A cross-sectional study assessing the alarming prevalence of burnout among physicians in a developing country facing a combination of a COVID-19 Pandemic and an economic collapse


 Background:

Burnout among physicians is a serious concern that cultivates its seeds during their education and matures in their daily practicing life. Lebanese physicians were particularly at high risk of burnout since they confronted a mixture of overlapping crises.
Objectives:

This study aimed to assess the prevalence of burnout among Lebanese physicians, to investigate its associated factors and to explore the combined effects of the pandemic and the economic crisis on burnout.
Methods:

A web-based cross-sectional study was conducted among Lebanese physicians over December 2020 using a snowball sampling technique. Self-reported data were collected electronically through an anonymous questionnaire that included information on socio-demographic characteristics, work-related variables, and 4 scales: fear of COVID-19, threat perception, InCharge Financial Distress/Financial Well-Being, and the Copenhagen burnout inventory (CBI). CBI scale cut-off score of 50 was used to evaluate the prevalence of burnout. Multivariable linear regression analyses were carried out to identify the factors associated with burnout.
Results:

A total of 398 physicians participated in the study. The majority of them were male (52.8%), married (60.1%), and aged between 40 and 49 years (43.2%). Burnout hits 90.7% of the Lebanese physicians where personal, work-related, and client-related burnout were detected among 80.4%, 75.63%, and 69.6% of them respectively. A strong association was found between the higher level of burnout and female gender, younger age, being single, having a dependent child, living with an elderly or having a family member with comorbidities, and insufficient sleeping hours. Physicians’ specialties, working in a public health facility, limited years of professional experience, lack of previous experience in pandemic and extensive working hours were also associated with increased burnout. Furthermore, low income, working in the frontline, higher threat perception, and fear of COVID-19 were contributing to higher burnout. However, financial wellbeing, altruism, having good health, and previous history of COVID-19 were significantly associated to lower burnout. The combined effect of threat perception and financial hardship significantly increased the level of burnout.
Conclusion:

The huge burnout level among Lebanese physicians raises alarm about the seriousness of the current situation and urges the health authorities to take prompt action to enhance the physicians’ wellbeing.

poor job performance [34,35]. In addition, it can also negatively impact the quality of care provided to patients by diminishing empathy toward patients hence increasing the risk for medical errors and eroding their professionalism [36][37][38].
In a world facing the COVID-19 pandemic, the prevalence of burnout among physicians is snowballing since they experienced ever-increasing pressure in their daily lives, particularly at their work [39][40][41]. This upsurge was reported in numerous studies conducted worldwide [42]. Similar to other countries, Lebanon experienced many challenges imposed by the COVID-19 pandemic on its healthcare system which was already in a fragile state even prior to the pandemic [43], the economic collapse [44], and the Beirut blasting [45]. It was overwhelmed by the humanitarian crisis revealed by the in ux of more than one million Syrian refugees making the country home to the highest number of refugees per capita in the world [46]. However, the COVID-19 pandemic overlapped with an economic crisis that has its roots in the aftermath of the civil war goaded by corruption and mishandling of the country's resources [47]. This economic crisis was ranked by the World Bank among the world's three worst crises since the mid-1800s affecting living standards where the Lebanese pound has lost more than 90% of its values since fall 2019 [48]. This was later followed by the devastating Beirut blast, one of the most powerful non-nuclear explosions in the history of the world which was coupled with a meteoric soar in infections and hospitalizations where ICU occupancy in the hospitals touched 95% in January 2021 [49]. Physicians, among other HCWs, bear the large toll of the pandemic [50]. The growing number of physicians diagnosed with COVID-19 unveiled gaps in policies and laws intended to warrant physician safety such as coverage for healthcare, disability, and death [51]. As a result of these consecutive and combined events, Lebanese physicians suffering from burnout and economic depletion are leaving to nd a better life elsewhere [52]. In such a typical context of multiple calamities that fueled mental health problems and burnout; it is of great interest to assess the level of burnout among Lebanese physicians and to understand its determinants in order to prevent and reduce such syndrome and to stop the forced immigration of the healthcare human assets. Of note, concerns about pandemic (Shuja et al., 2020) [53] and nancial wellbeing (Cooper, 2011) [54] could both instigated psychological illnesses and could interact and increase the level of burnout among physicians.
The present study aims to assess the prevalence of burnout levels among Lebanese physicians stranded among the mixture of crises, along with how sociodemographic factors, work-related factors, economic factors, and pandemic-related factors affect the intensity of burnout. Besides, we targeted to assess the combined effects of the COVID-19 pandemic and economic crisis on burnout among Lebanese physicians.

