1
Introduction
Children’s behavioral functioning is associated with many important outcomes, including
concurrent and future mental health, academic success, and social well-being (Duncan
et al., 2007, Liu et al., 2011, Reardon and Portilla, 2016). Child behavior encompasses
both internalizing (i.e., inward behaviors) and externalizing behaviors (i.e., outward
behaviors) (Achenbach et al., 2016), and behavior problems in early childhood may
place children at risk for later mental health problems, academic problems, and social
problems (van Lier et al., 2012, Weeks et al., 2016). Thus, understanding sources
of variability in child behavior is crucial for developing strategies aimed at equipping
children with the behavioral resources necessary to thrive. Recent research has reported
increases in child behavior problems during the ongoing COVID-19 pandemic, highlighting
the threat the pandemic poses to children (Colizzi et al., 2020, Patrick et al., 2020,
Saurabh and Ranjan, 2020, Whittle et al., 2020). In this study, our goal is to examine
whether change in child behavior during the pandemic varies as a function of two important
risk factors: parental mental health and preterm birth.
Two factors previously associated with increases in child behavior problems include:
1) parental mental health difficulties (Slomian et al., 2019), which have been shown
to increase during environmental stressors (e.g., pandemics, natural disasters) (Brock
et al., 2015, Harville et al., 2010), and 2) preterm birth (birth occurring prior
to 37 weeks gestation) (Allotey et al., 2018). Up to 1 in 10 children in the United
States are born prematurely, and it is estimated that the annual societal economic
impact of preterm birth is $26.2 billion due to health and developmental problems,
including behavior problems (Institute of Medicine (US) Committee on Understanding
Premature Birth and Assuring Healthy Outcomes, 2007). Moreover, not only are parental
mental health difficulties more common among parents of preterm children (Vigod et
al., 2010), but preterm children may also be more sensitive to environmental stressors,
like parental stress, than full-term children (Gueron‐Sela et al., 2015). Per the
differential susceptibility hypothesis (Belsky et al., 2007), preterm children may
be more susceptible than their full-term counterparts to both the consequences of
negative environmental exposures and to the benefits of positive ones (Gueron‐Sela
et al., 2015; Shah et al., 2013). Taken together, children of parents with mental
health difficulties and children born prematurely might be particularly vulnerable
to increases in child behavior problems during the pandemic. However, several important
gaps in the literature still exist.
Researchers have documented concurrent change in parental mental health and child
behavior during the current pandemic (Patrick et al., 2020, Whittle et al., 2020).
However, whether and how a global stressor relates to changes in child behavior via
changes in parental mental health remains unclear. Relatedly, most literature focuses
on the role of parental mental health disturbances. Less is known about the role of
parental well-being. That is, most literature focuses on the presence vs. absence
of unpleasant emotional experiences but largely ignores the presence vs. absence of
pleasant emotional experiences. If preterm children are indeed more sensitive than
full-term children to both negative and positive exposures, it is important to examine
the continuum of positive mood in addition to negative mood. Furthermore, evidence
for the differential susceptibility hypothesis is mixed. While some studies have found
support for it (Gueron‐Sela et al., 2015), others have not (Hadfield et al., 2017),
perhaps in part due to variations in the characteristics (e.g., chronological age,
gestational age) of the children included.
In the current study, we examined how change in parental mental health during the
pandemic relates to change in child behavior and whether preterm children are more
behaviorally susceptible than full-term children to changes in parental mental health.
Specifically, we examined the following questions:
1.
How has parental depression, anxiety, and well-being changed since the COVID-19 pandemic
began? Based on previous findings, we hypothesize that parental depression and anxiety
symptoms will have increased. However, due to limited prior literature, we do not
have specific hypotheses about changes in well-being.
2.
How have children’s internalizing and externalizing symptoms changed since the COVID-19
pandemic began? Based on previous findings, we hypothesize that children’s internalizing
and externalizing symptoms will have increased.
3.
Is change in parental depression, anxiety, and well-being associated with change in
children’s internalizing and externalizing symptoms? Does prematurity moderate that
relation? Based on previous findings, we expect increases in parental depression and
anxiety symptoms to be associated with increases in children’s internalizing and externalizing
symptoms. However, based on mixed findings, our questions about the role of parental
well-being and the role of prematurity will be exploratory.
