Differential impact of COVID-19 lockdown on physical activity in younger and older adolescents – prospective study

Abstract

Background:
Insufficient physical activity levels (PAL) during adolescence is a major public health concern, which is even more pronounced during the COVID-19 pandemic due to restricting movement opportunities. This study aimed to identify PAL changes and examine the age-specific determinants of PAL in younger and older adolescents during the COVID-19 lockdown.

Material and methods:
This study included 859 high-school students from Bosnia and Herzegovina stratified into 2 age groups: younger (N = 420, 14-16 years of age), and older adolescents (N = 439, 16-18 years of age). Participants were tested over 2 testing waves: before the COVID-19 lockdown (January 2020) and during the COVID-19 lockdown (April 2020). Variables included PAL assessed by the Physical Activity Questionnaire for Adolescents, sports factors, and parental factors.

Results:
The PAL significantly declined as a result of lockdown measures in the total sample (from 2.76±0.79 to 2.50±0.82). Larger absolute and relative decline of PAL was evidenced in younger adolescents. Sport participation positively influenced PAL before lockdown, with no significant influence during the lockdown. Older adolescents whose mothers were better educated were less likely to be in high risk group with regard to a large decline of PAL as a result of COVID lockdown (OR = 0.50, 95% CI: 0.21-0.84).

Conclusions:
Results of the study suggest that parental education influences health-related behaviors and that parental education is a protective factor against a decrease in PAL during the COVID-19 pandemic. Main educational agents (i.e., school and parents) should pay more attention to provide children and adolescents adequate information and develop their health literacy, which will hopefully positively impact children's PAL even in challenging situations similar to COVID-19 lockdown. Med Pr. 2021;72(6).

Full text

DIFFERENTIAL IMPACT OF COVID-19 LOCKDOWN ON PHYSICAL ACTIVITY
IN YOUNGER AND OLDER ADOLESCENTS – PROSPECTIVE STUDY
Marijana Geets Kesic1, Barbara Gilic1,2, Ivana Cerkez Zovko3, Patrik Drid4, Darinka Korovljev4, Damir Sekulic1
1 University of Split, Split, Croatia
Faculty of Kinesiology
2 University of Zagreb, Zagreb, Croatia
Faculty of Kinesiology
3 University of Mostar, Mostar, Bosnia and Herzegovina
Faculty of Science and Education
4 University of Novi Sad, Novi Sad, Serbia
Faculty of Sport and Physical Education
A
Background: Insucient physical activity levels (PAL) during adolescence is amajor public health concern, which is even more
pronounced during theCOVID-19 pandemic due to restricting movement opportunities. is study aimed to identify PAL changes
and examine theage-specic determinants of PAL in younger and older adolescents during theCOVID-19 lockdown. Material
and Methods: is study included 859 high-school students from Bosnia and Herzegovina stratied into 2 age groups: younger
(N = 420, 14–16 years of age), and older adolescents (N = 439, 16–18 years of age). Participants were tested over 2 testing waves:
before the COVID-19 lockdown (January 2020) and during the COVID-19 lockdown (April 2020). Variables included PAL assessed
by the Physical Activity Questionnaire for Adolescents, sports factors, and parental factors. Results: ePAL signicantly declined
as aresult of lockdown measures in thetotal sample (from 2.76±0.79 to 2.50±0.82). Larger absolute and relative decline of PAL was
evidenced in younger adolescents. Sport participation positively inuenced PAL before lockdown, with no signicant inuence
during thelockdown. Older adolescents whose mothers were better educated were less likely to be in high risk group with regard to
alarge decline of PAL as aresult of COVID lockdown (OR= 0.50, 95% CI: 0.21–0.84). Conclusions: Results of the study suggest that
parental education inuences health-related behaviors and that parental education is aprotective factor against adecrease in PAL
during theCOVID-19 pandemic. Main educational agents (i.e.,school and parents) should pay more attention to provide children
and adolescents adequate information and develop their health literacy, which will hopefully positively impact children’s PAL even
in challenging situations similar to COVID-19 lockdown. MedPr.2021;72(6)
Key words: pandemic, youth, sport participation, familial factors, physical activity, health literacy
Corresponding author: Damir Sekulic, University of Split, Faculty of Kinesiology, Teslina 6, 21000 Split, Croatia, e-mail: dado@kifst.hr
Received: July 14, 2021, accepted: October 1, 2021
ORIGINAL PAPER
https://doi.org/10.13075/mp.5893.01180
INTRODUCTION
Engaging in physical activity (PA) in adolescence is one
of themain determinants of health status because it has
been reported that PA can prevent many non-communi-
cable diseases. Specically, sucient PA in adolescence
is associated with areduced risk of metabolic syndrome,
obesity, coronary heart disease, and mental disorders [1].
However, it is aworrying fact that 81% of adolescents
do not have adequate PA levels (PAL) globally, which is
amajor public health concern [2]. Specically for Bosnia
and Herzegovina, recent study recorded that chil-
dren aged 14 years had 18% lower PAL compared with
10-year-old children, indicating alarming decline of PAL
from childhood to adolescents, and thefact that that ad-
olescents should be considered as vulnerable group for
decreasing PAL [3]. erefore, in order to be able to cre-
ate adequate programs to increase PAL, numerous stud-
ies have examined which factors inuence PAL among
adolescents. Inbrief, previous studies identied various
correlates of PA in adolescents including demograph-
ic factors (gender and age), behavioral factors (previous
PA, sports participation, smoking and drinking alcohol),
social factors (parental and friends support), and envi-
ronmental factors (population density and living envi-
ronment) [4]. Precisely, positive associations with PA
were continuously conrmed for themale gender, urban
living environment, sports participation, parental edu-
cation, and parental and friends support [5].
Adolescence is theperiod in which themost pro-
nounced decrease in PAL in life span occurs. Inare-
view study involving children and adolescents aged
Medycyna Pracy2021;72(6)
http://medpr.imp.lodz.pl/en