Methods:
Study design and population: A quantitative cross-sectional study enrolled 398 Lebanese physicians practicing in all active hospitals located in the eight Lebanese provinces. It was conducted over December 2020. In compliance with social distancing restrictions, participants were electronically invited to participate. A snowball sampling technique was used to select participants from all Lebanese governorates (Bekaa, Baalbeck-Hermel, South, Nabatyeh, Akkar, North, Beirut, and Mount Lebanon). Participants were identi ed via professional groups and health facilities.
Physicians working in health facilities were contacted via phone call and noti ed about the survey and its purpose. Upon their agreement to participate, an online questionnaire using a Google form was sent to them via emails or WhatsApp as per their preference. They were invited if possible to disseminate the link of the survey among their colleagues. The link of the study included a brief explanation of the study and the objective of the survey, in addition to speci c instructions for lling the questionnaire. All practicing physicians of either gender currently working in active hospitals and who had access to the internet were eligible to participate in this study. This study excluded physicians who are not practicing currently, those who were out of the country at the time of the survey, retired physicians, trainees of residency programs, interns, and those who refused to give informed consent. Since the study has no foreseeable risks, written consent was obtained in an electronic format. No reward was received by respondents in return for participation. All methods were performed following the relevant guidelines and regulations.

Sample size calculation
Using an estimated population of 10,918 physicians [55], an expected response of 50%, a 95% con dence level and an estimated absolute error of 5%, the requisite sample size was calculated using the RAOSOFT digital sample size calculator. All previous information were used to calculate the sample size for this study which yielded the least required sample size of 372 participants which was achieved at an early stage of the study before the closure of response acceptance (January 1st, 2021). Of note, the achieved sample size was 398 physicians.

Ethical consideration
Written informed consent was obtained for each participant. They were reassured that the participation is voluntary and that they were free to withdraw at any time. In addition, all information were gathered anonymously and handled con dentially. The study design assured adequate protection of study participants and do not imply any risk for them. None of the survey questions asked for information that could harm the participant in any way.

Instrumentation:
A questionnaire was developed in the Arabic and the English languages through an online platform via Google forms. The utilized scales used were translated into Arabic, except for the CBI-A already validated and available in the Arabic language (Youssef et al. submitted) following the guidelines for forward and backward translation. A consensus was used to resolve inconsistencies between the original and translated versions. A pilot survey was also conducted on 15 physicians, and some reformulations for some questions were made throughout its course. The answers to the pilot survey were excluded from the nal data of this study. The nalized anonymous, self-administered questionnaire was comprehensible and took 10 to 13 minutes to complete. The questionnaire included mainly closed-ended questions. It consisted of four sections: (a) introduction and informed consent (b) basic sociodemographic characteristics, (c) work-related and exposure to COVID-19 variables, and (d) the measurements The rst section consisted of a brief introductory paragraph stating the aims of the study, declaration of anonymity and con dentiality, and mandatory informed consent followed by the items of the survey. The second section collected sociodemographic data of the participants, including gender, age, marital status, specialty, urbanicity, health status, and living conditions. It also included questions about the history of medical illnesses and the health status of people living with the participant. Participants were also asked about the type of health facility and its geographical location. The third section covered the topic of exposure to COVID-19 in addition to work-related variables. Physicians were queried to answer on whether they have worked in the frontline, treating or caring for COVID-19 patients, (b) been tested for COVID-9 (c) been diagnosed as COVID19 case, (d) had a family member relative or colleague infected by COVID-19. Each of these variables was answered on a yes or no basis.
The fourth section consisted of 4 validated scales to objectively assess nancial well-being, threat perception, fear of COVID-19, and burnout among the participants. The scales were used after requesting permission from their copyright owners when required.
1-The perceived threat and altruistic acceptance of risk questionnaire: This tool was developed by Chong et al to assess the risk perception of COVID-19 among HCWs, the perceived threat questionnaire [56]. It consisted of 10 items where nine of these items described HCWs' perception toward COVID-19 threat and one item related to altruistic acceptance of COVID-19 risk. Ratings were given based on a ve-point Likert scale (1=strongly disagree, 2= disagree, 3 = neutral, 4=agree, 5=strongly angry). Responses were dichotomized into positive responses 'agree' or 'strongly agree', while 'strongly disagree', 'disagree', and 'not sure' were considered negative. The Cronbach alpha of this scale was equal to 0.703.

