https://orcid.org/0000-0001-9032-4540
Figures
Abstract
Background
Rates of burnout are currently at record high levels, and GPs experience higher burnout than many other specialties. Organisational interventions may reduce burnout, but few studies have investigated these in primary care.
Aim
The current study investigated whether breaks, both with and without social interactions, were associated with burnout and patient safety perceptions in GPs.
Method
Participants completed questionnaires at baseline measuring demographic variables, burnout and patient safety perceptions. They then completed a questionnaire in the evening each day for a week which captured whether they had taken a break that day, whether it involved a positive social interaction, burnout (comprising subscales of disengagement and exhaustion), positive and negative affect and patient safety perceptions. The data were analysed using hierarchical linear modelling to assess same-day and next-day associations.
Results
We included 241 responses from 58 GPs for analysis. Taking at least one break (involving any or no social interactions) was associated with lower disengagement that day and lower exhaustion the next day. Taking at least one break involving a positive interaction was associated with 1) lower disengagement, exhaustion, overall burnout and negative affect on the same day, as well as higher positive affect and improved perceptions of patient safety, and 2) lower exhaustion and improved patient safety perceptions on the next day.
Citation: Hall LH, Johnson J, Watt I, O’Connor DB (2024) Could breaks reduce general practitioner burnout and improve safety? A daily diary study. PLoS ONE 19(8): e0307513. https://doi.org/10.1371/journal.pone.0307513
Editor: Collins Atta Poku, Kwame Nkrumah University of Science and Technology, GHANA
Received: July 25, 2023; Accepted: June 28, 2024; Published: August 27, 2024
Copyright: © 2024 Hall et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: We are unable to share this data or upload it to a repository due to restrictions within our ethical approvals. To contact the ethics committee who approved the study, the email address is: psyc-ethicssubmissions@leeds.ac.uk.
Funding: This report is independent research supported by National Institute for Health and Care Research Yorkshire and Humber ARC [under grant NIHR200166]. The views expressed in this publication are those of the authors and not necessarily those of the National Institute for Health Research or the Department of Health and Social Care. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Rates of physician burnout have increased internationally since the onset of the Covid-19 pandemic and are now at their highest on record [1, 2]. Burnout is a syndrome characterised by disengagement and emotional exhaustion, and levels are consistently higher in general practitioners (GPs) than the mean across physicians specialties [3]. A recent UK survey found that 68% of GPs reported high exhaustion and 48% reported high disengagement [4].
Elevated burnout is linked with organizational outcomes including absenteeism, intentions to leave, poorer patient satisfaction and poorer patient safety [5–7]. Interventions for burnout can be categorized as those which are person-directed, aiming to treat an individual without reference to their context, and those which are organization-directed, aiming to improve an individual’s wellbeing by making organizational changes [8]. Person-directed interventions include, for example, mindfulness, stress-management workshops and exercise interventions. Organization-directed interventions include, for example, work-scheduling changes, improved cafeteria facilities and the delivery of professional training. Studies suggest both types of interventions can be effective, but more research has focused upon person-directed interventions, and effect sizes are generally small [8–10].
While organizational interventions may be effective for reducing burnout, few studies have investigated these in primary care settings. One possible intervention is the use of shared breaks. A focus group in GPs indicated this is a feasible intervention which reduces isolation and creates opportunity for positive interactions [11]. However, no quantitative evaluations have been conducted, and in general, there have been a lack of studies in this area. A review of 32 studies into breaks in doctors found only two investigated standard 30-minute breaks [12]. A further two investigated sleep interventions, four investigated yoga and mindfulness, six investigated ‘microbreaks’, seven were cross-sectional survey studies, four were prospective or retrospective cohort studies, four were qualitative studies and one was a mixed-methods study [12]. Only two studies included burnout as an outcome measure, and both investigated yoga or mindfulness rather than general social breaks. Only three of the included studies included primary care physicians, and these did not measure burnout as an outcome measure. Furthermore, despite consistent evidence of a link between burnout and patient safety [13], none of the included studies measured any patient safety outcome indicators [12]. It is also unclear whether it is breaks of any sort that are associated with physician wellness or the opportunity for positive social interactions these can provide.
The present study addressed these issues using a daily diary method, which allows comparisons within individuals across multiple days. We had four key research questions:
- Is taking a break associated with burnout and wellbeing that day?
- Is taking a break associated with stress, burnout, and wellbeing the next day?
- Is taking a break associated with patient safety that day?
- Is taking a break associated with patient safety the following day?
Within each of these questions, we investigated whether there were associations between breaks of any sort and the outcome variables, and breaks including a positive social interaction.
