A Qualitative and Longitudinal Study on the Impact of Telework in Times of COVID-19

: Mandatory teleworking has become a major tool of public authorities for mitigating the spread of COVID-19 when work activity allows it. A lexical and morphosyntactic analysis was performed with 22 employees interviewed at home by phone on two occasions during and after a national lockdown (April 2020 and March 2021). The results indicate that the organizational changes initially implemented became sustainable and highlight a change in work practices. Changes in working time structure were observed and led to a feeling of intensiﬁcation and/or increase in working hours. The preservation of the professional bond through informal exchanges required a deliberate communicative effort. The lack of face-to-face social relations deprived employees of both their usual ways of working and the meaning they found in them. Finally, the continuation of the health crisis (phase 2) and the multiple reorganizations generated a decrease in wellbeing (mental wear and tear). Employees feared that the company’s management would retain a working model based mainly on remote working.


Introduction
Since the beginning of the COVID-19 pandemic, many private and public companies have introduced teleworking. Work has been urgently rethought and redesigned by organizations, both in terms of organizational and collective management and in terms of task performance and content [1,2]. Teleworking as previously defined and practiced has undergone a transformation. Indeed, in France as in many countries, a massive, immediate, unanticipated, and forced deployment of teleworking was implemented during the first lockdown (March-May 2020) when the sector and the work activity allowed it. This period marked a change in the distribution of teleworking in France. For example, before the pandemic, only 4% of employees worked remotely at least one day a week, compared with 27% in January 2021 [3].
After an initial phase of mandatory working from home, an alternation of "stop and go" measures over the last two years has allowed employees to either return to their habitual workplace or to work from home. A new lockdown was imposed and then several curfews followed. The latter measures left companies freer to relax their teleworking arrangements. They then, depending on the health restrictions, organized themselves to ensure that their activities continued, and on several occasions, rearranged their organization.
Although it was not widespread in France until the pandemic, teleworking as an organization of professional activity has over the last few decades become an increasingly investigated subject in the field of research [4,5]. However, there is no consensus as to its definition. Some adopt a broad view of teleworking that includes all work done remotely by means of technology [6,7]. Others add various characteristics to it: the place(s) where the

Spatio-Temporal Reorganization of Work Activity
The unpredictability of the crisis and the suddenness of the government's health measures disrupted the routines established by companies and employees. Those companies that were able to maintain partial or full teleworking activity nevertheless had to reorganize their activities [14]. Contrary to expectations, a decrease in workload was not always observed in the literature. Indeed, the number of hours worked by employees during this period varied greatly depending on the organization, sector, and occupation of the employees. Several studies point to this finding, indicating that the number of working hours increased for some employees [15,16], and remained the same or decreased for others [10,15]. The decrease in the perceived workload was explained by the decrease in the volume of activity and part-time work due to the crisis [10]. The increase was more related to an additional amount of work, and/or an increase in the complexity of work management [17]. It should be noted that prior to the crisis, an increase in working time had been observed among teleworkers [18,19] for whom travel time was at least partly reinvested in work activity in order to deal with emergencies and unforeseen events [18]. Finally, this perceived increase in working time during the pandemic can also be explained by the increase in working time linked to family constraints [20,21]. Indeed, the closure of schools, day-care centers, nurseries, and other childcare facilities during the initial lockdown particularly altered the coordination of different spheres of life. The support that parents had to provide for children (school, homework, care) [1], and the maintenance of professional activities made the coordination of different spheres of life more complex [14]. The home became a place of multiple uses [22], where personal and professional spheres and the spatial and temporal requirements they entail overlapped. Many employees say that they succeeded in reconciling these spheres of life by introducing rules on sharing space and tools [23,24]. However, tensions and conflicts have arisen [25,26] due to the organization of a "double day" combining work and domestic demands [27]. The spatial configuration of the home has also been a consideration since the beginning of this crisis. Homes with an open space (e.g., garden or balcony) improved the feeling of satisfaction during the first lockdown [28]. Employees who benefited from an isolated and quiet place (e.g., to concentrate, to work without being disturbed) and from an appropriate equipment set-up for the work activity (an adapted workstation, necessary computer tools) were able to adapt to the situation [11]. This was more conducive to the reconciliation of work and private life [27]. However, some workers used personal equipment and tools that were not well suited to teleworking. These needs were also simultaneously coupled with those of other family members, making forward planning and the organization of material and digital resources necessary, something previously unknown [1].