2-The fear of COVID-19 scale:
This tool that consisted of 7 items was developed by Ahorsu et al., 2020 [57] to assess the extent of fear of the COVID-19, scored on a 5-point Likert scale from 1 (strongly disagree) to 5 (strongly agree). The total score is calculated by summing the answers to all questions and ranges from 1 to 35. Higher scores point out a large extent of fear from COVID-19 infection. The Cronbach's alpha for this scale was 0.769.

3-The InCharge Financial Distress/Financial Well-Being scale (IFDFW)
This tool was developed by Prawitz et al., 2006 [58] including eight items measuring the perceived nancial distress/ nancial well-being on a linear scale from 1 to 10. Higher scores re ect lower nancial distress and higher well-being. The Cronbach's alpha for this scale was 0.85 [59]. Permission was requested from the authors to use it in our study.

4-The Arabic version of Copenhagen Burnout scale A-CBI:
The validated Arabic version of the CBI that consisted of 19 items was used (Youssef et al. submitted). It evaluates personal-related (6 items), work-related (7 items), and client-related (6 items) burnout. Participants were asked to rate how often they felt exhausted. Ratings were given based on a ve-point Likert scale. Each item was scored from 0-100 (0=never, 25= Seldom, 0 = Sometimes, 75= Often, 100= Always). Of note, some questions were answered using another ve-point Likert scale (to a very high degree, to a high degree, somewhat, to a low degree, to a very low degree). Mean items score was calculated per scale. Each scale score depicts the direction indicated by its name. A cut-off of 50 was used to assess the prevalence of burnout among physicians. A score of more than 50 is considered high burnout level whereas a score less than 50 signify a low burnout level. The score was valid and reliable according to many previous studies [59]. In our study, the Cronbach's alpha of this scale was equal to 0.879. In fact, the questions of CBI are mixed with questions on other topics in order to avoid stereotyped response patterns. Of note, a reverse coding was performed to item number 7 "Do you have enough energy for family and friends during leisure time" in the work-related burnout score.

Statistical analysis:
The generated data on an excel spreadsheet was transferred to the statistical software IBM SPSS® software (Statistical Package for Social Sciences) version 24.0 for analysis. Given that the response to the majority of questions was mandatory, the missing data constituted < 10% of the total database, then it was not substituted. For descriptive analysis, frequency and percentage were used for categorical variables, the mean and standard deviation for quantitative variables. The normality distribution of CBI scale items was con rmed by calculation of skewness and kurtosis values which are lower than 1 [60]. The Student's T-test was used to compare the means between 2 groups whereas one-way analysis of variance ANOVA to compare between three groups or more, after checking for homogeneity of variances. A Spearman's correlation was applied to link used scores with burnout subscales. To estimate effect size, the Eta squared was used to compare means. All variables that showed a p-value< 0.2 in the bivariate analysis were included in the multivariable analysis as an independent variable. Four linear regressions using the stepwise method were conducted to identify the correlates of each of the CBI scales, after checking the absence of multicollinearity. P< 0.05 was considered statistically signi cant. To assess the interaction between the threat perception and the nancial wellness (IFDFW) scales, the estimated marginal means were calculated for burnout among subjects according to their perception of COVID-19 threat and IFDFW (high/low categories).The dichotomization of the two variables (perception of threat and IFDFW) into high and low categories was done according to the median of each scale.