2 M. Geets Kesic etal. Nr 6
10–19 years, it was recorded that thepeak in PAL oc-
curs at age 13, and PAL later begins to decrease by 7%
annually [6]. Given that adecrease in PAL was record-
ed in pre-adolescents aged 10–14 years [3], due to dra-
matic changes in life habits (i.e.,drop-out from sports,
social inuences, increased scholastic duties), an even
greater decrease in PAL was expected in older adoles-
cents. Supportively, astudy by Miljanovic Damjanovic,
etal. [7] conrmed this hypothesis. Prospective studies
in southeastern Europe additionally conrmed thede-
cline in PAL in adolescents from therst to thesecond
grade of high school (14–15 years of age)[8]; rst to the
second grade of high school (14–16 years of age) [9];
and third to the fourth grade of high school (16–18
years of age) [10]. Also, age-specic correlates of PAL
in younger and older adolescents were observed.
Specically, positive changes in the PAL of younger
adolescents were determined mainly by higher paren-
tal education [9], while parental conict and therural
living environment determined alarger decline in PAL
among older adolescents [11].
e proclamation of the COVID-19 pandemic in
March 2020 has led to theintroduction of social dis-
tancing measures and reduced movement opportunities.
Precisely, stay-at-home restrictions, school closures, and
cancellation of sports club practices led to aworldwide
decrease in PAL among children and adolescents. Not
surprisingly, studies in southeastern Europe recorded
adecrease in PAL in adolescents during COVID-19 lock-
down [12,13]. Apart from evidencing thechanges in PAL,
authors examined factors that have inuenced changes
in PAL among adolescents. Briey, agreater decrease in
PAL was observed in boys than in girls; thereduction in
PAL was greater in urban than in rural adolescents, con-
ict with parents determined alarger decline in PAL, ad-
olescents with better tness status before pandemic had
higher PAL during thepandemic [13], adolescents that
participated in sports activities before thepandemic were
more likely to have higher PAL during the pandemic
[12], while adolescents who smoked cigarettes had lower
PAL levels during thepandemic [12,13].
From theprevious literature review, it is evident that
studies extensively examined changes in PAL as a result
of theCOVID-19 pandemic and factors which could inu-
ence such negative changes. However, practically all stud-
ies done so far examined adolescents of arelatively large
age span (i.e.,14–18 years old, high-school students). More
precisely, previous studies done in regular circumstanc-
es conrmed that younger adolescents had higher PAL
than older adolescents and that dierent factors inuence
changes in PAL among older and younger adolescents [11].
Supportively, astudy on Chinese adolescents noted that
younger adolescents (12–15 years) had higher PAL than
older adolescents (15–18 years) during the COVID-19
pandemic [14]. Finally, it seems that trends of higher PAL
in younger age groups is characteristic even for lockdown
period, since recent study evidenced that Canadian chil-
dren (5–11 years) were more physically active than youth
(12–17 years) during lockdown [15].
Collectively, there is alack of prospective data about
changes in PAL during theCOVID-19 pandemic in dif-
ferent age groups of adolescents (i.e.,younger adolescents
aged 14–16 years vs. older adolescents aged 16–18 years).
Also, to thebest of the authors’ knowledge, no study ex-
amined factors that may inuence changes in PAL, which
occurred during theCOVID-19 pandemic, specically in
younger vs. older adolescents. erefore, theaim of this
study was to examine thechanges in PAL which occurred
as a result of theCOVID-19 pandemic, and to exam-
ine the age-specic socio-demographic-, familial-, and
sport-factors as correlates of PAL among younger and
older adolescents during theCOVID-19 lockdown. e
authors hypothesized that observed predictors will pos-
itively inuence PAL of studied adolescents before and
during theCOVID-19 lockdown, irrespective of partici-
pants’ age. is data is vital in developing targeted inter-
ventions in terms of improving PAL during COVID-19
lockdown and similar crises in dierent age groups.
MATERIAL AND METHODS
Participants and design of thestudy
is research included 859 students attending high
schools from several regions in Bosnia and Herzegovina
(Tuzla, Herzegovina-Neretva, and West Herzegovina
Canton). For thepurpose of this study, thetotal sample
was stratied into two subsamples: younger adolescents
(N= 420, 14–16 years of age at study baseline) and old-
er adolescents (N= 439, 16–18 years of age at study base-
line). Before thestudy initiation, students regularly at-
tended physical education classes at least twice aweek.
e response rate (based on responses at follow-up –
please see later for details on design), was 86%. All par-
ticipants gave their informed consent, and aparent or
alegal guardian gave their approval for those participants
younger than 18 years. einvestigation was approved
by theEthical Board of theUniversity of Split, Faculty of
Kinesiology, Croatia (EBO: 2181-205-05-02-05-14-005).
is prospective study was commenced over 2 test-
ing waves: 1) baseline, representing the measurement

