Burnout in Spanish Security Forces during the COVID-19 Pandemic

Since the beginning of the COVID-19 pandemic in Spain, members of the State Security Forces and the Armed Forces have been mobilized to guarantee the security and mobility of the population and to support health institutions by providing personnel for care, creating field hospitals, transferring the sick and the dead, etc. The objective of this study was to determine the levels of burnout in these professionals using the Maslach Burnout Inventory (MBI) scale, both in its different subscales and its total value. The study was developed using a quantitative methodology through a simple random sample (n = 2182). An ad hoc questionnaire was administered including variables related to: (a) socio-demographic issues, (b) subjective perceptions about their working conditions and the need for psychological and psychiatric treatment, and (c) the Death Anxiety Scale developed by Collett–Lester, and the MBI. The results show high levels of burnout (28.5%) in all its subscales: emotional exhaustion (53.8%), depersonalization (58.0%), and lack of personal development (46.3%). The logistic regression verifies a series of predictive variables that coincide in each of the subscales. These data indicate the need to implement prevention and treatment measures for workers so that their, stress, and anxiety to which they are subjected during their professional activity does not become a norm that can have negative repercussions for them, especially given the risk of new pandemic waves.


Introduction
In Spain, three entities integrate the set of corps and forces linked to different ministries that monitor the defense, protection, and security of the country and all its citizens: the Armed Forces (Fuerzas Armadas: FFAA), the State Security Forces and Corps (Fuerzas y Cuerpos de Seguridad del Estado: FFCCSE), and the Security Forces and Corps belonging to the Ministries of the Interior and Defense (Fuerzas y Cuerpos de Seguridad pertenecientes a los Ministerios del Interior y de Defensa: FFCCS). The Armed Forces (FFAA), "integrated into the Ministry of Defense, are [an] essential element of the national defense, and make up a unique entity that is conceived as an integrating set of the specific forms of action of each of its components: the Army, the Navy and the Air Force" [1]. They are organized in an organic structure designed to prepare the Force and other operative entities dedicated to the missions The forces experienced all this while still being required by the government [24]. The figures at the beginning of the pandemic showed that over 2300 national police and civil guards (1583 national police and 734 civil guards) were quarantining in their homes for symptoms compatible with COVID-19, 94 had confirmed positive for COVID-19 (60 in the police and 34 in the civil guard) as of 19 March 2020, and 19 died in the FFCCS according to official government figures as of 29 April 2020 [25]. However, the figures reported by the AUGC [23] on 16 April 2020 provided infection numbers of 1483 civil guards, 751 national police, and 868 local police. In terms of the number of isolated individuals, on that date, there were 2384 civil guards, 2360 national police, and 1671 local police. These figures continued to increase.
The stress level experienced by protection and advocacy professionals may have altered during the COVID-19 pandemic when intervening with persons with mental health problems exacerbated by fear of contagion, resource scarcity, economic uncertainty, and social alienation [19].
Since the end of the confinement in Spain on 21 June 2020, the relaxation of the citizenry, including non-compliance with health rules and restrictions, has increased the number of police interventions and the mass contact with possible transmitters [18]. Likewise, the second wave of the pandemic brought an increase in active cases and hospital admissions exceeding the cases located during the first wave in a one-month period.
All situations of psychosocial risk [26][27][28][29][30], which has coincided in Spain with numerous environmental disasters such as fires and floods, could once again generate situations of maximum stress. In a very short time period, both the civilian population and the professionals who compose the FFCCS have had to face a heavy workload, with the constant threat of possible contagion by the SARS-CoV-2, without rest since March 2020.

Objectives
The unprecedented situation faced by the members of the State Security Forces and Corps and the armed forces makes it necessary to consider these professionals, showing how the current health crisis has affected them from the dimension of burnout [8]. In this research, the general objective (GO) was to take a deep approach to the phenomenon of burnout in these professionals, contextualized in the temporary period during the first wave of COVID-19 in Spain. This GO is broken down into the following specific objectives (SOs): to quantify the general level of burnout and the set of subscales of the Maslach Burnout Inventory (MBI) scale (SO1), and to determine the predictive variables that influence these levels of burnout in these professionals (SO2).