Methods
Design and materials
A within-subjects, interval contingent, daily diary design. Participants completed a questionnaire battery at baseline, followed by a stress questionnaire before work in the morning and a longer questionnaire battery in the evening before bed, for seven consecutive days. Questionnaires were online and participants received reminders via text and email.
Recruitment
Participants were recruited from 1) GP organisations and media outlets, 2)
social media, 3) local practices, 4) local GP continuing professional development events and 5) pre-existing contact lists from prior research and personal network between 1 August 2017 and 1 January 2018.
Ethical considerations
This study received ethical approval from the University of Leeds, School of Psychology Ethics Committee (ref #17–0185, date: 27/06/17) and Health Research Authority approval (IRAS # 216260). Participants were fully informed of the types of questions that would be involved in the study prior to providing their written informed consent to take part. They were also informed that they may skip any questions they wished to, and that they may withdraw from the study at any point before, during, and after taking part, up until the point of analysis (one month post study completion). They were also informed that their responses would be kept completely anonymous through the use of a unique participant code, with any contact information stored securely and separately to their participant code and survey responses. Finally, contact information for relevant support/helplines were provided at the end of the study in the event that participants became distressed as a result of the potentially sensitive nature of the study.
Baseline questionnaire
Demographic variables (age, gender, ethnicity, location) and occupational variables (workload, practice list size, number of hours spent on various work tasks etc.) were measured. The GHQ-12 [14] was used to measure mental distress, with scores above 3 indicating a possible psychiatric disorder [15]. It showed good internal consistency in our sample (α = .820). The Quality of Life (QoL) was used to measure a broader sense of participants’ overall wellbeing [16–18]. The four-item shortened version of the PSS (the PSS-4) captured participants’ stress over the last month [19]. Cronbach’s alpha in our sample demonstrated good internal consistency (α = .809). Burnout was measured using the 16-item Oldenburg Burnout Inventory (OLBI) [20]. This measures burnout on two subscales; Exhaustion, Disengagement. Cronbach’s alpha in our sample demonstrated good internal consistency (α = .878). The Safe Practitioner Measure [21] captured perceptions about safety of personal practice using a single item. This measure has been validated in survey and daily diary study research [22–24]. Practice level support was measured with a visual analogue scale (VAS). This question has been previously used within a GP sample and found to be an important contributor to wellbeing, burnout, and patient safety [23].
Morning questionnaire
Prior to work, participants were asked to complete one question on stress; “How stressed do you feel this morning?” on a 10-point scale.
Evening questionnaire
Only data from work days was included; non-work days were excluded from the study.
Positive and negative affect.
Daily levels of wellbeing were measured using positive and negative affect adjectives (details in S1 Methods in the S1 Data).
Burnout.
Three items from the OLBI with adapted wording for use at the daily level. An example item is, ‘Today I found new and interesting aspects in my work’. Cronbach’s α = .818 demonstrated good internal consistency.
Safety perceptions.
The daily version of the safe practitioner measure was used [21]. A previous study which used this found good variability across days [22].
Adverse events & near misses.
Participants were asked whether they had been responsible for any 1) adverse events (AEs) and 2) near misses (NM) that day. Further questions were asked in order to classify the (potential) outcome of the AE/NM; “Minor reversible patient harm, Minor irreversible patient harm, Major reversible patient harm, Major irreversible patient harm”, the type of AE/NM; “Diagnostic error/near miss, Medication or prescription error, Equipment error, Communication, Monitoring error, Other” and the contributors to them; “System issue, Degree of fatigue, Lapse in concentration, Lapse in judgment, Lack of knowledge, Degree of stress or burnout, Other”. For inferential analyses, AEs and NMs were collapsed and coded as 1 (for either present), and 0 (for neither present).
Data preparation and analysis
The data set was created by matching participants’ morning responses to their evening responses. Responses from days off work were excluded for the purposes of this study. Survey time points that were completely missing were deleted, along with their corresponding morning responses (n = 8). Survey time points that were completed on the wrong day (i.e. backfilled) were excluded from the analysis (n = 16) along with one participant who did not complete any evening surveys. The final number of participants was 58, with 241 daily entries.
For the lagged analyses (to determine next day effects), participants were included if they had completed three or more consecutive days at work. To create the data file, participants’ burnout, wellbeing, stress rating, and safety scores from the subsequent day were mapped onto the previous day’s row, and the final day’s scores from that participant deleted. This resulted in 41 participants with a total of 128 days, giving an average of 3 days per participant.