Nr 6 COVID-19 and physical activity in adolescents 3
before implementation of the social distancing mea-
sures due to theCOVID-19 pandemic (January 2020),
and 2) follow-up, representing themeasurement during
theperiod when social distancing measures were im-
plemented (i.e.,COVID-19 lockdown; late April 2020).
Baseline measurement included socio-demographic
factors, parental/familial factors, sport factors, and
baseline PAL. Follow-up measurement was conduct-
ed during thelockdown and included only follow-up
PAL. Inorder to preserve participants’ anonymity and
collect more honest and precise answers in theques-
tionnaire, students were asked to create their personal
code and use it at baseline and follow-up measurement.
edesign of thestudy and detailed information about
thestudy period are presented in Figure 1.
Variables
is study included socio-demographic factors (age,
gender), parental/familial factors, sport factors (inde-
pendent variables), and PAL (dependent variable).
e Physical Activity Questionnaire for Adolescents
(PAQ-A) was used for assessing baseline and follow-up
PAL [16]. PAQ-A has been frequently used for assess-
ing PAL at similar samples of adolescents and has dis-
played appropriate validity and reliability [7,10]. It is
aself-administered questionnaire that assesses thefre-
quency of PA during thelast 7 days. Itconsists of 9items,
rst 8 assess PA during spare time, during physical edu-
cation classes, aer school, during evenings, and during
theweekend, and are scaled 1–5, representing low and
high PA, respectively. e 9th item questions wheth-
er participants were injured or ill and is not included
in thenal score, which is calculated as thearithmetic
mean of therst 8 items. Inthis study, the authors ob-
served crude results of PAL at baseline (PALBL), and PAL
at follow-up (PALFU). Next, crude PAL was also observed
as thebinomial variable with 2 categories: results; insuf-
cient/inadequate vs. sucient/adequate PAL (scores
≤2.73 and >2.73, respectively), as suggested previous-
ly [17]. Further, to quantify thechanges in PALBL and
PALFU, the authors calculated crude numerical dier-
ence between these 2 values (PALΔ = PALBL – PALFU).
Next, the authors calculated relative changes in PAL be-
tween baseline and follow-up (in %) using thecalcula-
tion: PALΔ%= (PALBL– PALFU) / PALBL × 100. For thepur-
pose of later statistical calculations, participants were
ordered according to their PALΔ%, and then grouped into
2 groups (0–50percentiles, and above). e participants
with agreater relative decline of PAL (ordered >50per-
centiles) were evidenced as “high risk group”, while those
placed in therst 50 percentiles were evidenced as “low
risk group”. Such dichotomization allowed the authors to
calculate logistic regression for PALΔ% as abinomial cri-
terion (please see later for details).
Parental/familial factors included thefollowing vari-
ables: self-reported socioeconomic status (below av-
erage–average–above average); maternal and paternal
education level (university degree-college degree-high
school-elementary school); conict with parents (nev-
er–rarely–from time to time-regularly/frequently); pa-
rental absence from home (never–rarely–from time to
time-regularly/frequently); and parental/familial care
(very poor care–low care–my parents/family care about
me–my parents/family care about me alot).
Sports factors were assessed by questions on partici-
pation in individual or team sports (with answers: nev-
er been involved–quit–currently involved), best results
in competitions (with answers on a4-point scale rang-
ing from “never competed” to “international compe-
titions”), and years of involvement in sports (with an-
swers: never involved–less than one year–two to ve
years–more than ve years).
JAN
2020
January 6–12
Baseline PAL
Parental factors
Sport factors
FEB
2020
February 11
WHO named
COVID-19
MARCH
2020
March 9
First confirmed
cases of COVID-19
in the Federation
of Bosnia
and Herzegovina
March 16
Government
implemented
restrictive
preventive
measures
APRIL
2020
April 9
WHO classified
Bosnia
and Herzegovina
as a country with
potential for community
transmission
April 20–26
Follow-up PAL
PAL – physical activity levels.
Figure 1. Timeline of thestudy done in Bosnia and Herzegovina and most important dates in rst half of 2020

4 M. Geets Kesic etal. Nr 6
Statistics
Normality of thedistribution was checked by Kol mo-
gorov Smirnov’s test, and descriptive statistics includ-
ed means and standard deviations (for numerical vari-
ables) and frequencies and percentages (for ordinal and
nominal variables).
Dierences between groups were evidenced by
t-test for independent samples (for quantitative vari-
ables), and Chi square (χ2). T-test for dependent sam-
ples was used to identify thechanges in PAL between
baseline (pre-pandemic period) and follow-up (lock-
down period).
Logistic regression (with Odds Ratio – OR, and
Condence Interval– CI reported) was applied to ev-
idence theassociation between predictors and catego-
rized PAL (insucient PAL– coded as “1” vs. sucient
PAL– coded as “2”) at baseline and follow-up.
Also, in order to identify theassociation which may
exist between predictors and changes that occurred in
PAL, the authors additionally calculated logistic regres-
sion with dichotomized PALΔ% as acriterion (low-risk
group - coded as “1” vs. high-risk group – coded as “2”).
All analyses were stratied for age groups
(e.g.,younger vs. older adolescents).
A p-value of 95% was applied, and theStatistica ver.
13.5 statistical package (Tibco Inc., CA) was used for all
calculations.
RESULTS
Age groups signicantly diered in self-perception of
socioeconomic status, with somewhat better familial so-
cioeconomic status among older adolescents (χ2= 9.05,
p < 0.01). Younger adolescents were more involved in
individual sports (χ2= 25.76, p < 0.01), and team sports
(χ2= 35.71, p < 0.01) than older adolescents, while sport
achievement was higher in older adolescents (χ2= 23.76,
p < 0.01) at baseline testing (Table1).
At baseline, 55% of younger and 43% of older ado-
lescents had sucient PAL, while only 35% of younger
and 28% of older adolescents had sucient PAL at fol-
low-up (during lockdown). Observing in crude PAQ-A
results, thePAL signicantly declined as aresult of lock-
down measures in thetotal sample (from 2.76±0.79 to
2.50±0.82, p < 0.001), and separately in older adoles-
cents (from 2.52±0.68 to 2.31±0.74, p < 0.01), and in
younger adolescents (from 3.02±0.82 to 2.71±0.85),
p < 0.001. Younger adolescents had higher PAL than
their older peers at baseline (before lockdown) and
follow-up (during lockdown) (Figure 2). Signicant
dierences between age groups were evidenced in PALΔ
(t-test= 1.98, p < 0.01), with larger decrease in young-
er adolescents, than in older adolescents (0.31±0.80 and
0.21±0.68 for younger and older adolescents, respec-
tively), with signicant dierence in PALΔ% between age
groups (t-test= 1.76, p= 0.04).
Table2 presents results of logistic regression analysis
for dichotomized PAL at baseline and follow-up as cri-
terion variable (insucient vs. sucient PAL). Inbrief,
for thepre-pandemic period, sucient PAL was more
likely to be achieved in males (for both age groups), in
adolescents whose fathers were better educated (on-
ly among younger adolescents), in adolescents whose
mothers were better educated (for both age groups), in
adolescents who participated in individual sports (for
both age groups), and those who participated in team
sports (for both age groups). Sucient PAL during
the lockdown period was more likely to be achieved
in males (for younger adolescents), adolescents whose
mothers were better educated (for theolder group), ad-
olescents who declared better parental care (for theold-
er group), those who participated in individual sports
(for younger age group), and those who achieved better
sport success (for younger age group).
Figure 3 presents associations between studied predic-
tors and PALΔ%, separately for younger and older adoles-
cents. eonly signicant predictor was thelevel of ma-
ternal education. More precisely, older adolescents whose
total 14–16 years old 16–18 year old
Participants
0
1
2
3
4
PAL score
*
¥
*
*
¥
pre-pandemic lockdown
Period:
Signicant (p < 0.001) dierences between (¥) and within groups (*).
Figure 2. Descriptive statistics, changes and t-test dierences
inphysical activity levels (PAL) before and during theCOVID-19
lockdown in Bosnia and Herzegovina in 2020