Elucidating relations between parental mental health, child behavior, and prematurity
has the potential to inform efforts aimed at supporting child behavior , especially
for children who have biological risk factors. If preterm children are indeed behaviorally
more susceptible to changes in parental mental health during the ongoing pandemic,
our findings can inform prevention and intervention efforts aimed at supporting preterm
children and their families. Additionally, they will inform theoretical discussions
on the differential susceptibility of children born prematurely.
2
Methods
2.1
Participants and Procedure
Participants in the current study included a subset of participants from an ongoing
study, who were recruited via electronic health records at a local university hospital.
The larger ongoing study, hereafter referred to as T1, aims to examine the role of
prematurity in children’s outcomes. Parent-child dyads were eligible for T1 if the
child was between the ages of 3 and 5 years old, and had no history of a genetic syndrome,
birth defect, or intellectual/ developmental disability. Parents were invited to participate
via phone, email, social media, and word-of-mouth. Enrollment for T1 began in June
2019. At the time of their T1 laboratory visit, 91.2% of dyads were from [Midwestern
State]. Parent-child dyads attended a laboratory visit, during which parents completed
consent forms, children completed assessments, and parents completed questionnaires.
At their T1 visit, parents indicated whether they would like to be informed of future
studies. In mid-March of 2020, we paused data collection due to COVID-19-related restrictions.
In April 2020, we invited parents who completed T1 and indicated they’d like to hear
about future studies to participate in a “COVID-19 Follow-Up” study, hereafter referred
to as T2. For T2, we contacted parents via phone and email. Of note, only parent-child
dyads who completed T1 were eligible for T2. This study received IRB approval from
the University of [State] (IRB ID#201805712, initially approved on October 30, 2018).
For T2, parents completed online questionnaires between April 2020 and October 2020.
An average of 273 days elapsed (SD = 93 days, range = 77-442 days) between T1 and
T2. A total of 48 parents completed T2, out of the 104 parents that had completed
T1. Of note, parents of two sets of twins provided data for T2; however, we only included
data for one twin per set in the final T2 dataset. Most parents at both T1 and T2
were Non-Hispanic White (87.1%, 87.5%, respectively), female (91.2%, 93.8%, respectively),
and had at least a bachelor’s degree (77.6%, 79.2%, respectively) (Table 1
). Of the parents that participated in T2, 86% were from [Midwestern State]. Other
states included Ohio, Illinois, and Texas. We did not find any significant differences
on T1 demographic characteristics, T1 child behavior scores, or T1 parental mental
health scores between those dyads that participated in T2 and those that did not.
Moreover, parents of preterm children and parents of full-term children did not differ
from each other in their parental mental health scores at T1 or T2, and preterm children
and full-term children did not differ from each other on their child behavior scores
at T1 or T2 (Table 2
).
Table 1
Demographic Information for the Study Population
T1
T2
M(SD)
M(SD)
Parent Age (years)
36.32 (5.01)
37.44 (4.51)
Child Age (years)
4.60 (0.62)
5.34 (0.84)
Child Gestational Age (weeks)- Full-Term
39.45 (1.39)
39.32 (1.09)
Child Gestation Age (weeks)- Preterm
30.33 (4.58)
30.40 (4.49)
Household Income (USD)
115,761 (74,984)
118,744 (64,032)
N(%)
N(%)
Parent Gender
Female
93 (91.2%)
45 (93.8%)
Male
8 (7.8%)
2 (4.2%)
Other
1 (1.0%)
