Burnout among Direct-Care Workers in Nursing Homes during the COVID-19 Pandemic in Spain: A Preventive and Educational Focus for Sustainable Workplaces

Due to their continuous contact with pain and death, healthcare workers have one of the most stressful professions. Pain and death are more common in nursing homes. During the health crisis associated with COVID-19, these work centers have been characterized as spaces of high vulnerability to infection for the elderly, with a high mortality rate. This research aims to determine how the health crisis associated with COVID-19 has influenced healthcare professionals working in nursing homes for the elderly. Using a quantitative and cross-sectional method, the research was developed in residential centers in the Region of Murcia (Spain) during the second wave of the pandemic. A survey design based mainly on the Maslach Burnout Inventory (MBI) was employed. MBI measures three subscales: emotional burnout, depersonalization, and self-fulfillment. The results show that 6.4% of the respondents were burned out, 53.8% of the participants were emotionally exhausted, 35.1% were found to suffer from depersonalization, and in the case of personal development, the respondents showed a low level of 15.6%. This study shows the need to consider the establishment, in the academic environment, of training programs for health professionals related to coping with, managing, and identifying stress, especially in adverse circumstances. Similarly, in the professional field, it is necessary to develop strategies to prevent stress and anxiety in the workplace. The development of training programs for this purpose is essential for achieving a sustainable work context.


Introduction
The appearance at the end of 2019 of the coronavirus (SAR-CoV-2) in Wuhan (China), which provoked the subsequent COVID-19 pandemic, necessitated a reorganization of all social, medical, political, cultural, and individual scenarios all over the world. Since 30 January 2020, when the World Health Organization called it a "public health emergency of international concern" [1], terror and panic gripped the world's population, mainly because of the rapid spread and virulence of the virus, especially among the most vulnerable healthcare population: the sick and the elderly. This means that many people were the hardest hit by COVID-19 in the first wave.
Recent studies [2][3][4][5][6] have shown that working in the COVID-19 pandemic is having a significant psychosocial impact on health professionals, especially those who work in more fragile contexts, with long working hours, without adequate Personal Protection Equipment, etc.
The main psychosocial risks are stress, burnout, and violence, which affect individuals physiologically, emotionally, cognitively, socially, or physically and can induce absenteeism and abandonment of work [7]. These risks guide the investigation of this social the second, but infections and deaths have continued to occur in smaller numbers, although they remain high.

Background
The healthcare profession, which includes residential care for the elderly, is the profession that generates the most work stress or burnout due to healthcare professionals' continuous contact with pain and suffering [25,26]. However, their performance of monotonous tasks, the excessive workload, the frequency and amount of social time, and deficient interpersonal relationships, together with the organizational climate and the characteristics of the job, are also predictive variables [27].
Burnout syndrome precipitates certain disorders that force the affected person to temporarily stop working [28], such as headaches or backaches, sexual and central nervous system dysfunctions, muscular tension, generalized musculoskeletal pain, and various cardiovascular and gastrointestinal problems [29].
Previous studies on burnout in health professionals have been able to relate certain symptoms of different disorders, such as mood, anxiety, and depression [30][31][32][33], affecting not only the workplace, but also the social and family environment [34,35].
In the geriatric context, there has been a lot of research on stress and burnout, since long-term care for dependent elderly people can negatively affect people's well-being and health [36], since it is a physically and emotionally demanding occupation.
Burnout is defined as a "prolonged response to chronic personal and relational stressors at work, determined by the dimensions known as burnout, depersonalization, and professional cynicism and inefficiency" [34,37].
Dr. Maslach and Dr. Jackson [37] spoke of a process of chronic contact stress, which was measured through an inventory developed by the aforementioned authors and used in the present study. It is composed of the following dimensions. Firstly, emotional fatigue, which is determined by the gradual loss of energy and exhaustion; that is, the experience of being emotionally exhausted by the demands of the work. Secondly, depersonalization, which indicates the degree to which each person recognizes attitudes and is marked by negative attitudes that result in a distancing from problems and even blaming the patients themselves for the problems suffered by the professional. Thirdly, lack of personal fulfillment, which is characterized by negative responses to oneself and one's work, with pseudo-depressive manifestations, tendencies to run away, physical and psychic exhaustion, and dehumanization. Thus, this third subscale evaluates feelings of self-efficiency and self-fulfillment at work.
Considering the context of the pandemic, and the problems associated with the work of nursing home healthcare providers, it is necessary to know the effects that nursing home workers may be experiencing. Thus, if the data are alarming, similar situations could be prevented in the future in the context of other possible waves of COVID-19 involving a high number of infected and dead people. Health professionals' levels of personal exhaustion, depersonalization, and self-fulfillment during the first wave exploded [20], which means that they suffered an increase in burnout syndrome.
The psychological health risk in these professionals must be faced not only when it occurs, but also from a preventive perspective, from which initial and continuous training programs are fundamental. At present, sustainable working environments must be developed that provide safety and well-being to workers in multiple dimensions, while enhancing their productivity even in extreme situations, such as those described, which must be contemplated in connection with these prevention strategies.
There has been little scientific literature related to the experiential processes of residential care workers, such as the development of burnout, both in Spain and in other countries, so researching this may help to improve the quality not only of the professionals themselves, but also of the care. If there is one good thing about the times of epidemics and pandemics, it is that they bring to the table the functioning of all categories of the population, showing their strengths and weaknesses at the structural and organizational, institutional, and even personal levels. Knowing this will help to improve our ability to face new events more efficiently.