Research Instrument
The research instrument was a questionnaire composed of several blocks to measure burnout in the State Security Forces and Corps and the Armed Forces based on the Maslach Burnout Inventory (MBI) developed by Maslach and Jackson [8] and validated in Spanish by García et al. [31]. This scale allows us to determine the general level of burnout, and to identify the existence of burnout through its three dimensions made up of subscales: Emotional Exhaustion (EE), Depersonalization (DP), and Personal Accomplishment (PA). The MBI is composed of 22 items with 7 response options through a Likert scale whose values range from 0 to 6. The AE scale is composed of 7 items, the DP of 5, and PA of 8. EE is considered low when its values are between 0 and 18, medium when they are between 19 and 26, and high when between 27 and 54. Regarding PA, low scores range between 0 and 5, average between 6 and 9, and high between 10 and 30. Finally, with regards to PA, a low score registers values between 0 and 33, medium between 34 and 39, and high between 40 and 56. For the overall burnout assessment, there must be high levels of EE and DP and a low level of PA.
Besides this burnout block, the questionnaire included: (a) questions regarding socio-demographic issues, (b) subjective perceptions about their working conditions and the need for psychological and psychiatric treatment, and (c) the Death Anxiety Scale developed by Collett-Lester [32] composed of 4 subscales on "Fear of one's Death", "Fear of the Process of Dying One's Own", "Fear of the Death of Others" and "Fear of the Process of Dying Others" with 7 items each, and a total of 28 items. The answers are of the Likert type from 1 (nothing) to 5 (a lot). This scale was included in the research as the increase in mortality during the first wave of the pandemic was very high in Spain and could provide a predictive variable of the burnout of these professionals.

Participants
There were 2182 participants in the research; their main characteristics are shown in Table 1. The majority of the sample was composed of men, representing 87.3% of the total, and women representing just 12.7%. In terms of age, the majority group was between 31 and 40 or 41 and 50 years old, at 36.6% each, totaling 73.2%. Regarding the workplace, 46.1% were National Police (Policía Nacional), 36.8% Civil Guards (Guardia Civil), 17.2% military, 3.9% local police, and 0.5% were other (autonomous and local police forces in the Basque Country and Catalonia or security agents).
In this descriptive study, the approval of the ethics committees of the Spanish universities of the researchers was not necessary (it is only required for experimental research). In our case, we subscribed to all the Codes of Good Practices of the Ethics Committees regarding research on human beings. The study was registered (code No. REPRIN-PEM-15B) and signed by the research team that performed the entire research process. The guidelines set by the Declaration of Helsinki were followed; all participants (n = 2182) provided their informed consent.

Variables
The dependent variable in this research was the burnout levels, both with the general index and of each one of its subscales: EE, DP, and PA.
For the independent variables, the following were used: (a) sex; (b) age; (c) professional body; (d) subjective perception of needing psychological/psychiatric treatment because of COVID-19; (e) assessment of whether work centers should offer psychological/psychiatric support as a consequence of COVID-19; (f) perception of need for psychological/psychiatric support; (g) assessment of whether the lack of PPE increased stress or anxiety level; (h) level of death anxiety (from the Collett-Lester scale); (i) whether they worked directly in COVID-19 tasks during the first wave of the pandemic in Spain; (j) whether they felt professional recognition during the first wave of the pandemic in Spain.

Procedure
The field research was conducted between 7 August and 7 September 2020. We accessed the participants through the collaboration of several professional associations and Spanish trade unions who were asked to participate in the questionnaire's dissemination. Initially, their collaboration was requested to participate in the research. Subsequently, these entities sent a link to access the questionnaire through a platform designed ad hoc by the University of Murcia for the preparation, administration, completion, and exploitation of statistical studies developed through surveys.
