METHODS

RESULTS

The literature review identified a total of 2 757 articles published between 2009 and 2020 in which the search criteria "Maslach Burnout Inventory" appeared together with the words "physician" or "physician" (Figure 1). As a result of the title and abstract reading, 2 603 articles were excluded leaving a total of 154 studies in analysis. One of the main reasons of removing articles at this stage was the duplication in the results from different databases. We also removed research published in other languages but English, Spanish and Portuguese, studies with fewer than 20 subjects or in which the study population did not include graduate physicians.

Figure 1 Flow diagram used in the identification, selection, eligibility and inclusion of articles. Diagram adapted from the PRISMA workflow¹⁰  

Of the 154 articles selected, those that did not use the MBI as a tool to diagnose BOS or used an adaptation of this tool that has not been validated were excluded. We also excluded papers that included all the personnel of the care centers and the data referring to medical professionals could not be extracted independently from the other groups studied.

Finally, a total of 65 articles were selected for eligibility (Figure 1). Of these, we found that 35 studies evaluated BOS in physicians alone or together with other health care professionals (nurses and/or physical therapists and/or ancillary staff), however it allowed data to be extracted from the evaluation of physicians independently. From the inspection of the full texts, 35 studies were selected to extract data on the prevalence of BOS and its relationship with extra-occupational sociodemographic factors.

Table 1 shows the results of the overall sample and the total number of selected articles published between 2009 and 2020. The years 2016 and 2018 stand out with 7 and 6 articles selected. The final selection does not present any research published in 2010.

A highly heterogeneous sample was obtained in the number of investigations selected by country (Table 1). Of the total number of articles included in the review, 4 corresponded to research carried out in hospital institutions in Argentina, 1 in Costa Rica, 1 in Chile, 2 in Colombia, 2 in Ecuador, 3 in Peru, 1 in Uruguay, 2 in Venezuela, with Brazil standing out with 19. However, the sample size does not correspond to the number of articles per country (Table 1). Countries such as Ecuador and Peru, with 2 and 3 articles selected respectively, present larger sample sizes (11 = 3012 and 1 = 2511, respectively) than Brazil (1=2377). The sample sizes of Ecuador and Peru respond to the fact that for both countries articles were found that conducted studies at the national level. The remain countries, the sample size is more closely related to the number of articles included: Argentina with a sample size of 404 individuals, Venezuela with 205, Colombia with 190, Costa Rica with 94, Uruguay with 82, and Chile with 45 (Figure 2).

Figure 2 Number of individuals included in studies published between 2009 and 2020 

The average age of the individuals surveyed in the selected studies ranged from 25 to 49 years (Table 1), with the median of 39 and a coefficient of variation (CV) of 20.3 %. The representation of each gender is balanced in most studies with 15 (46.8 %) studies containing 40-60 %o female sex; 6 (18.7 %>) with 60-80 %> and 8 (25 %>) with 20-40 %. Some studies with an asymmetrical representation of both genders stand out. Such is the case of the study by De Novais et al. 2016, with 4.7 % of females in a sample of 43 cases, and the investigations by Da Silva et al. 2017 and Martins et al. 2011, with 87 % and 81 % of females in a total of 78 and 74 individuals included in the study, respectively (Table 1).

The prevalence of BOS varied considerably, with interval values from 2.4 % to 83.3 % (Table 1). No correlation was found between sample size (N) and BOS, despite transforming the data to obtain a normal distribution (p=0.993, Figure 3A). Similarly, significant heterogeneity is observed in the prevalence of impairment in the subscales of the MBI with frequencies between 3.4 % and 79 % for EE; between 1 % and 68 % for DP, and between 1.4 % and 94.6 % for PA (Table 1). A positive correlation was found in the prevalence of the three subscales of the MBI with r2 values =0.72 between EE and DP (p<0.001); 0.39 for EE and RP (p = 0.029), and 0.53 for DP and RP (p=0.0021) (Figure 3B).