The Proportion of Mediated Work in Individual Activity and Professional Relations
The abrupt entry into the health crisis left little time for organizations to anticipate employee needs. Those who had already experimented with telecommuting were better prepared for the situation [29]. Employees who had already teleworked prior to the pandemic had access to a large enough technical and hardware environment to work from home [30]. For the latter, the adjustment of telework to this early period of crisis was experienced as easier and generated some satisfaction [12]. Moreover, several studies had already shown the role of ICT in the implementation of telework [31] and in contributing to the temporal and spatial flexibility of telework [32][33][34]. This observation is similar to those made since the beginning of the crisis. Indeed, this transition to teleworking was more easily adopted by large companies already equipped with IT tools [29]. However, although some people had already teleworked, no one had done so under such conditions. The massive use of ICT and the multiplicity of information received [13] may have generated an overload and intensification of work, a finding already highlighted in the pre-pandemic literature [35][36][37]. The pandemic study by Molino [38] found an association between technostress and work-family conflict. This period brought an intrusion of work into the personal and family sphere through ICT. This led to a spillover into the different life spheres and conflicts of work and family roles, confirming the work already produced on this subject [38,39]. However, even though ICTs appear to be central in a teleworking organization, technological resources are not enough to create adequate remote working conditions [1]. The use of technologies and the implementation of mediated work requires training and appropriation time for all members of the company as well as technical and organizational support [40].
Observations before the pandemic underlined some negative effects of telework on professional relations and group work [5,41,42]. In times of crisis and lockdown, mediated and remote work had rather beneficial effects. Indeed, the maintenance of professional ties and the performance of all or part of the work activity is based in part on the essential role played by technology [24]. Efforts were made to recreate times for group sharing so as to maintain the links for which managers were responsible [24]. Several studies indicate preservation of the link between the team, the circulation of information, and a unified work team [40], as well as a perception of support between colleagues and from their superiors [15]. For some employees, the willingness of management to adapt management styles to the context has resulted in an increase in operational leeway and the fostering of close relationships [1], thus experienced as a sign of support [43]. For others, managers have increased their level of control and sometimes employed awkward, intrusive practices (e.g., excessive use of virtual meetings), which limit employees' autonomy and recovery, thus promoting burnout [30,44].
In addition, there was an increase in the frequency of exchanges via asynchronous or instant messaging and videoconference meetings [23,24], particularly during the first lockdown. Although these tools helped to mitigate the physical and psychological distance of the teams, they did not fully compensate for face-to-face relationships and the cameras and microphones that were often left off during meetings helped to limit the spontaneity of exchanges [45]. Moreover, as pre-pandemic studies had highlighted [9,46], the lack of face-to-face interaction and the decrease in informal communication with colleagues were underlined and reinforced by social restrictions [10,11,47]. A decrease in informal exchanges has also been reported since the beginning of the pandemic [10,11,47].
Moreover, the absence of face-to-face contact was often not appreciated by employees, as it helps to give meaning to the job [21,47], regulate conflicts, and resolve relational and professional problems [48].

Changes in Employee Wellbeing since the Start of the Pandemic
Positive effects classically mentioned in teleworking outside times of crisis were also underlined during this period. These were the reduction in fatigue linked to the absence of commuting, a reduction in interruptions in work [10], and better concentration and flexibility [47], allowing employees to be more efficient in the tasks to be carried out [10,47].
The maintenance of telework within an organization served as a resource for coping with the lockdown period, especially by preserving psychological wellbeing [43]. Several studies indicate a positive perception of employees regarding working life in lockdown [15,29]. In this specific context, the increase in perceived workload was not always negatively experienced. Indeed, for some employees, it was interpreted as a guarantee of professional efficiency or confirmed the feeling of job security and the economic viability of their company during the crisis [11], and helped to reassure employees.
Moreover, even when employees reported less support at a distance than in person [49], they reported a level of wellbeing that was as high as outside the crisis period. The maintenance of the level of wellbeing reported by employees can also be explained by favorable lockdown and working conditions [50]. For example, the number of square meters per person reduced the risk of deleterious effects on psychological health [51].
The coordination of different spheres of life was dependent on the family situation of employees and the setup of the home (presence vs. absence of a separate place to work) [47]. The reorganization of spaces, temporalities, and places between the different confined members was mentioned as a difficulty of the situation [23]. The difficulty of reconciling work and private life and health and professional uncertainty made adaptation to teleworking difficult and generated stress [11].
Other studies indicate impacts on people's wellbeing, such as a decrease in life satisfaction compared with surveys conducted before the crisis [15,52]. The presence of a feeling of physical and/or social isolation, and the development of psychosocial risk factors such as feeling a loss of control, uselessness, or loss of meaning in one's work were noted during the first period of lockdown [22]. A decrease in wellbeing related to the reduction in face-to-face communication, the fragmentation of the work collective, and the reduction in autonomy have also been highlighted [2].
Finally, anxiety related to the pandemic, lockdown, and social restrictions have increased the risk of burnout [2,53]. Furthermore, wellbeing indicators of the French people have deteriorated since the beginning of the crisis, indicating more anxious states and sleep disorders [54].