Nr 6 COVID-19 and physical activity in adolescents 5
Table1. Descriptive statistics and dierences between age groups of adolescents from Bosnia and Herzegovina in early 2020
Var ia b l e
Participants
(N= 859)
χ2p
14–16 years old
(N= 420)
16–18 years old
(N= 439)
n%n%
Socio-demographic and familial/parental status
gender 0.79 0.37
males 244 58.10 239 54.44
females 176 41.90 195 44.42
socio-economic status 9.05 0.01
below average 46 10.95 28 6.38
average 368 87.62 405 92.26
above average 1 0.24 6 1.37
paternal education 0.07 0.96
elementary school 28 6.67 31 7.06
high school 303 72.14 317 72.21
college/university degree 89 21 91 21
maternal education 0.19 0.90
elementary school 98 23.33 108 24.60
high school 269 64.05 276 62.87
college/university degree 53 12.62 55 12.53
parental conict 0.12 0.98
never 162 38.57 174 39.64
rarely 162 38.57 165 37.59
from time to time 82 19.52 85 19.36
oen 14 3.33 15 3.42
parental absence 0.23 0.97
never 78 18.57 87 19.82
rarely 92 21.90 95 21.64
from time to time 157 37.38 163 37.13
oen 92 21.90 93 21.18
parental care 0.05 0.99
very poor care 6 1.43 11 2.51
low care 11 2.62 6 1.37
they care about me 144 34.29 148 33.71
they care alot 259 61.67 274 62.41
Sport factors at baseline testing
individual sport participation 25.76 0.01
currently involved 112 26.67 94 21.41
quit 83 19.76 156 35.54
never been involved 221 52.62 189 43.05

6 M. Geets Kesic etal. Nr 6
mothers were better educated were less likely to be in “high-
risk group” with regard to alarge decline of PAL as aresult
of COVID lockdown (OR= 0.50, 95% CI: 0.21–0.84).
DISCUSSION
e main aim of this study was to investigate theimpact
of mainly scholastic factors on PAL, and changes in PAL
during theCOVID-19 pandemic in younger and older
adolescents. Apart from theexpected decrease of PAL as
aresult of COVID-19 lockdown, there are several most
important ndings of this study. First, alarger absolute
and relative decrease of PAL was evidenced in younger
adolescents. Second, maternal education was astronger
factor of inuence on PAL in older (16–18 years) ado-
lescents. ird, alower decrease of PAL was evidenced
in older adolescents whose parents were better educat-
ed. erefore, the authors’ initial hypothesis might be
partially accepted.
PAL changes in younger and older adolescents
e total sample of participants declined their PAL.
is is in accordance with previous studies conducted
on adolescents globally, including theadolescents from
Bosnia and Herzegovina and neighboring countries
[12–14,18]. However, thenovel nding of this study is
that younger adolescents had higher PAL than older ad-
olescents in thepre-COVID period, with no signicant
dierence between age groups in PAL for thepandemic
period. Almost certainly, thelack of dierences in PAL
between age groups during COVID-19 lockdown can be
attributed to alarger decline in PAL due to COVID-19
Var ia b l e
Participants
(N= 859)
χ2p14–16 years old
(N= 420)
16–18 years old
(N= 439)
n%n%
Sport factors at baseline testing – cont.
team sport participation 35.71 0.01
currently involved 168 40.00 102 23.23
quit 106 25.24 181 41.23
never been involved 144 34.29 156 35.54
sport experience 7.70 0.07
never been involved 109 25.95 106 24.15
<1 year 67 15.95 101 23.01
2–5 years 150 35.71 129 29.38
>5 years 89 21.19 103 23.46
sport achievement 23.76 0.01
never competed 193 45.95 215 48.97
lower rank competition 149 35.48 190 43.28
regional competition 50 11.90 25 5.69
national/international level 21 5.00 5 1.14
Table1. Descriptive statistics and dierences between age groups of adolescents from Bosnia and Herzegovina in early 2020 – cont.
sport achievement
sport experience
team sport
individual sport
parental care
parental absence
parental conflict
maternal education
paternal education
socio-economic status
male gender
012
OR (95% CI)
3
14–16 years
Participants age:
16–18 years
Figure 3. Associations between independent variables and relative
changes in physical activity levels (PALΔ%) in younger and older
adolescents (with group of higher risk for large decrease of PAL
asreference category) from Bosnia and Herzegovina in 2020