Objectives
The increase in care pressure in residential centers due to the high mortality rate of residents in Spain may have had repercussions in the personal sphere of professionals and in their work dynamics, according to the dimensions of Maslach and Jackson [37]. In this context, two objectives are raised. The first is to determine the general burnout level and the set of subscales of the Maslach Burnout Inventory (MBI) (O1). The second is to determine the predictive variables that influence the burnout phenomenon in these professionals (O2).
The hypothesis guiding the present investigation is that because of the high mortality in residential centers for the elderly since the beginning of the health crisis, and since this population group is one of the most vulnerable in the COVID-19 pandemic, burnout may be endemic in their workers due to occupational stress and prolonged contact with death.

Instrument
The instrument selected to achieve the stated objectives was the MBI, developed by Maslach and Jackson [37], adapted by Gil-Monte [38], and applied to professionals working in residential homes for the elderly. This scale allows a general index of burnout and a set of subscales that compose it to be obtained, including Emotional Exhaustion (EE), Depersonalization (DP), and Personal Accomplishment (PA). The MBI is composed of 22 items through a Likert scale, whose response options range from 0 to 6. The EE scale is composed of seven items, the DP is composed of five, and the PA is composed of eight. In the case of EE, low levels are values between 0-18, medium levels are between 19-26, and high levels are between 27-54. Regarding DP, low levels are between 0-5, medium levels are between 6-9, and high levels oscillate between 10-30. Concerning PA, low levels comprise the values of 0-33, medium levels 34-39, and high levels 40-56. To interpret the existence of burnout, a triple condition must be met: a high level in EE and DP and a low level in PA.
Concerning the Cronbach's Alpha means, α (or the original version [37] of the alpha coefficient) was obtained for the EE dimension (0.89), for the DP dimension (0.77), and the PA dimension (0.74). In our version adapted to Spanish [38], a Cronbach's Alpha mean of 0.85 was obtained for the EE dimension, 0.58 for the DP dimension, and 0.71 for the PA dimension.
In addition to this burnout block, the questionnaire was made up of (a) a block of socio-demographic issues, and (b) subjective perceptions about their working conditions and the need for psychological and psychiatric treatment.

Sample
Concerning the participants in the research, the number of residential care workers was n = 296, and the main characteristics are shown in Table 1. Most of the participants were female, representing 86.1% of the total, with the percentage of men remaining at only 13.9%. This reflects the feminization of care, both in formal and informal spaces. In terms of age, the majority group was up to 30 years old and between 41-50 years old, registering values of 31.8% and 28.7%, respectively, together adding up to 60.5%.
Concerning the professional category, Nurse Assistants stand out, being 88.8% of all participants. It should be noted that, in the case of Spain, Nurse Assistants are not an extension of Nurses, since they are different professional categories. While a Nurse requires a university degree within the European Higher Education Area, the same is not true for Nurse Assistants, whose required degree is not a university degree and who perform functions related to personal hygiene, meal support, room preparation, etc. In residential centers, most of the staff are Nurse Assistants and Nurses, who administer medication, monitor vital signs, follow the guidelines established by the physician, etc. Within the healthcare setting, physicians have a residual representativeness, given that when there is an important medical problem requiring hospital care, these patients are transferred to these centers. Therefore, the sample includes more Nurse Assistants. Concerning the type of center, 26.7% work in public residences and 73.3% in residences subsidized by the public administration, where there is financing from the public administration in the management of the residential space. No residential center took part that was entirely privately funded, which has a residual character in the overall care of dependent persons in Spain. Finally, of the group of participants, only 18.4% were experiencing an active COVID-19 outbreak in their residential center at the time of completing the questionnaire.

