Assessing Changes in Lifestyle and Health-Related Behaviors During the COVID-19 Pandemic: A Cross-Sectional Web-Based Survey Among Lebanese Adults

The novel coronavirus disease 2019 (COVID-19) has quickly become a worldwide threat to human health and has signicantly impacted peoples’ lives and changed their lifestyle and health behaviors (LHB). This study aims to assess changes in lifestyle and health-related behaviors such as daily habits, dietary habits, physical activity, sleep patterns, psychological problems, and sexual behaviors among the general population in Lebanon during the COVID-19 pandemic. The study questionnaire was on included questions about and (smoking, alcohol and and of in Data analysis

In Lebanon, the COVID-19 outbreak occurred together with an unprecedented economic crisis, with the rst con rmed case reported on February 21, 2020. As of this date, the Lebanese government has implemented preventive measures to curb the spread of the disease. However, the steadily increasing numbers of COVID-19 cases have mandated several sterner interventions, including curfews and total lockdowns [23].
Several studies were conducted to assess the effects of the COVID-19 pandemic on Lebanese adults, covering mental health, domestic violence, weight change, stress, and anxiety [24][25][26][27]. Their ndings revealed that the fear of COVID-19 was associated with a negative weight perception [25] and higher stress and anxiety, particularly in women [26]. Furthermore, extended con nement was associated with a higher weight change perception [25].
These studies have addressed some of the factors affected by the COVID-19 pandemic; however, none has evaluated lifestyle and health-related behavior changes among Lebanese adults. Understanding the associations between lifestyle risk factors and COVID-19 is essential to identify people at risk of exhibiting unhealthy behaviors during the lockdown since it is hypothesized that lifestyle measures and health-related behaviors are altered during the COVID-19 pandemic.
Therefore, this study aims to assess changes in lifestyle and health-related behaviors such as daily habits (hygiene, smoking, alcohol, and work), dietary habits (intake, meal pattern, and snack consumption), physical activity (duration and type), sleep patterns (length and quality), psychological problems (physical and emotional exhaustion, irritability, and tension), and sexual behaviors among the Lebanese adults during the COVID-19 pandemic.

Study design and sampling
A cross-sectional web-based survey conducted between January 9 and January 28, 2021, enrolled 403 Lebanese adults using the snowball sampling technique. All people above 18 with access to the internet were eligible to participate. The study questionnaire developed on Google Forms was self-administered, anonymous, and available in two languages English and Arabic. The survey link was distributed through various social media platforms (WhatsApp, Facebook, and Instagram) to reach participants from all Lebanese Governorates/Districts (Beirut, Mount Lebanon, North, Beqaa/Baalbeck/Hermel, and South/Nabatieh). Participants were encouraged to share it with friends and people they know.
The rst page of the questionnaire included explanations of the study objectives and the following informed consent statement: Completing the questionnaire requires 10 to 15 minutes and indicates your consent to participate. Participation in this study was voluntary, and participants received no incentive in exchange for their participation.

Sample size calculation
The Epi info software (Centers for Disease Control and Prevention, Epi Info™) was used to calculate the required sample size [28]. The minimum sample calculated was 350 participants based on a population size of 3,413,308 Lebanese adults (according to the population estimates of 2019-2020 from the Central Administration of Statistics) [29], an alpha error of 5%, a power of 80%, a con dence level of 95%, and an expected frequency of behavior changes due to COVID-19 of 35% in our population (according to a recent study on eating habits and lifestyle behaviors during the COVID-19 pandemic in 18 countries from the Middle East and North Africa region, including Lebanon that represented 11.5% of the overall sample) [30]. The nal sample size comprised 403 participants to allow for adequate power for multivariable statistical analyses.

