rss
Emerg Med J doi:10.1136/emj.2009.082594
  • Original Article

Emergency physicians accumulate more stress factors than other physicians–results from the French SESMAT study

Press Release
  1. C Prudhomme8
  1. 1Department of Occupational Medicine, SCMT, Hôtel-Dieu, Assistance Publique-Hôpitaux de Paris, Paris, France
  2. 2Department of Anaesthesia,CHU, Caen, France
  3. 3Department of Pharmacy, Hôpital Charles Richet Assistance Publique – Hôpitaux de Paris, Paris, France
  4. 4Emergency Unit, Hôpital Saint-Antoine Assistance Publique – Hôpitaux de Paris, Paris, France
  5. 5Emergency Unit, CH, Corbeil-Juvisy, France
  6. 6Department of Occupational Medicine, Hôpital de Hagueneau, Haguenau, France
  7. 7Palliative Care Unit, SCMT, Hôtel-Dieu, Assistance Publique-Hôpitaux de Paris, 1 Paris, France
  8. 8Emergency Unit, Hôpital Avicenne, Assistance Publique–Hôpitaux de Paris, Paris, France
  1. Correspondence to Madeleine Estryn-Behar, SCMT, Hôtel-Dieu, Assistance Publique – Hôpitaux de Paris, Parvis Notre-Dame, 75004 Paris, France; madeleine.estryn-behar{at}sap.aphp.fr
  1. Contributors MEB conceived the study, designed the study and obtained research funding. MAD and IA supervised the conduct of the study and data collection. KG and CF managed the data, including quality control, and all authors contributed substantially to the revision of the paper. MEB takes responsibility for the paper as a whole.

  • Accepted 29 March 2010
  • Published Online First 1 December 2010

Abstract

Introduction France is facing a shortage of available physicians due to a greying population and the lack of a proportional increase in the formation of doctors. Emergency physicians are the medical system's first line of defence.

Methods The authors prepared a comprehensive questionnaire using established scales measuring various aspects of working conditions, satisfaction and health of salaried physicians and pharmacists. It was made available online, and the two major associations of emergency physicians promoted its use. 3196 physicians filled out the questionnaire. Among them were 538 emergency physicians. To avoid bias, 1924 physicians were randomly selected from the total database to match the demographic characteristics of France's physician population: 42.5% women, 57.5% men, 8.2% <35 years old, 33.8% 35–44 years old, 34.5% 45–54 years old and 23.6% ≥55 years old. The distribution of physicians in the 23 administrative regions and by speciality was also precisely taken into account. This representative sample was used to compare subgroups of physicians by speciality.

Results The outcomes indicate that the intent to leave the profession (ITL) was quite prevalent across French physicians and even more so among emergency physicians (17.4% and 21.4% respectively), and burnout was highly prevalent (42.4% and 51.5%, respectively). Among the representative sample and among emergency physicians, work–family conflict (OR=4.47 and OR=6.14, respectively) and quality of teamwork (OR=2.21 and OR=5.44, respectively) were associated with burnout in a multivariate analysis, and these risk factors were more prevalent among emergency physicians than other types. A serious lack of quality of teamwork appears to be associated with a higher risk of ITL (OR=3.92 among the physicians in the representative sample and OR=4.35 among emergency physicians), and burnout doubled the risk of ITL in multivariate analysis.

Conclusions In order to prevent the premature departure of French doctors, it is important to improve work–family balance, working processes through collaboration, multidisciplinary teamwork and to develop team training approaches and ward design to facilitate teamwork.

Throughout Europe, healthcare institutions encounter difficulties recruiting and retaining highly qualified physicians with adequate specialisation.1 Dissatisfaction has been considered an important factor underlying intention to quit and early retirement.2–8 It is widely acknowledged that since the mid-1980s, restructuring and cost reduction initiatives have resulted in decreased patient stays. At the same time, patient acuity has risen, and the medical profession has become more technical and complex. Consequently, working conditions have generated a great deal of dissatisfaction.9–11 Emergency physicians have to compensate for the lack of available general practitioners in certain areas of the country and after normal working hours.12

In France, studies describing levels of career satisfaction among physicians are scarce.13 Most surveys in other countries indicate that career satisfaction and dissatisfaction vary across specialities as well as by age, income, region and site of practice.14–20

