High Caregiver Adverse Childhood Experiences Are Associated With Pediatric Influenza and COVID-19 Vaccination Uptake

Abstract

Background

Factors surrounding vaccine uptake are complex. Although anxiety, which could influence vaccination decisions, has been associated with adverse childhood experiences (ACEs), little is known about links between caregiver ACEs and pediatric vaccine uptake. We evaluated associations between caregivers’ ACEs and decisions to vaccinate their children with influenza and coronavirus disease (COVID-19) vaccines.

Methods

A cross-sectional study of caregivers of patients ≥6 months at one pediatric primary care center (PPCC) was performed. Caregivers completed a 19-question survey examining caregiver ACEs, influenza vaccine acceptance and beliefs, and intention to vaccinate their child with the COVID-19 vaccine. Demographic characteristics, social risks (eg, housing and food insecurity), and vaccination data for children present with each caregiver were extracted from the electronic health record (EHR). Statistical analyses included χ² tests for categorical variables and t-tests for continuous variables.

Results

A total of 240 caregivers participated, representing 283 children (mean age of 5.9 years, 47% male). Twenty-four percent (n = 58) had high ACEs (≥4). Of those with high ACEs, 55% accepted pediatric influenza vaccination compared with 38% with low ACEs (P = .02). Those with high ACEs had more positive attitudes toward influenza vaccine safety and efficacy (P ≤ .02). Those with high, compared with low, ACEs were also more likely to accept COVID-19 vaccination (38% vs 24%; P = .04).

Conclusions

Pediatric influenza vaccination rates and intention to vaccinate children against COVID-19 differed between caregivers with high and low ACEs: those with more ACEs were more likely to vaccinate. Further studies assessing the role of caregiver ACEs on vaccine decision-making are warranted.

To the best our knowledge, this is the first study to find a link between caregiver adverse childhood experiences (ACEs) and pediatric influenza and coronavirus disease (COVID-19) vaccine uptake in pediatric primary care, with higher intent to vaccinate among those with higher ACEs.

INTRODUCTION

Adverse childhood experiences (ACEs) are traumatic, stressful life events that occur before age 18 [1]. Increasingly, investigators are evaluating the impact of ACEs on child health, with early childhood adversities like physical abuse and neglect correlating with children attending to negative more than positive stimuli [2, 3]. This may prompt anxiety and greater reactivity to negative events, shaping lifelong stress responses. We do not know whether this stress response influences a parent’s approach to vaccination choices, or the perceived threat of vaccine-preventable diseases. A retrospective cohort study found that caregiver ACEs were not associated with delayed or missed immunizations in children ≤2 years of age [4], though this study did not examine decisions surrounding influenza vaccination and was prior to the coronavirus disease (COVID-19) pandemic.

Childhood vaccination uptake has decreased since the COVID-19 pandemic. Greater concerns for vaccine effectiveness have emerged, including that for the influenza vaccine [5]. Misinformation and disinformation have amplified vaccine hesitancy [6]. Since Spring 2020, influenza vaccination rates have declined in our pediatric primary care center (PPCC) from 39% in the 2018–19 influenza season to 30% in the 2021–22 influenza season [7]. Reasons for decreased uptake of COVID-19 and influenza vaccines are of particular interest given vaccine completion challenges in marginalized populations [8], prompting us to explore factors potentially related to vaccine uptake, like caregiver ACEs.

Given that parental stress has increased during the COVID-19 pandemic [9], understanding the interplay of caregiver trauma and its impact on vaccination rates could inform future interventions to promote preventative health strategies for vulnerable populations. Therefore, we sought to evaluate the impact of caregiver ACEs on pediatric influenza and COVID-19 vaccine uptake and decision-making in one PPCC, which cares for a predominantly publicly insured, non-Hispanic Black population.

METHODS

Study Design

We performed a prospective study of caregivers with children ≥6 months to determine associations between caregiver ACEs and influenza vaccine acceptance and beliefs, and intention to receive the influenza and/or COVID-19 vaccines. Our study was approved by the Cincinnati Children’s Hospital Medical Center (CCHMC) Institutional Review Board.

Study Site and Population

We conducted our study at the CCHMC’s PPCC that serves ~17 000 patients; of which, 80% identify as Black and 95% non-Hispanic, and 95% have public insurance or self-pay status.