1 (2.1%)
Parent Race/Ethnicity
Non-Hispanic White
88 (87.1%)
42 (87.5%)
Hispanic White
3 (3.0%)
0 (0.0%)
Non-Hispanic Black
6 (5.9%)
2 (4.2%)
Non-Hispanic Asian
4 (4.0%)
4 (8.3%)
Parent Education
High School
3 (2.9%)
0 (0%)
Some college
6 (5.8%)
3 (6.3%)
Associate’s Degree
14 (13.6%)
7 (14.6%)
Bachelor’s Degree
43 (41.7%)
19 (39.6%)
Graduate Degree
37 (35.9%)
19 (39.6%)
Child Gender
Female
50 (48.5%)
22 (45.8%)
Child Race/ Ethnicity
Non-Hispanic White
86 (82.7%)
41 (85.4%)
Hispanic White
5 (4.8%)
1 (2.1%)
Non-Hispanic Black
5 (4.8%)
2 (4.2%)
Hispanic Black
2 (1.9%)
0 (0.0%)
Non-Hispanic Asian
5 (4.8%)
4 (8.3%)
Non-Hispanic Other
1 (1.0%)
0 (0.0%)
Child Premature (<37 weeks gestation)
Yes
36 (35.0%)
20 (41.7%)
Note: For T1, N=104; For T2, N=48
Table 2
T1 and T2 IDAS-II and CBCL Scores for Preterm and Full-Term Children
T1
Preterm
Full-Term
t-test
CBCL- Internalizing
9.72
6.67
t(43)=-1.39, p=0.17
CBCL- Externalizing
12.22
9.85
t(43)=-0.81, p =0.42
IDAS- General Depression
35.61
34.36
t(41)=-0.38, p =0.71
IDAS- Social Anxiety
8.75
8.07
t(46)=-0.69, p =0.49
IDAS- Well Being
24.00
26.11
t(44)=1.15, p =0.26
T2
Preterm
Full-Term
t-test
CBCL- Internalizing
9.33
10.00
t(44)=0.24, p =0.81
CBCL- Externalizing
11.94
13.14
t(43)=0.38, p =0.71
IDAS- General Depression
36.29
39.96
t(40)=1.02, p =0.31
IDAS- Social Anxiety
7.53
8.30
t(42)=0.99, p =0.33
IDAS- Well Being
19.41
21.19
t(42)=0.95, p =0.35
2.2
Measures
Timepoint.
Timepoint refers to T1 (pre-pandemic) and T2 (during pandemic). Timepoint was dummy-coded
with T1 as the reference point.
Days Elapsed.
Days elapsed refers to the number of days between when participants completed T1 and
when they completed T2.
Prematurity.
Parents reported their child’s gestational age at T1. We considered those born at
<37 weeks gestation preterm and those born at ≥37 weeks gestation full-term.
Parental Mental Health.
Inventory of Depression and Anxiety Symptoms, Expanded Version (IDAS-II).
Parents completed the IDAS-II (Watson et al., 2007, Watson et al., 2012) at both T1
and T2. The IDAS-II is a 99-item inventory. Parents reported on the extent to which
they experienced various depression and anxiety symptoms during the previous two weeks,
using a 5-point Likert scale. We used three subscales from the IDAS-II to reflect
parental mental health: 1) General Depression, 2) Social Anxiety, and 3) Well-Being.
It should be noted that this was not a clinical sample (Stasik-O’Brien et al., 2019).
Using the screening cut-offs from Stasik-O’Brien et al. (2019), 14% of parents at
T1 and 16.7% of parents at T2 scored above the screening cut-off for Major Depressive
Disorder on the IDAS-II General Depression Scale; 35.4% of parents at T1 and 22.2%
parents at T2 scored above the screening cut-off for Social Phobia on the IDAS-II
Social Anxiety Scale.
Child Behavior.
Child Behavior Checklist (CBCL).
Parents completed the CBCL (Achenbach & Rescorla, 2000) at both T1 and T2. The CBCL
is a 100-item inventory. Parents reported on how true statements were of their child’s
behavior, using a 3-point Likert scale. We used two subscales from the CBCL to reflect
child behavior problems: 1) Internalizing and 2) Externalizing. It should be noted
that this was not a clinical sample. Using CBCL t-scores, 24.4% of children at T1
and 22.2% of children at T2 scored in the borderline clinical to clinical range for
internalizing symptoms; 14% of children at T1 and 20.5% of children at T2 scored in
the borderline clinical to clinical range for externalizing symptoms. As this was
a non-clinical, community sample, we used raw scores for analyses involving the Internalizing
and Externalizing subscales to preserve the full range of variability in scores (Thurber
& Sheehan, 2012).