Results:
Baseline information of the participants: A total of 398 physicians participated in the survey. Table 1 displayed the baseline characteristics of the surveyed physicians. The majority of them were male (52.8%); married (60.1%), aged between 40-49 years old (43.2%), and residents of Mount-Lebanon province (34.7%). Around half of participants had currently a dependent child (47.7%) or were living with the elderly (53%) or a family member with comorbidities at home (53.8%). More than two-thirds (69.85%) of surveyed physicians had a professional experience larger than 10 years and a previous experience in working in pandemic or emergency incidents (74.12%). The highest percentage of respondents were working in the frontlines (62.1%) and 51.9% of them were caring of COVID-19 cases. Only 15.3% of them had a previous history of COVID-19. However, 44.2% of the participants had a family member diagnosed with COVID-19 and 90.2% of them had a colleague diagnosed with COVID-19. Of note, the majority of surveyed physicians (39%) were specialized in internal medicine. Figure 1 detailed the specialties of the surveyed physicians. The normality of the all used scales was assumed since skewness and kurtosis were lower than 1 for all, and the sample size larger than 300. All the used scales showed good reliability; IFDFW (α = 0.85); FOC (α = 0.769); threat perception (α = 0.703) and CBI (α = 0.879). The lower scores of IFDFW reported in all items of the scale re ected higher nancial distress and lower well-being.
The highest burnout level was shown in the dimension related to work (71.5±16.33) followed by the one related to personal burnout (64.8±17.32). A detailed description of the scales is presented in Table 2.

Prevalence of burnout among Lebanese physicians
Using a cut-off of 50 for CBI, burnout was detected among 90.1% of surveyed physicians, where 71.6% had a moderate burnout and 19.1% had a high level of burnout. Personal burnout (PB) ranked rst among other burnout aspects, where we can found that 80.5% of physicians suffered from PB with 45.8% of them reported high levels. As for work-related burnout (WB), it was detected among in its moderate and high level among more than three-quarters of physicians (75.6%), where 60.3% of them exhibited a high level of WB. In regards to client-related burnout (CB), it was noticed among 69.6% of respondents ( Figure 2).

Socio-demographic characteristics and burnout
For the overall burnout scale, female gender, younger age, being single or divorced, and physicians who had a dependent child and those who live with elderly and family members with comorbidities had a signi cantly higher level of burnout. Similarly, all these sociodemographic variables were signi cantly associated with a high level of personal burnout, work-related burnout, and client-related burnout except the age which was not signi cantly associated with the work-burnout dimension. On other hand, residency of the physician was found to be not signi cantly associated with overall burnout and either of its subscales. The largest effect size was observed in age, marital status, and presence of a dependent child at home (Table 3).

Economic characteristics and burnout
Surveyed physicians who subjectively classi ed themselves as having currently a low socioeconomic status following the COVID-19 pandemic and the economic crisis had signi cantly higher burnout in all its aspects (personal, work, and client-related burnout. In addition to those currently labeling themselves in this category, physicians earning less than 2 Million Lebanese lira per month had also a high level of burnout. Similarly, physicians who considered that pandemic or economic crisis highly impacted their monthly income had higher burnout levels, while those who perceived the minor impact of the pandemic and the economic collapse had the lowest burnout level. Having private health coverage was associated with a higher burnout level of burnout, while nancial well-being was negatively correlated with high burnout. Regarding stress, the largest effect size was seen in low economic status after the COVID19 pandemic and economic crisis, major impact of the economic crisis on the income and nancial well-being (Table 4). Table 5 displayed the association between occupational factors and burnout subscales. As for occupational factors, physicians who practiced in public hospitals had signi cantly higher levels of burnout. The increase in anxiety was similar, whether due to the economic situation or to the COVID-19 crisis. Those who have working professional experience of fewer than 10 years and those who lacked previous experience in working during pandemics or emergencies had signi cantly higher levels of burnout compared to those who had large work experience and had practiced during a previous pandemic. Participants whose enterprise temporarily closed and whose income decreased by 75% had signi cantly higher mean anxiety scores. Furthermore, insu cient sleeping hours and extensive working hours increased the overall burnout among respondents. Finally, physicians who considered that the pandemic has a major impact on their work had signi cantly higher burnout. However, the location of the hospital didn't affect the level of burnout. Sleeping hours, extensive working hours, and the major impact of the pandemic on workload had a large effect size related to the overall burnout. Similar occupational factors were associated with a high Level of BP except for extensive working hours. In addition to previously identi ed professional factors increasing the level of burnout among physicians, physicians working in hospitals located in urban areas had higher work-related burnout. In terms of client-related burnout, a higher level among physicians was found to be associated with health facility type, previous experience during a pandemic, and extensive working hours (Table 5).    In terms of health-related characteristics and exposure, having good health status, previous history of COVID-19, and accepting the risk of taking care of COVID-19 cases signi cantly reduced burnout in all its aspects. As for fear of COVID-19 and higher perception of COVID-19 threat, these scales were correlated with higher burnout levels among physicians. Physicians who considered that pandemic had a major impact on their daily life and their familial relationship reported higher levels of burnout. The largest effect size was found for the perception of COVID-19 threat, altruistic and impact of COVID-19 on familial relationships. In terms of higher PB, it was associated with the same factors as the overall burnout in addition to the impact of the COVID-19 pandemic on the participants' social relationships. Altruism was signi cantly associated with a reduction of burnout in all its aspects ( Table 6).