Research Objectives
Initially, mandatory telecommuting was part of new government safety regulations and mandatory behaviors (e.g., wearing a mask or social distancing). During an initial experience with mandatory telecommuting, employees had to deal with many organizational disruptions. They had to become familiar with a new home-based work environment where technology enabled remote personal contact with members of their organization. After this initial experience, many employees were able to return to work full-time or part-time at the company offices, even though COVID-19 remained a global threat. New episodes of mandatory telecommuting appear as a "stop and go" cycle.
From a global perspective, our study allows us to understand employees' perceptions of the evolution and impact of organizational changes on their professional activity, their professional relationships, and their mental health during the two phases of the study. H1. We expect that working conditions and environments in the context of public health restrictions alter the temporal and spatial structure of work activity.
H2. We hypothesize that: (a) If the business continuity plan is likely to favor an increase in the frequency of remote professional interactions (b) the mediatization of these relationships may affect their quality.
H3. We expect that, as pandemic and sanitary measures persist, employee mental wellbeing will decline.

Study Procedure
Exploratory research was carried out among French teleworkers using a longitudinal within-subject qualitative approach. The first phase of data collection took place during the first lockdown in France (17 March to 10 May 2020) and all participants were confined to their homes. The second collection phase was carried out in April 2021, when the French population was subject to a curfew with variable hours depending on the department (6 p.m. or 7 p.m. to 6 a.m.). Access to the company was severely restricted, subject to social distancing measures and guidelines. By the time of the second phase of the study, the employees were working mainly from home, except for some jobs where they had to travel to customers' premises. None of them benefited from partial unemployment these two phases and continued to work full-time at home (phase 1) or partially (phase 2) depending on their jobs and the government health measures imposed.
This study was carried out in an insurance company that agreed to distribute the call for participation to its employees via their internal network. As confidentiality clauses were signed, we cannot give any further information on this subject. Regarding the collection procedure, employee volunteers were invited to contact the researchers conducting the study directly. It was stated that the study could include several parts. Volunteers had the option of participating in only one part of the research. Thus, in the first instance, 31 employees agreed to be interviewed, and 22 of them volunteered to take part in a second interview. In this article, we present the results of the analyses based on the two data collections from the 22 employees. Although the sample size remains modest, the first period of lockdown and the successive waves of pandemic peaks were particularly demanding and it was not possible, despite our reminders, to recruit a larger number of participants.
All the interviews were conducted by phone and lasted on average 1 h and 30 min for the first phase and 40 min for the second phase. The semi-structured interviews were carried out at both times and structured identically. After introducing themselves, the employees were asked to describe their work activity before lockdown (phase 1 only), and then to describe comparatively during both phases of the study: (a) the changes in the work organization in carrying out their work activity, the objectives to be achieved, and the time structure of their day; (b) the changes in professional relations with their manager (n + 1) and their colleagues, both in terms of the frequency of exchanges and in terms of the quality of the information transmitted; (c) their perception of their physical and psychological health; (d) predictions for possible new work modalities in the future (see Appendix A). Recorded interviews were translated by five research assistants and supervised by one of the authors. The final corpus length was 109,904 words.

Participants
Our survey was carried out on a convenience sample of 22 employees of the same company. All participated in both phases of the study. Their average length of service was 24 years in the company (from 8 to 38 years). The interview population consisted of 11 women and 11 men. They were on average 50 years of age (age varied between 38 and 61). Overall, 4 lived alone, 17 were part of a couple, 3 of whom had no children, 9 had children Sustainability 2022, 14, 8731 6 of 20 aged 13 and over, 5 had children under 12, and 1 was alone with 2 children over 13. Among the sample, 7 employees had already practiced teleworking, formalized by a company agreement that authorized one day of teleworking per week. In addition, 12 employees, who were sales representatives, were involved in so-called nomadic teleworking. This is characterized by the fact that workers move around and carry out their work activity in different locations outside the company premises [5,55]. Finally, 3 had never formally teleworked from home.
We can also point out that during the first phase, the employees were all under lockdown and worked from home. Eighteen of them lived in a house, and four lived in a flat. The employees interviewed worked in the same company, but in different jobs: 12 were salespeople, 4 were underwriting technicians, 3 were internal trainers, 2 were recruiters and 1 was an accountant. It should also be noted that, even when the jobs were identical, the employees were not necessarily part of the same team.

Data Analysis Method
A lexical and morphosyntactic analysis was carried out on the interviews using the Alceste software ("Analyses of Co-occurring Lexemes in the Sample Enunciations of a Text") [56]. This method of statistical lexical analysis makes it possible to identify the "lexical worlds" of the corpus. The corpus is divided into text units and then into elementary contextual units (ECUs). Descending Hierarchical Classification is divided into several parts of results. The first part corresponds to the textual statistics of the corpus (e.g., number of cases, number of ECUs classified on the whole corpus). The second part consists of the dendrogram (see Figure 1), representing the partition into different classes and indicating the size of each class according to the proportion of the corpus analyzed (percentages). Then a table represents the characteristic forms or lemmas (words) of each class. Chi2 values, as well as a p-value for each word, are indicated to appreciate the strength of the link between the word and the class. Thus, the highlighting of "lexical worlds" by means of Descending Hierarchical Classification carried out by the software also makes it possible to establish correspondences between (illustrative) variables previously coded in the corpus and their significant preponderance in a class (p < 0.05). The illustrative variables retained for our study were: phase of the study (1 or 2), gender, occupation, and family situation (with children under 12 years of age).

