Changes in physical activity during the COVID-19 lockdown based on the sociodemographic profile of 5569 students and academic staff of Austrian universities

Objectives The COVID-19 pandemic and the associated restrictions/lockdowns have influenced the health and lifestyle behaviour of university students and staff, particularly their physical activity (PA) habits. However, the association between the direction of PA changes and sociodemographic characteristics has not been studied. This nationwide study was conducted in Austrian colleges/universities and aimed to identify the magnitude and direction of changes in PA levels during the COVID-19 restrictions and compare participants who had unchanged vs decreased vs increased PA based on sociodemographic characteristics (sex, age, body mass index, study level, living area, nationality and Austrian regions). Study design This was a cross-sectional study to measure the association between the explanatory variable and outcome measures. Methods A total number of 4528 students (mean age 24.9 years) and 1041 academic staff (mean age 46.4 years) participated in an online survey and provided self-reported data on sociodemographic characteristics and PA change during the pandemic. Results A total of 41.3% of students and 37.5% of academic staff reported a decrease in PA level, whereas PA levels increased in 36.3% and 27.9%, respectively. In students, all sociodemographic variables (including sex, age, body mass index, study level, living area, nationality and Austrian regions) were significantly associated with the direction of PA changes (P < 0.05). Living area and Austrian region were found to be significant indicators of direction of PA changes in academic staff (P < 0.05). Conclusions These findings suggest that the inconsistency of previous studies in PA change during the pandemic may at least partly be explained by differences in the sociodemographic characteristics of the participants.


Introduction
Physical activity (PA) is a substantial part of a healthy lifestyle, which includes engagement in sports, physical exercise and movement habits. 1 It has been well documented that regular engagement in PA can be a supplementary or even an independent 'medicine' to control several health problems such as noncommunicable diseases 2,3 and the associated risk factors, including obesity. 4 According to the World Health Organization (WHO) 5 and based on conclusions from a systematic review, 6 regular PA is associated with a lower likelihood of adverse COVID-19 outcomes. In this regard, it has been shown that the immunological benefits of regular PA 7 are linked to a protective effect against respiratory infections. 7,8 Therefore, PA is considered an effective strategy in the prevention and management of several diseases associated with COVID-19 (e.g. cardiovascular and metabolic diseases), 9 which are known as predictors of COVID-19 infection, severity and mortality. 10 In addition, PA seems to have been an effective strategy at reducing psychological distress during the COVID-19 pandemic social lockdowns. 11 The COVID-19 pandemic has presented the world with an unprecedented challenge affecting individual and public health. In efforts to control the spread of the disease, social restrictions and lockdowns were frequently implemented by governments, which proved to be highly effective. 12 The restrictions involved social distancing, encouragement to stay at home and avoid unurgent travels and limiting access to public services, including schools and universities, sports facilities and leisure spaces. 12,13 In addition to the unfavourable impacts on education and community life, 14 data show that the COVID-19 pandemic and the associated restrictions have influenced health and lifestyle behaviour, particularly PA habits. 15,16 Emerging adulthood, commonly defined as the period from the late teens to the mid-20s, 17 is a critical period of life to develop and shape lifestyle habits and the associated health-related consequences. 18,19 University life, which typically links with emerging adulthood, is independently associated with psychological stressors, which can cause further unfavourable health impacts. 20 Data from large-scale studies show that most university students and professors do not meet the existing PA guidelines. 21e23 It seems that the COVID-19 pandemic and restrictions have had a considerable influence on the PA habits of university students and academic staff. 24,25 However, the direction and magnitude of PA changes differ markedly across the available investigations. For instance, previous studies show contradictory results regarding PA changes during the pandemic in the form of a decrease, 26,27 increase 28,29 or no change, 30,31 among university students. Findings from a systematic review indicate that students who met the PA recommendations before the restrictions also met the recommendations during the pandemic. 24 Another study reports while the PA level of university students decreased, their PA behaviour improved during the COVID-19 restrictions. 32 These data, together with the abovementioned inconsistencies in research findings, appear to be associated with differences in sample size and/or the sociodemographic characteristics of participants. 6 Thus, there is a necessity to consider potential moderators of change in PA, and this area remains an open field of research.
Despite the existence of various research reports on university students and their PA change during the COVID-19 pandemic, 26e31 to date, no studies have examined the direction of PA change as an independent variable to test the potential associations with sociodemographic characteristics. Therefore, the present study aimed to identify the magnitude and direction of change in PA levels during the COVID-19 restrictions and compare participants who had unchanged vs decreased vs increased PA to find associations with sociodemographic factors in a large group of Austrian university students and academic staff.