3
Analytic Plan
To address Research Question 1, we examined change in parental mental health between
T1 and T2 with multi-level mixed models fit using the lmer function in the Lme4 package
(Bates et al., 2014) in R (R Core Team, 2018). Using AIC to compare models, nested
models were evaluated to determine the maximal random and fixed effects structure
supported by the data for each parental mental health domain. To address Research
Question 2, we followed the same procedure to examine change in child externalizing
and internalizing behavior between T1 and T2. The most complex models evaluated for
both parental mental health and child behavior included the following predictors:
timepoint, days elapsed, prematurity, and a random intercept for participant. The
most complex models for child behavior also included child age at T1. The interaction
between prematurity and parent or child variables was added to the model after the
most parsimonious model was selected. To address Research Question 3, we ran multiple
linear regressions to determine whether change in parental mental health was associated
with change in child behavior. The most complex models evaluated included the following
predictors: change in parental mental health, parental mental health at T1, days elapsed,
and prematurity. Given the small sample size, the interaction between prematurity
and change in parental mental health was added to the model after the most parsimonious
model was selected. All results remained the same when we used a continuous measure
of gestational age (in weeks) instead of categorical prematurity. To account for the
possibility that the relation between parental mental health and child behavior is
bidirectional, we also conducted exploratory analyses in which change in child behavior
was associated with change in parental mental health.
4
Results
4.1
How has parental depression, anxiety, and well-being changed since the COVID-19 pandemic
began?
The best fitting models for parental depression symptoms, parental anxiety symptoms,
and parental well-being included a random intercept for participant and fixed effects
for timepoint and prematurity. Timepoint was significantly associated with change
in parental depression symptoms, t(38.61) = 2.58, p = .01, and in parental well-being,
t(41.03) = -6.49, p < .001, but not in parental anxiety symptoms, t(44.63) = -0.38,
n.s. Parental depression symptoms were significantly greater at T2 than they were
at T1, and parental well-being was significantly lower at T2 than at T1. Prematurity
was not significantly associated with change in parental depression symptoms, anxiety
symptoms, or well-being. Moreover, an interaction factor between timepoint and prematurity
did not improve the model fit and was therefore excluded.
4.2
How have children’s internalizing and externalizing symptoms changed since the COVID-19
pandemic began?
The best fitting model for children’s internalizing symptoms and externalizing symptoms
included a random intercept for participant and fixed effects for timepoint and prematurity.
Results suggested that timepoint was significantly associated with change in children’s
internalizing symptoms, t(39.53) = 2.21, p = .03, and children’s externalizing symptoms,
t(40.91) = 2.08, p = .04. Children’s internalizing symptoms and externalizing symptoms
were significantly greater at T2 than they were at T1. Child age at T1 improved the
model fit for children’s internalizing symptoms but not externalizing symptoms, with
child age at T1 positively associated with internalizing symptoms. Prematurity was
not significantly associated with children’s internalizing symptoms or externalizing
symptoms. Moreover, an interaction factor between timepoint and prematurity did not
improve the model fit for children’s internalizing symptoms or externalizing symptoms
and was therefore excluded from both models.
Is change in parental mental health disturbances and parental well-being associated
with change in child behavior? If so, does prematurity moderate that relation?
Change in parental depression symptoms was positively associated with change in children’s
internalizing symptoms and externalizing symptoms. Neither change in parental anxiety
symptoms nor change in parental well-being was associated with change in children’s
internalizing symptoms or externalizing symptoms (Table 3
).
Table 3
Linear Regressions Predicting Child Behavior Change
Estimate
SE
95 CI% LL
95 CI% UL
t-value
p-value
Adjusted R2
Child Externalizing Change
IDAS GD Change
0.40
0.17
0.05
0.75
2.32
0.03*
0.12
IDAS WB Change
-0.09
0.27
-0.63
0.45
-0.34
0.74
-0.05
IDAS SA Change
-0.31
0.48
-1.30
0.67
-0.65
0.52
0.13
Child Internalizing Change
IDAS GD Change
0.49
0.11
0.27
0.71
4.54
< 0.001*
0.37
IDAS WB Change
-0.01
0.20
-0.41
0.39
-0.05
0.96
-0.07
IDAS SA Change
0.06
0.40
-0.75
0.87
0.16
0.88
-0.07
Note: *p<.05. Statistics are from six different models for, each including an IDAS
subscale and child behavior subscale. Each model includes the following covariates:
parental mental health at T1 and prematurity. Days elapsed and the interaction factor
between parental mental health and prematurity did not improve the model, so they
were not included as covariates.
Parental mental health at T1 did not emerge as a significant predictor in any model
for children’s internalizing symptoms or externalizing symptoms. Prematurity did not
moderate the effect of change in parental depression symptoms, anxiety symptoms, or
well-being on children’s internalizing symptoms or externalizing symptoms. Figure
1, Figure 2
represent individual trajectories for change in children’s internalizing symptoms
and change in externalizing symptoms, respectively. On each figure, for visualization
purposes only, we divided the children into two groups using a median split on change
in parental mental health: high versus low change on the three IDAS-II subscales.