Questionnaire
The questionnaire consisted of three sections, including closed-and open-ended questions.
The rst part of the questionnaire clari ed the sociodemographic features of participants: age, gender, marital status, education level, family monthly income, the number of people living in the same house, the number of rooms in the house, occupation status, the region of residence, and religion. The household crowding index was calculated by dividing the number of persons living in the house by the number of rooms, excluding bathrooms and kitchen. The family monthly income, in Lebanese pound (LBP), was divided into four levels: no income, low (< 1500000 LBP), intermediate (1500000-3000000 LBP), and high income (> 3000000 LBP). Two additional questions, rated on a 5-point Likert scale from 0 (not worried at all) to 5 (extremely worried), assessed nancial distress by measuring how worried participants are about their current nancial situation and being unable to meet regular monthly living expenses.
The second part consisted of COVID-19-related questions, such as direct or potential contact with someone with COVID-19, having been diagnosed or tested positive with COVID-19, family history of COVID-19, and the changes in hours spent working/studying/using electronic devices (screen time), salary/income, and weight during the COVID-19 pandemic.
The third part of the questionnaire consisted of three scales and questions about health-related behavior changes as follows.
Lifestyle and Health Behavior Changes questionnaire  This 17-item newly constructed questionnaire (LHBC-17) is a short, concise, and user-friendly tool based on a validated questionnaire [31] used to assess the lifestyle and health-related behavior changes during the COVID-19 pandemic. It covers all essential aspects of lifestyle and health-related behaviors, including dietary habits (intake, meal pattern, and snack consumption), physical activity (duration and type), sleep (length and quality), and psychological health (irritability, stress, and anxiety).
Diet-related items (1 to 9) assess the consumption of main meals, snacks, healthy and balanced diet (including whole grains, fruits, vegetables, eggs, nuts), and unhealthy foods (fried food, fast food, sugarsweetened products). Physical activity-related items (10 to 12) assess the participation in aerobic exercise, household-related activities (cooking, laundry, cleaning), and leisure-related activities (indoor/outdoor activities, walking, and gardening. Two questions (13 and 14) assessed sleep length and levels of stress and anxiety. The last three questions (15 to 17) evaluated psychological health changes, i.e., irritability and physical and emotional exhaustion. All items were graded on a 5-point Likert scale: (a) Signi cantly increased (b) Slightly increased (c) Grossly similar (d) Slightly decreased (e) Signi cantly decreased.
The total score was calculated by adding responses to all items, depending on change outcomes; grades could be negative (tending for the worse) or positive (tending for better).
Item 3 that assessed changes in meal and snack portions during COVID-19 is rated, assuming that the person had regular portions of meals and snacks before the pandemic. Also, item 13 that evaluated changes in hours of sleep during COVID-19 is rated, assuming that the individual was getting an adequate 6-8 hours' sleep before the pandemic.
Thus, higher LHBC-17 scores indicate a change to the better in lifestyle and health-related behaviors (LHB), while lower LHBC-17 scores indicate worsening of LHB among participants (α Cronbach = 0.71).

WHO-5 Well-Being Index
The WHO-5 Well-Being Index developed by the World Health Organization (WHO) is a 5-item tool that measures mental well-being during the past two weeks [32,33]. In this study, the Arabic validated version (WHO-5-A) was used [34]. The raw score calculated by summing the ve answers ranges from 0 to 25, with higher scores indicating increased well-being (α Cronbach = 0.95). The WHO-5-A is a helpful screening tool to detect depressive episodes among Lebanese adults at a cut-off point of less than 13. Therefore, a score below 13 re ects poor well-being and is an indication for testing for possible depression [34].

Fear of COVID-19 scale (FCV-19S)
The FCV-19S is a 7-item scale developed to assess the fear of COVID-19 among the general population [35]. In this study, the Arabic validated version of the FCV-19S was used [36]. Items are graded on a 5point Likert scale from 1 (strongly disagree) to 5 (strongly agree). The total score is calculated by adding up each item score, where higher scores indicate a greater fear of COVID-19 (α Cronbach = 0.91) [35].

Questions related to other behaviors changes
Questions related to risky behaviors, such as smoking, alcohol consumption, and sexual behaviors, were included to assess the changes that occurred during COVID-19 among the study population. The questions were: "Did your smoking behavior change during the COVID-19 pandemic?"; "Did your alcohol drinking behavior change during the COVID-19 pandemic?"; "During the COVID-19 pandemic, have your sexual and intimate behaviors (dating, kissing, cuddling, spending private time with your partner, having sexual intercourse with your partner…) changed?"; and "During the COVID-19 pandemic, have your sexual desire changed?". These questions were assessed as follows: (a) Signi cantly increased (b) Slightly increased (c) Grossly similar (d) Slightly decreased (e) Signi cantly decreased. The option "I prefer not to answer" was added for sexuality-related questions. For smoking and alcohol consumption, two additional answers per behavior were possible, i.e., "I have recently started smoking/drinking alcohol", and "I don't smoke/drink alcohol".
Moreover, four questions evaluated hygiene behavior changes during the COVID-19 pandemic, including handwashing, surfaces and materials disinfection, respiratory etiquette (the use of face mask and cough etiquette), social distancing, and avoiding crowded places. Two questions assessed changes in nutritional supplements and immunity-boosting foods intake (lemon, turmeric, garlic, citrus fruits, and green leafy vegetables) during the COVID-19 pandemic. All these questions were also graded on a 5-point Likert scale as follows: (a) Signi cantly increased (b) Slightly increased (c) Grossly similar (d) Slightly decreased (e) Signi cantly decreased.