English-speaking adult (≥18 years) caregivers of PPCC children ≥6 months with decision-making ability for the patient were eligible to participate. Caregivers of children <6 months were excluded, as patients were ineligible for influenza vaccination. Although COVID-19 vaccine eligibility changed during the study, caregivers of children ≤12 years of age were included. We sought to understand caregiver’s perspectives for these ages since COVID-19 pediatric vaccination was novel and future intention to vaccinate was highly relevant.

Survey Administration and Data Collection

Study participants completed a 19-question survey which included: (1) the standardized 10-question ACE questionnaire [1], (2) an adapted, Centers for Disease Control and Prevention (CDC)-validated 6-question survey about beliefs surrounding the influenza vaccination and infection [10], and (3) a de novo 3-question survey about intention to receive the influenza and COVID-19 vaccines for themselves, and for their child to receive the COVID-19 vaccine once approved (Supplementary Appendix 1). We identified positive or concerned attitudes about influenza vaccination or infection (for self and child), via survey response [10].

A convenience sample of caregivers were approached after checking into their child’s appointment, and informed consent was obtained before study participation. Written surveys were self-completed by participants between June and December 2021 and a research assistant entered responses into the REDCap^© database. Demographic and vaccination data were extracted from the electronic health record (EHR) for each patient present that day, including siblings. Influenza vaccine status of pediatric patients was obtained from the EHR in May 2022, the end of the 2021–22 influenza season, to ensure complete data acquisition. We also extracted COVID-19 vaccine status for patients who were ≥5 years of age in May 2022, those eligible for the vaccine at the time. Caregiver influenza and COVID-19 vaccine status was based on self-report.

Primary and Secondary Outcomes

Our primary outcome was the proportion of children who received the influenza vaccine among those surveyed. A child’s influenza vaccine status was determined based on the presence of a documented influenza vaccine in our EHR or via a statewide vaccine registry. The Ohio Impact Statewide Immunization Information System is a statewide vaccine registry used by hospitals, ambulatory care sites, and other vaccine delivery sites (eg, pharmacies, grocery stores). If eligible, we assessed COVID-19 vaccine status similar to influenza.

Our secondary outcome was the proportion of caregivers with a positive or concerned attitude about influenza vaccination or infection, respectively. Additional outcomes analyzed the proportions of caregivers who intended to receive the influenza and COVID-19 vaccine themselves. Since the COVID-19 vaccine was approved for children ≥5 years old during the study period, we completed a sub-analysis investigating the proportion of those eligible who intended to accept the COVID-19 vaccine for their child who ultimately vaccinated their child against COVID-19.

Predictor Variables and Covariates

Our primary predictor was the caregiver ACE score. A high ACE score was defined as ≥4 ACEs, given studies showing significant differences in health outcomes in those with ≥4 compared with <4 ACEs [11]. Additional predictor variables and covariates were obtained from the EHR. Specifically, we included demographic characteristics like patient age (continuous variable defined by mean/standard deviation), gender (number/percentage), race (number/percentage of those self-identifying as Black, white, other, or unknown/refused), insurance type (number/percentage of those with public insurance, private insurance, or self-pay), and information from a standardized social risk questionnaire used in our PPCC (number/percentage with a positive screen). The questionnaire assesses food insecurity, issues with public benefits (eg, supplemental nutrition assistance program, medical insurance, daycare vouchers), housing conditions or insecurity, transportation issues, and caregiver depression (Supplementary Appendix 2). We also determined whether a social work note and/or referral to our medical-legal partnership was placed within 3 days of a visit to assess intervention to address social risks during the encounter occurred. Specific diagnoses (eg, asthma, prematurity, developmental delay) were extracted from each patient’s problem list. We evaluated each diagnosis individually and assessed whether children had any of the included diagnoses. We chose this set of diagnoses as the CDC’s Advisory Committee on Immunization Practices highlights each as increasing the risk of influenza-related complications [12].

Data Analysis

Demographic characteristics and identified social risks were presented descriptively and compared between groups with high and low ACE scores. Statistical analyses included χ² tests for categorical variables and t-tests for continuous variables. A two-tailed P-value of <.05 was considered significant.