Nr 6 COVID-19 and physical activity in adolescents 7
lockdown in younger adolescents. On theother hand,
both ndings (higher pre-pandemic PAL, and greater
decrease of PAL in younger adolescents) may be associ-
ated with participation in sports.
In brief, the results of the study conrmed higher in-
volvement in sports among younger adolescents (please
see Results for details). e participation in sports con-
tributes to alarge part of thetotal PAL in children and
adolescents [19]. Itwas repeatedly reported that adoles-
cents involved in organized sports activities have sig-
nicantly higher PAL than their peers who do not par-
ticipate in sports [20]. Further, it is known that there is
thelargest sport drop-out rate at theage of 14–15 years
[21]. Evidenced dierences in sports participation be-
tween younger and older adolescents are likely to occur
due to increased commitments, lack of free time, focus
on academic achievement [6], socialization with friends
in sedentary activities, and employment in older ado-
lescents [20].
It is also important to note that thehigh-school period
is known for dropping out of sport as adolescents are
mainly oriented towards competitive success in sports.
Atthe same time, adolescents are not suciently edu-
cated about theimportance of PA and its’ positive im-
pact on health [22]. Supportively, astudy on adolescents
from Portugal suggested that thedegree of participation
in sports is related to competitive sport achievement,
which limits participation only to students who have bet-
ter physical predispositions and who are talented [23].
Meanwhile, another Portuguese study highlighted insuf-
cient education about theimportance of theimpact of
PA on health as an even more likely cause of giving up
sports [24].
Collectively, older adolescents in this study have
lower PAL most likely as aconsequence of dropping
out from organized sports, compared to younger ad-
olescents who are not yet at thepoint of thedecision
to continue or stop sports participation. Putting it al-
together, it is absolutely logical that younger adoles-
cents reduced their PAL to a greater extent during
theCOVID-19 lockdown simply because they did not
have theopportunity to participate in sports activities
during follow-up (i.e.,sports clubs, facilities, and or-
ganized recreational activities were closed and banned
during thelockdown period). Finally, as younger ado-
lescents generally had higher PALs than older adoles-
cents in thepre-lockdown period, it is understandable
that their PAL decreased to alarger extent than thePAL
of their older peers.
Parental factors and PAL
before and during lockdown
Parental education is known to be an important factor
of inuence on children’s PAL [25], and the authors’ re-
sults of the study support such considerations. eas-
sociation can be explained through thebackground of
general parents’ inuence on thebehavioral patterns of
their children. Precisely, parents with their attitudes,
Table2. Associations between independent variables and physical activity levels (sucient physical activity– reference group)
before and during COVID-19 lockdown in younger and older adolescents from Bosnia and Herzegovina in early 2020
Var ia b l e
Pre-pandemic period Lockdown period
14–16 year old participants
(N= 420)
16–18 year old participants
(N= 439)
14–16 year old participants
(N= 420)
16–18 year old participants
(N= 439)
OR 95% CI OR 95% CI OR 95% CI OR 95% CI
Male gender 1.25 1.09–1.4 1.2 1.01–1.44 1.11 1–1.21 1.09 0.88–1.38
Socio-economic status 0.99 0.5–1.5 0.98 0.6–1.42 0.99 0.44–1.6 1.02 0.55–1.59
Paternal education 1.61 1.25–1.88 1.44 0.98–1.88 1.32 0.8–1.76 1.3 0.45–2.32
Maternal education 1.55 1.21–1.86 1.74 1.44–2.05 1.32 0.88–1.71 1.66 1.11–2.2
Parental conict 0.78 0.44–1.11 0.8 0.5–1.1 0.88 0.45–1.11 0.81 0.32–1.44
Parental absence 0.75 0.44–1.09 0.8 0.61–1.02 0.91 0.77–1.33 0.95 0.46–1.41
Parental care 0.89 0.65–1.21 0.9 0.7–1.11 1.21 0.98–1.56 1.31 1.02–1.61
Individual sports 2.09 1.5–2.61 1.66 1.33–1.99 1.45 1.03–1.88 1.21 0.81–1.65
Team sports 2.21 1.95–2.55 2 1.75–2.28 1.28 0.71–1.76 1.11 0.45–1.88
Sport experience 1.92 1.71–2.21 1.44 0.97–1.98 1.09 0.51–1.65 1 0.87–1.36
Sport achievement 1.84 1.11–2.5 1.22 0.74–1.75 1.54 1.11–1.91 1.21 0.66–1.89