A large number of studies have shown how the quality of teamwork is associated with increased physicians'21 and nurses' job satisfaction,22 quality of care23 24 and reduced burnout25–27 or mental health problems.28 Moreover, improved team communication and decision sharing are factors associated with decreased overtime, decreased emotional distress and decreased turnover. Constructive relationships between nurses and physicians, more control over practice and autonomy are associated with attracting and retaining qualified physicians.29 30

Despite this body of knowledge, cost reduction initiatives in hospitals and reorganisations resulting in a lack of nurses have led to managers using floating assignments and reducing the time available for in-depth discussions between physicians and nurses. As a consequence, hostile feelings develop between occupational groups, and information circulates poorly with possible consequences for patients' safety, healthcare quality and career development opportunities for nurses.

Our study follows the Nurses' Early Exit Study (NEXT) research projecti conducted in ten European countries in which we studied the quality of teamwork in healthcare and investigated its impact upon nurses' intent to leave.32 33 The NEXT study addressed factors such as work environment, influence at work, exposure to violent events with patients and personal factors such as age, sex and burnout in order to better understand how these are related to the shortage of nurses. After the national conference communicating the study results, physicians associations asked for the same study to be conducted among physicians.

Method

Study design and sample

Our study investigated physicians working in France on a salaried basis. A comprehensive questionnaire was prepared using established scales measuring: various aspects of working conditions, satisfaction and health of salaried physicians and pharmacists. It was available online (www.presst-next.fr/SESMAT/), and the two major associations of physicians promoted its use by their journal (Intersyndicat national des praticiens hospitaliers and Coordination des praticiens hospitaliers). Four thousand, seven hundred and ninety-nine visited the third page of the questionnaire where questions about professional life started, between March 28th of 2007 and April 30th of 2008. Three thousand, one hundred and ninety-six physicians filled in the questionnaire making a 66% response rate. Among them, 538 were emergency physicians. To avoid bias, 1924 physicians were randomly selected from the total database to match by quotas the demographic characteristics of France's physician population: 42.5% women, 57.5% men, 8.2% <35 years old, 33.8% 35–44 years old, 34.5% 45–54 years old and 23.6% ≥55 years old. The distribution of physicians in the 23 administrative regions and by speciality was precisely taken into account. This representative sample was used to compare subgroups of physicians by speciality.

Across participating institutions, our total sample comprised different subgroups of specialities: emergency units (N=538), intensive care and anaesthesiology departments (N=565), medicine (N=576), surgery (N=331), psychiatry (N=179), geriatric (N=106), biology and radiology (N=217), preventive medicine (N=285) and pharmacy (N=155).

The questionnaire

The questionnaire for the NEXT research project was developed by a group of experts (medical doctors, nurses, psychologists and statisticians) drawing on previous research, interviews with nursing staff in three countries and three pilot studies. Validated scales were used (see table 1). The translation–back-translation method was used by all country members using the basic English version.34 The survey contained about 260 questions31 comprising occupational biography, private life, social work environment, work organization, work demands, individual resources and future occupational plans. The adaptation to physicians was conducted with specialists representing the physicians' associations. Over two thirds of the questions were kept with no changes to the original study. The questions that did not apply to physicians were replaced by physician-specific questions. The study was carried out between April 2007 and May 2008.

Table 1

Outline of the predictor variables

Measures

Dependent variables

Burnout was measured using the six items of the Copenhagen Burnout Inventory (CBI).36 Example item: ‘Do you feel tired?’ The answer categories ranged from: ‘never/almost never’ to ‘(almost) every day’. This variable was dichotomised with a split between 2.99 and 3.00.

ITL was measured by one item: ‘How often during the course of the past year have you thought about giving up the medical profession?’ The scale anchors were: ‘never’, ‘sometimes in a year’, ‘sometimes in a month’, ‘sometimes in a week’ and ‘every day’. We interpreted ‘sometimes in a month’ or more often as an indicator of frequently considering leaving the profession (17.4% of the total sample).

Predictor variables

For clarity, we described the scales in table 1. They are validated scales in this field and were revalidated during the previous NEXT study among nurses. We considered scales in two or three levels.

