCL-5 scores) and resilience scores on the dichotomized outcome variables of PQoL sub-scores, adjusting for gender, race, age, and role at the HSO. Firth’s correction (Firth, 1993) for logistic regression was used for handling categories (e.g., race, role, and gender) when sub-samples were small (n < 5). Analyses were repeated using the original, continuous Likert-scoring scales for the PQoL scale sub-scores; a semiparametric cumulative probability model was fit to these scores (Liu et al., 2017). For the moderation analysis (RQ2), an interaction term for trauma and resilience scores were used to estimate the moderation effect of resilience on the relationship between trauma and all 3 PQoL outcome variables (compassion satisfaction, compassion fatigue, and burnout).

Results

Descriptive statistics

There were 114 participants: 29% were Black and 64% were cisgender female. The median age was 38 years (interquartile range [IQR] of 28-53). Organizational roles included paid employees (89%) and regular volunteers (11%). See Table 1.

Table 1.

Descriptive statistics.

Variables	Survey participant results (N = 114)
Race
White	67% (73)
Black	29% (32)
Asian/Multi-racial/Other	5% (6)
Role
Paid employee	89% (101)
Regular volunteer	11% (12)
Gender identity
Cisgender Female	64% (72)
Cisgender Male	36% (41)
Age at survey (Median, IQR)	38 (29–53) (42 ± 14)
ACE score (Median, IQR)	3.5 (2.0–6.0) (3.9 ± 2.8)
Trauma score (PCL-5; Median, IQR)	43 (38–46) (41.4 ± 6.2)
Resilience score (Median, IQR)	40 (33–42) (38 ± 7)
Compassion Satisfaction (Median, IQR)	42 (35–46) (39.9 ± 7.7)
Compassion Satisfaction Yes/No
Yes	46% (53)
No	54% (61)
Compassion Fatigue (Median, IQR)	11 (6–18) (13.0 ± 8.9)
Compassion Fatigue Yes/No
Yes	45% (51)
No	55% (63)
Burnout (Median, IQR)	20 (15–25) (20.2 ± 7.5)
Burnout Yes/No
Yes	47% (54)
No	53% (60)
ACEs score ≥ 4
Yes	38% (43)
No	62% (71)
ACEs score ≥ 3
Yes	50% (57)
No	50% (57)

Note: Results present median (interquartile range of 25th percentile and 75th percentile) (mean ± standard deviation); N is the total number of participants; numbers after percentages are frequencies; ACE is Adverse Childhood Experiences questionnaire and PTSD Checklist 5 (PCL-5).

The median ACE score was 3.5 (IQR of 2-6), with 38% having greater than or equal to 4 ACE score, and 50% having ACE score greater than or equal to 3 (See Supplementary Table 1). The median trauma score was 43 (IQR 38-46), and the median resilience score was 40 (IQR 33-42). Trauma effects and resilience responses are reported in full in Supplementary Table 2. Notably, resilience items showed 95% had stress management techniques and ~93% were frequently able to maintain healthy self-care; trauma effects items showed 24% reported some drug or alcohol “abuse”; 48% often felt jumpy or easily startled, and three-quarters (72%) had difficulty with sleep. The median score for compassion satisfaction was 42 (IQR 35 46), for compassion fatigue 11 (IQR6 18), and for burnout 20 (IQR 15 25). Missingness for demographic variables was low, with 5% missing age, 5% missing race, < 1% missing gender, and <1% missing role.

Research question 1: professional quality of life

A higher resilience score (adjusted odds ratio [aOR] per 1-unit increase in scale 0.86; 95% confidence interval [CI] 0.78, 0.94) and older age (aOR per 1-year increase 0.96; 95% CI 0.92, 1.00) were associated with lower odds of burnout. Increased trauma (aOR per 1-unit increase in scale 0.82; 95% CI 0.72, 0.92), increased resilience (aOR per 1-unit increase in scale 0.90; 95% CI 0.82-0.98), and older age (aOR per 1-year increase 0.92; 95% CI 0.87, 0.96) were associated with lower odds of compassion fatigue. Men experienced less compassion fatigue compared to women (aOR 0.24; 95% CI 0.07, 0.68). As resilience increased, the odds of compassion satisfaction increased (aOR 1.09 per 1-unit increase in scale; 95% CI 1.02, 1.18). See Table 2. Results were similar when PQoL was analyzed as original continuous scales, although gender identity was no longer statistically associated with compassion fatigue. Regular volunteers had lower burnout and compassion fatigue than paid employees, and Black personnel had higher compassion satisfaction than White personnel (Supplementary Table 3).