Variables
The dependent variable of this research corresponds to the burnout levels, both in the case of the general index and each one of its subscales: EE, DP, and PA.
While other questions could have been raised, all independent variables are related to the MBI and provide information on: (a) the environment in which the healthcare professionals work and their feeling of belonging to their institution and (b) their feelings and perception of their own and collective mental health as a consequence of the pandemic.

Procedure
The field research was conducted from 9 November to 27 November 2020, during the peak of the second wave of the COVID-19 in Spain. The study focused on a single territory (in this case, the Region of Murcia), on the one hand, given the difficulties of implementing it in the whole of Spain, and on the other, the need to obtain data on the implications of emotional stress and anxiety in workers in residential centers at such a crucial time of the pandemic. This research had the benefit of collaborating with the Murcian Institute of Social Action (Instituto Murciano de Acción Social), which administered the questionnaire to its workers.
The research work met all the ethical requirements in the scientific field. All the participants (n = 296) gave their informed consent, following the Declaration of Helsinki. While in Spain, no official approval of the ethical committees of the universities for studies of this nature, i.e., descriptive ones (it is required only in the experimental ones) is needed, in this study, we applied all the Codes of Good Practice of Research in Human Beings of the universities. The work was registered and signed by the research team, which includes the authors of the work, with the code nº REPRIN-PEM-21, abiding the rules of all Ethics Committees.
The following phases were followed in the process of data exploitation and analysis using the SPSS v.24 program (IBM). Initially, a frequency analysis was performed to determine the existence of the general level of burnout and each one of its subscales. Then, to obtain information about the variables linked to this phenomenon, a cross-table analysis was performed, taking into consideration a chi-square p ≤ 0.05. Finally, to determine the predictive variables of the phenomenon, binary logistic regressions were performed, taking as variables the existence of the general burnout level and the reference level of each subscale to determine the existence of burnout.
The variables used in the binary logistic regression are shown in the following table (Table 2).