Hierarchical linear modelling in HLM7 was conducted [25]. This allows the data to be assessed both within and between levels. Relationships can be established at a daily level and more accurate conclusions can be drawn through this type of analysis compared to only between-participant analyses, as each participant essentially acts as their own control. The data from this study was across two-levels. Level 1 contained within-participant variation (e.g. daily burnout) with predictor variables centred around the group mean. Level 2 contained between-participant variability (e.g. baseline levels of burnout) with variables centred around the grand mean [26, 27]. Each model controlled for participants’ age, gender, and years in practice. For further data preparation and analysis information, including the equations used for calculating the Pseudo R2 to provide an estimate of effect sizes, see the Supplement (S2 Methods in the S1 Data) [28].
Results
Descriptive statistics
Fifty-nine GPs took part. Descriptive data (Table 1 and S1 Table in the S1 Data) was based on the dataset used for same-day analyses (n = 58 with missing data imputed). Fifty-one (87.9%) participants scored as having a possible minor psychiatric disorder on the GHQ-12, 50 (86%) were classed as having mild (29%) or severe (57%) exhaustion and 44 (76%) were classed as having mild (40%) or severe (36%) disengagement. The average length of breaks was 13 minutes (median = 20, range = 5–120), and there were more days when GPs took breaks than did not take a break (165 vs. 76, respectively; S2 Table in the S1 Data). We measured the number of breaks with a negative interaction with a colleague, but did not include this in any models as there were not enough for statistical power (only 5% of cases in the same day analyses, and 3.9% in the lagged analyses).
The most common outcome for AEs and NMs was (potential for) ‘minor reversible patient harm’ (92% of AEs, 76% of NM). The most common AE and NM types were medication or prescription (42% and 38%, respectively), followed by communication (25% and 35%, respectively). Lapse in concentration was the most cited contributor for the AE or NM (27%), followed by fatigue (24%) and stress or burnout (15%).
1. Is taking a break associated with burnout and wellbeing that day?
1a. Taking a break and burnout.
Having a break of any sort (including any interaction type) had a significant negative association with disengagement (Tables 2 and 3). This remained significant when controlling for baseline disengagement (β = -.433, p = .040, pseudo R2 = .017). There was no significant association between breaks and total burnout or exhaustion.
Breaks with a positive interaction had a significant negative association with total burnout. This held when controlling for baseline total burnout (β = -.927, p = .032, pseudo R2 = .038). When using the subscales of burnout independently, breaks with a positive interaction were significantly negatively associated with both disengagement and exhaustion. Disengagement remained significant when controlling for baseline disengagement scores (β = -.639, p = .001, pseudo R2 = .040). Exhaustion was no longer significant when controlling for baseline exhaustion scores (β = -.432, p = .111, pseudo R2 = .024).
1b. Taking a break and wellbeing.
Having a break of any sort during the day was not significantly associated with positive or negative affect on the same day.
Breaks with a positive interaction(s) were significantly positively associated with positive affect and significantly negatively associated with negative affect. These remained significant when controlling for baseline wellbeing scores (GHQ-12: β = 3.258, p = .010 for positive affect, β = -6.621, p = .002 for negative affect, and QoL: β = 3.45, p = .011 for positive affect, β = -6.603, p = .002 for negative affect).
2. Is taking a break associated with stress, burnout, and wellbeing the next day?
2a. Taking a break and next day morning stress.
Having a break of any sort, and having a break with a positive interaction both had no association with morning stress the following day (Tables 3 and 4).
2b. Taking a break and next day burnout.
Having a break of any sort, and specifically having a break with a positive interaction were both significantly associated with lower exhaustion the following day. However, when controlling for baseline exhaustion, neither remained significant (β = -0.443, p = 0.106, pseudo R2 = .019; β = -0.474, p = .122, pseudo R2 = .016, respectively). No significant associations were found for disengagement and overall burnout the following day.
3. Is taking a break associated with patient safety that day?
3a. Taking a break and safety perceptions.
Breaks with a positive interaction had a significant negative association with safe practitioner scores that day (Tables 3 and 5), meaning they were more able to act as a safe practitioner that day. This association remained significant when controlling for baseline safety perceptions (β = -0.343, p = .008, pseudo R2 = -.003). Breaks of any sort were not significantly associated with safe practitioner scores.
4. Is taking a break associated with patient safety on the following day?
4a. Taking a break and next day safety perceptions.
Taking a break with a positive interaction was associated with perceptions of safer practice on the following day (Tables 3 and 6). This association remained significant when controlling for baseline safety perceptions (β = -0.697, p = .001, pseudo R2 = .162). No significant association was found for breaks of any sort.
