Exploring Global Research Trends in Burnout among Nursing Professionals: A Bibliometric Analysis

Nursing professionals are constantly exposed to several risk factors and high levels of stress that can affect their mental, emotional, and physical health, which can trigger burnout syndrome. This article aims to use bibliometric analysis to investigate burnout research trends among nursing professionals worldwide and to compare the contributions of different countries/institutions, scientific journals, authors, keywords, and citations. A bibliometric study was performed using the Scopus and Web of Science databases, in the period up to November 2021, aiming to search original and review articles in the English language regarding burnout in nursing professionals. The analysis was performed with a sample of 1406 articles. The most cited article indicated that 43% of nurses had high burnout scores, and a similar percentage were dissatisfied with their work. The most productive and most cited country in the world was the United States of America. Regarding the 10 most cited documents, there were no studies that could provide interventions to reduce burnout in nursing professionals, which can result in a need to develop studies on prevention capable of mitigating the problem, in view of the impacts generated on their mental, emotional, and physical health.


Introduction
Professionals working in the health area, in general, are exposed to high levels of occupational stress [1], especially nursing professionals who perform their activities under a high level of pressure and risk [2]. These professionals need to deal with different demands that can affect their emotional and psychological state and lead to burnout syndrome (BS).
The term burnout, described by Herbert Freudenberger in the 1970s, is a psychological syndrome resulting from a delayed response to chronic interpersonal stressors present in the work environment. This response encompasses some dimensions, which can be characterized by emotional/overwhelming exhaustion, depersonalization, feelings of cynicism, detachment from work, sense of ineffectiveness, and lack of accomplishment [3,4].
In addition, it is important to highlight people who struggle with depression. This group may also experience burnout, and can underestimate the severity of their condition, not seeking medical help when self-identifying as "burned out". Therefore, the line between burnout and depression must be carefully considered and not be distinguished [5].
Furthermore, BS is included in the eleventh revision of the International Classification of Diseases (ICD-11), being a problem related to employment or unemployment, resulting from chronic workplace stress not successfully managed [6].

Study Design
The bibliometric study aimed to map and investigate the scientific production of burnout syndrome in nursing professionals. The bibliometric analysis provided statistics that summarized publications and analyzed the literature [35]. It consisted of a tool to explore information in scientific studies through quantitative approaches [36].

Study Source
The databases selected to carry out the research were Scopus and Web of Science. The choice of these databases was due to the number of indexed journals and because of the broad coverage of international publications relevant to the subject.

Search Strategy
The sample included the articles of the databases up to 9 November 2021, using search terms in English combined through Boolean operators. The databases were used to retrieve the relevant publications using the following search string: (TITLE: "burnout" OR Healthcare 2021, 9,1680 3 of 16 "burn-out" OR "burn out" OR "professional burnout" (MeSH Terms) OR "occupational burnout" (MeSH Terms) OR "burnout, occupational" (MeSH Terms) OR "career burnout" (MeSH Terms) OR "burnout, career" (MeSH Terms)) AND (TITLE: ("nurse" OR "nursing" OR "nurses" (MeSH Terms)). The research query conducted in the titles aimed to minimize the risk of obtaining false-positive results (irrelevant ones) [33].

Data Collect
The study included: (i) original and review articles; (ii) articles in the English language only; (iii) articles that comprehended any aspect of BS in nursing professionals. Exclusion criteria consisted of publications that did not meet the inclusion criteria, as well as duplicated publications. The articles were retrieved from the databases by two researchers to avoid biases in the selection of articles to be evaluated.
Meetings were held for debate and agreement among researchers about the inclusion or exclusion of studies. In case of disagreements that could not be resolved by means of a consensus, a third researcher was called to improve decision-making. After eliminating nonrelevant studies, the dataset for bibliometric analysis was formulated. Figure 1 presents a flowchart with the steps of the process for selecting articles aligned with the research question. In the eligibility process, a total of 3076 articles (1705 on Scopus and 1371 on Web of Science) were identified. After inserting the filters according to the inclusion criteria, 2408 articles (1379 on Scopus and 1029 on Web of Science) were retrieved. A preliminary analysis of the studies was developed based on the reading of titles and abstracts, with 1002 documents being excluded because they were not related to the research topic or they were duplicated articles. At the end of the steps, this bibliometric review included 1406 studies to develop the analysis, as seen in the following flowchart.