Figure 3 Correlation analysis. (A) Sample size (N) and prevalence of BOS found for each study. (B) Relationship in the prevalence of the MBI subscales. EE, emotional exhaustion. DP, depersonalization. PA, personal development 

Twenty-five studies (71.4 %) address the relationship of extra-occupational sociodemographic factors in relation to the development of BOS (Table 1). In 14 (56 %) of them, one or more extra-occupational factors are found to be prevalent, either as stressors or protectors. The extra-occupational factors studied include age, sex, marital status, children, the practice of sports and physical activities, living with family, professing a religion, hours of relaxation practicing a hobby or meeting with friends, place of residence (metropolitan area or not) and home ownership or not (Table 2).

DISCUSSION

BOS is a syndrome lead by poor management of workload and work environment ¹. For healthcare personnel, the emotional burden derived from the constant interaction with people suffering from illnesses, sometimes have low chances to be solved. All these factors make healthcare personnel more prone to suffer this syndrome ⁴⁶.

Previous studies have shown differences of features in the BOS development between physicians and nurses in the same hospital institution ⁴⁷. This shows that the healthcare workforce professionals with different job profiles and training may perceive and respond differently to the same stressors. Hence the importance of conducting studies focused on each of the professional profiles will help to establish more effective strategies to prevent and control the occurrence of this syndrome.

Within the healthcare workforce, physicians have a high burden of patient interaction and direct responsibility for outcomes of care. They also have bigger risks to develop BOS within hospital institutions. In fact, it is estimated that one in three physicians experience burnout at any given time ⁴⁶.

In general physicians play a central role in the quality of healthcare services. A high prevalence of BOS in medical staff may not only interfere with their own well-being, but also with patient satisfaction and safety. Additionally, the complications of the medical staff's service results in significant economic impacts for the afflicted person and for the institution. As an example, the cost per year associated with BOS in physicians in the United States has been estimated at $5 billion ⁴⁸.

Consequently, it is essential to focus the present work to identify the different sociodemographic and extra-occupational factors that contribute the development of BOS in medical personnel. When sociocultural differences are recognized in Latin American countries, many development programs at the regional level are aimed to create conditions more conducive to commercial exchange and, in general, to economic and social integration. Therefore, it is important to recognize the common points that affect the performance of different society sectors to contribute the development of programs to improve working conditions that have a regional scope. This is a particular interest at the present time when the COVID-19 pandemic is creating great pressure on all health systems and their professionals ⁴⁹.

Considering the sample size of all the sources consulted, the present work summarizes the prevalence rates identified for a total of 8949 health professionals, specifically medical personnel from nine Latin American countries: Argentina, Brazil, Chile, Costa Rica, Colombia, Ecuador, Peru, Uruguay, and Venezuela (Table 1). In general, the results show a high prevalence of BOS in the period 2009-2020 and confirm the relevance of conducting studies aimed to determine factors that trigger the development of this syndrome in medical personnel.

Seven studies from Argentina (2 studies), Brazil (3), Colombia (1) and Peru (1), reported a prevalence of BOS in medical staff between 25-50 %; similar to figures reported in another regions ^50-52. However, a total of sixteen studies among publications from different countries, Argentina (1), Brazil (10), Chile (1), Costa Rica (1), Uruguay (1) and Venezuela (2), found a prevalence of more than 50 % in medical personnel. Only ten studies reported less than 25 % prevalence (Table 1).

The differences found in the different institutions and countries of development of the prevalence of BOS in different studies should be interpreted carefully since they may not reflect real differences in the frequency of occurrence of this disease. The efforts to obtain comparable results that would reflect the prevalence of BOS in relation to sociodemographic factors in different countries and places in Latin America. The methodology used established the inclusion of only those studies that used the MBI scale to diagnose BOS. However, even in the application of this scale, there are important differences.