Due to the nature of certain variables collected, the following have missing data: (1) positive social risk screen (patient-completed and could be skipped), (2) influenza beliefs/attitudes and caregiver intent to receive influenza/COVID-19 vaccination for themselves/child (not everyone completed all questions), and (3) pediatric influenza vaccination accepted last year (not every child was eligible for influenza vaccination the prior year). Some children did not have race identified in their charts and were labeled as “unknown/not reported” in their respective chart. Participants with missing data for analyses using that value were excluded. Missing data in Tables 1 and 3 are denoted as footnotes.

Table 1.

Influenza and COVID-19 Vaccination Beliefs, Acceptance, and Intentions to Vaccinate by ACE Score

	Total (n = 240)	High (≥4) ACEs (n = 58)	Low (<4) ACEs (n = 182)	P-value
Demographic features
Number of children, n (%)				.93
1	200 (83%)	48 (83%)	152 (84%)
2	37 (15%)	9 (16%)	28 (15%)
3	3 (1%)	1 (2%)	2 (1%)
Age of children in years, mean (SD)	5.9 (4.2) [n = 283]	6.0 (4.4) [n = 69]	5.8 (4.2) [n = 214]	.75
Gender children, n (%)				.64
Female	149/283 (53%)	38/69 (55%)	111/214 (52%)
Male	134/283 (47%)	31/69 (43%)	103/214 (48%)
Race of children, n (%)				.80
White	46/283 (16%)	12/69 (17%)	34/214 (16%)
Black	229/283 (81%)	55/69 (80%)	174/214 (81%)
Asian	2/283 (1%)	0	2/214 (1%)
Unknown/not reported	6/283 (2%)	2/69 (3%)	4/214 (2%)
Diagnosis of asthma, prematurity, or developmental delay (%)	108/283 (38%)	32/69 (46%)	76/214 (36%)	.11
Asthma	45/283 (16%)	12/69 (17%)	33/214 (15%)	.70
Prematurity	33/283 (12%)	9/69 (13%)	24/214 (11%)	.68
Developmental delay	47/283 (17%)	14/69 (20%)	33/214 (15%)	.34
Socioeconomic characteristics
Positive social risk screen*, n (%)	75/232 (32%)	19/54 (35%)	56/178 (31%)	.61
Medicaid insurance, n (%)	219/240 (91%)	52/58 (90%)	167/182 (92%)	.62
Social work note or Medical Legal Aid Referral placed within 3 days of clinic visit, n (%)	23/240 (10%)	5/58 (9%)	18/182 (10%)	.77
Flu vaccine outcomes and beliefs
Flu vaccine accepted during the 2021–2022 season, n (%)	101/240 (42%)	32/58 (55%)	69/182 (38%)	.02
Flu vaccine accepted last season*, n (%)	73/204 (36%)	22/49 (45%)	51/155 (33%)	.13
Parent planned to obtain flu vaccine during the 2021–22 season*, n (%)	108/234 (46%)	29/55 (53%)	79/179 (44%)	.26
Positive attitude toward influenza vaccine safety*, n (%)	158/232 (68%)	46/57 (81%)	112/175 (64%)	.02
Positive attitude toward influenza vaccine efficacy*, n (%)	147/232 (63%)	44/56 (79%)	103/176 (59%)	.01
Concerned attitude toward influenza infection*, n (%)	111/232 (48%)	35/57 (61%)	76/175 (43%)	.02
COVID-19 vaccine intent
Parent planned to receive COVID-19 vaccine for themselves*, n (%)	97/234 (41%)	28/54 (52%)	69/180 (38%)	.08
Planned for child to receive COVID-19 vaccine when eligible*, n (%)	62/229 (27%)	20/52 (38%)	42/177 (24%)	.04
COVID-19 vaccine eligible, n (%)	136/240 (57%)	36/58 (62%)	100/182 (55%)	.34
Among eligible COVID-19 vaccine received, n (%)	41/136 (30%)	15/36 (42%)	26/100 (26%)	.08

^*Variable with missing data due to incomplete participant survey responses.

Bolded values represent significant values.

Table 3.