The following phases were followed in the process of data analysis: Initially, a frequency analysis was performed to determine the general level of burnout and each one of its subscales. Later, to obtain information about the variables linked to this phenomenon, a cross-table analysis was performing considering a chi-square p ≤ 0.05. Finally, to determine the variables predictive of the phenomenon, binary logistic regressions were performed, taking the existence of the general burnout level, and the medium-high level of each one of the EE, DP, and PA subscales as reference variables.
The variables used in the binary logistic regression are shown in Table 2. Table 2. Variables used in binary logistic regression.

Results
First, the burnout levels of the Security Forces and Corps are high, both in terms of the subscales and the general burnout level. In the first case, the Emotional Exhaustion subscale showed a high level in 53.8% of participants (70.1% with medium-high level). The Depersonalization subscale showed levels even higher than EE, at 58.0% (reaching a high percentage if we take the medium-high level of 85.5% as a reference). The Personal Accomplishment subscale was low (taken as a reference), reaching 46.3% (72.9% if considering the low-medium level). Finally, 28.5% of the total were found to be suffering from burnout, showing high levels of emotional exhaustion and depersonalization, and low levels of personal accomplishment (Table 3).  For the rest of the variables, 69.1% of the participants reported high levels of death anxiety. A total of 26.2% reported the subjective perception of needing psychological treatment. These data are similar to the percentage of people with burnout levels on the general scale (28.5%). A high percentage of respondents (88.2%) considered it necessary to provide psychological or psychiatric treatment from work centers because of COVID-19, and 52.5% reported they may need new psychological treatment in the future due to COVID-19, which is twice as many as those who state that they need it at present (26.2%).
A total of 87.2% considered that the lack of PPE during the most virulent weeks of the pandemic in the first wave in Spain increased their level of stress and/or anxiety. Almost 9 out of 10 people saw their level of stress and/or anxiety increase at work as a result of not having the basic supplies for their personal protection. Of the participants, 73.9% worked directly in a COVID-19 environment and 89.6% reported that they have not felt recognized in their professional activity by the institution in which they work. These percentages within this section of subjective assessment are high and demonstrate the position and disposition that the members of these corps may show in their professional activity, and how the context and working conditions may affect their psychological and social status.
Concerning the binary logistic regression applied in all the MBI subscales and to the total burnout index, adequately adjusted, robust, and reliable models were obtained. Each of the models is developed below.
The binary logistic regression of the Emotional Exhaustion subscale was found to be a statistically significant model (χ 2 = 301.911, p < 0.000). The model explained 18.9% (Nagelkerke's R 2 ) of the variance of moderate-high consumption and correctly classified 74.2% of the cases. The Hosmer-Lemeshow test showed that there were no significant differences between the observed and predicted results in the model (p = 0.800).
The variables included in the equation were: age, professional body, subjective perception of needing psychological/psychiatric treatment as a consequence of COVID-19 (NPPS), perception of whether they believe they may need psychological/psychiatric support (PPSN), assessment of whether the lack of PPE increased their level of stress or anxiety (PPE), high level of anxiety in the face of death (DA), and whether they felt professional recognition during the first wave of the pandemic in Spain (FWR).
Regarding age, the interval between 41 and 50 years had an odds ratio (OR) of 1.422 (95% CI, 1062-1904, p = 0.018). Those in the middle age group reported up to 1.4 times more emotional exhaustion than younger people (up to 30 years). As this was the only value within the age category that showed representativeness, we found this age group was at the highest risk of suffering emotional exhaustion. All professional categories showed high levels concerning the reference category (others). The results were high in all cases. For the military, OR = 6678, 95% CI; 1175-37,941, and p = 0.032. For the National Police, OR = 9342, 95% CI: 1660-52,589, and p = 0.011. For the Guardia Civil, OR = 7670, 95% CI: 1362-43,216, and p = 0.021. Finally, the local police recorded an OR of 8457, a 95% CI of 1400-51,092, and p = 0.018. Therefore, these professional bodies reported high levels compared to the reference level (others), with the national and local police standing out as suffering from emotional exhaustion at 9.3 and 8.4 times higher levels, respectively. The Guardia Civil and the military registered 7.6 and 6.6 times higher possibilities of suffering emotional exhaustion than the reference value, respectively.