Participants
Our survey was carried out on a convenience sample of 22 employees of the same company. All participated in both phases of the study. Their average length of service was 24 years in the company (from 8 to 38 years). The interview population consisted of 11 women and 11 men. They were on average 50 years of age (age varied between 38 and 61). Overall, 4 lived alone, 17 were part of a couple, 3 of whom had no children, 9 had children aged 13 and over, 5 had children under 12, and 1 was alone with 2 children over 13. Among the sample, 7 employees had already practiced teleworking, formalized by a company agreement that authorized one day of teleworking per week. In addition, 12 employees, who were sales representatives, were involved in so-called nomadic teleworking. This is characterized by the fact that workers move around and carry out their work activity in different locations outside the company premises [5,55]. Finally, 3 had never formally teleworked from home.
We can also point out that during the first phase, the employees were all under lockdown and worked from home. Eighteen of them lived in a house, and four lived in a flat. The employees interviewed worked in the same company, but in different jobs: 12 were salespeople, 4 were underwriting technicians, 3 were internal trainers, 2 were recruiters and 1 was an accountant. It should also be noted that, even when the jobs were identical, the employees were not necessarily part of the same team.

Data Analysis Method
A lexical and morphosyntactic analysis was carried out on the interviews using the Alceste software ("Analyses of Co-occurring Lexemes in the Sample Enunciations of a Text") [56]. This method of statistical lexical analysis makes it possible to identify the "lexical worlds" of the corpus. The corpus is divided into text units and then into elementary contextual units (ECUs). Descending Hierarchical Classification is divided into several parts of results. The first part corresponds to the textual statistics of the corpus (e.g., number of cases, number of ECUs classified on the whole corpus). The second part consists of the dendrogram (see Figure 1), representing the partition into different classes and indicating the size of each class according to the proportion of the corpus analyzed (percentages). Then a table represents the characteristic forms or lemmas (words) of each class. Chi2 values, as well as a p-value for each word, are indicated to appreciate the strength of the link between the word and the class. Thus, the highlighting of "lexical worlds" by means of Descending Hierarchical Classification carried out by the software also makes it possible to establish correspondences between (illustrative) variables previously coded in the corpus and their significant preponderance in a class (p < 0.05). The illustrative variables retained for our study were: phase of the study (1 or 2), gender, occupation, and family situation (with children under 12 years of age).  Finally, as the software offers a statistical distributional analysis, the analysis produced is independent of the meaning [57]. It is then up to the researcher to give meaning to each of the classes and interpret them in relation to each other [58].

Results
The Descending Hierarchical Classification (DHC) carried out by the Alceste software indicated the presence of four classes in the corpus (44 interviews) which were significantly representative of all the interviewees (93.48% of the corpus), i.e., 2868 ECUs classified out of 3068.
The attached dendrogram ( Figure 1) shows an opposition between Class 1 and Classes 2, 3, and 4.
The first class (Class 1) is centered on the spatial and temporal organization of the work activity since the beginning of the pandemic. Classes 2 and 3 are interrelated and describe the quality of professional relationships, the impacts related to the work activity carried out mainly at a distance and the communication tools that made it possible to mediate these relations. Finally, Class 4 deals with production targets and the perception of quality of work and health.
3.1. Class 1-Spatial, Temporal, and Material Organization of Work Activity and Work-Life Balance Class 1 comprises 668 ECUs (i.e., 23.99% of the corpus analyzed) and represents all the interviewees' discourse. In fact, no illustrative variable is significantly associated with this class. This class is related to the spatial and temporal structure of work and the articulation between professional and personal life (see Table 1).  At first lockdown, the hasty introduction of teleworking did not allow adequate material working conditions for employees to be deployed. In fact, not all of them had a room where they could work in isolation or adequate material conditions to work comfortably at home. Of the 22 employees interviewed, half (n = 11) indicated that they had a separate office or dedicated workspace. The remaining employees reported that they sat at their kitchen or living room table and had to move their belongings every day. In the second phase of the study, the majority of employees were able to obtain additional equipment to improve working conditions and some were able to set up a dedicated space. "What I rarely had was a lunch break. That means that I tended to keep working, I stopped to eat or make the children eat because they asked me to, but I didn't even share the meal with them". T1-07 At the second stage of the interviews, when the restrictions had been eased and the schools reopened, 12/22 perceived a lengthening of working hours and/or an increase in the pace of work, compared with 10/22 who considered that they had returned to similar pre-pandemic hours.
"But before, I didn't have this range of hours [...] I never really started work before 9 a.m. and now, from 8 a.m. until 8 p.m. [...] so yes, the range of hours is much greater". T2-11 "It's really an incredible intensification of meeting time which risks leading to more burn-out, more exhaustion". T2-21 "The workload, I would say, is back to where it was before the pandemic". T2-18 In addition, regardless of the period, several employees said that they limited their break time when teleworking.
"We don't take breaks anymore. At noon we eat quickly without a break, we go back to the office quickly and in the evening, we stay there until midnight". T1-13 The frequency of meetings has remained high since the start of the pandemic and is causing spillover in activity.
"So, it's any time, at any time [...] Today, they land you with a Teams meeting at 6.30 p.m. because they think you're at home... that's not possible. That's a destructive element". T2-15 "Before, there was a lunch break, now with all the meetings, they'll put training meetings and communication interviews between noon and 2 p.m.". T2-03 Finally, the sales representatives highlighted an activity that was restricted and prevented at the second stage of the study because of the curfew.
"We finish at 6 p.m., we were asked not to make appointments too late so as not to incur a penalty, a fine". T2-04 "We not only had to reduce the time slot we eliminated one of the best times to do business". T2-15 Sustainability 2022, 14, 8731 9 of 20 * T1 and T2 correspond to time 1 and 2 of the study. The numbers from 01 to 22 correspond to the interviewees.