Correlates of burnout and its subscales: a multivariable analysis
In the multivariable analysis displayed in Table 7, higher burnout was associated with female gender, younger age, being specialized in infectious diseases or internal medicine, working in public hospitals, higher perception of COVID-19 threat, insu cient sleeping hours, low income, extensive working hours, having currently a dependent child, having a family member with comorbidities, having limited professional experience. However, being married, nancial wellbeing, good health status, previous history of COVID-19, altruistic, and previous experience of working during pandemic were signi cantly associated with lower burnout. The full model could explain 76.1% of the overall burnout. In terms of higher personal burnout (PB), it was associated with younger age, female gender, having a single or divorced marital status, presence of elderly or child at home, and living with a family member with comorbidities. In addition, higher threat perception, fear of COVID-19, sleeping disturbance, extensive working hours, and low income were associated with higher PB. However, factors such as nancial wellbeing, altruistic and good health condition were associated with lower PB levels. The full model could explain 67.2% of the PB. As for work burnout (WB), similar factors were associated with higher burnout in addition to the type of hospital. The full model could explain 58.4% of the WB. In terms of client-related burnout (CB), we found that younger age, higher perception of threat, fear of COVID-19, and low income were associated with higher CB. Similar to the previous burnout dimensions, altruistic and large professional experience and nancial wellbeing were associated with a decreased level of CB.  Interaction between threat perception of COVID-19 and nancial wellness score