Discussion
Summary
This is the first study to quantitatively investigate the association between taking breaks with burnout and patient safety in GPs. Taking at least one break during the working day, involving any type of social interaction (including no interactions), was associated with lower levels of disengagement that day and lower exhaustion on the following day. Taking at least one break involving a positive interaction was associated with lower disengagement, exhaustion, overall burnout and negative affect on the same day, as well as higher levels of positive affect and improved perceptions of patient safety. Taking at least one break involving a positive interaction was also associated with lower exhaustion and improved patient safety perceptions on the following day. Taking breaks was not associated with reported stress levels or risk of patient safety incidents, and it should be noted that the next-day associations between breaks and exhaustion were not maintained when controlling for baseline emotional exhaustion.
Strengths and limitations
A strength of the study was its use of a daily diary study approach, which enabled the analysis of associations both within the same day and across days, and enabled participants to act as their own controls. It may have been limited by a possible response bias, which could have caused physicians who were burnt-out to be more likely to want to respond to study advertisements.
Comparisons with existing literature
The present study addresses concerns about the erosion of breaks in healthcare settings [29] by reporting the first quantitative evidence of an association between taking breaks, burnout and patient safety perceptions in doctors. A previous systematic review of the literature on breaks and wellbeing in doctors found 32 studies, but concluded that further research was needed to clarify whether breaks are beneficial [12]. This review also highlighted a lack of studies measuring burnout or patient safety perceptions as outcome measures, and identified a need for more studies in primary care physicians [12]. Our findings complement previous qualitative work which has identified breaks as a realistic and beneficial intervention in primary care settings [11]. They also complement previous cross-sectional survey studies in hospital doctors, which indicate that taking breaks is associated with lower work stress and faster reaction times [12]. More broadly, these results are consistent with a large literature which has linked higher fatigue with higher burnout and poorer patient safety [30, 31]. The present study investigated breaks, a potential antidote to fatigue [32], and highlights breaks as a potential intervention. It also extends this literature by identifying the opportunity for positive interactions as an important feature of effective breaks.
Implications for research and practice
Having a break during the day where GPs can interact with their colleagues may help improve wellbeing, alleviate burnout and improve patient safety. While this solution may not be the ultimate answer to current staffing shortages, it is practical, feasible, and could be more assertively and routinely implemented with relatively little difficulty compared to alternative solutions. These findings have relevance for healthcare managers; the finding that the opportunity for positive social interactions was important suggests that breaks should be organised at the practice or team level, so staff can take them together.
Supporting information
S1 Methods. Daily measurement of positive and negative affect.
https://doi.org/10.1371/journal.pone.0307513.s002
(DOCX)
S2 Methods. Further information regarding data preparation and analysis.
https://doi.org/10.1371/journal.pone.0307513.s003
(DOCX)
S1 Table. Descriptives for daily Level 1 data, at work only.
https://doi.org/10.1371/journal.pone.0307513.s004
(DOCX)
S2 Table. Total number of breaks and patient safety incidents from Level 1 data, same day analyses.
https://doi.org/10.1371/journal.pone.0307513.s005
(DOCX)
References
- 1. General Medical Council, National training survey 2022 results. 2022.
- 2. Shanafelt T.D., et al. Changes in burnout and satisfaction with work-life integration in physicians during the first 2 years of the COVID-19 pandemic. in Mayo Clinic Proceedings. 2022. Elsevier. pmid:36229269
- 3. Shanafelt T.D., et al., Burnout Among Physicians Compared With Individuals With a Professional or Doctoral Degree in a Field Outside of Medicine. Mayo Clinic Proceedings, 2019. 94(3): p. 549–551. pmid:30832797
- 4. Whitehead I.O., et al., A national study of burnout and spiritual health in UK general practitioners during the COVID-19 pandemic. Plos one, 2022. 17(11): p. e0276739. pmid:36322555
- 5. Hall L.H., et al., Healthcare staff wellbeing, burnout, and patient safety: a systematic review. PloS One, 2016. 11(7): p. e0159015. pmid:27391946
- 6. Firth H. and Britton P., ‘Burnout’, absence and turnover amongst British nursing staff. Journal of Occupational Psychology, 1989. 62(1): p. 55–59.