Search Strategy
The sample included the articles of the databases up to 09 November 2021, using search terms in English combined through Boolean operators. The databases were used to retrieve the relevant publications using the following search string: (TITLE: "burnout" OR "burn-out" OR "burn out" OR "professional burnout" (MeSH Terms) OR "occupational burnout" (MeSH Terms) OR "burnout, occupational" (MeSH Terms) OR "career burnout" (MeSH Terms) OR "burnout, career" (MeSH Terms)) AND (TITLE: ("nurse" OR "nursing" OR "nurses" (MeSH Terms)). The research query conducted in the titles aimed to minimize the risk of obtaining false-positive results (irrelevant ones) [33].

Data Collect
The study included: (i) original and review articles; (ii) articles in the English language only; (iii) articles that comprehended any aspect of BS in nursing professionals. Exclusion criteria consisted of publications that did not meet the inclusion criteria, as well as duplicated publications. The articles were retrieved from the databases by two researchers to avoid biases in the selection of articles to be evaluated.
Meetings were held for debate and agreement among researchers about the inclusion or exclusion of studies. In case of disagreements that could not be resolved by means of a consensus, a third researcher was called to improve decision-making. After eliminating nonrelevant studies, the dataset for bibliometric analysis was formulated. Figure 1 presents a flowchart with the steps of the process for selecting articles aligned with the research question. In the eligibility process, a total of 3076 articles (1705 on Scopus and 1371 on Web of Science) were identified. After inserting the filters according to the inclusion criteria, 2408 articles (1379 on Scopus and 1029 on Web of Science) were retrieved. A preliminary analysis of the studies was developed based on the reading of titles and abstracts, with 1002 documents being excluded because they were not related to the research topic or they were duplicated articles. At the end of the steps, this bibliometric review included 1406 studies to develop the analysis, as seen in the following flowchart.

Bibliometric Analysis
Data were collected and exported to software called bibliometrix R-package. The package offered a set of tools that allowed the generation of descriptive analysis, statistical graphs, and science mapping, along with the development of research and quantitative analysis in bibliometrics [36].
Bibliometrics is a type of analysis based on identifying information in the literature about a particular area [37,38]. Bibliometric methods are used to quantitatively evaluate scientific production [39]. It generates analysis, including statistical methods that use specific indicators to obtain information about research activity [40]. It allows the discovery of emerging trends in articles and journals, making it possible to explore knowledge in a given area from the analysis of existing literature [41].
This analysis can identify the most productive authors, institutions, countries, and journals in an area of knowledge, can assess the impact of journals, can identify citation patterns, and can identify research topics and research trends based on literature. [42]. In recent years, this method has drawn considerable attention from nursing researchers [43].

Results
Our study selected the following scientific indicators to analyze the documents: volume and types of publication; annual scientific production; top 10 most relevant sources; top 10 source impact; top 10 most global and local cited publications; authors' production over time; scientific production by country; top 10 most relevant institutions; most cited country; mapping of the 10 most frequent keywords of the authors and databases; and trending topics.

Main Information about Data
The search query found 1406 publications. The most recurrent document type was article (n = 1313; 93%) and review (n = 93; 7%). These documents showed a total of 46,066 references and were published in 580 sources (journals, books, etc.). The average citation per document was 32.07 in the period from 1978 to 2021, and they were written by 4343 authors, with a mean of 0.324 documents per author, according to Table 1.    Figure 2 shows the number of articles published per year until 2021. The first publications in the databases took place in 1978. In 2021, 203 studies (191 articles and 12 reviews) were published, corresponding to the largest production in the period analyzed. The annual growth rate was 11.63%, indicating an exponential growth rate, especially in the last 3 years, with 124 papers published in 2019, 178 papers published in 2020, and 203 papers published in 2021. The number of publications per year remained below 50 publications until 2013, when there was a considerable increase in the number of publications. In 2021, the number of publications was more than double in comparison to 2017, and it was more than quadruple compared to 2012. The first review was published in 1988. The years 2020 and 2021 received the highest number of reviews: 19 and 12, respectively.

Top 10 Most Relevant Sources
The retrieved published papers gathered 580 different sources. Figure 3 shows the top 10 most relevant sources, that is, the ones that most contributed with published documents. The number of publications that the 10 leading journals produced was 299, representing 21.27% of the total number of publications by 2021. The three most productive sources were the "Journal of Advanced Nursing", with 55 articles published in the period of 1978-2021, followed by the "Journal of Nursing Management" (n = 49; 3%) and the "International Journal of Nursing Studies" (n = 46; 3%). Those three journals are within the subject area of nursing.