Translation procedure
The LHBC-17 scale was translated from English into Arabic using the forward and backward translation method. One of the authors performed the translation from English into Arabic, and another author did the back-translation. Discrepancies were resolved by consensus between the original English version and the translated one.

Statistical analysis
Data analysis was done using the SPSS software version 21. A "weighting" variable was created to adjust for the composition of our sample (especially the over or under-representation of gender and Governorate distribution), aiming at re ecting that of the Lebanese adult population.
Cronbach's alpha was calculated for the reliability analysis of all scales. Scales with a coe cient value of 0.7 or higher are considered internally consistent [37].
Descriptive statistics were performed to represent the participants' characteristics and LHB changes during the COVID-19 pandemic and were expressed as frequencies and percentages for categorical variables and as means and standard deviations (SD) for continuous variables.

Descriptive analysis
Our sample consisted of 403 participants; the mean age was 29.74 ± 10.81 years, with 51.2% females and 48.8% males. Most participants were single (64.8%), had a university degree (85.2%), and lived in an urban area (68.6%). The mean household crowding index was 1.11 ± 0.88. Of the total sample, 34.2% had low to no income, 25.5% had intermediate income, and 40.3% had a high income. Moreover, 44.4% reported a decrease in salary during the COVID-19 pandemic, and 34.6% declared being very or extremely worried about their current nancial situation. About half of the participants had full-time jobs (48.8%) and were mainly working/studying from home (53.2%) during the lockdown period; 37.2% were health care professionals, and 44.5% had direct/potential contact with con rmed/suspected COVID-19 patients.
More than half of the participants (51.6%) had a normal weight range (BMI = 24.17 ± 4.41 kg/m 2 ), and only 11.3% were obese (BMI ≥ 30 kg/m 2 ). Also, 32.4% reported weight gain during the COVID-19 pandemic, 31.7% lost weight, and 28.1% maintained weight. Most participants stated they had a signi cant (52%) and slight (23.7%) increase in sitting and screen time (TV and electronic devices). When queried about their overall health state, the majority reported a good health state (49.5%), and only 2.7% reported a poor state of health.
Only 26.2% of the participants were previously diagnosed with COVID-19 while 55.5% reported a history of COVID-19 in the family. Moreover, 9.2% were diagnosed/screened for depression/anxiety due to COVID-19 by a specialist/doctor and only 13.4% had chronic diseases.
The associations between the LHB changes and the participants' characteristics and associated health factors were compared with the Pearson correlation for continuous variables and the Student t-test or ANOVA F test for categorical variables with two or more levels, respectively. The Bonferroni test was used for multiple comparisons to control the overall signi cance level for some set of inferences performed as a follow-up to ANOVA.
A multivariate analysis was conducted using a three-level nested model analysis taking the LHBC-17 scale as the dependent variable. The rst level included sociodemographic-and work-related variables. In the second level, the LHBs were added to the signi cant variables found in the rst level. The third level consisted of all the COVID-19-related variables added to the variables found signi cant in the rst and second levels. In each level, the stepwise method was used simultaneously to remove the weakest correlated variables and come up with a model that best explains the distribution. All variables that showed a p-value < 0.2 in the bivariate analysis were included as covariates in the related levels. Standardized beta and their 95% CIs were used to quantify the associations between variables and LHB changes. A p-value less than 0.05 was considered signi cant.
In this study, 24.3% of the surveyed individuals smoked cigarettes, 20.4% smoked nargileh, 1.4% started smoking during the pandemic, and 61.8% of those who already smoked slightly and signi cantly increased smoking. Moreover, 43.2% of participants drank alcohol, 1.1% of the participants started drinking during the pandemic, and 37.6% of those who already drank slightly and signi cantly increased alcohol drinking behavior.
Finally, 29.6% of the surveyed individuals reported they had slightly and signi cantly decreased their sexual behaviors during the COVID-19 pandemic, such as dating, kissing, cuddling, spending intimate time, or having sexual intercourse with their partner and 15.5% had slightly and signi cantly decreased sexual desire.
Description of the scales used in the study Table 1 describes all the scales used in this study in terms of mean, standard deviation (SD), median, minimum, maximum, and Cronbach Alpha values.
Furthermore, 63.9% of the study participants had a WHO-5-A score below 13, therefore considered to have poor mental well-being.