Table 2.

Adjusted Odds Ratios (aOR) for ProQol sub scales using Firth’s logistic regression.

Variables	Burnout aOR (95%CI)	Compassion Fatigue aOR (95%CI)	Compassion Satisfaction aOR (95%CI)
ACE (per unit)	1.01 (0.85–1.20)	0.98 (0.82–1.17)	1.08 (0.91–1.31)
Trauma (per unit)	0.91 (0.82–1.01)	0.82 (0.72–0.92)	0.99 (0.90–1.09)
Resilience (per unit)	0.86 (0.78–0.94)	0.90 (0.82–0.98)	1.14 (1.05–1.25)
Age (per year)	0.96 (0.92–1.00)	0.92 (0.87–0.96)	1.04 (1.00–1.08)
Male (vs. Female reference)	1.80 (0.70–4.81)	0.24 (0.07–0.68)	1.10 (0.40–3.20)
Race
White (reference)	1	1	1
Black	1.28 (0.47–3.59)	1.06 (0.37–3.1)	0.32 (0.08–1.01)
Asian Pacific Islander/Other	1.74 (0.16–27.1)	0.36 (0.04–3.08)	0.91 (0.09–13.32)
Regular Volunteer (vs. Paid Employee reference)	0.19 (0.02–1.12)	0.34 (0.03–1.89)	0.76 (0.16–4.58)

Note: Burnout, Compassion Fatigue, and Compassion Satisfaction are dichotomized based on above (“yes”) versus below or equal to (“no”) median scores; Scores for ACEs, trauma, resilience, and age are continuous.

Research question 2: moderation effect

The moderation effect of resilience on the relationship between trauma and dichotomized PQoL outcomes was assessed using logistic regression. No significant interaction effect was observed between trauma and resilience for each of the PQoL outcomes (p > 0.05). Lower levels of trauma remained a good predictor of higher compassion satisfaction, regardless of resilience level, while there was little statistical evidence of an association between trauma and burnout and between trauma and compassion fatigue at all resilience levels. See Table 3.

Table 3.

Moderation analysis: the association between trauma and ProQol sub-scales for different values of resilience.

	Burnout	Compassion Fatigue	Compassion Satisfaction
p-value, test for interaction between trauma and resilience	0.64	0.64	0.05
Resilience value	aOR for trauma (per unit increase) (95% CI)
35	0.91 (0.77–1.01)	0.81 (0.68–0.93)	1.10 (0.99–1.27)
40	0.92 (0.83–1.02)	0.82 (0.72–0.93)	1.03 (0.93–1.13)
45	0.94(0.81–1.09)	0.84 (0.73–0.97)	0.96 (0.84–1.09)

Note: Resilience and trauma scores are continuous variables; a range of resilience value scores were applied to assess an interaction effect, with 40 as the median resilience value.

Discussion

Limited research focuses on how psychological trauma and resilience influence the PQoL of HIV care personnel. Findings from the current study uncover prevalence rates of trauma (e.g., childhood exposure and adult effects) and psychological resilience in a southern community-based organization and indicate a negative relationship between resilience and burnout, and a positive relationship between resilience and compassion satisfaction. Trauma results suggest personnel had high rates of trauma exposures and effects. To our knowledge, these are the first results showing associations between primary trauma and PQoL among HIV care personnel. Yet, there was no evidence to support that psychological resilience moderated the relationship between trauma and professional quality of life.

Observed composite ACE scores were higher than national estimates. Where we found 50% had experienced three or more ACEs, only 22% of Americans endorsed these levels of ACEs (Giano et al., 2020). Other samples, of social workers (Steen et al., 2020) and child welfare workers, reported 24%–25% had four or more ACEs (Howard et al., 2015), but we found 38%. However, responses by items were similar to a study of substance use counselors (Bride et al., 2009): we found a 39% physical abuse and 27% sexual abuse rate; they found 34% and 30%, respectively.