Top 10 Most Relevant Sources
The retrieved published papers gathered 580 different sources. Figure 3 shows the top 10 most relevant sources, that is, the ones that most contributed with published documents. The number of publications that the 10 leading journals produced was 299, representing 21.27% of the total number of publications by 2021. The three most productive sources were the "Journal of Advanced Nursing", with 55 articles published in the period of 1978-2021, followed by the "Journal of Nursing Management" (n = 49; 3%) and the "International Journal of Nursing Studies" (n = 46; 3%). Those three journals are within the subject area of nursing.

Top 10 Source Impact
The retrieved publications received 45,085 citations. The average number of citations per document was 32.07. Table 2

Top 10 Source Impact
The retrieved publications received 45,085 citations. The average number of citations per document was 32.07. Table 2

Top 10 Most Global and Local Cited Publications
This study investigated global and local citation of publications. The global citation refers to the number of citations an article received considering the entire database. On the other hand, local citation measures the number of citations a paper received considering the data included on the survey [44]. Table 3 shows the top 10 most global and local cited publications. It is important to mention that the ranking was made based on the local citations. The three most globally cited articles received 3178, 603, and 435 citations, respectively. The three most local cited

Top 10 Most Global and Local Cited Publications
This study investigated global and local citation of publications. The global citation refers to the number of citations an article received considering the entire database. On the other hand, local citation measures the number of citations a paper received considering the data included on the survey [44]. Table 3 shows the top 10 most global and local cited publications. It is important to mention that the ranking was made based on the local citations. The three most globally cited articles received 3178, 603, and 435 citations, respectively. The three most local cited publications obtained 127, 86, and 85 citations, respectively. The ranking of the two most cited articles is the same, both in the local and global citations.
The  Among the 10 most cited papers analyzed, it was possible to identify studies about the association between different variables and factors related to nurse retention [45]. Another study found that the mediation model of burnout was able to predict nurse turnover intentions [47]. There were studies aiming to identify the effect of work environment on nurses' burnout and its impact on patient satisfaction [46]; to assess burnout levels among nurses, as well as related variables [48]; and to explore the relationship between exhaustion of nurses and the quality of care [49].
Besides that, other studies had the objective of verifying the prevalence of burnout in emergency nurses and identifying its specific determinants [10]; of identifying the determinants of BS in nurses acting in the Intensive Care Unit (ICU) [50]; of comparing burnout with general and specific aspects related to job satisfaction, in addition to evaluating levels of burnout and job satisfaction [51]; and of evaluating the applicability of the Maslach Burnout Inventory (MBI) in nursing studies [52] Lastly, one study revealed that certain explanatory selected variables were responsible for 41.9% of emotional exhaustion, 16.4% of depersonalization, and 25.6% of personal accomplishment in the study sample randomly selected, consisting of 510 psychiatric nurses [53]. Figure 4 shows the production of the main authors over time. The size or diameter of the circle represents the number of articles, and the dark blue color identifies the most cited article per year. The first circle in the line means that, when the author started to publish on the analyzed topic, the larger the circle, the greater the number of papers published per year by author. Figure 4 shows the production of the main authors over time. The size or diameter of the circle represents the number of articles, and the dark blue color identifies the most cited article per year. The first circle in the line means that, when the author started to publish on the analyzed topic, the larger the circle, the greater the number of papers published per year by author.