Study design and sampling
This study is a part of the project Sustainably healthy e From Science 2 Highschool & University, which has been conducted in tertiary education settings with a cross-sectional study design 33 following the original study on secondary schools. 34,35 The boards of college/university deans of all 102 public and private colleges/universities in Austria were directly contacted and received relevant information about the study goals, procedures and instructions. Any individual who was either enrolled as a student or employed as a lecturer, researcher or other academic staff were eligible to participate in the study, which equated to a total population of 376,050 students and 69,310 academic staff. After approval of the study protocol by the ethics board of the associated educational entities, students and academic staff from 52 public and private colleges/universities participated in the present study.

Procedures
Participants received a Web link to take part in an online survey. The associated instructions were included, and participants provided written informed consent to participate in the study. The survey was based on self-reports (LimeSurvey, version 3.25.15) and was provided in German and English languages and separately for students and academic staff. Participants were able to complete the survey within approximately 20 min. Data collection was done from April to July 2021. It is noteworthy to indicate that during the data collection period, a series of restrictions and measures had been implemented in Austria to combat the spread of the COVID-19. The restrictions included mandatory maskwearing and social distancing between individuals in public indoor spaces, limitations on gatherings, closure of non-essential businesses and a nighttime curfew in some areas. In addition, a majority of university classes were administered online, and classes that required in-person attendance were held in small groups to guarantee the mandated 2-m distance between attendees at all times. Further methodological information is available in the study protocol of the Sustainably healthy e From Science 2 Highschool & University study, 33 which the interested readers are kindly referred to.

Measures and calculations
The standardised questionnaire consisted of questions on sociodemography (nationality: Austrian or international; age; sex: females, males or diverse; living area: urban or rural; federal state: Vienna, Burgenland, Lower Austria, Styria, Carinthia, Upper Austria, Salzburg, Tyrol or Vorarlberg), biometric data (height and body weight), current academic/professional status (including the study level of students: undergraduate level or graduate level) as well as health-related data. To assign participants in the study groups based on PA change during the COVID-19 restrictions, participants were specifically asked to answer a question on their change in PA level during the COVID-19 lockdown with the response options of 'No', 'Yes, I did less' or 'Yes, I did more'. Body mass index (BMI) was calculated using body weight and height values. The WHO cutoff points for BMI were used to assign participants in BMI subgroups, including underweight (<18.5 kg/m 2 ), normal weight (18.5e24.9 kg/m 2 ), overweight (25.0e29.9 kg/m 2 ) and obese (>30.0 kg/m 2 ). 36,37 Statistical analysis Statistical tests were performed using SPSS version 26.0 (SPSS Inc., IBM Corp., Armonk, NY, USA). Exploratory analysis was conducted by descriptive statistics, and data are reported as prevalence (percentage; for nominal data) or mean with standard deviation (M ± SD; for continuous data). Chi-squared tests (c 2 ) were conducted to examine differences between the study groups (based on PA change) by sex, weight status, study level, living area, nationality and Austrian region. Analysis of variance (F values) was used to examine age differences between the study groups (PA change). When significant differences for multilevel variables were found, post-hoc tests (using Bonferroni adjustment) were also conducted to determine the origin of differences within subgroups. The level of statistical significance was set at P 0.05.

