This study is the first with the aim of producing an operational concept of professional burnout that enables classification into clinical subgroups. This concept was a need felt by clinicians because not all individuals with burnout present the same characteristics and prognosis. Analysis of the selected items and resulting scales for each profile has confirmed the factor validity and high reliability of the model. All of the operational definitions were faithful to the meanings contained in the Farber's theory.
The frenetic scale was composed of the involvement, ambition and overload dimensions. The high scores and low variability obtained in the items belonging to the involvement factor suggest that these responses may be influenced by social desirability, an aspect that should be considered when establishing anchoring points on a scalar level in later studies. The frenetic profile generally presented significant relations with exhaustion and with efficacy in a positive sense. These subjects are affected by burnout, given that this is what they express in therapy sessions when manifesting their psychological distress [3, 10]. However, judging from their characteristics and relations, they seem closer to the concept of workaholics [27–29]. Nevertheless, this addiction is associated with burnout , and may be one of the possible causes of it [31, 32] due to exhaustion of the individual's energy resources. Highly committed subjects typically show a great likelihood of developing burnout [2, 5], as their commitment and addiction are related by means of the absorption factor , making the employee a captive of his or her own activity [33–35]. Consequently, by learning to keep a certain distance from work and prioritising self-care, individuals could avoid excessive involvement and prevent burnout .
The underchallenged profile comprised the indifference, lack of development and boredom dimensions. This last factor, despite fulfilling Kaiser's criteria, presented a low percentage of explained variance, likely due to its high association with the other two factors. However, this factor should be included in the model because its content clearly differs from that of the other two factors. We observed relations between the underchallenged profile and exhaustion, lack of efficacy and, particularly, cynicism. Underchallenged employees have lost interest in the tasks involved in their work, have become cynical, and consequently seem to be affected by preliminary stages of burnout, such as dissatisfaction, limited variety and absence of feedback in tasks [15, 37]. In other works, it has been observed that individuals' perception that other jobs would better acknowledged their talents, lack of interest or gratification, and monotony could precede burnout [11, 13–15]. Specifically, the perception of minimum likelihood of personal development in a job predicts burnout in three years . Efforts aimed at increasing employees' personal and career development and reducing boredom and apathy appear to lower levels of stress and exhaustion .
The worn-out profile is characterised by neglect, lack of control and lack of acknowledgement. The low scores and lower variability for items belonging to the neglect factor suggest that social desirability may have influenced subjects' responses. The worn-out type presents significant relations with exhaustion, cynicism and lack of efficacy, and therefore appears to be the profile that best fits the definition of burnout provided by Maslach, Schaufeli and Leiter . Their neglect and/or apathy are associated with a lack of efficacy and may be inversely related to drive, participation and absorption , aspects considered diametrically opposed to burnout . The desperation caused by absence of control over results has been related to high levels of stress, exhaustion, emotional fatigue and depersonalisation [38, 39, 42, 43], which is in line with our results. The current study also shows that the perception of lack of acknowledgement is strongly associated with cynicism. Moreover, this appears to produce dissatisfaction and burnout in general . Greater acknowledgement seems to have a positive influence on the work climate of an organisation, reducing exhaustion and raising quality of life at work [44, 45].
Structural conditions, such as the temporary nature of work contracts, accentuate the development of some types of burnout. According to our results, temporary employees exhibit significantly higher scores for the frenetic subtype, associated with excessive dedication. Permanent employees displayed significantly higher scores for the underchallenged and worn-out subtypes, characterised by lower dedication. Significant differences were also found in the involvement, ambition, indifference, boredom, lack of acknowledgement and neglect dimensions, with the first two being higher in temporary workers, and the remaining dimensions higher in permanent employees. The structural condition of the temporary nature of work contracts appears to be associated with the type of burnout experienced, perhaps owing to differential involvement in work tasks. On the contrary, there were no significant differences by gender.
Although the characteristics of the subtypes may comprise determining factors for burnout syndrome, not all profiles fit the definition of Maslach, Schaufeli and Leiter  in the same way. These results can be explained if we interpret the burnout subtypes as different stages in the development of the syndrome, as proposed by Montero-Marín et al. . The development of burnout syndrome is arranged longitudinally by degree of dedication at work, which progresses from more to less (from enthusiasm to apathy) [5, 10, 46, 47]. Therefore, burnout appears to develop at a time of excessive involvement and commitment, typical of the frenetic profile [2, 5, 10, 38]. Given that it is not easy to maintain this level of activity without becoming exhausted or affected , workers will adopt a certain distancing to protect themselves, behaving with indifference and cynicism [48, 49]. While alleviating excess activity owing to excessive involvement, this distancing produces the type of frustration and stress suffered by the underchallenged profile . Distancing also erodes the perception of efficacy in the long run by leading to passive coping strategies, such as neglect of responsibilities and emotional venting, which are typical of the worn-out profile [51–55].
Subtypes are affected by different sources of stress and discontent at work, depending on the level of dedication with which they cope with obstacles and difficulties. Consequently, in order to efficiently adapt treatment strategies for burnout syndrome, we must specifically consider the burnout subtype experienced in each case. From a clinical perspective, exclusive consideration for the most recent manifestations of the syndrome, as performed in current evaluation standards, are insufficient. In order to overcome this limitation, it is necessary to have a more extensive definition for burnout syndrome that takes into account the level of involvement with which subjects cope with their work as part of the syndrome development process.
This study has several limitations. First, a low response rate was obtained. However, the rate is quite similar to those found in previous studies using internet surveys [17, 18]. This low rate could produce a bias in assessing point prevalence values, but does not affect the assessment of relationship patterns among different variables . In addition, differences in response rates based on occupational level could decrease the representativeness of the sample; however, all of the various jobs showed the expected response rate values [17, 18]. Another limitation is the sample selection, which was exclusively composed of workers from the University of Zaragoza. However, the sample was big and multi-occupational, as individuals in several jobs were included, improving the external validity of the study. Finally, this was exclusively a psychometric study; therefore, the predictive validity of the model has not yet been demonstrated. One of the main strengths of this study is that data quality was controlled by eliminating possible errors in the questionnaire transcription process through the use of purpose-designed software.