Main Findings
Our ndings demonstrate a signi cant and huge burnout level among physicians during these unprecedented times in the context of the COVID19 pandemic and economic crisis. A strong association was found between sociodemographic variables and burnout such as female gender, younger age, being single, having a dependent child or living with an elderly or having a family member with comorbidities and reduced sleeping hours. Similarly, occupational factors such as physician specialty, working in a public health facility, limited years of professional experience, lack of previous experience in a pandemic, and extensive working hours were associated with a high level of burnout. Economic factors and COVID-19 exposure factors such as low income, threat perception of COVID-19, fear from COVID-19, and working in the frontline were also contributing to a high level of burnout. However, nancial wellbeing, altruism, having a good health status, and being diagnosed as COVID-19 were signi cantly associated with lower level of burnout. The analysis of the combined effect of the COVID-19 pandemic and nancial wellness (IFDFW) demonstrated that the presence of both threat perception COVID-19 and nancial hardship signi cantly increased the level of burnout.
In terms of burnout prevalence, our ndings revealed that burnout hits more than 90% of the Lebanese physicians and around 20% suffered from a high level of burnout. In addition, more than the third quarter of them expressed personal burnout (mean=64.8) and work-related burnout (mean=71.5). As for clientrelated burnout (mean=58.7), it was detected among 69.6% of respondents (mean=58.7). Before the pandemic, physician burnout and its effects have been Of note, it was not possible to compare our results directly with the ndings of a previous study conducted among Lebanese physicians in 2013 before the pandemic [67], due to the use of different scales to assess burnout. However, comparing to other studies using the CBI scale, our ndings seem to be much higher than the percentage reported in such studies before or even after the pandemic. For example, a study conducted among emergency physicians (EPs) in Bahrein (N=116) showed that those physicians reported a prevalence rate of 81.0% for personal burnout (mean=63), 69.8% for work-related burnout (mean=60), and 40.5% for patient-related burnout (mean=43) [68]. Another study conducted among general practitioners in Germany showed that one-third of the participants suffered from PB symptoms, one quarter showed WB while only 12% of them reported a high prevalence of patient-related burnout symptoms This crippling effect on mental health revealed by the huge prevalence of burnout among Lebanese physicians could be explained as the upshot of such typical context that cumulate the traumatic effect of the COVID-19, one of the deep existential crises revealed by the COVID19 pandemic [70] and the Lebanese unprecedented economic downfall that deteriorate the nancial and psychological wellbeing of the physicians. Hence, urgent measures that tackle this tragedy are required to save an already ailing health sector.
In terms of sociodemographic characteristics associated with a higher level of burnout among physicians, our ndings showed that higher burnout was associated with the female gender. However, when it comes to which gender is most affected by burnout, there have been contrasting results with some studies nding no gender differences whereas other studies found that female surgeons experienced more burnout compared to male counterparts [2].
However, our ndings were consistent with the results of studies conducted by McMurray et al. [71] found that physicians women had increased odds of reporting burnout when compared to men and by Kannampallil et al. who found that there was a higher prevalence of burnout amongst women during the pandemic [72]. Furthermore, this study highlighted the association between younger age and a high level of burnout. Such nding is comparable with the results of a study among general practitioners and residents in Hungary that considered younger age as the strongest predictor of burnout in its emotional exhaustion aspect [73]. Another study conducted among physicians in Portugal and assessing burnout during the pandemic reported that younger age and female gender were independent determinants of burnout, similar to our results [74].
Another important aspect of burnout, noticed in our study was that being married or having a partner decreased the level of burnout. Our results were comparable to the results of a study conducted by Shanafelt et al [28] who supported that having a partner or being married was associated with a decreased risk of burnout McMurray et al [6]. This could be explained by the fact that physicians who are supported or feel supported by their partners or loved ones at home experience less burnout when compared to those who do not. In another study, it was found that support by a spouse decreased burnout by 40% [3].
Notably having a dependent child, or living with a family member with comorbidities were associated with higher burnout levels having limited professional experience. Our results were comparable to those reported by Koh et al. and Maunder et al. both suggest that having children is a predisposing factor to burnout [4,5]. However, McMurray et al. found that women physicians who had young children to look after reported a decrease in burnout by 40%, if there was a spouse, supporting colleague, or signi cant other to balance work and home issues [3]. In summary, concerns about contracting the disease and about family members getting it was also linked to higher stress and anxiety [75].
In terms of pandemic-related factors, a higher perception of COVID-19 threat was also associated with a higher level of burnout. It is well recognized that Intense fear and threat perception when people experience physical and psychosomatic disorders lead to such anxiety, depression, burnout, and emotional exhaustion [76-78] which can shape the greatest behavioral changes. In addition, the uncertainty surrounding the pandemic in terms of healthcare policy reform and compensation changes could instigate a higher level of burnout.