Multivariable Logistic Regression of Factors Associated with Influenza Vaccine Uptake in the 2021–22 Influenza Season

	Odds Ratio	95% Confidence Interval
ACE score, per point	1.18	1.00–1.39 *
Race
Black	1.53	0.54–4.35
Non-Black (White, Asian, unknown/not reported)	Ref
Positive social risk screen**	0.82	0.19–3.65
Insurance
Medicaid/self-pay	1.54	0.45–5.24
Private	Ref
Influenza vaccine accepted for child last season**	5.66	2.57–12.48
Parent planned to obtain Influenza vaccine during the 2021–22 season**	7.76	3.59–16.79

^* P = .0494.

^**Variable with missing data, due to incomplete participant survey responses.

Bolded values represent significant values.

We created multivariable logistic regression models to test independent associations between predictor variables and pediatric influenza/COVID-19 vaccine uptake. For pediatric influenza vaccine uptake during the 2021–22 season, we examined the following variables: ACE score, Black race of child, positive social risk screen, Medicaid insurance, pediatric influenza vaccine acceptance for the prior season, and parental acceptance for the influenza vaccine during the current season. Low sample size precluded our ability to create a multivariable model for intention to accept the pediatric COVID-19 vaccine. Models yielded odds ratios (OR) and 95% confidence intervals (CIs).

Mediation analyses were undertaken to decompose the total effect of caregiver ACE score with the outcome of pediatric influenza vaccine uptake into direct and indirect effects through mediator variables of positive attitudes toward influenza vaccine efficacy and safety and concern for influenza infection. We utilized the Mplus software [13] to create a path model fitted with logistic regression, suitable for the dichotomous outcome and bootstrap 95% CIs for model estimates.

RESULTS

Participation Rate

We approached 338 caregivers in the PPCC. Two were ineligible; one did not speak/read English and one did not have decision-making authority. Of the remaining 336 eligible caregivers, 71% (n = 240) consented to participate.

Demographic Characteristics and Prevalence of ACEs

These 240 caregivers presented with 283 children for office visits (Table 1). The children were an average age of 5.9 years, and 47% were male. Ninety-one percent were publicly insured, and 32% indicated one or more social risk on our standardized questionnaire.

Twenty-four percent of caregivers reported ≥4 ACEs. The most frequently encountered ACE was caregiver’s parents’ divorcing during childhood. There were no significant differences in patient age, gender, race, diagnosis of asthma, prematurity, or developmental delay, insurance status, social risk presence, or social work referrals between those with high and low ACE scores.

ACEs and Pediatric Influenza Uptake, and Pediatric COVID-19 Vaccine Intent

There was no statistically significant difference in pediatric influenza vaccine uptake between caregivers with high and low ACEs during the 2020–21 influenza season. For the 2021–22 influenza season, however, children of caregivers with high ACEs were more likely to receive the influenza vaccine than children of caregivers with low ACEs (55% vs 38%, P = .02). Specific ACEs involving verbal abuse (25% vs 14%, P = .04), lack of family support (28% vs 16%, P = .03), and parental abuse (21% vs 11%, P = .03) were associated with increased pediatric influenza vaccine uptake.

In caregivers with high ACEs, there was increased intent for their children to receive the COVID-19 vaccine compared with those with low ACEs (38% vs 24%, P = .04). No individual ACE question was associated with pediatric COVID-19 vaccination intent. Although we noted similar trends in influenza and COVID-19 vaccine uptake in caregivers themselves, analyses did not reach statistical significance.

Caregivers with high ACEs, compared with those with low ACEs, reported positive attitudes toward influenza vaccine safety (81% vs 64%, P = .02) and efficacy (79% vs 59%, P = .01), and more concern about influenza infection (61% vs 43%, P = .02). There were indirect effects of positive attitude toward influenza vaccine safety (OR 1.06, CI 1.02–1.11) and efficacy (OR 1.06, CI 1.03–1.11) as mediators in the association between ACE score and influenza vaccine uptake (Table 2). Figure 1 depicts the causal associations of caregiver ACEs, influenza vaccine attitudes, and pediatric influenza vaccine uptake.

Table 2.

Caregiver’s Attitudes Mediating the Association Between ACE Score and Influenza Vaccine Uptake

	OR (95% CI)
ACE score, total effect	1.13 (1.01, 1.27)
Positive attitude toward influenza vaccine safety mediator
ACE score, direct	1.06 (0.95, 1.19)
ACE score, indirect (mediated effect)	1.06 (1.02, 1.11)
Concerned attitude toward influenza infection mediator
ACE score, direct	1.13 (1.00, 1.28)
ACE score, indirect (mediated effect)	1.00 (0.97, 1.02)
Positive attitude toward influenza vaccine efficacy mediator
ACE score, direct	1.06 (0.96, 1.19)
ACE score, indirect (mediated effect)	1.06 (1.03, 1.11)

Figure 1.