Confounding factors

In order to control for confounders, we included personal factors such as age (there are four age categories), sex and family situation. We also considered organisational factors such as having hierarchical responsibilities.

Statistical analysis

First, we conducted bivariate analyses, using Pearson's χ2 test, to determine the association of predictors with burnout and ITL separately for each of the specialities. Second, we conducted multivariate analyses using SPSS V.12.0. Overall, all the different questions contained less than 10% missing data, with the exception of 34 questions out of the 486 questions and sub-questions. For some variables, missing values comprised several items, explaining the difference in sample sizes.

Results

Personal characteristics in relation to burnout and intent to leave the profession

Among participating physicians, ITL was higher among emergency physicians: 21.4% compared to 17.4% among the representative sample (p<0.05; see table 2). There were fewer female physicians among emergency physicians (37.6% vs 42.5%; p<0.05). More young physicians worked in emergency units (only 23.5% were 45–54 years old and 4.5% were ≥55 years old versus, respectively, 34.5% and 23.6%; p<0.001). No significant differences were found for family situation, place of work, place of birth or satisfaction with pay. Emergency physicians got the same number of proposals for another job as physicians from the representative sample.

Table 2

Personal characteristics in relation to burnout and intent to leave the profession (ITL) among emergency physicians and among a representative sample of French physicians

Female physicians had a high burnout score (CBI) more frequently among the representative sample and emergency physicians (49.1% of female physicians versus 37.5% of male physicians of the representative sample and 65.5% of female physicians versus 43.2% of male emergency physicians). Physicians from the representative sample had the highest burnout rate in the two age groups between 35 and 44 and between 45 and 54. The trend was a high burnout score among young emergency physicians, but this was non-significant, due to the small number of physicians in the older group. Physicians dissatisfied with their pay among the representative sample declared more burnout compared to those who were satisfied. This relationship was even stronger for emergency physicians. Physicians born in another country had a high burnout score more often than among the representative sample. Having received proposals for other jobs was also high among factors linked with a high burnout score for the representative sample and for the emergency physicians. Family situation and place of work were not linked with burnout among the representative sample or among the emergency physicians.

Physicians aged ≥55 in the representative sample declared more ITL compared to those <35 years old. The trend was similar for emergency physicians but non-significant. Physicians in the representative sample either living alone or with another adult without children declared more ITL compared to those living with children. The trend was similar for emergency physicians but not significant. Physicians in the representative sample who were dissatisfied with their pay declared more ITL compared to those who were satisfied. This relationship was even stronger for emergency physicians. Having received proposals for other jobs was highly linked with ITL for the representative sample and the emergency physicians. Sex, place of work and place of birth are not linked with ITL either for the representative sample or the emergency physicians.

Health characteristics in relation to burnout and intent to leave the profession

Among participating physicians, high general burnout scores (CBI) were more frequent among emergency physicians: 51.5% compared to 42.4% of the representative sample (p<0.001; see table 3). High specific burnout scores (patient-related burnout) were more frequent among emergency physicians: 33.0% compared to 23.1% of the representative sample (p<0.001). Musculoskeletal disorders were more frequently declared by the representative sample which was also older (15.3% are treated compared to 10.9% of emergency physicians; p<0.001). No significant differences were found for perceived health, mental health disorders and sport practice. Other health habits that differed significantly between the representative sample and the emergency physicians, with more inadequate habits among the latter, were: fewer social activities, more tobacco smoking, eating fewer fruits or vegetables per day and taking fewer meal breaks during the work day. However, these young emergency physicians would like to take care of their health, as shown by the fact that more of them wished to have an occupational health consultation and were correctly vaccinated against hepatitis B.

Table 3

Health characteristics in relation to burnout and intent to leave the profession (ITL) among Emergency physicians and among a representative sample of French physicians

Among health habits, only sport was studied in relation to burnout. Absence of practice was highly linked to burnout for the representative sample and for emergency physicians.

Burnout (general and patient-related) was highly linked to ITL for the representative sample and for emergency physicians. Musculoskeletal disorders were linked to ITL in the representative sample, and this trend was similar for emergency physicians.