Results
First, it can be stated that the burnout levels were not excessively high, from the point of view of the set of subscales and concerning the general burnout level. Second, the EE subscale recorded the highest values, reaching 53.8% of the participants. Concerning DP, the high level of the subscale showed values lower than that of EE, standing at 35.1%, which is similar to the low-level values of EE, which stood at 35.8%. For the subscale PA, the low level (reference value for measuring burnout) reached 15.6%. Finally, concerning Burnout, 6.4% of workers who showed high levels of EE and DP and low levels of PA suffered from it. Regarding NPPS, 21.6% stated that they needed it at present; however, 41.0% stated PPSN. One fact stands out, namely, that 90.4% stated PPSS, which is a high figure (Table 3). Then, the technique of crossed tables was applied, using each subscale as dependent variables: EE, DP, and PA, with a chi-square significance level ≤ 0.05 (See Appendix A). The most outstanding results are the following. First, no significant differences are observed in sex in the set of subscales. We did observe some differences in age, especially in the younger age group in PA. This subscale obtained the highest values among people up to 30 years of age and the lowest values in older people. Therefore, the lack of PA decreases as age increases in residential professionals.
Concerning the professional category, there are also differences. The highest levels of EE and DP are found among Nurse Assistants, while Nurses showed the highest PA. The participation of other professional categories (including Physicians) was irrelevant in the sample.
In terms of residential center ownership (public/private), among public residential staff, we found high values of PA (31.8%), followed closely by EE (29.0%). Concerning staff in private residences, the DP scale was over-represented (75.0%), followed by EE (71.0%).
Significant data are obtained concerning NPPS; although they were very similar between subscales, we obtained significance, and the need for psychological treatment was determined by the values obtained in the set of subscales. Thus, it was possible to establish an association between the levels of burnout and the need for this type of treatment. Moreover, there is a significant association (in terms of chi-square) between the PPSN variable and EE. In this way, the EE situation can determine PPSN.
As for PPSS, the values were high in the set of subscales, but they stood out especially in EE and, again, had a relevant level of significance in terms of chi-square. Finally, concerning ACOW, it had similar values in the set of subscales (around 20%), so the existence of an active outbreak of COVID-19 in residences did not predispose people to a high level of burnout.
Subsequently, the technique of binary logistic regression was applied to all the subscales of the MBI and to the burnout index. Only adequately adjusted, robust, and reliable models were obtained for EE and the burnout index. Each of the models is developed below.
The binary logistic regression of EE presented a statistically significant model (X 2 = 50.600, p < 0.001). The model explained 22.5% (Nagelkerke's R 2 ) of the variance and correctly classified 67.5% of the cases. The Hosmer-Lemeshow test showed that there were no significant differences between the observed and predicted results in the model, with a p = 0.998.
The In this way, the chance of suffering from burnout of those who consider that attention should be paid to their work centers in the context of COVID-19 increased in the PPSS subgroup by almost 3.5 times. Concerning the type of center, workers in subsidized centers showed an OR = 1.863 (CI 95% [1.021-3.400], p = 0.043). Therefore, when working in a private residence, the staff suffered up to two times more EE.
If the Burnout index was taken into consideration, we observed that the binary logistic regression presented a statistically significant model (X 2 = 10.091, p < 0.017). The model explained 4.9% (Nagelkerke's R 2 ) of the variance of moderate-high consumption and correctly classified 93.4% of the cases. The Hosmer-Lemeshow test showed that there were no significant differences between the results observed and those predicted in the model, with a p = 0.424. While the model explains a reduced percentage based on Nagelkerke's R 2 , it should be noted that the percentage of workers suffering from burnout is very low, at 6.4%; however, given that the model is properly adjusted, it is interesting to consider these results.
The variables included in the equation are: (a) Work (Nurses) and (b) NPPS. Regarding Nurses, an OR = 25.954 is obtained (CI 95% [1.478-455.647], p = 0.026). Therefore, being a nurse increased the risk of suffering burnout by up to almost 26 times, with respect to the reference category (others). The NPPS subgroup presented an OR = 3.835 (CI 95% [1.392-10.564], p = 0.009). Thus, the subjective perception of needing psychological treatment increased the chances of suffering burnout by almost four times (Table 4).