Bivariate analysis
The results of the bivariate analyses taking the LHBC-17 scale during the COVID-19 pandemic as the dependent variable are summarized in Tables 3 and 4.
The mean hours per day spent using electronic devices signi cantly increased (from 5.33 ± 4.13 before the lockdown to 7.88 ± 4.48 during the lockdown) (p < 0.001). Also, participants who reported negative changes in physical health during the con nement were those who signi cantly increased the intake of nutrition supplements (72.3%) and immunity-boosting foods (77.9%) (p < 0.001).
Details of the associations between sociodemographic and COVID-19-related variables and the LHBC-17 scale are shown in Tables 2 and 3.

Multivariable analysis
The nested model analysis taking the LHBC-17 scale as the dependent variable has shown several signi cant associations in each of the three models ( Table 4).
The rst model included the socio-demographic variables and showed that living in an urban region  Other sociodemographic factors such as being a Lebanese permanent resident and worrying about the economic situation were associated with unfavorable lifestyle changes, whereas having a job and living in an urban region were signi cantly associated with positive changes. Such results can be explained by the economic hardship that adds to the sanitary crisis in Lebanon. Indeed, Lebanon is witnessing an unprecedented economic crisis and was recently downgraded from a high-income to upper-middleincome country by the World Bank [43]. This critical nancial situation was ampli ed with the inability to control the exchange rates of USD versus Lebanese Pounds that dramatically skyrocketed in less than a year and has led to massive demonstrations, strikes, and temporary bank closures [44] [45]. Hence, Lebanese residents, particularly the unemployed or those worrying about their nancial situation, cannot introduce good practices and afford healthy products; they would exhibit a limited and weakened motivation to change habits positively and enhance a healthy lifestyle.
As for the general health-related factors, an increased BMI and gaining weight during COVID-19 were signi cantly associated with lower LHBC-17 scores. Almost one-third of our sample reported weight gain, in agreement with previous ndings showing that 31% of 1012 subjects reported weight gain since lockdown due to COVID-19 began in the United Arab Emirates [30]. Several other studies, including Lebanese ones, had demonstrated weight gain during the COVID-19 home con nement [2,25,[46][47][48], likely due to boredom, anxiety, and stress reactions to COVID-19 leading to changes in eating habits, such as increased consumption of highly energetic foods (rich in sugar and fat) [49].
Our results revealed increased smoking among participants. Studies on changes in tobacco use during the pandemic have yielded controversial results [22,50,51]; some studies have reported increased tobacco use [52], while others have recorded decreased smoking [50]. Smokers usually intensify their consumption as a coping mechanism during the pandemic, given its detrimental consequences on mental health [53].
When exploring COVID-19-related factors, our results revealed that most participants reported a longer screen time, in line with previous ndings [54][55][56][57], showing a 65-70% increase in screen time during the COVID-19 pandemic [54,56]. This increase is explained by the imposed lockdown, where most people had to work or study from home and thus were more likely to spend long screen periods reading, homeschooling, playing, meeting, or watching movies [57]. . Health-promotion strategies should be implemented to assess negative changes in physical or mental health and maintain as much as possible positive health-related behaviors, particularly in young populations [58].

Limitations and strength
Our study has several limitations related to its cross-sectional design and data collection process. Even though our sample was weighted to adjust for the over or under-representation of gender and Governorate distribution, our population consisted mainly of young people with a university level of education, thus expected to have high computer literacy and internet access. Therefore, our results might not be generalized to the whole population. Moreover, a possible information bias could be likely since the questionnaire was self-administered, with no possibility for clarifying confusing questions.
However, despite these limitations, to the best of our knowledge, this study is the rst to evaluate the effect of sociodemographic features, general health-related behaviors, and COVID-19-related factors on lifestyle changes among the general Lebanese population, almost a year after the COVID-19 outbreak.
Furthermore, the use of a standardized questionnaire with validated scales with very good to excellent reliability is likely to reduce possible information bias.

Conclusions
Lebanese adults reported several unfavorable lifestyle changes and psychological problems during the lockdown due to COVID-19. Health-promotion strategies are needed to assess negative changes both on physical or mental health and maintain as much as possible positive health-related behaviors among the Consent for publication