Professional status and work schedules in relation to burnout and intent to leave the profession

Emergency physicians worked less frequently in university hospitals, were less often in hierarchical positions and had the opportunity to teach less frequently than did the physicians in the representative sample (table 4). However, they attended continuous education more frequently; about 10 days within the last 12 months. Fewer of them devoted more than 20% of their working days to tasks outside direct care. A higher percentage of them worked 50 h per week or more (59.8% vs 49.9% of the representative sample), as well as more nights per month (48.3% worked eight nights or more versus 14.6% of the representative sample). They also had high work/family conflicts scores more often (50.1% vs 43.3% of the representative sample).

Table 4

Professional status and work schedules in relation to burnout and intent to leave the profession (ITL) among Emergency physicians and among a representative sample of French physicians

Emergency physicians who did not attend any continuous education had a high burnout score more frequently as did those who had to devote more than 20% of their working days to tasks outside of direct care. A dose–response increase exists between work/family conflict and burnout (17.6%, 39.9% and 68.9% with high burnout scores for, respectively, low, medium and high work/family conflict scores). For the representative sample of salaried physicians, continuous education, work/family conflicts, workweek duration and the number of nights worked per month were significantly linked with burnout. For emergency physicians, continuous education and work/family conflicts were linked with the intent to leave the profession. This relationship between tasks outside of direct care, workweek duration and number of nights was also found in the representative sample.

We analysed factors linked with work/family conflict and showed that workweek duration and changing work schedules with short notice were highly linked with work/family conflict. Indeed, physicians of the representative sample who worked 50 h per week or more were 57.6% with a high work/family conflict compared to 24% of those working less than 45 h per week (N=1580; p<0.001 result not shown). These figures are, respectively, 59.4% and 33.3% among emergency physicians (N=475; p<0.001). Also, physicians of the representative sample who had to change their work schedules with short notice 3 to 5 times per month or more than 5 times per month, respectively, 51.5% and 69.4% experienced high work/family conflict compared to 32.4% of those not facing these changes (N=1868; p<0.001). The figures are, respectively, 66.7%, 83.3% and 39% of emergency physicians (N=533; p<0.001 result not shown).

Working environment in relation to burnout and intent to leave the profession

Emergency physicians declare more frequently low influence at work, high quantitative demand and tense relations with administration (table 5). Their quality of teamwork score was more frequently low. They declare being subjected to violence from patients or their relatives twice as often as the doctors in the representative sample—monthly or more (69.3% vs 27.5%). Higher percentages of them were often or always worried about making mistakes. Emergency physicians were found to be more frequently dissatisfied with staff handovers when shifts change (62.5% vs 39.3%). Very few of them found their working space adequate (24.8% vs 45.9% of the representative sample).

Table 5

Work environment in relation to burnout and intent to leave the profession (ITL) among Emergency physicians and among a representative sample of French salaried physicians

All of the working environment risk factors were highly linked to burnout and intent to leave the profession for the representative sample and for emergency physicians; only working space was not a significant factor for emergency physicians.

Multivariate analysis of factors linked with burnout

All demographic and professional variables found significantly linked with burnout in bivariate analysis were included and removed step by step when not significant. Finally, work–family conflict seems to be the highest-risk factor for burnout in the representative sample as well as for emergency physicians, with four and six times as much burnout when high work–family conflict scores were found (respectively, OR=4.47, 95% CI 3.05 to 6.54 and OR=6.14; 95% CI 2.89 to 13.04). We observed an exposure–outcome gradient for this risk factor (see table 6).

Table 6

Multivariate analysis of factors linked with general burnout (Copenhagen Burnout Inventory) among emergency physicians and among a representative sample of French physicians

Quality of teamwork was the second risk factor significantly related to burnout for the representative sample of French salaried physicians and for emergency physicians with, respectively, twice as much and five times as much burnout when low quality of teamwork was found (representative sample OR=2.21, 95% CI 1.61 to 3.03 and emergency physicians OR=5.44, 95% CI 2.81 to 10.53). We observed an exposure–outcome gradient that was significant for emergency physicians and with the same trend for the representative sample.

For the representative sample of French salaried physicians, quantitative demand was linked with three times as much burnout when quantitative demand was high (OR=3.32, 95% CI 1.84 to 5.99).