Results
A total number of 4528 college/university students (mean age: 24.9 ± 6.3 years) and 1041 academic staff (mean age: 46.4 ± 11.5 years) completed the questionnaire. The final sample size of the present study included 1.2% of student population and 1.5% of eligible academic staff enrolled or employed in Austrian colleges/universities. This response rate successfully met the initial anticipation of 1.0%. Among students, 65.8% were females, and 0.7% reported a diverse gender. Among academic staff, 51.6% were females and 0.2% reported a diverse gender. Classification of participants is displayed in Fig. 1.
While 7.4% of students were underweight (BMI <18.5 kg/m 2 ), the general prevalence of overweight/obesity was 17.8%, with a higher rate in male students (24.6%) compared with females (14.3%). On the other hand, the prevalence of underweight was higher in female students (9.4%) compared with males (2.9%). The prevalence of overweight/obesity among academic staff was 32.2%, whereas only 2.9% were underweight. Table 1 presents the anthropometric characteristics of the participants based on sex groups.
Across the sample of students, 22.4% of participants reported their PA level had not changed during the COVID-19 restrictions, whereas 41.3% experienced a decrease and 36.3% reported an increase in their PA level. Accordingly, sociodemographic analyses were conducted based on PA change groups: no change in PA, less PA and more PA. A significant sex difference was found between the study subgroups (P < 0.001), where female students were more likely to report an increase in PA level during the COVID-19 restrictions compared with male and diverse students (P < 0.001); on the other hand, male and diverse participants reported a higher likelihood to decrease their PA level compared with female students (P < 0.001). Age significantly differed (P < 0.001) between subgroups, where students with increased PA were generally younger than their peers in the two other groups of PA change (P < 0.001). Sex-specific analyses, however, showed that this age difference was only present in females (P < 0.001). A significant difference between PA change subgroups was also found in BMI categories (P < 0.001), with overweight and obese participants being more likely to report a decrease in PA during COVID-19 compared with normal weight or underweight participants (P < 0.001). Analysis of the study level showed that undergraduate students were more likely to report an increase in PA during the COVID-19 restrictions than graduate students (P < 0.001). Rural participants were more likely to increase their PA level, whereas urban participants were more likely to decrease their PA level (P < 0.001). A significantly higher prevalence of participants reporting a decrease in PA level was found among those from eastern Austria (states: Vienna, Burgenland, Lower Austria) compared with centre (states: Styria, Carinthia, Upper Austria) and western Austria (states: Salzburg, Tyrol, Vorarlberg; P < 0.001). It was also found that international students had a greater likelihood to decrease PA during the pandemic compared with Austrian students (P < 0.001). Further data are presented in Table 2.
Across the sample of academic staff, 34.7% reported no change in PA level during the COVID-19 restrictions, whereas 37.5% and 27.9% reported a decrease and increase in PA level, respectively. Analyses of sociodemographic parameters showed that there were significant differences between the PA change subgroups in 'living area' and 'Austrian region'. Specifically, those who lived in   urban areas were more likely to decrease their PA level, whereas rural participants were more likely to maintain their PA level during the COVID-19 pandemic (P < 0.001). In addition, participants in the eastern region of Austria were more likely to report a decline in PA during COVID-19 compared with participants living in the west or centre of Austria (P ¼ 0.018). No significant difference was found between the study groups in other sociodemographic variables, including age, sex, BMI and nationality (P > 0.05). Table 3 presents further data regarding the association between PA changes and different sociodemographic variables in Austrian academic staff.