Class 2-Mediated Professional Relationships and Work Activity at a Distance
Class 2 contains 566 ECUs (i.e., 19.74% of the corpus analyzed). The discourse is mostly represented by employees working as internal trainers and underwriting technicians (respectively χ 2 = 30.95; p < 0.01 and χ 2 = 4.68; p < 0.03) and men (χ 2 = 5.07; p < 0.02), but it does not exclude the discourse of other employees interviewed. Neither phase of the study is over-represented in this class.
The discourse significantly associated with this class refers to the frequency of exchanges since the beginning of the pandemic as well as the evolution of the quality of professional relations, the informal dimension of exchanges as well as the effects of the absence of face-to-face interactions in the work activity (see in Table 2).  During the first lockdown and the introduction of forced teleworking, the company management asked the managers to maintain the link with their teams. To do this, managers increased the frequency of communication. This was positively perceived by employees and interpreted by most as support, although some had doubts about the motivation behind this.
"The fact that we can be in contact with our manager or the team on a daily basis [...] that was a very good initiative, I didn't feel abandoned, it was very good". T1-19 "There's an order given to the managers I think, to call their team every day [...] maybe they also wanted to see if we were actually working, there's also that, well it was also to maintain the link". T1-21 Employees said that the quality of their relationships with their manager and colleagues had not been altered and appeared identical to the pre-pandemic period. "With colleagues we were able to maintain relations, we are a very close team". T1-19 "The social link has been maintained and rather... how shall I put it... It's rather unchanged, we are a team that is neither united nor disunited". T1-06 Informal relationships had not disappeared and were maintained in both the periods investigated through attempts to recreate moments of conviviality organized around virtual coffee, games, etc. Some of these informal exchanges took place before more formal meetings.
"On Fridays we played games, we had, we did quizzes, we did blind tests, we really kept this link and it went really well". T1-19 "Before Christmas, she had a Christmas jumper meeting, so we all dressed up in Christmas jumpers. We all put on our cameras and for almost an hour we were joking around, it was fun, we had a laugh". T2-01 However, the distance tended to limit informal exchanges (in duration and frequency) and was relatively costly to set up over time.
"But it's clear that being at a distance has been a handicap because it limits the informal contacts we can have". T2-21 "I set up a challenge in the team [...] it required time [...] it didn't last long; I think it's something that requires a lot of energy to motivate everyone". T2-10 Finally, a lack of face-to-face interaction was acknowledged for almost all (n = 21) employees at the first stage of the study and persisted for a majority (14/22) at the second stage. However, only those in jobs with strong interpersonal skills, such as trainers, recruiters, and sales staff, emphasized the direct impact on their activity and changes in work practices.
"I think that in face-to-face meetings we understand each other better and we can go further. On the other hand, in remote mode we work faster [...] But face-to-face you gain more relevance, more finesse in establishing and evaluating needs, etc. And there is the non-verbal part which can also give us some indications". T1-12 "As there are fewer humans, fewer exchanges, fewer gestures [...] It doesn't last as long as it used to. And yes, it takes less time to complete a task [...] Having to sell a job over the phone to a salesman who has 30 job offers, yes, it's more complicated than having him face to face". T1-11 It should be noted that, during the second data collection, 7/22 employees stated that they had changed manager. One department (training) was completely restructured involving the merger of two jobs and involved 3/22 employees of our study population.