Scientific Production by Country
Eighty-one countries-according to all author affiliations-contributed to the development of research on the subject. These countries are geographically distributed in North America, South America, Asia, Europe, Africa, and Oceania. Figure 5 shows the scientific production by country. The map was built through "Biblioshiny", a web interface of the bibliometrix package. Several shades of blue can indicate different productivity rates. Therefore, the darker the shade of blue, the greater the productivity. On the other hand, the gray color represents that there are no articles. The three most productive countries were the United States of America (USA) (n = 527), China (n = 401), and Spain (n = 234). America, South America, Asia, Europe, Africa, and Oceania. Figure 5 shows the scientific production by country. The map was built through "Biblioshiny", a web interface of the bibliometrix package. Several shades of blue can indicate different productivity rates. Therefore, the darker the shade of blue, the greater the productivity. On the other hand, the gray color represents that there are no articles. The three most productive countries were the United States of America (USA) (n = 527), China (n = 401), and Spain (n = 234).  Figure 6 lists the top 10 most relevant institutions. University of Granada was the most relevant institution (n = 80; 6.0%), followed by Shiraz University of Medical Sciences (n = 30; 2.0%) and University of Pennsylvania Central (n = 20; 1.0%). Three of the top ten most relevant institutions were in Iran, two in China, one in the USA, one in Australia, one in Belgium, one in Spain, and one in Canada.   Figure 7 shows the top 10 most cited countries that contributed to the development of research on the topic. The three most cited countries were the United States of America (n =11,203 citations), Canada (n = 4772 citations), and China (n = 2817 citations). The number of citations in the USA was more than double that in Canada, and nearly four times the total citations from China.   Figure 7 shows the top 10 most cited countries that contributed to the development of research on the topic. The three most cited countries were the United States of America (n =11,203 citations), Canada (n = 4772 citations), and China (n = 2817 citations). The number of citations in the USA was more than double that in Canada, and nearly four times the total citations from China.  Table 4 shows the mapping of the 1hn0 most frequent keywords from authors and databases (keywords plus). The most common keyword used by the authors was "burnout" (n = 795), followed by "nurses" (n = 284). On the other hand, the keywords provided by the databases with greater frequency were "burnout" (n = 1772), followed by "female"   Table 4 shows the mapping of the 1hn0 most frequent keywords from authors and databases (keywords plus). The most common keyword used by the authors was "burnout" (n = 795), followed by "nurses" (n = 284). On the other hand, the keywords provided by the databases with greater frequency were "burnout" (n = 1772), followed by "female" (n = 1203). The selection of keywords aimed to reflect the focus of the entire study performed and can be identified as future research trends [54]. Therefore, the keywords pro-vided bring an idea of the aspects covered in the studies, such as stress, compassion fatigue, job satisfaction, and job burnout, among others. 3.11. Trending Topics Figure 8 shows the trending topics. The size or diameter of the circle represents the frequency of the terms provided by the authors in the keywords, following the Biblioshiny parameters of five words minimum frequency and five words per year. In 2019, terms such as nurse (n = 85), burnout syndrome (n = 39), China (n = 19), patient safety (n = 18), and turnover intention (n = 17) were observed in the surveys. In 2020, it was possible to identify the following terms: mental health (n = 34), emotional exhaustion (n = 32), resilience (n = 27), occupational burnout (n = 25), and depression (n = 22). In 2021, the highlighted terms were COVID-19 (n = 30), emergency department (n = 6), Spain (n = 5), qualitative research (n = 5), and COVID-19 pandemic (n = 5). In March 2020, the World Health Organization (WHO) declared the COVID-19 pandemic [55]. The studies produced may reflect the impacts of this situation, considering that many health professionals work directly with patients affected by the disease. Recent studies indicate that, while facing the pandemic, some health professionals developed fatigue, exhaustion [56], and psychological distress [57,58]. The lack of knowledge about how the pandemic can impact on the mental state of health professionals is a research gap [59].

Discussion
Initially, 3076 documents were counted in the databases using the search terms. After verifying the articles obtained based on the inclusion and exclusion criteria, 1,406 papers were used to perform the bibliometric analysis. Among the publications retrieved for this study, 93% were articles and 7% were review.
Concerning the research developed in the sample, some considerations can be observed and extracted. The first articles were published in 1978. The annual scientific production increased over the analyzed time , indicating an exponential growth rate (11.63%). Therefore, it is clear that there is a concern to produce knowledge on the subject, considering that the BS can influence the mental health of the professional, the work performed, and the safety of the patient cared for in different health institutions.
In order of importance, the five most productive countries were as follows: USA, China, Spain, Iran, and Brazil. The five most cited included USA, Canada, China, Spain, and Netherlands. In the bibliometric indicator of the ten most relevant institutions, three institutions were located in Iran, two in China, one in the USA, one in Australia, one in Belgium, one in Spain, and one in Canada. Although Spain appeared with only one institution, it was the most productive affiliation, contributing with 72 articles.
The results showed no standardization for the selection of keywords provided by