Discussion
This Austrian nationwide study aimed to examine PA changes during the COVID-19 pandemic in college/university students and academic staff, focusing on potential influences of sociodemographic characteristics on behaviour change. In general, it was shown that (1) the majority of students (77.6%) experienced a change in PA level during the COVID-19 restrictions in form of decrease (41.3%) or increase (36.3%); (2) among academic staff, 37.5% reported a decrease and 27.9% experienced an increase in PA level; (3) analysis of sociodemographic characteristics of students showed that females (compared with male and diverse students), younger (compared with older students), underweight and normal weight (compared with overweight and obese students), undergraduate (compared with graduate students), rural (compared with urban students) and Austrian students (compared with international students) were more likely to report an increase in PA during the restrictions compared to their peers; (4) among academic staff, a significant association between PA change and two sociodemographic parameters (i.e. living area and Austrian region) were observed, where urban (compared with rural participants) and participants from the eastern region of Austria (compared with those who live in centre and west of Austria) reported a greater decrease in PA level during the pandemic.
Consistent with the present findings, another study concluded a range between 50 and 90% of university students changed PA level during the COVID-19 pandemic. 38 However, there has been an inconsistency in the direction of PA changes between studies in terms of decrease, 26,27 increase 28,29 or unchanged. 30,31 Data from a comparable study on German university students show that half of the students indicated a decrease and one-third of students reported an increase in PA level during the pandemic; however, a general decrease in their daily walking time was found. 39 This contradiction may highlight that sedentary time (which is associated with the current sharp increase in usage of online platforms) is not necessarily translated to daily PA levels. In this regard, a study reports that although university students' sedentary behaviour (in terms of daily sitting hours) increased during COVID-19 restrictions, a considerable number of them experienced an increase in PA level. 28 PA changes may also be related to previous PA background because it has been reported that highly active students had a greater change in their PA behaviour before and during the restrictions. 25 Conclusions from a review study indicated that while most investigations on university students reported a decrease in PA level during the pandemic, the intensity of PA seems unrelated to PA change. 24 In the comparison of study groups, students in the present study showed a slightly greater PA change compared with academic staff (77.6% vs 65.4%). Consistently, another study found that the PA level of university students was more affected by the COVID-19 pandemic compared with the university employees, 25 which may be explained by the steadiness of daily routines in university staff, mostly because of maturity and/or stable life. Previous studies showed that young adults, including university students, are prone to experience many unexpected changes in their life (because of variations in, e.g., academic, social and occupational situations), which may profoundly affect their lifestyle. 18 This is in line with the general fact that environmental or institutional barriers are accountable for 90% of predictors of lifestyle behaviour among university students. 20 There are contradictory reports in the literature regarding sex differences in PA change following the lockdowns in terms of the predominance of female 25,26,40 or male students 41e43 in the magnitude of change. The present findings show that sex was a significant indicator for direction (not magnitude) of PA change in students but not academic staff, where female students reported a greater likelihood to increase PA and male (as well as diverse) participants had a greater likelihood to decrease PA during the COVID-19 restrictions. Similar results were found in a Spanish study on university students. 44 Consistently, the results from studies on general populations show that increases in PA levels were more commonly reported by females than males, whereas decreases in PA levels were more prevalent in males. 45,46 It has also been reported that being female is associated with the probability of meeting the PA recommendations during the pandemic. 47 A possible explanation may be that females are more likely to be able to adopt their movement habits based on online opportunities/ platforms, whereas males may be more dependent on organised club activities that were shut-down during COVID-19 restrictions. Sex differences in health-related motives, particularly the greater health and weight consciousness of females, may also play a significant role in health behaviour, including PA. 48 In the present study, there was a significant association between the direction of the PA change and age as well as BMI in students (but not academic staff). We observed that students with a greater likelihood to increase PA during COVID-19 restrictions were generally younger than the two other groups (i.e., unchanged and decreased PA). Consistently, it has been shown that being younger than 22 years of age is positively associated with regular engagement in PA during the restrictions. 47 However, data in general populations show that mature adults were significantly more likely to keep and even increase PA levels during the COVID-19 time compared with younger adults. 49 While the prevalence of overweight/obesity in the present study was 17.8% in students and 32.2% in academic staff, overweight/obese students (but not staff) were more likely to decrease their PA during the pandemic compared with the students with BMI <25 kg/m 2 . This can be partially justified with previous research, indicating that students who met the PA recommendations before the pandemic were more likely to meet the recommendations during the pandemic. 24 The distribution of weight categories in the present study was in line with the global data, 50 where the majority of both samples had normal weight (74.8% of students and 64.9% of academic staff). However, the low prevalence of obesity/overweight in the sample of students (17.8%) is in accordance with the data from the WHO European Regional Obesity Report 2022, in which the Austrian people present among the lowest rates of overweight and obesity in Europe. 51 Living environment was also found to be a significant indicator for PA change in the present study. In both groups, university students and academic staff, urban participants were more likely to decrease their PA level compared with rural members. This finding might be explained by the greater severity of COVID-19 restrictions in cities and towns and/or more outdoor opportunities in rural areas for PA. Interestingly, we found that students and academic staff from the eastern parts of Austria were more likely to reduce their PA during COVID-19 compared with participants living in the centre or west of Austria (which is widely encompassed by highlands, including mountains and forests). In this regard, it has been documented that geographical area is a remarkable indicator of the health-related consequences of COVID-19. 52 Another study found that health behaviour of university students during the pandemic varied by the country of residence, 53 which is in line with the present finding where international students were more likely to decrease PA level compared with Austrian students. According to European Observatory on Health Systems and Policies (EOHSP) reports, Austrian adults are among the most physically active in Europe, 54 which is comparable with the previous data indicating that the majority of Austrian university students are physically active. 55 Altogether, these findings suggest that cultural and environmental situations need to be taken into account when interpreting changes in health-related behaviours.
The present study includes some limitations. The cross-sectional study design and the associated self-reported data carry the risk of misreporting, particularly socially desired statements from overreporting (e.g. in PA engagement) and/or underreporting (e.g. in body weight). In addition, there might be confounding factors, including direct and indirect lifestyle parameters, which may potentially influence the findings and the associated interpretations. Another limitation of this study is the use of a self-reported assessment of PA using a single-item question, which may be subject to recall bias and/or social desirability bias. Because of practical considerations, self-reported assessment of PA was the only option for this study. The large sample size of the study may also minimise the likelihood of the abovementioned limitation. Evidence suggests that single-item questionnaires can be considered valid and are consistently used in the social and behavioural sciences, as they may significantly reduce the problems associated with lengthy surveys. 56 Nevertheless, interpretations of the findings should be done with caution. Despite the aforementioned limitations, the nationwide nature of the sample and the large sample size in both groups of students (n ¼ 4528) and academic staff (n ¼ 1041) are considerable strengths of the present study. Research-based evidence in the area of public health provides a basis for health administrations to upgrade and establish policies and guidelines. Particularly, efforts towards understanding of the predictors of health behaviours, including PA patterns during the COVID-19 pandemic are crucial for policymakers at educational settings. Such evidence may ultimately help increase the health status of university students in such an important period of life e called emerging adulthood e which eventually leads to the promotion of public health.