Trauma response rates were relatively difficult to compare, given the lack of extant data on this topic (Forkus et al., 2023). Comparisons of trauma response rates show a sample of substance use counselors (Bride et al., 2009) reported markedly lower rates than our study. While they found 15% felt they were constantly reliving events, we found 28%; while they reported 24% had difficulty concentrating, we found 70%; while they found 30% reported irritability, we found 73%; while they found 26% reported sleep disturbances, we found 72%; finally, while they found 30% experienced emotional numbing, we found 47%. Our results suggest the HIV care personnel experienced high rates of trauma effects and underscore the importance of adequate trauma services for HIV care personnel.

Scores for PQoL were compared with average PQoL scores from a seminal study among general health care personnel (e.g., clinicians and administrators), child and family workers (e.g., residential and child protective services), and school personnel (e.g., teachers, counselors, and administrators). While we observed a median for compassion satisfaction of 42, with an IQR of 35 and 46, the aforementioned study observed an average of 37, with an IQR of 33 and 42; hence, the 50th percentile in our study aligned with only the top 25th percentile in the seminal study, indicating we observed very high levels of compassion satisfaction. For compassion fatigue, we observed a median score of 11, with an IQR of 6 and 18, which was also lower than the average observed in the seminal study (13, IQR of 8 and 17). Finally, for burnout, we observed a median of 22, IQR of 18 and 27, which is lower than the average of 20 observed elsewhere, with an IQR of 15 and 25. While our observed burnout rates (47%) appear slightly lower than that observed among some physicians (60-67%) (Lacy & Chan, 2018; Rotenstein et al., 2018) general healthcare workers, children and family workers, and teachers (Stamm, 2009) but higher than other samples of providers from palliative care (17%) or even another southern-based HSO (which found 37% met the criteria for secondary traumatic stress) (Ali et al., 2023; Parola et al., 2017).

The PQoL scale sub-dimensions were shown to have several significant relationships with both psychological trauma and resilience, with age and gender presenting as covariates. Observed relationships between trauma, resilience, and PQoL present mixed findings compared with other literature (Howard et al., 2015; Padmanabhanunni, 2020). Similar to Howard et al. (2015) we found that as resilience scores increased, the odds of compassion satisfaction increased, and the odds of reporting burnout decreased. However, we observed that this was also true when trauma scores were higher. While this finding is not entirely clear, there is literature indicating that exposure to childhood trauma has been connected to increased empathy in adulthood (Greenberg et al., 2018). Thus, it is possible that the higher levels of trauma exposure contributed a greater sense of empathy or sensitization that acted as a buffer to compassion fatigue.

Contrary to our hypothesis, psychological resilience was not found to be a moderator of the relationship between trauma and professional quality of life. This finding is inconsistent with literature showing resilience moderated attributes of burnout (e.g., emotional exhaustion and cynicism) and psychological health overall (Chen et al., 2022). However, other studies have shown that psychological resilience acts as a mediator but not a moderator (García-Izquierdo et al., 2018), that it mediated the pathway between ACEs and PQoL (Milne et al., 2024), and acted as a robust primary predictor of PQoL (Chung & Chun, 2015; Hegney et al., 2015; Wong et al., 2022). However, findings from a study of lay counselors, showing organizational – and family – level resilience predicted burnout (Padmanabhanunni, 2020), suggest the utility of broadening definitions of resilience beyond a psychological dimension. For example, recent studies show the importance of assessing resilience using multi-dimensions along the socio-ecological health context (Dulin et al., 2018) rather than simply as a unidimensional concept that rests solely upon psychological attributes, as persons may possess high levels of psychological grit and fortitude but still carry a heavy allostatic load based on other factors in their environment, including lack of social support, quality care at the interpersonal and organizational levels, poor access to services at the neighborhood level, etc. Hence, our finding likely reflects the narrowness with which resilience was measured only at the psychological level. Alternatively, this finding could be specific to our sample. Both scenarios should be explored in future research to test these hypotheses.