In terms of economic factors, a current low socioeconomic status and income, in addition to negative nancial wellbeing, were associated with a higher burnout level. Our results were consistent with a study reporting that lack of job security perception appeared to be the most important predictive factor for exhaustion. Of note, a previous higher socioeconomic status and a current fear of poverty were associated with higher stress and burnout, whereas current nancial wellness was correlated with lower burnout. Such piercing association in low-and middle-income countries leading to several mental disorders [79] is typical for the Lebanese context where even physicians with savings in the country's banks were unable to reclaim their money. Moreover, owing to the enormous devaluation in the country's currency the total loss in physicians' income total loss was more than 80% [8]. The current situation results in alarming consequences, including increased trends in the prevalence of burnout, and psychiatric illnesses [80,81] in addition to an uphill of physicians who left the country searching for stability, nancial wellbeing, and safety for themselves and their families. This exodus is frightful since many of these physicians left despite they worked in well-established and recognized university hospitals where they both practiced and educated future physicians. Rising poverty and economic insecurity are associated with stress [54] (Cooper, 2011).which in turn, can lead to burnout and demission. In a country in freefall where the economic crisis is expected to escalates, health facilities were in danger of laying off employees, postponed some services, or completely closing their doors.
In terms of occupational factors, the rst concept to be discussed was the specialty of the physician. Our ndings showed that internal medicine and infectious diseases specialties of physicians were associated with higher levels of burnout compared with other specialties. The difference of burnout by specialties was also highlighted by the meta-analysis conducted by Lee, et al. (Correlates of physician burnout across regions and specialties: a metaanalysis. The role of specialties as a contributor to burnout in our study may be partly due to differences in exposure to COVID-19 cases as ID specialists, and internal medicine physicians such as pulmonologists and cardiologists were more involved than other physicians in the treatment of COVID-19 cases. Besides, our ndings highlighted that burnout rates were highest amongst physicians involved in frontline care. This was expected since their job deals with uncertainty all the time and they are in direct exposure to COVID19 cases. This nding was in line with the results of a study conducted by Kannampallil et al. who reported that physicians who were exposed to COVID-19 tested patients had a higher prevalence of burnout (46.3%) compared to those who were not exposed (33.7%) [14].
However, other studies conducted among physicians found different aspects [65]. For example, Wu et al and found that medical staff working on the front line had a lower level of burnout compared to those working on usual wards explaining this unexpected trend, by suggesting that frontline workers may have felt a greater sense of control over the situation. Similar to other studies, our ndings showed that insu cient sleeping hour and extensive working hours were associated with a higher level of burnout.
One peculiar nding in our study was that working in public hospitals was found also associated with higher burnout. This could be because public hospitals were rstly designated by health authorities to receive, treat and isolate COVID-19 cases, so physicians working in these health facilities were more exposed to COVID-19. In addition, due to economic collapse, the government, short on funds, and was unable to support alone hospitals with much-needed resources and supplies as the pandemic surged. This was dependent on the support of foreign and local non-governmental aid to import essential supplies and equipment, including personal protective equipment (PPE).
In regards to extensive working hours and sleep deprivation, numerous studies highlighted that sleep de ciency is a key risk factor for burnout among physicians [82]. With the soar of COVID-19 cases, physicians are facing intense workload, extensive working hours, which eventually impact physicians sleeping hours. A study conducted prior to the pandemic showed that 33% of the HCWs were screened positive for the sleeping disorder [17] and this was associated with 4-fold bigger odds of burnout.
In addition to the above, limited work experience was associated with higher burnout levels. Our ndings are also consistent with the results of a Portuguese study that found that healthcare providers with larger experience were less affected by burnout [83] and with those of a study evaluating the prevalence of burnout using the CBI scale among hospital physicians in Lithuania, which found a signi cant reverse relationship between work-and patient-related burnout and length of employment [84]. On the contrary, previous experience during a previous pandemic or emergency decreased the level of burnout. This can be due to their skills acquired from previous comparable situations. Such previous experience will provide the physician with a sense of con dence and control over the situation and lessen his worries and concerns when dealing with patients. We also found that physicians with good health status and previous history of COVID-19 experienced a lower level of burnout. Their good health status could lessen their perception of susceptibility and severity of COVID-19 and history of COVID-19 instigate their sense of being immune naturally.
The role of altruism in decreasing the level of burnout was supported in our study as we found that the burnout level of physicians who accept the risk of caring of COVID19 cases and who choose that willingly was lower than the burnout level of the physicians who didn't accept this risk. Our results were compatible with the ndings of a study conducted among Turkish physicians which reported that the burnout level of physicians who were actively involved in the ght against COVID-19 was lower than their counterparts who are not actively involved [85].
Lastly, the combined effect of the threat of COVID-19 pandemic and nancial hardship supported that the dual presence of COVID-19 fear and economic collapse signi cantly increased burnout levels among physicians. Despite the scarcity of previous studies tackling such a topic, a review supported the effect of economic uncertainty on mental health in the era of COVID-19 [86]. Despite that such topic was not tackled previously among physicians, the increased risk of burnout among Lebanese physicians necessitate a combined approach addressing the stressors resulting from both of the pandemic and economic crisis.
It highlights needed measures to reduce the nancial strain on physicians and puts forward recommendations to support the psychological and nancial wellbeing of physicians.