Discussion
Concerning the burnout rate, it is surprising that only 6.4% of the workers registered it, despite the collapse of these centers, the high mortality rate of the residents, and the media pressure on the public management of the residential centers. These figures are much lower than those found in other studies conducted on the pandemic in Spain because of COVID-19 on health workers or the Security Forces, which are essential sectors during the pandemic [20,39].
A characteristic of care is the fact that, if we make an approximation to the subscales, we observe a significant increase in the records of the three subscales that make up the MBI in comparison with other studies also carried out in this territory [40]. Likewise, the highest percentage of the subscales is concentrated in workers who suffer from EE, which stands at 53.8%. As for DP, 35.1% is concentrated in professionals working in homes. However, the lowest levels are found among people with a lack of PA, which only reaches 15.6% of workers. Here, it is important to note that it mainly affects younger people up to 30 years of age, as noted in previous studies [41]. Thus, we consider that the lack of personal fulfillment decreases as age increases in residential care professionals. However, in other previous studies, no conclusive results were found for the age variable [38,[42][43][44].
Concerning the rest of the variables introduced in the research, the NPPS only reaches 21.6%, while the subjective perception of PPSN doubles to 41.1%. However, it reaches 90.4% of PPSS. The paradox is that, although NPPS is not perceived at the moment at an individual level, nor is there a disproportionate increase in the short term, the work climate, social problems, and the collective sphere of work affect the perception of the need for psychological or psychiatric attention to the workers affected by the COVID-19 pandemic.
A major controversy, introduced into the public debate as a result of the high mortality rate, has been the assumption that the public administration should take over residential facilities, especially after the process of medicalizing these facilities. However, significant differences can only be observed in terms of PA, and it is precisely in the workers in public residences where higher levels are recorded. Among the personnel of public residences, considering all the subscales of the MBI, the lack of PA (31.8%) stands out above EE and DP.
Regarding the subjective variables analyzed through the cross-tables, they are closely connected with elevated levels of EE, which are especially disposed to NPPS, PPSN, and PPSS. Thus, a high EE can significantly influence these aspects. No differences are observed between the ACOW and the records obtained in the different burnout subscales. Thus, ACOW does not cause increased levels of stress and/or anxiety, partly because the risk is very high and could be found in any workplace, despite not being detected. This assumption is unlike that employed in other studies, which relate stress to active cases [20].
While the applied binary logistic regressions only obtained significant data in EE and Burnout, the results to be considered are shown. As for EE, workers who consider PPSN and PPSS are at a greater risk. This finding is unlike that presented in other studies, which show that this subscale is determined by the professional category, mainly affecting those who have more responsibility for medical treatment [20,45]. Additionally, in this case, there are differences between workers in public nursing homes and those who work in private nursing homes, showing that the former have a greater risk of suffering from EE.
As for burnout, high values are obtained for nurses, whose levels are up to 26 times higher than the reference value. If we add to this the fact that the highest levels of DP are found in young people and in nurses, the data seem to show that young nurses suffer from a high PA, which, together with the other subscales, leads to burnout. It should also be considered that young nurses represent the highest percentage of professionals that are permanently present in residential centers, which is not the case for center directors or physicians. This responsibility can act as a stressor for young nurses who have developed their professional activity in this context, even though it is one of the fastest-growing professions [46]. Similar results have been found in other research on health personnel, where the highest burnout was seen in older professionals [20]. Moreover, the subjective perception of NPPS is also a predictive variable for suffering burnout. In this sense, some authors obtained results in both directions [47,48]. For example, some authors found that the less professional seniority workers have, the more likely they are to develop the syndrome, while other authors obtained a negative relationship between this variable and burnout [42,49].
One characteristic of care is the feminization of the personnel working in these centers [50][51][52][53], either in the informal space of the private sphere or, in this case, in residential care services.
In this context, therefore, it is more urgent than ever to develop specialized training programs, both initial and continuing, for professionals in this field [54][55][56][57][58]. In particular, one should focus on exceptional situations that, as in the case of the pandemic, may increase the psychological impact that workers may suffer [59][60][61][62][63]. Today, the concept of sustainable work environments must be analyzed in the academic world, in initial training, and in professional training proposals [64][65][66]. In the broad framework that this implies, seeking to reconcile work and family life, gender equality, flexibility, team building, welfare, workspaces, etc. [65], we determine measures for the care and prevention of health problems, among which psychological problems represent a great risk for healthcare workers [67][68][69].
In the context of all the current measures that seek the productivity and welfare of workers and that, in recent years, have been directed to improvements and progress towards a more collaboration, flexibility, justice, etc. [65], as well as more sustainability in the short term, exceptional events such as epidemics, natural disasters, economic crises, etc., must also be considered, as such events exacerbate the risk factors faced by workers. Educational strategies are fundamental [70,71], especially from a holistic perspective [72,73]. They are the basis of prevention, and it is, therefore, urgent that they are integrated, as has been pointed out, into the initial and ongoing processes of these professionals. Only in this context can true sustainability be achieved.

Conclusions
The results of the research show low levels of burnout; however, they show excessively high levels of Emotional Exhaustion. This may be a risk factor in the development of mental disorders in the medium and long term in professionals working in residential centers. Considering the results, it is necessary to establish, both in the academic environment and in the management of nursing homes, training related to coping, managing, and identifying stress through health sciences grades. Likewise, in the professional field, developing strategies to prevent stress and anxiety at work, such as promoting improved interpersonal relationships, is crucial to maintaining good personal health and a productive and sustainable work environment.
The prevention of stress must be adapted to the characteristics of the job, considering the optimization of the job, social support, and compensation for efforts that workers make [51]. In this sense, the creation, development, and consequent application of training programs are essential. It is not in vain that research, training, and education in health matters are supported by the Framework Directive on Health and Safety at Work (Directive 89/391 EEC) of the European Union.
Concerning the limitations of the study, we experienced difficulty in accessing the respondents, because after an intense period of living in the context of the first wave and then in the second wave, with more infections and deaths than the first, the predisposition to answer a questionnaire was low. Moreover, including a qualitative perspective would allow us to determine how burnout influences both the work dynamics and the individual sphere of the participants.
Another important limitation is the type of study, which is cross-sectional. In the future, it would be interesting to deepen this object of study by including the participants themselves to see how burnout levels develop from a longitudinal perspective.
Finally, it should be added that, in the future, the sample should be enlarged to include a greater representation of the professional groups of nursing homes.