Consistent with the analyses, the figures show that children whose parents exhibited
above median change on parental depression symptoms exhibited greater change in internalizing
and externalizing symptoms (light grey line), compared to children whose parents exhibited
below median change on parental depression symptoms (dark grey line).
Figure 1
Individual trajectories of change in children’s internalizing symptoms as a function
of change in parental mental health: a) parental depression symptoms b) parental anxiety
symptoms, c) parental well-being.Note: High and low change groups are defined by a
median split on each parent IDAS-II scale measuring parental mental health. Light
grey represents change in children’s internalizing symptoms for children whose parents
scored above the median. Dark grey represents change in children’s internalizing symptoms
for children whose parents scored below the median. The solid line represents the
Loess curve fit on the observation.
Figure 2
Individual trajectories of change in children’s externalizing symptoms as a function
of change in parental mental health: a) parental depression symptoms b) parental anxiety
symptoms, c) parental well-being. Note: High and low change groups are defined by
a median split on each parent IDAS-II scale. Light grey represents change in children’s
externalizing symptoms for children whose parents scored above the median. Dark grey
represents change in children’s externalizing symptoms for children whose parents
scored below the median. The solid line represents the Loess curve fit on the observation.
To account for the possibility that the relation between parental mental health and
child behavior is bidirectional, we also conducted exploratory analyses in which child
behavior predicted parental mental health. We found that, when accounting for child
externalizing symptoms at T1, change in child externalizing symptoms was associated
with change in parental depression and change in parental anxiety but not parental
well-being. Moreover, we found that, even when accounting for child internalizing
symptoms at T1, change in child internalizing symptoms was associated with change
in parental Depression but not parental anxiety or parental well-being.
5
Discussion
In this study, we examined whether parental mental health and child behavior changed
during the COVID-19 pandemic. Importantly, we also examined how changes in parental
mental health related to changes in child behavior and whether preterm children were
more behaviorally susceptible than full-term children to changes in parental mental
health.
In line with our hypotheses, parental depression symptoms, children’s internalizing
symptoms, and children’s externalizing symptoms significantly increased, and parental
well-being significantly decreased during the pandemic. Notably, for these variables,
adding timepoint improved the model fit but adding days elapsed did not, suggesting
that change in these variables was not necessarily related to the mere passage of
time. In other words, although we cannot conclude that the pandemic is what caused
a decline in parental mental health and child behavior, our results suggest that parental
mental health and child behavior declined between T1 (pre-pandemic; June 2019-March
2020) and T2 (during pandemic; April 2020 to October 2020) even when accounting for
the mere passage of time.
Recent studies with larger samples have reported concurrent changes in parental mental
health and child behavior during the pandemic (Patrick et al., 2020). Our study expanded
upon this work, and our results suggest that the pandemic might impact child behavior
via parental mental health. Moreover, our results suggested that increases in specifically
parental depression symptoms, but not increases in parental anxiety symptoms or reductions
in parental well-being, were associated with increases in child behavior problems.
These findings contribute to the literature on identifying the specific aspects of
parental mental health that most strongly relate to change in child behavior. Relations
remained robust even when accounting for parental depression symptoms at T1; that
is, the baseline severity of parental depression symptoms at T1 did not account for
the relation between change in parental depression symptoms and change in child behavior.
Moreover, few parents scored above the screening cut-off on the IDAS-II General Depression
Scale. Thus, even changes in subclinical levels of symptoms were associated with child
behavior, regardless of baseline differences in parental mental health.
Why might changes in parental depression symptoms be related to change in child behavior?