Concerning subjective issues, NPPS presented an OR of 1.642, 95% CI: 1207-2234, and p = 0.002. Needing psychological/psychiatric support increased the risk of suffering from emotional exhaustion by 1.6 times. The PPSN registered an OR of 2.002 (95% CI: 1.570-2.552, p = 0.000); so, people who reported that they may need future psychological/psychiatric treatment were up to two times more likely to suffer emotional exhaustion than those who did not report this need. PPE had an OR of 1.585, 95% CI: 1.187-2.116, and p = 0,002. Therefore, people who reported that the lack of PPE has influenced their health were up to 1.5 times more predisposed to suffering EE. DA registered an OR of 2.759 (95% CI: 2.238-3.402, p = 0.000). This is one of the highest values within this subscale, along with the professional categories, and showed how DA increases the possibility of suffering EE by up to 2.7 times. FWR had an OR of 0.678, 95% CI: 0.494-0.932, and p = 0.017. Because the B was negative (−0.388), people who did not feel recognized for their work in the first wave of the pandemic were more likely to suffer.
The binary logistic regression of the Depersonalization subscale showed a statistically significant model (χ 2 = 127,948, p < 0.000). The model explained 10.1% (Nagelkerke's R 2 ) of the moderate-high consumption variance and correctly classified 83.0% of the cases. The Hosmer-Lemeshow test showed that there were no significant differences between the observed and predicted results in the model (p = 0.411).
The variables included in the equation were: sex, perception of whether the respondent believed they may need psychological/psychiatric support (PPSN), high level of death anxiety (DA), and whether they had felt professional recognition during the first wave of the pandemic in Spain (FWR).
Regarding sex, men had an OR of 1.744 95% CI: 1269-2398, and p = 0.001. Thus, men were almost twice as likely to suffer depersonalization as women. However, these data should be treated with caution since the sample is mainly composed of men. PPSN registered an OR of 1380, 95% CI: 1083-1758, and p = 0.009. Therefore, a connection exists between depersonalization and the perception of needing psychological/psychiatric treatment; people with this perception were almost 1.4 times more likely to report high values in this subscale. DA had an OR of 2.220, 95% CI: 1.740-2.832, and p = 0.000. Thus, the risk of suffering depersonalization was 2.2 times higher among people who had high values of AM. Finally, FWR recorded an OR of 0.653, 95% CI: 0.463-0.922, and p = 0.015. As with the Emotional Exhaustion subscale, its B was negative (-0.426), showing that people who did not felt recognized during the pandemic were more likely to suffer depersonalization.
For the Personal Accomplishment subscale, the binary logistic regression presented a statistically significant model (χ 2 = 160,786, p < 0.000). The model explained 9.7% (Nagelkerke's R 2 ) of the variance of moderately high consumption and correctly classified 62.4% of the cases. The Hosmer-Lemeshow test showed that there were no significant differences between the observed and predicted results in the model (p = 0.525).
The variables included in the PA equation, as with the Personal Accomplishment subscale, were: sex, age, subjective perception of needing psychological/psychiatric treatment because of COVID-19 (NPPS), subjective perception of whether psychological or psychiatric support should be offered from workplaces (PPSS), perception about whether they believe they may need psychological/psychiatric support (PPSN), whether the lack of PPE increased their level of stress or anxiety (PPE), high level of anxiety in the face of death (DA), and whether they have felt professional recognition during the first wave of the pandemic in Spain (FWR).
With respect to sex, men had an OR of 0.727, 95% CI: 0.553-0.957, and p = 0.023. Their negative B (-0.318) indicated that being a woman reduces the possibilities of personal fulfillment, although the values were small and the female population is significantly less. In relation to age, people between 31 and 41 years old had an OR of 0.679, 95% CI: 0.523-0.881, and p = 0.004. As with the sex variable, B was negative (-0.519), so people in this age range were more likely not to suffer from a lack of professional fulfillment compared to the reference value (age up to 30 years).