Schematic of Mediators of Caregiver ACE score and Pediatric Influenza Vaccination Uptake.

During the study period, 136 of patients ≥5 years were eligible for the COVID-19 vaccine; 129 had complete data regarding vaccine decision-making. Of these 129, 44 (34%) had caregivers with positive intention to vaccinate the child against COVID-19, and 28 (64%) received the COVID-19 vaccine. In those with high ACEs who indicated intent to vaccinate, 73% (11 of 15) received the vaccine compared with 59% (17 of 29) in those with low ACEs.

In a multivariable logistic regression model, ACE score (per point, OR (CI) 1.18 (1.00–1.39)), pediatric influenza vaccine acceptance in the prior season, and caregiver intent to receive that season’s influenza vaccine, were each independently associated with pediatric influenza uptake (Table 3).

DISCUSSION

In our PPCC, which serves a predominantly non-Hispanic Black, publicly insured, population, one of the four caregivers reports high ACEs (≥4), a rate greater than the general U.S. population [14]. Contrary to our hypothesis, caregivers reporting high ACEs had increased rates of pediatric influenza vaccine uptake and increased intent for their child to receive the COVID-19 vaccine. This finding highlights a new factor—ACEs—to consider related to vaccine acceptance in children.

We found with each incremental increase in the number of ACEs, the adjusted odds of pediatric influenza vaccination increased by 18%. Our findings run counter to our initial hypothesis presuming that caregivers with higher ACEs would have decreased rates of vaccine uptake. We hypothesized this given many individuals with high ACEs have more anxiety and perceived threat of negative stimuli [2, 3], which we thought would translate to perceived threat of vaccination side effects. Instead, we found those with higher ACEs reported more fear of disease and more positive attitudes of safety and efficacy of influenza vaccines. We demonstrated that positive attitudes toward influenza vaccine safety and efficacy mediate the relationship between caregiver ACEs and influenza vaccine uptake. This could potentially reflect a heightened trust in preventative therapies in caregivers with high childhood adversity, though this warrants further investigation to fully understand this relationship. Interestingly, concern for influenza infection was not a mediator of vaccine uptake, which may indicate less of a role of parental anxiety surrounding adverse health outcomes for their own children in caregivers with high ACEs. However, this potential relationship warrants further examination. Because of lacking published data on how caregiver ACEs influence healthcare decision-making and medical information-seeking for their children, particularly since the COVID-19 pandemic, a new area of exploration to understand associations is warranted.

Our study is the first, to the best our knowledge, to examine how caregiver ACEs relate to childhood influenza and COVID-19 vaccination status. Eismann et al, [4] reported that caregiver ACEs were not significantly associated with missed immunizations prior to 2 years of age, though this study did not examine influenza vaccine uptake and occurred prior to the COVID-19 pandemic. Bellis et al, [15] examined how adults’ ACEs affect their own decisions and beliefs surrounding getting the COVID-19 vaccine for themselves. They found that those with ≥4 ACEs had a 4-fold higher rate of vaccine hesitancy (defined as responding “no” or “unsure” to a question about intent to receive the COVID-19 vaccine) compared with those with no ACEs (19% vs 5%) [15]. Another study, though not directly examining ACEs, investigated how mental health diagnoses like post-traumatic stress disorder (PTSD) influenced intentions for caregivers’ children to receive the COVID-19 vaccine. Milan and Dau [16] found that mothers with PTSD had higher rates of COVID-19 vaccine hesitancy for themselves and their children than individuals with other mental health diagnoses due to a general distrust of the healthcare system. Our study addresses knowledge gaps by assessing how caregiver ACE influence decisions to vaccinate their children with influenza and COVID-19 vaccines, helping us to understand the relationship between caregiver ACEs and decisions to vaccinate their children in primary care.