8 M. Geets Kesic etal. Nr 6
values, and knowledge guide their children to form
their own behavioral patterns, including participating
in any form of PA [25]. Parents are acting as role models
to their children, meaning that parents who are physi-
cally active and who promote PA with their behavior
positively inuence PA habits of their children.
is phenomenon occurs according to the theo-
ry that children imitate parents and their habits, which
means that children are likely to adopt PA habits as well.
Besides theimportance of role modeling and parental
encouragement, attitudes and tangible support are con-
sidered as important determinants of parental inuence
on children’s PAL. us, parents with higher education-
al levels may better perceive thehealth-related benets
of PA, which will provoke greater parental support and
encouragement for PA and inuence theoverall forma-
tion of healthy habits of their children.
Previous studies noted somewhat dierent inu-
ences of maternal and paternal education on their chil-
dren’s PAL. Specically, fathers are more likely to act as
role models for PAL and sports activities compared with
mothers. Furthermore, fathers more frequently engage
in PA and are actively involved in PA with their chil-
dren in terms of co-physical activities [26]. On theother
hand, mothers have agreat role in their children’s educa-
tional achievement, which includes health-related educa-
tion. Supportively, astudy on Nigerian children noted that
mothers with higher education levels had better knowl-
edge of PA [27]. Itis assumed that better educated and in-
formed mothers possess higher health literacy skills that
they transmit to their children and create apositive inu-
ence on making healthier decisions and encouraging chil-
dren to participate in PA and other health behaviors [28].
erefore, the authors’ nding that maternal edu-
cation positively inuences PAL during theCOVID-19
lockdown in older adolescents could be explained in
thecontext of thesituation and period when thestudy
was conducted. Namely, during thepre-pandemic pe-
riod, adolescents regularly participated in sports, and
consequently, fathers probably had astronger inuence
on their PAL. Meanwhile, during theCOVID-19 lock-
down, adolescents were required to stay at home due to
movement restrictions (school and sports facilities clo-
sure). erefore, they were probably more exposed to
thecontrol and inuence of mothers.
e study recorded that the inuence of maternal
education on PAL in children was more pronounced
in older adolescents. A recent study evidenced that
younger adolescents (14–16 years of age) with better-
educated mothers and fathers had ahigher likelihood
of having sucient PAL in regular circumstances [9].
However, astudy on older adolescents (16–18 years of
age) evidenced that paternal education was correlated
with adolescents’ PAL aged 16 years old, but at theage
of 17 years, the correlation started to weaken [11].
Moreover, prospective studies on older adolescents
(16–18 years old) reported that adolescents with more
educated mothers are more likely to incline PAL during
late adolescence [7,10]. is altogether indicates that
theinuence of paternal education on their children’s
PAL weakens during late adolescence, while thelevel of
maternal education remains the signicant inuential
factor on PAL even in late adolescence.
is is one of therst studies to prove that paren-
tal education acts as apreventive factor against ade-
crease of PAL during aCOVID-19 lockdown. ere-
sults of this study can be supported by recent research
on Spanish children, which recorded a more evident
PAL decline during the COVID-19 lockdown in chil-
dren (10–14 years) whose mothers were less-educated
[29]. However, the authors’ results of this study extend
theprevious knowledge regarding factors of inuence on
changes in PAL during theCOVID-19 lockdown, since
the authors investigated older children/adolescents (16–
18 years of age) than authors’ Spanish colleagues, and
showed asignicant inuence of both maternal and pa-
ternal educational levels on changes in PAL.
e fact that maternal education prevents the de-
cline of PAL leads the authors to conclude that children
of better-educated mothers are physically active more
because of health-related issues and not because of in-
volvement in competitive sports. Also, it is reasonable
to hypothesize that older adolescents had better “health
literacy” and that they knew how important it was to
maintain appropriate PAL even during thelockdown. If
the authors dene health literacy as “the ability to selec-
tively access and analyze information and to take action
to promote personal and someone else’s health behav-
iors” [30], thepreventive eect of better maternal edu-
cation on thedecrease of PAL during lockdown is un-
derstandable.
Supportively, previous studies noted that adoles-
cents who had better tness status and were involved in
sports before thepandemic eciently maintained PAL
during thepandemic [12,13]. eauthors explained this
nding with theterm “physical literacy” precisely, with
theassumption that adolescents of advanced tness sta-
tus have sucient knowledge and motivation to create
exercise programs themselves even with limited equip-
ment and space during thelockdown. Health literacy