Regarding the subjective variables, the NPPS recorded an OR of 0.595, 95% CI: 0.470-0.753, and p = 0.000. As with the previous variables, its B was negative (−0.519), showing that people who do not need psychological or psychiatric support had a greater chance of reporting positive values on this subscale. The PPSS had an OR of 1.593, 95% CI: 1.190-2.132, and p = 0.002. Thus, people who reported the need for psychological/psychiatric treatment in the workplace showed higher levels, up to 1.5 times more in this subscale, than those who thought this type of treatment should not be provided. PPSN registered an OR of 0.718, 95% CI 0.577-0.894, and p = 0.003. As in previous cases, its B was negative (-0.331), indicating that people who claimed not needing psychological/psychiatric treatment were more likely to be represented in this subscale. PPE recorded an OR of 0.611, 95% CI: 0.454-0.823, and p = 0.001. Its B was also negative (-0.492), so people who did not consider the absence of PID as increasing their stress/anxiety levels were more likely to be represented in this subscale. DA had an OR of 0.604, 95% CI: 0.495-0.738, and p = 0.000. Their B, as in most cases in this subscale, had negative values (-0.504), showing that the professional performance of those who did not report suffering from DA may have been more affected. Finally, FWR showed an OR of 1.304, 95% CI: 1.046-1.624, and p = 0.018. Thus, the professional performance of those who reported feeling recognized for their work during the first wave of the pandemic may have been affected up to 1.3 times more than the rest.
Finally, approximating the Total MBI, we observed that the binary logistic regression presented a statistically significant model (χ 2 = 160,829, p < 0.000). The model explained 10.4% (Nagelkerke's R 2 ) of the moderate-high variance in consumption and correctly classified 71.0% of the cases. The Hosmer-Lemeshow test showed that there were no significant differences between the observed and predicted results in the model (p = 0.656).
The variables included in the equation were: age, NPPS, PPSS, PPSN, and DA. Those aged 31-40 years recorded an OR of 1606, 95% CI: 1194-2159, and p = 0.002. NAPP presented an OR of 1888, 95% CI: 1479-2409, and p = 0.000. Thus, NPPS increases the chances of suffering burnout by people who consider that they need psychological/psychiatric support by almost two times PPSS showed an OR of 0.678, 95% CI: 0.489-0.940, and p = 0.020, although B was negative (-0.389). This result showed that people who consider that psychological/psychiatric support should not be provided in their workplaces have a greater predisposition to suffer burnout. Subsequently, PPSS recorded an OR of 1630, 95% CI: 1277-2082, and p = 0.000. Those who reported that they may need psychological or psychiatric treatment were at 1.6 times greater risk of burnout. Finally, DA presented an OR of 2.236, 95% CI: 1.762-2.837, and p = 0.000. This result indicated that people who have a high level of DA experience up to 2.2 times more burnout. For an overview, see Table 4.

Discussion
One consequence of the COVID-19 health crisis has been the proliferation of studies related to clinical research to generate methods for the disease's treatment including the contextual and psychological effects derived from the pandemic. In this sense, we must determine the consequences of the pandemic on other levels [33][34][35][36], including the emotional and psychological levels.
Our findings showed the concern of security professionals regarding their mental health status. Of our respondents, 26.2% felt that they needed psychological treatment in the first wave of the pandemic, and 52.5% felt that they might need psychological treatment again.
We found that 28.5% of those surveyed reported a high level of burnout and high levels on the subscales of Emotional Exhaustion (53.8%) and Depersonalization (58.0%), and low levels of lack of personal fulfillment (46.3%). Similar results were found in a study in Italy [37], where there was no analysis of the protective elements that could protect these workers from the effect of burnout or of the weakening of the reasons for living and dehumanization.
Despite the measures implemented to contain mobility and improve hygiene, these groups have had to face the situation with caution and adopt new protocols [38] as in other countries [39]. One of these newly implemented safety standards was the use of PPE. Our study showed that 87.2% of the participants reported their level of stress and/or anxiety increased in their work because of not having the basic supplies for their personal protection. This is one of the reasons why the levels of death anxiety and burnout were elevated, as also reported in previous studies [5,21].