Conclusions
Taking a large sample into account (n ¼ 5569), the present Austrian-wide study is the first to map the association between the direction of PA change during the COVID-19 time and sociodemographic characteristics of college/university students and academic staff. Most students (77.6%) and academic staff (65.4%) experienced a change in PA level (in terms of decrease or increase) following the COVID-19 restrictions. In the sample of students, almost all sociodemographic variables (including sex, age, BMI, study level, living area, nationality and Austrian regions) were significantly associated with the direction of PA change; however, only living area and Austrian region were found to be indicators of increased or decreased PA among academic staff. These findings provide the first explanatory evidence specifying that the inconsistency between data from previous studies may be, at least partially, because of differences in sociodemographic characteristics of university students and staff.

Author statements
Author contributions K.C.W., together with G.R., W.K. and C.D. contributed to conceptualisation and study design. M.M. and C.D. contributed to methodology, formal analysis and writing original draft preparation. K.C.W., C.D., G.R., D.R.T. and W.K. contributed to critical review and editing. G.W. contributed to technical support. All authors have read and agreed to the published version of the article.

Ethical approval
This study was conducted in accordance with the medical professional codex, the Helsinki Declaration as of 1996, Data Security Laws and good clinical practice guidelines. Informed consent was obtained from all participants involved in the study. Participation in the study was voluntary and could be terminated at any time without providing reason or negative consequences.

Acknowledgements
This Austria-wide college and university study is supported by the Austrian Federal Ministry of Education, Science and Research (BMBWF e Bundesministerium für Bildung, Wissenschaft und Forschung, Abteilung I/7 e Schul-und Universit€ atssport), as well as by the Austrian Students' Union ( € OH e € Osterreichische Hochschü-ler_innenschaft): https://uni.science2.school/en/#Partners. The study investigators are grateful for the administrative, organizational and personal support, help and engagement of partners, experts and governmental and educational authorities.

Funding
The Austria nationwide study Sustainably healthy e From Science 2 Highschool & University is funded by the Tiroler Wissen-schaftsf€ orderung (TWF, Innsbruck, Austria; reference number: F.30,976/6-2021), and the present specific study (which is a part of the abovementioned project) is funded by the PhD scholarship granted by the University of Innsbruck, Innsbruck, Austria, Vice Rectorate for Research (Doktoratsstipendium aus der Nach-wuchsf€ orderung 2021 der Universit€ at Innsbruck). However, there has been no impact from the funding agency on the study design and conduction, data collection and analysis, as well as presentation/publication of the findings.