First, increases in parental depression symptoms may have led to changes in parenting
behavior, in turn leading to increases in children’s internalizing symptoms and children’s
externalizing symptoms (Network, 1999, Reising et al., 2013, Slomian et al., 2019,
Sohr-Preston and Scaramella, 2006). Indeed, some literature during the pandemic has
already reported changes in parenting behavior, including increases in the use of
harsh parenting and reductions in parent-child relationship closeness (Chung et al.,
2020). It is also possible that parental depression symptoms increased as a result
of increasing child behavior problems rather than vice versa (Herring et al., 2006)
and that the relation between parental depression symptoms and child symptoms was
bidirectional (Neece et al., 2012). Indeed, our results suggested that the relation
between parental depression symptoms and child symptoms may be bidirectional; child
internalizing and externalizing symptoms were associated with parental depression
symptoms in an exploratory regression model. Finally, parents with more mental health
problems may rate their children’s behavior more negatively than parents with fewer
mental health problems. However, research that has accounted for parental reporting
errors has still suggested a positive association between parental mental health disturbances
and child behavior problems (Vallotton et al., 2016). Regardless of the specific mechanism,
our findings suggest that parental mental health and child behavioral problems changed
together during the pandemic.
Finally, in our study, inconsistent with the differential susceptibility hypothesis,
prematurity did not moderate change in parental mental health, change in child behavior,
or the effect of change in parental mental health on change in child behavior. The
children in our study were older than the children in other studies that found support
for the differential susceptibility hypothesis (Gueron‐Sela et al., 2015). Perhaps
by the time preterm children are preschool age, the strength of their susceptibility
to environmental factors dissipates. Although many studies have documented prematurity-related
behavioral difficulties well into childhood (Arpi & Ferrari, 2013), other studies
have similarly failed to find support for an interaction between parental mental health
and child outcomes in older preterm children (Hadfield et al., 2017). It is also possible
that, because the majority of our sample was from a higher socioeconomic status, they
benefited from resources that buffered the effects of prematurity on parental mental
health and child behavior (ElHassan et al., 2018). Finally, our sample was small,
and we may have been underpowered to detect statistically significant interaction
effects. Although we found similar results when using continuous gestational age rather
than binary categorical prematurity, some studies have reported significant interactions
when using more than two categories of prematurity (e.g., very early preterm, early
preterm, moderate preterm, late preterm, and term) (ElHassan et al., 2018). We did
not have the power to parse our preterm sample beyond two categories.
Our study has limitations that should be addressed in future studies. First, the subsample
that participated in T2 was small, although the response rate was similar to other,
larger studies on the effect of the pandemic (Patrick et al., 2020). Second, our sample
was predominantly White and of a high socioeconomic status, limiting the generalizability
of our findings. However, of note, the fact that we observed pandemic-related changes
in parental mental health and child behavior even in a relatively advantaged population
highlights the widespread influence of the pandemic. Third, there was variability
in days elapsed between T1 and T2. The stressors associated with the pandemic have
been dynamic, so the time at which the parent completed questionnaires may have influenced
their answers. Due to the small sample for T2 and the variability in days elapsed
between T1 and T2, we were not able to account for the various aspects of the pandemic
that may have influenced answers (e.g., school closures, mask mandates). Future studies
should pinpoint which aspects of the pandemic may be most strongly associated with
changes in parental mental health and child behavior. Finally, we relied on self-report
data for both parent and child measures; however, we had data prior to the pandemic,
so we were able to avoid recall bias.
To our knowledge, this was the first study to examine the relation between parental
mental health and child behavior in a group of children with biological risk during
a global stressor. We found increases in parental depression symptoms, child internalizing
symptoms, and child externalizing symptoms and decreases in parental well-being. Increases
in parental depression symptoms across time were associated with increases in child
internalizing and externalizing symptoms across time. This relation was not moderated
by prematurity. Our findings have the potential to inform prevention and intervention
efforts aimed at equipping children with the behavioral resources necessary to thrive.
Uncited references
Achenbach, xxxx, Gueron-Sela et al., 2015, Institute of Medicine (US), 2007.
CRediT authorship contribution statement
Francesca Scheiber: Conceptualization, Methodology, Formal analysis, Investigation,
Data curation, Writing – original draft, Visualization, Funding acquisition. Paige
M. Nelson: Conceptualization, Methodology, Formal analysis, Investigation, Data curation,
Writing – review & editing, Visualization, Funding acquisition. Allison Momany: Conceptualization,
Methodology, Formal analysis, Investigation, Data curation, Writing – review & editing,
Visualization, Funding acquisition. Kelli K. Ryckman: . Ö. Ece Demir-Lira: Conceptualization,
Methodology, Formal analysis, Writing – review & editing, Visualization, Funding acquisition,
Supervision.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal
relationships that could have appeared to influence the work reported in this paper.