Considering the parameters of the Maslach Burnout Scale [8], we found that 28.5% of the people interviewed presented high levels of burnout with high values on the subscales of Emotional Exhaustion and Depersonalization, and low values in Personal Achievement. This suggests that although there were high scores in the previous subscales, the sense of achievement and self-efficacy were high, indicating great professional involvement. This result agrees with those of other European studies [37,40,41] considering the existence of burnout, although the percentage we found is much higher, suggesting that the context of the COVID-19 pandemic accentuates the number of people affected. Our findings contradict those of other studies such as one conducted in another context and in other countries, which find low job satisfaction in a group of police officers [42].
Many sociodemographic variables, such as subjective perceptions, increase the possibilities of suffering burnout in the security forces including its different subscales. We highlight the finding of (1) the need for psychological and psychiatric treatment; (2) the subjective perception of needing such treatment because of a new outbreak coronavirus, a situation that is currently being faced; (4) the need to establish psychological or psychiatric support services from work centers; (5) the absence of PPE; (6) anxiety in the face of death; and (7) the lack of recognition by the organizations for which they work.
The professionals reporting high levels of anxiety in the face of death, both in the set of subscales and for burnout, were found to be up to 2.2 times more likely to suffer from burnout. This shows the relationship between mortality, anxiety, and burnout. During the first wave of the pandemic in Spain, security and law enforcement professionals had to perform tasks that required contact with infected and dead people, as they were forced to disinfect centers, build field hospitals, transfer patients or bodies, provide health care, or have contact with infected and dead colleagues. They have also been direct witnesses to the enormous damage caused by the SARS-CoV-2, which has produced horrific scenes. It is therefore necessary to seek preventive actions against this problem [66,67], especially since it has been shown that these professional groups are more psychologically affected than the general population [68].
It is also significant the apparently paradoxical fact that the professionals who have felt most recognized by their organization during the first wave of the pandemic show a greater predisposition to feel a lack of professional fulfillment. It is not strange. There have also been cases of decorated wartime veterans who later indicated this circumstance, sometimes in the context of post-traumatic stress [69][70][71][72][73]. It may be, precisely, a consequence of this pandemic that is modifying so many patterns in our society, and that shows the impact it is having on these professionals. This is a topic that must continue to be analyzed in the future and for which more research is needed.

Conclusions
The novel processes and preventive measures that have been implemented to mitigate, as much as possible, the effects of the coronavirus have been named the new norm. In the groups studied here, measures must be taken to ensure that the high levels of stress experienced during the first wave of the pandemic do not also become the new normal. Working in situations of extreme stress should not be considered acceptable. This could cause the work performed by these professionals, in the context of the COVID-19 pandemic, to result in occupational disease. So, paradoxically, while these professionals protect the population by ensuring their safety and well-being, they may become ill in the eyes of the public institutions for which they work, with none of them taking decisive action for their protection.
Our findings highlight the relevance of conducting longitudinal research to report the existence of burnout syndrome over time, of implementing programs to reduce stress and promote healthy living habits, and of establishing preventive programs to identify professionals who present burnout and the phase in which they are within the police and military professionals.
Finally, this research has several limitations, including the difficulty of access, since it was only possible to approach these bodies through associations and unions in Spain. This situation may explain the lack of studies conducted with this group during the health crisis caused by COVID-19. However, it was necessary to establish a map of the current situation of the FFAA, FFCCS, and FFCCSE, whose role has been, along with the group of health workers, essential as they experienced a greater exposure to the virus. Another limitation found was that the proportion of women participating in the study was significantly less than that of men to establish a reliable gender comparison. Although this is a difficult issue to address because of the traditional disproportion of the percentage that exists in the professional group studied.
In short, the impact of the COVID-19 pandemic is being very important in the professional work of the group studied. Further studies will be necessary to determine its evolution along the successive waves. Only in this way will it be possible to obtain the most useful information to establish the best actions for the prevention of psychological problems in scenarios such as the one described. Which must always be based, of course, on an improvement in working conditions and the provision of sufficient material and human resources.