Discussion
Initially, 3076 documents were counted in the databases using the search terms. After verifying the articles obtained based on the inclusion and exclusion criteria, 1406 papers were used to perform the bibliometric analysis. Among the publications retrieved for this study, 93% were articles and 7% were review.
Concerning the research developed in the sample, some considerations can be observed and extracted. The first articles were published in 1978. The annual scientific production increased over the analyzed time , indicating an exponential growth rate (11.63%). Therefore, it is clear that there is a concern to produce knowledge on the subject, considering that the BS can influence the mental health of the professional, the work performed, and the safety of the patient cared for in different health institutions.
In order of importance, the five most productive countries were as follows: USA, China, Spain, Iran, and Brazil. The five most cited included USA, Canada, China, Spain, and Netherlands. In the bibliometric indicator of the ten most relevant institutions, three institutions were located in Iran, two in China, one in the USA, one in Australia, one in Belgium, one in Spain, and one in Canada. Although Spain appeared with only one institution, it was the most productive affiliation, contributing with 72 articles.
The results showed no standardization for the selection of keywords provided by both authors and databases, that is, words with the same meaning. This is the case, for in-stance, for the keywords provided by the authors: "burnout" and "burnout syndrome". The "keywords plus" provided by the database showed studies about female and male professionals.
The The article that obtained the highest number of citations in both sample (local) and databases (global) was published in 2002 [45]. Among the results of this study, it is note-worthy that nurses working in hospitals with higher patient-to-nurse rates are more than twice as likely to suffer from work-related burnout and almost twice as likely to be dissatisfied with their jobs compared to nurses working in hospitals that have lower rates. In addition, the study indicated that 43% of the nurses had high burnout scores and a simi-lar percentage were dissatisfied with their work.
Another study pointed out that 32.8% of the nursing staff members who work in the Intensive Care Unit trigger severe BS. The ICU is an extremely stressful place and may contribute with a high rate of BS to the team [50]. Burnout syndrome is a complex problem and an important predictor in modern workplaces, substantially increasing its prevalence in the recent years [47,52], especially among nursing professionals [48].
Prevalence was measured in a study performed with nurses from six countries: USA, Canada, United Kingdom (Scotland and England), Germany, New Zealand, and Japan. Japan had the highest burnout levels in nurses. On the other hand, the lowest burnout levels were with the German nurses. Nurses from Canada, the United Kingdom, and New Zealand had lower burnout levels than nurses from Japan and the USA, but higher burnout than the German ones [49].
It was highlighted that BS is multi-factorially determined [53]. With this in mind, some risk factors can trigger BS, such as work overload, value conflicts, injustice, lack of support, and inadequate rewards [10,47]. In this sense, the role of the nurse team managers is crucial to keep the work environment less hostile [60,61].
The concern with the worsening of nurse shortages and the intention to leave was mentioned in some studies [45][46][47]. Some changes in the work environment can reduce professional burnout and the intention to leave, as well as produce positive effects on patient satisfaction [46]. When nurses are exhausted and dissatisfied with working conditions, patient satisfaction becomes much lower [51].
From 2019 onwards, it was possible to observe the trend of studies addressing terms in keywords, such as occupational burnout, patient safety, intention to leave, COVID-19, emergency room, resilience, emotional exhaustion, mental health, and depression. It was noteworthy that these terms demonstrate what could be developed in the study and may indicate future research trends. In the case of the term COVID-19, there is likely an increase in the number of publications evaluating the impact of this pandemic on the professionals' mental health.
After a month working in the frontlines of the COVID-19 pandemic, some nurses experienced moderate to severe emotional exhaustion and depersonalization. [76]. During the outbreak, the levels of anxiety and depression presented by health professionals were significantly higher [77]. However, it becomes a challenge to offer support or interventions that can contribute to diminishing stress and exhaustion, since there is a lack of time due to high work demands [78].
There are some limitations in the current study; for instance, the research included only articles in English in the sample. Even though English is a language mostly used in international databases, there could certainly be articles in other languages with more contributions. Besides that, only the Scopus and Web of Science databases were used to carry out the study. It would be interesting to develop this analysis on other databases to verify if the obtained results are similar.

Conclusions
This study analyzes the research and scientific production trends on burnout syndrome among nursing professionals. Based on bibliometric indicators, publications were investigated at a global level indexed in the Scopus and Web of Science databases. This research indicates that, although the first article was published in 1978, research on BS had a substantial growth from 2007 onwards.
Most of the retrieved studies are inclined to the prevalence, predictors, associated factors, and determinants. In the sample, there are few investigations that provide strategies for prevention, which indicates the importance of developing specific studies on interventions in nursing teams, especially the ones trying to reduce stress and burnout in these professionals.
Currently, COVID-19 appears as the most used term in research, demonstrating the possibility of developing studies that associate burnout with the pandemic. The analyzed sample displays studies that evaluate nursing professionals during the pandemic. Moreover, there were few studies during the outbreak addressing the topic.
For future research, it is interesting to use the methodology in other databases, thus making the results broader and disseminating what has been produced, thereby enabling the increase in the evaluated sample. There is also the possibility of identifying other aspects and analyses that were not considered in this study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.