Class 3-Uses, Benefits, and Ambiguities of New Technological Tools
Class 2 included 568 ECUs (i.e., 19.8% of the corpus analyzed). It characterized employees in a commercial profession (χ 2 = 6.43; p < 0.01) and women (χ 2 = 4.26; p < 0.04) but did not exclude the discourse of those in another profession or of men. The discourse of this class concerned the place of new technological tools and their uses (see Table 3). In this class, employees talked about the changing use of collaborative communication applications such as Skype and Teams. The company switched from Skype to Teams during the study, but both have the same general functionality (i.e., audio and video calls, instant messaging, screen and document sharing). As a result of the pandemic, this type of tool has become the primary means of communication.
"We are more often on the Teams tool than on the phone now". T2-13 "So, we learned from this situation, from this period. We have become familiar with the tools, Teams has become second nature for us". T2-12 Employees mentioned some benefits of these tools such as maintaining relationships and overcoming sticking points, but that they did not fully replicate face-to-face exchanges.
"With colleagues, it's less human, because it's through the tools, that's one of the criticisms we can make of this mode". T1-14 "This type of remote relationship, we'll say... even if it doesn't replace physical contact, which remains limited in any case, of course, it allows us to solve many problems". T2-08 The use of tools (Skype and Teams) in a significant proportion was supposed to be temporary and oriented towards meetings aimed at maintaining social links during lockdown. In reality, employees found that their massive use tended to become entrenched in their professional practice and have an impact on their activity. In fact, at the second stage of the study, there were still many meetings (see Class 1) focused on professional information, leading to an excess of work activity.
"On the other hand, we do a lot more team meetings. In fact, we have one team meeting after another, it's unbearable [...] it doesn't stop". T2-02 "At 6 pm-8 pm, we have time slots for meetings that we haven't had time to do during the day. It can push personal work very late into the night". T2-03

Class 4-Production Targets, Perception of Work Quality, and Wellbeing
Class 4 includes 1046 ECUs (i.e., 36.47% of the corpus analyzed). No illustrative variable is significantly associated with this class. It therefore represents the discourse of all the interviewees.
The comments reported in this class concerned the production targets set for employees, the quality of the work performed and the changes in perceived health since the start of the pandemic (see Table 4). In the first phase of the study, the employees reported that they were calm about adapting their targets and taking the lockdown into account.
"The targets will be reviewed, that's for sure". T1-08 "In lockdown, it has become complicated to achieve the targets requested, but the targets have been reviewed". T1-19 On the other hand, in the second phase of the study, they underlined the lack of consideration given to the pandemic in the new defined targets. This lack of consideration was not well received by employees and suggests that the company did not recognize the efforts made to maintain production since the beginning of the pandemic.
"The fact is that the targets set for 2021 curiously fail to take into consideration the fact that we are still in a very particular situation". T2-05 "So yes, the results for 2020 were exceptional, but because the sales staff went to see the customers despite everything [...] We do our job, and we have the impression that it is not recognized". T2-19 The employees drew attention to the current working arrangements, which worked, but in which they saw a possible deterioration in the quality of work in the long term.
"There is certainly a loss of quality in knowledge of the customer and the proximity necessary to the customer in the long term". T1-12 "We used to be able to do a recruitment interview in an hour and a half, now it takes 40, 45 min, why? Because there is more of a human side, more of a relationship side. I don't say that it doesn't work, but I find that there is less quality". T2-11 In terms of wellbeing, whereas the employees evaluated the first lockdown positively, a deterioration was observed in the second period. Indeed, the wear and tear linked to the pandemic, combined with the reorganizations necessary to maintain professional activity, led to a deterioration in wellbeing. The multiple adaptations were costly (in terms of resources) and the lasting compromises led to fears of a loss of meaning.
"At the beginning it was fine, I was getting there. But now, the more it goes on, the less I can do. It's becoming very difficult. I find it more difficult to concentrate in fact. There are times when I have to go back to a task several times because I find it hard to concentrate. I have a certain mental lassitude". T2-01 "We have been able to adapt, but at what cost? [...] It's been enormous, a huge job, some people are very tired, I'm tired". T2-15 Fears relating to decisions that the company might take to maintain a mainly remote working model were present and mentioned by several employees. "Insofar as we had results during the lockdown, [...] I am afraid that I will be told, listen, since it went so well, you are going to be home-based for example. There's all this human side that will be taken away". T1-11 "What I fear is that interpersonal relations will change a lot, which will have an impact on the commercial part of my job, not the figures, the part about how to do my job". T1-08 The final table summarizes the results at the two measurement times (see Table 5).