Nr 6 COVID-19 and physical activity in adolescents 9
and physical literacy, although not thesame, are con-
sidered to be lifelong health-promoting behaviors [30].
erefore, it is likely that adolescents whose mothers
were better educated had better health literacy, recog-
nized theimportance of PA for their health, and strived
to maintain PAL even during COVID-19 lockdown.
Considering all theabove mentioned, it is not surpris-
ing that the authors have found astronger inuence of
maternal education on PAL in older than in younger ado-
lescents. Itcan be assumed that mothers inuence adoles-
cent habit formation and that older adolescents will more
likely recognize and use it. On theother hand, younger
adolescents are still in theprocess of forming their hab-
its and may not yet be able to identify and choose what
is good for them in terms of healthy behavior. Inother
words, younger children are more likely to participate in
PA because of fun and parental supervision, while older
children are more likely to participate in PA because they
know what thehealth benets of thePA are [10].
Limitations and strengths
Most of thevariables were self-reported, which could be
considered as themain limitation of thestudy. Precisely,
students might answer some questions not completely
honestly and provide more socially desirable answers.
However, participants lled in questionnaires anony-
mously (for thepurpose of repeated testing they use
condential codes); therefore, leaning towards social-
ly desirable answers was less likely. Inthis research re-
sults on PAQ-A were clustered into 2 groups, which un-
doubtedly allocated participants with similar results
into opposed groups. Inthefuture studies division into
three groups is therefore suggested.
is is one of the rst studies investigating the
changes and correlates of PAL during theCOVID-19
lockdown in relation to age groups of adolescents
(younger and older adolescents). erefore, this study
expands theknowledge on physical activity in adoles-
cents during theCOVID-19 pandemic and other simi-
lar situations and creates thebackground for future re-
search. Additionally, this study used a measurement
instrument (PAQ-A) which is widely used in studies on
similar participants; therefore, theresults of this study
could be compared with other similar studies.
CONCLUSIONS
e greater decline in PAL during theCOVID-19 lock-
down was evidenced in younger adolescents com-
pared to older adolescents, which was attributed to
higher pre-pandemic PAL of younger adolescents and
younger adolescents’ higher engagement in sports in
thepre-pandemic period.
is study’s results can be considered great value for
creating public-health policies for improving PAL in
adolescents in regular circumstances and in situations
similar to COVID-19 lockdown. Specically, in order
to achieve proper PAL, younger adolescents should be
encouraged to continue participating in sports with an
emphasis on thehealth benets of PA and not on com-
petitive achievements.
Schools or sports clubs should promote sports as
health-enhancing behavior and teach children to partic-
ipate in sports for their health regardless of competitive
achievement. Itwill have direct positive repercussions
on PAL in situations similar to COVID-19 lockdown
when circumstances will limit regular sports participa-
tion. Insuch situations, adolescents with better health
literacy will be able to recognize theimportance of PA
and will try to maintain sucient PAL. Further studies
are needed in order to:
■
evaluate aspecic insight into various components
of PAL during lockdown (i.e., physical education,
sports, and free time),
■
overview thechanges which occur in each compo-
nent, and
■
investigate the factors of inuence on changes in
each component.
ACKNOWLEDGMENTS
Authors are particularly grateful to all children who volun-
tary participated in thestudy. Special thanks goes to school
authorities who approved theinvestigation and helped in re-
search.
REFERENCES
1. Pavičić ŽeželjS, Kenđel JovanovićG, KrešićG. easso-
ciation between the Mediterranean diet and high physi-
cal activity among theworking population in Croatia. Med
Pr. 2019;70(2):169–176. https://doi.org/10.13075/ mp. 5893.
00773.
2. GutholdR, StevensGA, Riley LM, BullFC. Global trends
in insucient physical activity among adolescents: apooled
analysis of 298 population-based surveys with 1.6 million
participants. Lancet Child Adolesc Health. 2020;4(1): 23–35.
https://doi.org/10.1016/s2352-4642(19)30323-2.
3. PojskicH, EslamiB. Relationship Between Obesity, Physical
Activity, and Cardiorespiratory Fitness Levels in Children

10 M. Geets Kesic etal. Nr 6
and Adolescents in Bosnia and Herzegovina: An Analysis
of Gender Dierences. Front Physiol. 2018;9:1734.
https://doi.org/10.3389/fphys.2018.01734.
4. Sallis JF, Prochaska JJ, TaylorWC. Areview of correlates
of physical activity of children and adolescents. Med Sci
Sports Exerc. 2000;32(5):963–975. https://doi. org/ 10. 1097/
00 0 0 5768-200005000-00014.
5. Bauman AE, Reis RS, Sallis JF, Wells JC, Loos RJ,
MartinBW. Correlates of physical activity: why are some
people physically active and others not? Lancet. 2012;
380 (9838):258–271. https://doi.org/10.1016/ s0140- 6736
(12) 60735-1.
6. DumithSC, GiganteDP, DominguesMR, KohlHW, 3rd.
Physical activity change during adolescence: a sys-
tematic review and a pooled analysis. Int J Epidemiol.
2011;40(3):685–698. https://doi.org/10.1093/ije/dyq272.
7. Miljanovic Damjanovic V, Obradovic Salcin L, ZenicN,
ForeticN, Liposek S. Identifying Predictors of Changes
in Physical Activity Level in Adolescence: A Prospective
Analysis in Bosnia and Herzegovina. Int J Environ Res
Public Health. 2019;16(14). https://doi.org/ 10. 3390/ ijerph
16 1 42573.
8. Štefan L, Sorić M, Devrnja A, Petrić V, Mišigoj-
DurakovićM. One-year changes in physical activity and
sedentary behavior among adolescents: the Croatian
Physical Activity in Adolescence Longitudinal Study
(CRO-PALS). Int J Adolesc Med Health. 2018;32(5).
https://doi.org/10.1515/ijamh-2017-0223.
9. MaricD, KvesicI, LujanIK, BiancoA, ZenicN, SeparovicV,
etal. Parental and Familial Factors Inuencing Physical
Activity Levels in Early Adolescence: AProspective Study.
Healthcare (Basel). 2020;8(4). https://doi.org/10.3390/
healthcare8040532.
10. SekulicD, RodekJ, SattlerT. Factors associated with phys-
ical activity levels in late adolescence: aprospective study.
Med Pr. 2020;71(6):637–647. https://doi.org/10.13075/
mp.5893.01012.
11. SekulicD, MaricD, VersicS, ZevrnjaA, TerzicA, ZenicN.
Familial and Parental Predictors of Physical Activity in
Late Adolescence: Prospective Analysis over aTwo-Year
Period. Healthcare (Basel). 2021;9(2). https://doi. org/ 10.
3390/healthcare9020132.
12. GilicB, ZenicN, SeparovicV, Jurcev SavicevicA, SekulicD.
Evidencing the inuence of pre-pandemic sports partic-
ipation and substance misuse on physical activity during
theCOVID-19 lockdown: aprospective analysis among old-
er adolescents. Int J Occup Med Environ Health. 2021; 34(2):
151–163. https://doi.org/10.13075/ijomeh.1896.01733.
13. Sekulic D, Blazevic M, Gilic B, Kvesic I, Zenic N.
Prospective Analysis of Levels and Correlates of Physical
Activity during COVID-19 Pandemic and Imposed
Rules of Social Distancing; Gender Specic Study among
Adolescents from Southern Croatia. Sustainability. 2020;
12 (10): 4072. https://doi.org/10.3390/su12104072.
14. Xiao S, YanZ, Zhao L. Physical Activity, Screen Time,
and Mood Disturbance Among Chinese Adolescents
During COVID-19. J Psychosoc Nurs Ment Health Serv.
2021;59(4):14–20. https://doi.org/10.3928/02793695- 20 2
0 1 104-04.
15. MooreSA, FaulknerG, RhodesRE, BrussoniM, Chulak-
Bozzer T, Ferguson LJ, et al. Impact of the COVID-19
virus outbreak on movement and play behaviours of
Canadian children and youth: a national survey. Int J
Behav Nutr Phys Act. 2020;17(1). https://doi.org/10.1186/
s12966-020-00987-8.
16. KowalskiKC, CrockerPR, DonenRM. ephysical activ-
ity questionnaire for older children (PAQ-C) and adoles-
cents (PAQ-A) manual. College of Kinesiology, University
of Saskatchewan. 2004;87(1):1–38.
17. Benítez-Porres J, Alvero-Cruz JR, Sardinha LB, López-
Fernández I, Carnero EA. Cut-o values for classify-
ing active children and adolescentes using thePhysical
Activity Questionnaire: PAQ-C and PAQ-ACut-o val-
ues for classifying active children and adolescents using
thePhysical Activity Questionnaire: PAQ-C and PAQ-A.
Nutr Hosp. 2016;33(5):564.
18. ŠtverákováT, JačiskoJ, BuschA, ŠafářováM, KolářP, Ko-
besováA. eimpact of COVID-19 on Physical Activity
of Czech children. PLOSONE. 2021;16(7):e0254244.
19. WickelEE, EisenmannJC. Contribution of youth sport to
total daily physical activity among 6- to 12-yr-old boys.
Med Sci Sports Exerc. 2007;39(9):1493–500. https://doi.
org/10.1249/mss.0b013e318093f56a.
20. Machado-Rodrigues AM, Coelho e Silva MJ, Mota J,
SantosRM, CummingSP, MalinaRM. Physical activity
and energy expenditure in adolescent male sport partici-
pants and nonparticipants aged 13 to 16 years. J Phys Act
Health.2012;9(5):626–33. https://doi.org/10.1123/ jpah. 9.5.
626.
21. EimeRM, HarveyJT, CharityMJ. Sport drop-out during
adolescence: is it real, or an artefact of sampling be-
haviour? Int J Sport Policy Politics. 2019;11(4):715–726.
https://doi.org/10.1080/19406940.2019.1630468.
22. VaaraJP, VasankariT, KoskiHJ, KyröläinenH. Awareness
and Knowledge of Physical Activity Recommendations
in Young Adult Men. Front Public Health. 2019;7(310).
https://doi.org/10.3389/fpubh.2019.00310.
23. MarquesA, Martins J, SantosF, Sarmento H, Carreiro
da Costa F. Correlates of school sport participation:
A cross-sectional study in urban Portuguese students.