Limitations:
However, there are several limitations to be acknowledged in our study. First, the design of our study was cross-sectional design which does not allow us to deduce causality. Selection bias is possible due to the snowball technique that was used to collect data which limits the generalizability of the ndings. The collected data was also based on self-reported information which makes it prone to social desirability and might cause a non-differential error and drive the results towards the null, leading to underestimation of some associations. Furthermore, because survey respondents voluntarily completed the survey, only those who may have had available time during the pandemic may have participated. This may have led us to capture less of a selection of physicians that had higher demands on the job during this period, possibly leading to underreporting of burnout during the pandemic. Another possibility was that physicians who are suffering from this syndrome were more interested to participate. In addition, fatigue related to increased computer and screen usage during the pandemic and lack of nancial incentive for completing the survey may have further contributed to refusal of participation. Third, since we did not have comparable data about burnout of physicians just prior to the pandemic, we were not able to evaluate any incremental effect of the COVID-19 pandemic on burnout. Finally, although taking into consideration of some potential confounders in the multivariable models, residual confounding is still possible. Further studies following up on the burnout of Lebanese physicians would be recommended in the future to con rm our results especially that the economic crisis escalates sharply since December 2020.

Implications for Clinical Practice and Research
The alarming level of burnout detected among Lebanese physicians represented only the tip of the iceberg of the crisis in Lebanon. Its negative impacts that begin to effervesce with the exodus of some physicians would not be restricted to those healthcare providers but would also affect the patient's quality of care

Discussion
The COVID-19 pandemic has aggravated the levels of burnout among health care providers, particularly among physicians who had to shoulder the burden of COVID19. Our present study aims to assess the level of burnout among physicians stranded amid a mixture of the crisis revealed by the ongoing COVID-19 pandemic and the economic crisis along with how sociodemographic, occupational, economic, and pandemic-related factors affect the intensity of burnout.
Lastly, it aimed to explore the combined effects of the COVID-19 pandemic and economic crisis on burnout among Lebanese physicians. It is believed that this paper is the pioneer study in Lebanon focusing on burnout during the context of double hit and investigated factors associated with this burnout and the combined effect of crises among physicians.

Conclusion:
After dealing with more than a year of the stressors of the COVID-19 pandemic combined with an unprecedented economic collapse, Lebanese physicians reached a crisis point and the problem is only getting worse in absence of urgent measures. The huge and serious prevalence of burnout among Lebanese physicians reported by our study called for collaborative efforts from all stakeholders in healthcare to adopt urgent measures and to develop and implement effective remedies for physician burnout. Taking into account the unveiled factors contributing to burnout in this study, interventions and effective coping strategies based on a combination of personal-level and system-level measures are required to regain resilience and to optimize work quality or quantity based on physicians' preference. Special focus should be accounted to improving physician nancial wellbeing and work satisfaction. It is time to make eliminating physician burnout a priority to preserve this country's most valued health care asset.

Declarations
Author Contributions: D.Y. was involved with study conception and design, data collection and analysis, drafted and revised the manuscript. Funding: This research did not receive any speci c grant from funding agencies in the public, commercial, or not-for-pro t sectors.

Informed Consent Statement:
Informed consent for participating in the study was obtained digitally through Google Forms from all subjects, and all methods were carried out in accordance with the relevant guidelines and national regulations for the Non-clinical studies. Speci cally, at the beginning of the questionnaire, participants were asked whether they agree to participate in the research in order to be included in the study. Participants were also informed that their participation was voluntary and that they had the right to leave at any time without providing any explanation. No incentives were provided to the study participants.

Data Availability Statement:
After publication, the survey data will be made available on reasonable request to the corresponding author. A proposal with a detailed description of study objectives and a statistical analysis plan will be needed for assessment of requests. Additional materials might also be required during the process of assessment.
Con icts of Interest and the healthcare organizations [19]. Due to the uncertainty of the length of the current pandemic and the ongoing economic crisis, one can only speculate the lasting impact to be considerable. However, to date, there were no realistic evidence-based interventions and tangible measures that focused on physician burnout in Lebanon. Hence, it is important to address factors identi ed by our study that potentially contributing to burnout among physicians identi ed by our study to mitigate the long-term negative consequences. More studies exploring possible interventions based on physicians' preferences and the feasibility of such interventions were recommended. These interventions could include a formalized burnout reduction program and the availability and accessibilities to helplines and counselors. Since our study highlighted the importance of partner existence, a supportive network from partners, peers, and dependents is needed. We suggest also developing speci c training integrating stress management methods and pandemic planning during medical school and residency education. Since gender difference was revealed in our study to affect burnout level, hence gender-based issues may require to be addressed. It is of high priority that government and health facilities start to recognize and roll out effective interventions to prevent and mitigate physician burnout. Lastly, addressing the physician's nancial hardship seems to be the rst matter to be targeted in such a situation.  Prevalence of burnout and its three dimensions among Lebanese physicians Figure 3 Estimated marginal means of burnout and its dimensions through categories of threat perception scale (low and high) and nancial wellbeing (IFDFW).