Discussion
This study attempts to understand to what extent the implementation of teleworking since the beginning of the COVID-19 pandemic has affected the spatio-temporal and organizational dimension of employees' work practices as well as their professional relationships and their wellbeing. The lexical and morphosyntactic analysis carried out revealed four classes and allows us to understand the impacts of the multiple reorganizations on the work activity and the experience of employees during a health crisis.
In line with hypothesis 1, our findings (see Class 1) corroborate the literature focusing on the first period of the crisis [11,27], but also provide details on the spatial arrangements that the employees could make (or not) to their home workspace. Indeed, the space at home as well as the possibility of isolating oneself enabled teleworkers to better adapt to the period in question. Our results also show that the employees, thanks to their own resources and organizational support, were able to improve the material conditions of teleworking between the first and second phases of the study. On the other hand, the delimitation of the workspace could only be achieved sustainably when housing conditions allowed and increased the physical/spatial encroachment of the work sphere on personal space. Thus, the ability to recreate a workspace that meets the correct working conditions is based on individual possibilities. These results corroborate prior studies [1,59,60] that highlight possible social inequalities that emerged between employees with a dedicated teleworking space and those without. Moreover, during the first lockdown, the superimposition of work and family spheres in a single location had effects on employees' working time structure. For employees with children to take to school or look after, teleworking during the first lockdown was punctuated by voluntary interruptions of tasks leading to fragmentation in work but aimed at facilitating the articulation of family and work spheres. In this sense, our results support the research in the field [20,21]. However, work that could not be completed during the day was made up early in the morning or later in the evening or at night. Half of our participants said that they had experienced an increase in workload since the beginning of this period.
As outlined in preliminary surveys [24,40], the continuation of the activity during this period, work performance, and the maintenance of socio-professional links were dependent on the availability of technology. In our study, no hypothesis had been made about ICTs and their uses, but the results shed light on changes in work practices. Although these new communication tools (Skype and Teams) had hardly been used, the crisis seems to have fast-tracked their use. Indeed, the discourse of the employees interviewed shows their massive introduction to working practices. As stated in previous investigations [40], these new collaborative communication applications have become fully-fledged communication channels and brought undeniable advantages during this period. Among them, we find the preservation of links, the dissemination of information, and/or the federating of work collectives. Nevertheless, the addition of a new communication channel to the more traditional ones (e.g., email, phone) and the increase in the number of meetings that need to maintain focal attention could generate fatigue as shown by Dolce et al. [44], whereas the ease of setting up meetings and involving other departments in the company tended to increase exchanges compared with pre-pandemic level. These effects were not well received and were highlighted in the second phase of the study. The gradual resumption of activity since the end of the first lockdown, coupled with the numerous meetings (at unusual times), added to the already existing workload, and sometimes led to a spillover of work activity. This phenomenon had also been reported in the existing literature [61].
Moreover, in line with what has been observed before the pandemic [18,19,36,62], our results indicate a feeling of intensification of working time, which persisted in the second stage of the study, resulting from the sequencing of tasks and the reduction in break times. A lengthening of working hours where the time initially planned for the journey (home to work) was now part of the working time was also observed. On the other hand, unlike what was indicated in the pre-pandemic literature [18], this time was not only used to deal with emergencies and unforeseen events but also represented (new) time for informational meetings, which tended to become entrenched in practice.
As highlighted in the literature produced during the first lockdown [15,23,24,63] the frequency of mediated meetings increased sharply during the first lockdown, enabling the management of activity and the maintenance of professional relations. Appreciated overall by the majority of employees and experienced as a mark of support, some employees expressed doubts about the real purpose, indicating that it could also be a means of control. In the second phase of the study, we expected (H2) a decrease in the frequency of meetings. Our results, however, indicate a continuing high level of meetings. This can be explained by the appearance of new coronavirus variants and the successive restrictions that pushed companies to practice an organizational modality in which teleworking remained the main mode of working. Companies have therefore favored mediated exchanges which are considered easier to set up because of the restrictions imposed on meeting face-to-face.
In addition, as observed in earlier scientific work [24,63], the use of technology enabled professional links to be maintained. However, in contrast to the pre-lockdown literature on work collectives [42], our study brings a nuance to the evolution of the quality of professional exchanges. Indeed, the socio-affective dimension of the professional relationship doesn't seem to have been impacted by our population of employees. We can suggest that the solid relationship built up before the pandemic survived due to the significant seniority of the participants (24 years of work experience on average). However, the collaboration within the context of the professional activity was altered. Our results indicate that the frequency of mediated exchanges did not compensate for the lack of face-to-face relationships [45]. At a distance, the employees indicate that they go directly and more quickly to the desired information, which generates short-term efficiency, but they lose the finesse of the description of the subject of exchange. Moreover, the non-verbal dimension is altered, limiting the reading of exchanges between interlocutors. The persistence of this loss of face-to-face contact during both phases of the study highlights two perspectives. The first relates to the lack of informal social relationships between members of the same team and to a broader collective (to other departments) and questions the impact of these working arrangements on a commitment to work, the values of the company, and the sharing of practices. This result is in line with what was demonstrated pre-pandemic [36]. The second perspective concerns certain occupations in which the relational and physical dimension is crucial for the performance of work activity. Our results finally indicate that the strategies usually used by employees become inoperative when the activity is carried out in a mainly remote setting. Thus, the compromises newly put in place since the pandemic alter not only the quality of work criteria but also the deep meaning that they attribute to their job, thus generating numerous dissatisfactions. The excessive mediatization of work activity alters one's relationship to their work and profession, in this case, also generating numerous dissatisfactions.
As the literature has pointed out [43], informal mediated exchanges decreased during the first lockdown, but less strongly than we thought (H2). In strict lockdown, exchanges were maintained through video-cafés, regular events contributing to the psychological support of employees. Our longitudinal approach enables us to observe the difficulties in maintaining informal exchanges despite the successive resumption of work on site. This finding reveals that maintaining mediated professional relations requires a communicative effort on the part of employees. The participants have to take time together and the coordination of this time appears less spontaneous than when employees meet physically in one place. The decrease in this type of exchange is harmful insofar as the permanent physical distance between employees calls into question not only the vitality of work collectives [2] but also the socializing function of work deployed in informal spaces [30]. On the other hand, these results underline the need to support informal exchanges and to complement them by institutionalizing discussion spaces with the aim of both exchanging information on the resources and constraints of organizations and preserving the health of employees.
Regarding the mental wellbeing of employees, our results show that despite the anxiety-provoking situation and the unprecedented context of the health crisis, a positive feeling about work life during the first lockdown appears in the discourses of the employees in our study, similar to previous studies [15,29]. This can be explained by the favorable lockdown conditions [50], the maintenance of full-time work and salary, and optimism regarding the rapid resolution of the health crisis and a return to "normal" activity. In contrast, the second stage of the study highlights, as we predicted (H3) [54], more pessimistic observations including a deterioration in mental wellbeing. First, the burden of the pandemic, the restrictions, and the multiple reorganizations required regular adaptations and have developed a feeling of wear and tear and weariness among the employees.
The support provided by the group during the first period of lockdown already found in previous studies [40,43] seems to be exhausted during the second stage. It would be more difficult to mobilize a professional entourage over time when most of the work is done at a distance. However, we know that social support is recognized as a resource that prevents exhaustion [64], particularly in the relational professions. Moreover, as the scientific literature in the field points out, both before and during the pandemic [13,[35][36][37]62], the massive use of technology and the sequencing of tasks leads to a feeling of intensification that could bring employees to premature burnout.
Finally, there are concerns about maintaining most of the remote work after the crisis that does not correspond to the usual work practices and strategies developed by employees to achieve their goals. Employees fear premature wear and tear due to the many resources mobilized to change work practices, but also a decrease in the quality of their work. This feeling is reinforced by new higher work targets. They see in this a lack of recognition of the efforts made during the first lockdown and the inability of the management to consider the still ongoing pandemic. These changes are extremely costly for employees and could lead to burnout or the development of psychological defenses such as disengagement from work.
Although this study deepens previous work in the field, several limitations can be highlighted. The convenience sample and the very small number of participants from a single organization limit the generalization of our observations. Another important limitation is the age of the employees. A younger population could have provided different results on the evolution of their relationship to work in this period. The same applies to the high seniority of employees in the company. Low seniority or employees newly recruited into the company could have brought different ways of mobilizing the professional network of contacts and different strategies for adjusting to the organizational changes inherent in the crisis. Furthermore, interviews conducted at each major change (e.g., multiple changes in health measures) would have allowed us to gain a more detailed understanding not only of the changes in practices (individual and collective) but also of the impact of the reorganizations on professional activity and wellbeing.