Links between pediatric influenza vaccine decision-making and caregiver ACEs appeared related to the most recent influenza season and not the prior season. We found that if a parent intended to receive the influenza vaccine for themselves and if the child had received an influenza vaccine the prior season, there was an increased odds of pediatric influenza vaccination for the 2021–22 season. Such vaccination decisions may be disproportionately influenced by the COVID-19 pandemic itself. The lack of access to healthcare resources during the pandemic could have motivated vaccination uptake, as ACEs disproportionately affect populations experiencing oppression historically and heightened disparities during the pandemic [17]. For example, persons of color experienced higher rates of cases and deaths from COVID-19 [18], which may have prompted those with high ACEs, who have higher rates of depression and anxiety [19], to seek out vaccination for their child to prevent poor health outcomes attributed to the virus. It will be important to understand how the pandemic continues to shape thoughts on pediatric vaccinations, including routine vaccinations, and how these perceptions relate to caregiver ACEs.

Our study had several limitations. We did not obtain caregiver sociodemographic information to minimize questions for caregivers. We assessed a family’s social context by analyzing the child’s insurance status and social risk screen. Such information provided some, though incomplete, evaluation of the complex relationship of ACEs and a family’s social, economic, and environmental context. Though our social risk screen asks caregivers about symptoms of depression and anhedonia, it is not a comprehensive evaluation of caregiver mental health. Additionally, we could not determine if those who declined to participate were significantly different than participants, though our cohort’s demographics are representative of the office’s population. We only recruited English-speaking caregivers to participate; thus, a future direction would be to include non-English speakers. Our study occurred at a single PPCC, which limits generalizability. However, we think that our population demographics may be applicable to academic medical centers within medium- to large-sized cities. Our study did not examine ethnicity as a factor impacting vaccine uptake as this is not reliably obtained. It is also possible that certain sites of influenza vaccination may not reliably update their data in the statewide registry, or we may not have captured all vaccinations completed out of state. We were unable to develop meaningful multivariable models for COVID-19 vaccine uptake due to low numbers. Finally, our study was conducted during the COVID-19 pandemic in the summer and fall of 2021, making it difficult to fully understand the impact of the pandemic itself on our results. However, we postulate the pandemic will have lingering and long-lasting effects on pediatric vaccine uptake, so evaluation at different timepoints may be useful.

Next steps include analyses, quantitative and qualitative, to understand linkages between caregivers’ intentions to vaccinate their children and ACEs, explore the underlying differences in pediatric vaccine uptake between caregivers with high and low ACEs, and examine relationships among caregivers’ demographics, medical history, and ACEs. Having a better sense of the fundamental mechanisms that link ACEs to vaccination uptake could lead to the creation of targeted messaging strategies and tailored clinical interventions. For instance, caregivers with low ACEs may need more information on the harm of influenza infection and the safety and efficacy of the pediatric influenza vaccine. We base this on our results showing relatively low rates of positive attitudes toward the influenza vaccine and less concern about influenza infection within this subgroup. Our research highlights the need to investigate how caregiver ACEs influence decision-making across child health, including inside and outside of infectious diseases.

CONCLUSION

Pediatric influenza vaccination rates and intention to vaccinate children with the COVID-19 vaccine differed between caregivers with high and low ACEs, with caregivers with ≥4 ACEs being more likely to vaccinate their children during the COVID-19 pandemic. Further studies exploring the influence of caregiver ACEs on vaccine decision-making are warranted. Such investigation could lead to trauma-informed ways for clinicians to discuss the importance of, and need for, childhood vaccination.

Financial support. This project was funded by the Research Innovation in Support of Excellence (RISE) resident research grant at Cincinnati Children’s Hospital Medical Center. We also utilized a REDCap database, funded by the National Center for Advancing Translational Sciences of the National Institutes of Health, under award number 5UL1TR001425-04.

Potential conflicts of interest. All authors: No reported conflicts.

All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

Contributors’ statements. M. E. D. conceptualized and designed the study, drafted the initial manuscript, and reviewed and revised the manuscript. D. G. and A. R. collected the data for the study and reviewed and revised the manuscript. H. S. carried out the statistical analyses and reviewed and revised the manuscript. E. P. S. and M. K. conceptualized and designed the study, and critically reviewed the manuscript for important intellectual content. A. F. B. and M. C. B. provided insight and feedback on study design, and critically reviewed the manuscript for important intellectual content. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.