The 10 studies that found a prevalence of BOS of less than 25 %, 8 used as criteria for the diagnosis of BOS that the three subscales of the MBI were altered simultaneously, i.e. EE, DP or PA. In contrast to this, the studies that found a higher prevalence of BBS took as a condition for diagnosis that the person presented only one of the three altered domains. The discordance in the criteria of the diagnosis of BOS has been found in other literature studies ⁵³, and it is recognized as one of the main limitations that exist to define a prevalence rate at regional levels ⁴⁶.

Another source of divergence in the included studies is sample size. A smaller sample size may result in lower sensitivity to detect the prevalence of BOS, as well as the impact of sociodemographic factors. However, the correlation analysis performed evidenced that there is no bias in the prevalence of BOS in relation to the number of people included in each study (Figure 3A).

On the other hand, the correlation obtained between the different components of the MBI scale (Figure 3B) implies that there is a higher chance that individuals in whom only one of the altered subscales was identified were individuals in early stages of BOS development, still subclinical, and may make detection through the interview difficult. For this reason, both the studies that made the diagnosis based on the alteration of one of the factors and those that made the diagnosis only when the three subscales were altered were considered.

To contribute the development of effective programs for the prevention of BOS it is required an intervention that encompasses the personal response, the organizational context, and the interaction between the organization and the individual. Hence, it is necessary to understand not only the factors inherent to the organization and work environment, but also those associated with culture, interpersonal relationships, family and home ⁴⁶.

The possible influence of extra-work sociodemographic factors such as gender and family on the development of BOS was recognized from the initial work on the implementation of the MBI scale ⁵⁴. In fact, sex, age, marital status and number of children were found to be among the most frequently analyzed factors, with values of 57 %, 42 %, 45.7 % and 31.4 % respectively, out of 25 papers that analyzed extra-occupational factors in our analysis.

It has been found that men and women experience burnout differently. Generally speaking, women tend to experience greater EE, while men are more likely to experience DP ⁵⁵. Age may influence the occurrence of burnout and maturity carries out better available strategies of emotional self-control, as a result, the person is less impressionable ⁵⁶. Marital status and having children can also influence, because professionals in this condition can find goals beyond work, allowing them to find strength in other areas of life and providing them motivation to cope their work ⁵⁷. However, when family life and work interfere with each other, BOS can occur ⁵⁸.

The particular interest that other factors included in a more limited number of studies such as physical activity (20 %), hours of relaxation (11.4 %), living together (25.7 %), and place of residence (17.1 %), showed influence on the development of BOS with frequencies between 25 % and 33 %; similar to those of the commonly analyzed factors. Physical activity has been identified as one of the best ways to cope BOS because it promotes physical benefits, controls stress, and reduces anxiety and depression levels ⁵⁹. Likewise, quality time off, time to share with friends as well as a well-balanced family relationship can contribute to offset the stress derived from work relationships.

The influence of the place of residence has been analyzed from several angles, including elements related to the proximity to the workplace and the quality of life it offers, i.e. satisfaction with basic services. Elements such as close relationships with family may favor greater emotional stability, considering this interaction as an element that contributes to reduce over-involvement in work. It has also been observed that the greater the proximity to the workplace, the lower the burnout index. Following this last observation, results of a study proved that 83.5 % of the individuals surveyed stated that they would not take a job at a distance of 80 km from their place of residence and only 16.5 % said that they would take it ⁶⁰.

A remarkably high prevalence of BOS is identified in the medical personnel of several countries of Latin America in the last decade, but sociodemographic factors outside of work are still underrepresented in the prevalence studies of BOS. There is no correspondence between the extra-work factors included more frequently in the studies (sex, marital status, and age), with the factors found with the greatest incidence in the development of BOS (marital status, physical activity, coexistence, and location of residence).

The results obtained evidence the need to assess the risk of BOS and explore possible influence on the development of this syndrome in medical staff including a broad spectrum of sociodemographic and extra-occupational factors. Studies on burnout in healthcare professionals conducted at the regional level may achieve standardization of the extra-occupational factors analyzed and greater information regarding their influence ⁵¹. All this could contribute to achieve more effective BOS detection and treatment programs by being based on well-established evidence and not only on the empirical context that is currently accumulating ♦