Conclusions
The results of our analyses help to shed light on the changes that have been made since the beginning of the pandemic to teleworking in times of crisis. Whereas some of the results are consistent with or confirm those highlighted at the beginning of the pandemic, particularly during the initial lockdown, the qualitative study carried out enabled us to gain a more detailed understanding of the psychosocial processes at work. Indeed, we were able to distinguish the effects of continuing to work during a pandemic on the socio-affective vs. operative dimensions of relations with the professional environment.
Our results also reveal the effects of the reorganization of work in times of crisis in the medium term and the way in which some of the processes or dimensions examined have evolved or become established over time. In particular, whereas the psychosocial resources mobilized during the first lockdown had the potential to contribute to the preservation of health in times of crisis, the prolongation and repetition of work reorganizations over time led to a deterioration in the quality of relations and in the health of employees.
Additionally, the profound change in the relationship to work that seems to be taking place invites further investigation. Indeed, in France, the health crisis has not only boosted teleworking, which was previously underdeveloped and sometimes poorly supported but has also led to the reorganization of workspaces in companies. However, these working arrangements were relatively unusual in Europe and in France before the pandemic. Therefore, it is necessary to understand the effects of these recent and sometimes brutal developments on relationships to work organization, work, and the work team, as well as on health, which had already been called into question by the COVID-19 crisis period. A more global approach to work integrating both teleworking and working in the office would lead to a better understanding of the issues and impacts of these new work modalities, at organizational and managerial levels, and of the functioning of individuals and work collectives. Finally, on a practical level, the results of empirical research carried out before and during the pandemic period, as well as the feedback that has been structured within work organizations, are precious sources for feeding reflections and thinking about the deployment and support systems for tomorrow's work organization arrangements. The results enable organizations responsible for the health and safety of their employees to identify risks to be prevented. Thus, they are able to design and experiment with work organization arrangements that are favorable to the quality of work and the wellbeing of employees. They demonstrate that telework refers to major issues both from the point of view of work organization, management methods, the life and functioning of work groups, professional relations, forms of socialization, as well as workload, working hours, and workplace health.