Emergency physicians who did not participate in continuing education during the past 12 months were three times as likely to have high burnout scores than physicians who attended continuing education (OR=3.14, 95% CI 1.23 to 8.00); a trend that was similar but not significant for the representative sample.

For both groups, being worried about making mistakes was linked with a large increase in burnout (representative sample OR=1.87, 95% CI 1.45 to 2.41 and emergency physicians OR=1.70, 95% CI 1.07 to 2.71).

Low satisfaction with pay was less predictive of burnout and was a significant factor, only for the representative sample of French physicians (OR=1.30, 95% CI 1.02 to 1.66).

Because female physicians appeared to have significantly more burnout compared with male physicians, we conducted a separate analysis for female and male emergency physicians. The importance of work–family conflict remained the highest risk factor for female emergency physicians and had a high predictive value for male physicians as less factors remained included in this smaller sample (female physicians OR=5.87, 95% CI 1.96 to 17.59 and male physicians OR=7.99, 95% CI 2.43 to 26.26). Low quality of teamwork was the highest risk factor for burnout among male emergency physicians (OR=14.31, 95% CI 4.71 to 43.43). Also, male emergency physicians who did not participate in continuing education during the last 12 months had a nearly 6-fold increase in burnout occurrence (OR=6.23, 95% CI 1.45 to 26.76), and those who did not practice exercise activities at least once a month had a 2-fold increase in burnout occurrence (OR=2.20, 95% CI 1.21 to 3.98). For female physicians, the only other significant factor was dissatisfaction with psychological support, an item included on the quality of teamwork index (OR=2.33, 95% CI 1.12 to 4.83), dissatisfaction with quality of care, another item included on the quality of teamwork index is also close to significance.

Multivariate analysis of factors linked with the intent to leave the profession

All demographic and professional variables found significantly linked with intent to leave the profession in the bivariate analysis were included and removed step-by-step when non-significant. Quality of teamwork was found to be the highest risk factor for ITL for the representative sample of French salaried physicians and for emergency physicians with, respectively, nearly a 4-fold increase in ITL and a 4-fold increase in ITL when quality of teamwork was low (representative sample OR=3.92, 95% CI 2.45 to 6.29 and emergency physicians OR=4.35, 95% CI 1.55 to 12.21; see table 7). The second major risk factor for ITL was burnout, with a stronger influence of general burnout for emergency physicians (OR=3.87 and only 1.43 for the representative sample) whereas patient-related burnout had a stronger influence on the representative sample (OR=2.31 and 2.06 for emergency physicians).

Table 7

Multivariate analysis of factors linked to intent to leave the profession (ITL) among emergency physicians and among a representative sample of French salaried physicians

For the representative sample of French salaried physicians, older age was linked with nearly a 3-fold increase in burnout when 55 years of age or over (OR=3.27, 95% CI 1.57 to 6.77). Physicians who received proposals for another job within the healthcare sector or, more importantly, outside the healthcare sector declared more ITL (respectively, OR=1.59 and OR=2.08). Other aspects of the working situation had an important influence: absence of continuing education within past 12 months was linked with a 1.5-fold increase in ITL (OR=1.59); being worried about making mistakes (OR=1.61) and tense relations with administration (OR=1.50). Dissatisfaction with pay had only a moderate influence on ITL (OR=1.33; p=0.06).

For emergency physicians, who tended to be younger, age was not linked with ITL. As most of them participated in continuous education, this factor was not a predictor for ITL. Emergency physicians who received proposals for another job outside the healthcare sector declared significantly more ITL (OR=3.14) than those who did not receive such proposals. The only other aspect of the working situation which had an important influence on ITL was reported harassment by superiors (OR=2.89; 95% CI 1.61 to 5.18).

We conducted a separate analysis for female and male emergency physicians. Quality of teamwork, burnout and proposals of another job outside the healthcare sector remained major factors for male emergency physicians. Harassment by superiors had a very high influence on female physicians (OR=4.04) as well as low influence at work (OR=3.44), and being worried about making mistakes was an important factor for male emergency physicians (OR=2.38). Female emergency physicians aged ≥55 years old had a much higher ITL than younger ones.