- 7. Zhang Y. and Feng X., The relationship between job satisfaction, burnout, and turnover intention among physicians from urban state-owned medical institutions in Hubei, China: a cross-sectional study. BMC health services research, 2011. 11(1): p. 235. pmid:21943042
- 8. Panagioti M., et al., Controlled interventions to reduce burnout in physicians: A systematic review and meta-analysis. JAMA Internal Medicine, 2017. 177(2): p. 195–205. pmid:27918798
- 9. Dreison K.C., et al., Job Burnout in Mental Health Providers: A Meta-Analysis of 35 Years of Intervention Research. Journal of occupational health psychology, 2016. 23(1): p. 18–30. pmid:27643608
- 10. West C.P., et al., Interventions to prevent and reduce physician burnout: a systematic review and meta-analysis. The Lancet, 2016. 388(10057): p. 2272–2281. pmid:27692469
- 11. Hall L.H., et al., Strategies to improve general practitioner well-being: findings from a focus group study. Family practice, 2017. 35: p. 511–516.
- 12. O’Neill A., et al., Impact of intrawork rest breaks on doctors’ performance and well-being: systematic review. BMJ open, 2022. 12(12): p. e062469. pmid:36517098
- 13. Hodkinson A., et al., Associations of physician burnout with career engagement and quality of patient care: systematic review and meta-analysis. bmj, 2022. 378. pmid:36104064
- 14.
Goldberg D.P. and Williams P., A user’s guide to the General Health Questionnaire. 1988, Windsor, England: NFER-Nelson.
- 15. Hardy G.E., et al., Validation of the General Health Questionnaire-12: Using a sample of employees from England’s health care services. Psychological assessment, 1999. 11(2): p. 159.
- 16. Fahrenkopf A.M., et al., Rates of medication errors among depressed and burnt out residents: prospective cohort study. British Medical Journal, 2008. 336(7642): p. 488–491. pmid:18258931
- 17. West C.P., et al., Association of perceived medical errors with resident distress and empathy: a prospective longitudinal study. JAMA, 2006. 296(9): p. 1071–1078. pmid:16954486
- 18. West C.P., et al., Association of resident fatigue and distress with perceived medical errors. JAMA, 2009. 302(12): p. 1294–1300. pmid:19773564
- 19. Cohen S., Kamarck T., and Mermelstein R., A global measure of perceived stress. Journal of Health and Social Behavior, 1983. 24(4): p. 385–396. pmid:6668417
- 20.
Demerouti E. and Bakker A.B., The Oldenburg Burnout Inventory: A good alternative to measure burnout and engagement, in Handbook of Stress and Burnout in Health Care, Halbesleben J.R.B., Editor. 2008, Nova Science: Hauppauge, NY.
- 21. Louch G., et al., The daily relationships between staffing, safety perceptions and personality in hospital nursing: A longitudinal on-line diary study. International Journal of Nursing Studies, 2016. 59: p. 27–37. pmid:27222447
- 22. Louch G., et al., A Daily Diary Approach to the Examination of Chronic Stress, Daily Hassles and Safety Perceptions in Hospital Nursing. International Journal of Behavioral Medicine, 2017. 24(6): p. 946–956.
- 23. Hall L.H., et al., Association of GP wellbeing and burnout with patient safety in UK primary care: a cross-sectional survey. 2019. 69(684): p. e507–e514. pmid:31015224
- 24. Johnson J., et al., Burnout mediates the association between symptoms of depression and patient safety perceptions: A cross-sectional study in hospital nurses. Journal of Advanced Nursing, 2017. 73(7): p. 1667–1680.
- 25. Raudenbush S.W., HLM 6: Hierarchical linear and nonlinear modeling. 2004: Scientific Software International.
- 26.
Bryk A.S. and Raudenbush S.W., Hierarchical models: Applications and data analysis methods. Atlanta, GA: Sage, 1992.
- 27. Kreft I.G., De Leeuw J., and Aiken L.S., The effect of different forms of centering in hierarchical linear models. Multivariate behavioral research, 1995. 30(1): p. 1–21. pmid:26828342
- 28.
Snijders T. and Bosker R., Multilevel Analysis. 1999, London: Sage.
- 29. Chatfield C. and Rimmer A., Give us a break. 2019. 364: p. l481.
- 30. Tawfik D.S., et al., Physician Burnout, Well-being, and Work Unit Safety Grades in Relationship to Reported Medical Errors. Mayo Clinic Proceedings, 2018. 93(11): p. 1571–1580. pmid:30001832
- 31. Gates M., et al., Impact of fatigue and insufficient sleep on physician and patient outcomes: a systematic review. 2018. 8(9): p. e021967. pmid:30244211
- 32. Blasche G., Arlinghaus A., and Crevenna R., The impact of rest breaks on subjective fatigue in physicians of the General Hospital of Vienna. Wiener klinische Wochenschrift, 2022: p. 1–6. pmid:34613476