Nr 6 COVID-19 and physical activity in adolescents 11
Science & Sports. 2014;29(4):e31–e8. https://doi.org/
https://doi.org/10.1016/j.scispo.2013.07.012.
24. MarquesA, MartinsJ, da CostaFC. Scholar sports par-
ticipation according to age and sex of students from mili-
tary schools. Br J Sports Med. 2010;44(14):i8. https://doi.
org/10.1136/bjsm.2010.078972.24.
25. GustafsonSL, RhodesRE. Parental Correlates of Physical
Activity in Children and Early Adolescents. Sports Med.
2006;36(1):79–97. https://doi.org/10.2165/00007256-20 0
6 3 6 0 10-00006.
26. LeeSM, NihiserA, StrouseD, DasB, MichaelS, HuhmanM.
Correlates of Children and Parents Being Physically
Active Together. J Phys Act Health. 2010;7(6):776. https://
doi.org/10.1123/jpah.7.6.776 10.1123/jpah.7.6.776.
27. Hammed AI, Usman SO, Ezekiel O, Duru JC, Ola di-
meji OJ, Jirho O. Inuence of maternal level of educa-
tion and socioeconomic status on maternal knowledge
of nutrition, physical activity and children’s bodyweight
of Nigerian school pupils. Exerc Qual LifeJ.2021;13(1):
39–44. https://doi.org/10.31382/eqol.210605.
28. ShiehC, HalsteadJA. Understanding theimpact of health
literacy on women’s health. J Obstet Gynecol Neonatal
Nurs. 2009;38(5):601–610; quiz 10-2. https://doi. org/ 10.
1111/ j.1552-6909.2009.01059.x.
29. MedranoM, Cadenas-SanchezC, OsesM, ArenazaL, Ama-
seneM, Labayen I. Changes in lifestyle behaviours during
the COVID-19 connement in Spanish children: A lon-
gitudinal analysis from the MUGI project. Pediatr Obes.
2021;16(4):e12731. https://doi.org/10.1111/ijpo.12731.
30. Lynch T, Soukup GJ. “Physical education”, “health and
physical education”, “physical literacy” and “health litera-
cy”: Global nomenclature confusion. Cogent Educ. 2016;
3(1):1217820. https://doi.org/10.1080/233118 6X. 2016. 12
1 7820.
is work is available in Open Access model and licensed under aCreative Commons Attribution-NonCommercial3.0Poland License– http://creative-
commons.org/licenses/by-nc/3.0/pl/deed.en.
Publisher: Nofer Institute of Occupational Medicine, Łódź, Poland