Discussion

According to our findings, burnout and ITL reach striking proportions among a representative sample of French salaried physicians and even more so among emergency physicians. Emergency physicians more frequently report poor working conditions and poor health. Our approach shows that working conditions may be more important than pay. Physicians (emergency physicians and the representative sample of all specialities) exposed to negative working conditions lacking positive team characteristics respond similarly.

Work/family conflict appears to be a major factor associated with burnout. Workweek duration and changing work schedules with short notice are situations which have to be addressed due to their influence on work/family conflict.

Quality of teamwork seems to be the second-most important factor associated with burnout for emergency physicians and for the representative sample of all specialities. A serious lack of teamwork appears to double the rate of burnout among the representative sample and quadruples it among emergency physicians.

Quality of teamwork appears to be a major factor associated with ITL for emergency physicians and for the representative sample. A serious lack of teamwork nearly triples ITL among the representative sample and quadruples it among emergency physicians.

Our results confirm the importance of teambuilding to reduce burnout and turnover, as shown by other authors reporting on specific healthcare settings. Factors that we found predictive of turnover intentions in our study are similar with those described in other countries.26–30 They are similar to our findings about nurses studied in 10 countries by the European NEXT.32 Safety initiatives in hospitals should focus on common healthcare interventions to promote wards organised in such a way that physicians and non-physician professionals work together within teams. The extent to which nurses and physicians have access to information, resources, support and opportunities in their work environment may have an impact on the perceived quality of collaboration with physicians and managers, the degree of job strain experienced in the work setting and, ultimately, their health.33

Shared governance structures have been shown to be highly successful in empowering nurses for professional practice in previous research.42 43 The application of an action research model aimed at managing change with active participation by all members has been found to contribute positively to the development and implementation of change efforts.44–50 Results indicate that the model promoted positive staff morale, open communication, lower staff turnover, better problem solving and improved goal attainment.

Burnout was found to be the second risk factor for ITL. A high burnout score appears to double the frequency of ITL in the representative sample and to quadruple it among emergency physicians. In previous studies, patients receiving care within units that nurses perceive as having adequate staff, good administrative support for nursing care and good relations between doctors and nurses were more than twice as likely to report high satisfaction with their care compared with other patients, and their nurses reported significantly lower burnout.51

We also found sex differences implying more burnout for female physicians but no differences in ITL. However, the predictors of burnout and ITL did not differ much among male and female emergency physicians.

Having work–family conflicts increase burnout but does not increase ITL significantly. This is in keeping with research on adverse outcomes (sickness absence) for employees struggling to combine their work and family life.52 In healthcare, many work situations require long work hours, frequent overtime, frequent shift change with short notice and involve understaffing. Physicians may, therefore, look for other work situations that would allow a better work–family balance.

Again, good teamwork and interpersonal relationships allow a better fit to the different team members' needs. Moreover, a better agreement on work organization and information sharing reduces ambiguity and interruptions and might limit the need for overtime work. It is of primary importance to establish improved processes through collaboration and multidisciplinary teamwork (among nurses, nursing aids, physicians and pharmacists) and to develop team training approaches and ward designs that facilitate high quality teamwork.

Physicians as well as nurses benefit from having a more collaborative work environment. In healthcare settings, individuals from different disciplines come together to care for patients. Although these groups of healthcare professionals are generally called teams, they need to earn true team status by demonstrating constructive teamwork,46 group cohesion and low turnover. Cohesive healthcare teams have five key characteristics: (a) clear goals with measurable outcomes, (b) clinical and administrative systems, (c) division of labour, (d) training of all team members and (e) effective communication.43–46 Empirical research on patient care teams suggests that teams with greater cohesiveness are associated with better clinical outcome measures, with higher patient satisfaction and with improved patient outcomes.47–50 56