The observed relationships between ACEs, adult trauma effects, resilience, and PQoL underscore the importance of trauma-informed workspaces in HIV care organizations and highlight the necessity for structured training programs for care providers. Findings also reinforce the urgency for healthcare systems to institutionalize trauma-informed care frameworks that address both systemic and individual-level resilience. Despite this need, a recent study with HSOs in the Southern U.S. shows less than half of sampled HSOs reported offering any trauma awareness training for personnel, and only 37% of those sampled receive training on vicarious trauma as an occupational hazard (Ali et al., 2023). Personnel require institutional resources to process how trauma exposure in the work environment impacts them personally as well as resources to process their own primary trauma, given that occupational trauma exposure can activate primary trauma effects. This would ideally involve both individual acknowledgement by personnel to work through past adverse experiences but also environments where this trauma awareness is realized and destigmatized and resources are provided. These findings highlight the strong potential for including personnel wellness components as part of trauma-informed interventions (Substance Abuse and Mental Health Services Administration, 2014) to be implemented throughout healthcare systems, as part of these initiatives involves discussing patient and personnel trauma exposure, resilience, and PQoL. Some mitigation tactics might include provision of onsite mental health support for personnel and formal opportunities for self-care and support (e.g., administrators offering focused solutions like self-awareness and stress reduction techniques and small-group programs that promote community and connectivity among colleagues) (Quenot et al., 2012; Sales et al., 2019). These positive self-care strategies can buffer trauma exposure response, reducing emotional exhaustion and depersonalization while enhancing personal accomplishment, spiritual and physical well-being, and positive occupational coping mechanisms to deal with workplace stress.

Future research should explore longitudinal effects of burnout and impacts of multi-level interventions on trauma effects (i.e., to investigate if there are pre/ post changes relative to symptoms such as sleep disturbances, hypervigilance, feeling trauma activation/ reliving experiences, etc.). Overall, there are a lack of rigorous empirical studies examining the connection between personnel burnout and patient outcomes, and research should develop and test causal models to better understand which domains of patient care are influenced by provider burnout (Kim et al., 2018; Rathert et al., 2018; West et al., 2018). Finally, given the potential discomfort in personnel responding to items from sensitive questionnaires, investigators conducting future research with personnel on the topic might consider using alternative questionnaires such as Positive Childhood Experiences (Bethell et al., 2019) (rather than ACEs) and assessing for more than just psychological resilience (e.g., a multidimensional resilience tool) (Harvey et al., 2003). Overall, there is need for more rigorous research on the predictors, consequences, and moderators of burnout in HIV care settings, and community-engaged research is needed to include healthcare personnel and people with lived experience to collaboratively develop and implement effective solutions (Bethell et al., 2019; Rathert et al., 2018; West et al., 2018).

Limitations

The current study relied on cross-sectional data with purposively sampled personnel from a community-based HSO; findings cannot be generalized beyond the current study population. Data were collected as part of a cross-sectional study, with data collected over several years, limiting our ability to assess the influence of temporality of relationships. It is possible that the cross-sectional study design did not fully capture personnel experiences fluctuating over time. Only childhood trauma exposures were captured, and adult exposures were not included. Because “don’t know” and “not sure” defaulted to “no” or zero, ACE scores may be deflated in the current study. The observed missingness/decline to answer responses for questionnaires such as ACEs is hypothesized to be, in part, due to participant discomfort in responding to sensitive questions in the workplace, underscoring the potential for pivoting to positive psychology instruments (i.e., probing for positive childhood experiences rather than adversity). Because PTSD was not assessed (i.e., full assessment diagnostics were not administered due to the nature of the workplace study setting, but rather a smaller number of trauma effects items were administered to pilot the concept), the study variable trauma effects cannot be confidently attributed to trauma exposures or definitively disentangled from general mental health attributes. Additionally, the construct of resilience was narrowly measured due to the need for brief surveys. Further, operationalization of resilience was limited to only psychological-level resilience; however, the field of research clearly indicates multiple levels of resilience (e.g., interpersonal, organization, community, etc.) play significant roles in shaping persons’ overall resilience (Wright et al., 2013).

Conclusion

Our results suggest that greater resilience was associated with a lower likelihood of burnout and compassion fatigue and a higher likelihood of compassion satisfaction. However, we found little evidence to suggest that psychological resilience moderated relationships between trauma and PQoL. Past literature indicates healthcare organizational culture plays a critical role in addressing burnout (Friganović et al., 2019); hence future research should explore resilience as a multi-dimensional construct. Specifically, levels of organizational trauma resilience (Brown et al., 2021) should be explored. Given the observed relationships between childhood adversity, adult trauma effects, and PQoL, future research should continue to include these variables, as they are likely important components of evaluating trauma-informed care/ organizational trauma resilience. Further research is needed to understand predictors and moderators of PQoL and how trauma-informed care might mitigate ill effects stemming from childhood adversity and trauma.