Previous research about nurses has indicated that serious problems in work design and workforce management threaten the provision of adequate healthcare. Despite this body of knowledge, cost reduction initiatives in hospitals and reorganisation related a lack of nurses have led to managers using floating assignments and reducing the time available for in-depth discussions. As a consequence, hostile feelings may develop between occupational groups, and there may be poor circulation of information, with possible consequences for patients' safety, healthcare quality and career development opportunities for nurses. While there is some legislation in European countries, such as the ‘Plan Hospital 2007’ in France, promoting floating assignments among nurses across departments, there is a lack of research into its effects. Organizational and managerial supports for the nursing profession appear to have a profound effect on nurse dissatisfaction and burnout and are directly and independently related to quality of care.53 Similarly, the outcomes of our study imply that the physical and emotional burden of physicians requires team discussions of work organization and equipment selection. Much more understanding and many more elaborate discussions are needed to find strategies that might positively impact the environment where nurses and physicians work.41 54 Frequency of burnout among French physicians is higher that the level observed by Cydulka and Korte57 in the USA. We took into account the impact of the internal and external labour markets, and ‘employment opportunities’ was used as a control variable, receiving a proposal for a new job highly influenced physicians' intent to leave.

Limitations of our study and recommendations for future research

Because we used self-reported measures for the predictor variables and for the dependent variables (ie, burnout and intent to leave), a common-method bias may exist.55 In order to increase the validity of the outcomes, physicians' self-assessments and assessments by others, such as superiors, ought to be considered in future research.

Another limitation of our study is that we included a large number of variables in our equations to control for confounding. This resulted in more missing data, which may have decreased our power to detect statistically significant differences. However, our large sample size, the magnitude of the different risk factors for burnout and ITL, the consistency of the findings across specialities and the narrow confidence intervals for the adjusted odds ratios for the major risk factors are strengths of our analysis.

Response bias is a significant potential limitation. Perhaps physicians with the strongest opinions about these issues were more likely to respond. However, the multivariate analysis allows us to pinpoint significant aspects which can efficiently be improved in order to reduce psychosocial risk factors and burnout among physicians and more specifically emergency physicians. Major factors to improve in order to retained skilled physicians in their profession are also clarified.

Our results may indicate that medical professionals work in some institutions with no adverse outcomes such as burnout and ill health and without resulting in a large proportion of dissatisfied physicians who want to leave their profession. Further analyses are needed in order to identify factors that can explain why some institutions are better at retaining physicians than others. However, physicians leaving uncollaborative organizations for organizations where physicians and non-physician professionals collaborate in teams are supported by our findings.

What this paper adds

  • Dissatisfaction has been considered an important factor underlying intention to quit and early retirement among physicians. Restructuring and cost reduction initiatives have generated a great deal of dissatisfaction. Emergency physicians have to compensate for the lack of available general practitioners in certain areas of the country and after normal working hours. Reward by better pay has been often considered as the main proposal.

  • According to our findings, burnout and intent to leave the profession reach striking proportions among a representative sample of French salaried physicians and even more so among emergency physicians. Emergency physicians more frequently report poor working conditions and poor health. Our approach shows that working conditions are more important than pay.

  • A serious lack of teamwork appears to double the rate of burnout among a representative sample of French physicians and quadruples it among emergency physicians.

  • The outcomes of our study imply that floating assignments have to be limited and that is of primary importance to establish improved processes through collaboration and multidisciplinary teamwork and to develop team training approaches and ward designs that facilitate high quality teamwork.

Acknowledgments

The NEXT study was initiated by SALTSA (Joint Program for Working Life Research in Europe) and financed by the European Union within the 5° framework. Key action n°6.3. The population and disabilities. (QLK6-CT-2001-00475). The extension to physicians (SESMAT study) was financed by Assistance Publique-Hôpitaux de Paris and the regional Council of Ile-de-France and the regional Council of Rhône-Alpes.

Footnotes

  • The content was presented orally at the Conference ‘Doctors' Health Matters’ in London 2008.

  • Funding The NEXT study was financed by the European Union within the 5° framework. Key action n°6.3: the population and disabilities (QLK6-CT-2001-00475). The extension to physicians (SESMAT study) was financed by Assistance Publique-Hôpitaux de Paris and the regional Council of Ile-de-France and the regional Council of Rhône-Alpes.

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the University of Wuppertal, Wuppertal, Germany.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • i The NEXT research project is financed by the European Commission within the Fifth Framework, Project ID: QLK-6-CT-2001-00475, Website: http://www.next.uni-wuppertal.de.

References

Relevant Article


Free sample
This recent issue is free to all users to allow everyone the opportunity to see the full scope and typical content of EMJ.
View free sample issue >>

Don't forget to sign up for content alerts so you keep up to date with all the articles as they are published.