Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Effect of burnout among physicians on observed adverse patient outcomes: a literature review

Effect of burnout among physicians on observed adverse patient outcomes: a literature review Background: Physician burnout has potentially harmful effects for both physicians and their patients. Despite relationships between physician burnout and lowered patient satisfaction and clinician-rated adverse patient outcomes, there is scarce literature regarding effects on objective patient outcomes. This study aimed to examine the relationship between physician burnout and observed adverse patient outcomes via a review of the literature. Methods: A search was performed on the MEDLINE, EMBASE and PsychINFO databases, using keywords and Medical Subject Headings. The identified studies were in English, published from 2007 to 2019, measured burnout among physicians using the Maslach Burnout Inventory (MBI), and included observed adverse patient outcomes. In total, 360 eligible articles were identified, and 11 were included in the final review. All included studies measured patient outcomes by observed clinical measures (e.g. quality of care and medical errors). Results: Four studies found a clear significant relationship between physician burnout and observed adverse patient outcomes, while 6 did not. One study found a significant relationship with one of the MBI subscales. Burnout was, in contrast to depression, only partly associated with observed patient outcomes. Conclusions: This review illustrates the need for a validation of physician burnout measured by MBI with respect to observed patient outcomes. Further studies are required to investigate the effects of physician burnout on observed quality of their patient care. Keywords: Physician burnout, Patient outcome, Medical error Background from 2015 distinguished between different types of nega- Burnout among physicians is a widespread phenomenon tive treatment outcomes, and described one category of and has been found to be of importance for both physi- more subjectively perceived outcomes, and two categor- cians and their patients [1]. Recent reviews suggest a link ies of objectively measured outcomes [5]: (I) The sub- between physician burnout and negative impact on pa- jectively perceived outcomes were the physicians’ and/or tient care, including increased medical errors [2–4]. A patients’ perception of quality of care: These were out- review study on psychosocial work stress and burnout comes related to patient centeredness of clinical activ- ities and the physician–patient relationship as well as to physician empathy, accessibility, reliability and exchange * Correspondence: reidar.tyssen@medisin.uio.no of information. (II) One category of objectively measured The first two authors (Mangory and Ali) share the first authorship. Kashan Yasin Mangory and Lavin Yadgar Ali contributed equally to this outcomes included assessments of treatment success work. and absence of complications. These outcomes were Department of Behavioural Medicine, Institute of Basic Medical Sciences, evaluated by different observed indicators of treatment Faculty of Medicine, University of Oslo, Oslo, Norway Full list of author information is available at the end of the article © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Mangory et al. BMC Health Services Research (2021) 21:369 Page 2 of 8 success (such as diabetes or blood pressure control), performance [10, 11]. In an editorial from 2018 titled clinical quality or lack of treatment complications. (III) “Physician burnout—a serious symptom, but of what?” Another category of objectively measured outcomes was Schwenk and Gold discuss the high but varying preva- prevalence of medical errors: These outcomes included lence of burnout reported in many studies and they observed or recorded prescriptions, drug administration strongly emphasize the need for a deeper understanding or surgical errors. of its origins and possible consequences, which have not Most of the reviews so far build mainly on the subject- been sufficiently explained [12]. It is therefore important ively perceived data, from physicians and patients them- to relate burnout levels both to the physician’s individual selves. There have been surprisingly few reviews of suffering, but also to impact of burnout in the clinical observational studies on the relationship between phys- setting and for patient treatment. ician burnout and objectively measured outcomes of The relationships between job demands, burnout, work sub-optimal care of patients or errors [4]. We therefore engagement, physician impairment and quality of patient need to study whether physician burnout is actually as- care are complex. High job demands have been found to sociated with observed adverse patient care and if it in- relate to emotional exhaustion, whereas lack of resources creases the risk of medical errors. has been related to disengagement [13]. However, the doc- The distinction between objectively observed out- tor’s perception of the type of demand can influence how comes and subjective perception of quality of care is it impacts burnout. Demands perceived as challenges were crucial for patients, physicians and their employers. We found to be positively associated with work engagement, have therefore reviewed relevant observational studies to while demands perceived as hindrances could increase provide insights into this issue. burnout. Engagement in work tasks could be developed Burnout is usually measured using the Maslach Burn- despite a certain level of exhaustion [14], and could thus out Inventory (MBI). In a review of burnout prevalence, mitigate consequences of burnout. Additionally, there is a covering the period from 1991 to 2018, including 185 strong moral imperative in the profession to “put patients studies from 45 countries, 86% of the studies had used a first”, articulated in the Declaration of Geneva’sPhysi- version of the MBI [6]. The MBI has been found to have cians´ Pledge [15], and this can urge doctors to preserve a replicable factor structure within and between profes- the quality of care even during heavy workload periods for sional groups and across countries [7]. Although there the individual physicians. are several other valid measures of burnout (such as In addition, a stressful working climate in some hos- OLBI—Oldenburger Burnout Inventory, Burnout Inven- pital units may lead to burnout also among health pro- tory, CBI—Copenhagen Burnout Inventory), we have, on fessionals other than physicians [16]. This applies also to this background, limited the present review to include primary health care settings in non-Western countries, only studies using the MBI. The MBI has, however, been as shown in recent studies [17, 18]. Burnout in whole modified in several places in order to accommodate a hospital units and departments may also impact quality variety of professionals and it has been shortened to in- of patient care [5]. clude fewer items per dimension, or even only a single Several studies have shown that burned-out physicians item [3]. themselves report that their condition affects their pa- The MBI includes three main components or subscales: tient treatment and care negatively [2, 4]. A recent emotional exhaustion (EE), which has been thought to be meta-analysis on the link between physician burnout the initial reaction, followed by depersonalization/disen- and patient safety found that “physician burnout may gagement/cynicism (DP) which represents an emotionally jeopardize patient care”, but that most studies were detached attitude towards work (understood as a way of cross-sectional and relied on outcomes self-reported by coping with exhaustion) and, finally, an experience of re- physicians [2]. The authors “failed to show significant duced personal accomplishment (or performance) (PA) links between physician burnout and patient safety out- [8]. Both cynicism and reduced accomplishment could in- comes recorded in the health-care systems (e.g., the dicate impaired physician functioning and subsequent health records of patients, monitoring etc.).” In line with subordinate patient care. this, another recent review found no significant link be- However, critique of this burnout model points to un- tween physician burnout and observed clinical patient certainties about these possible relationships. Firstly, the outcomes in any of the five studies identified; however, MBI is not a diagnostic measure. Maslach surveyed per- there were associations between physician burnout and sons “at work”, defining the third with highest score as outcomes reported by physicians and patients [4]. “high burnout”. This has later been used as a cut-off In this literature review, the definition of “adverse pa- score. Those who were off sick or not able to be at work tient outcomes” includes observed measures for failure were excluded from definition [9]. The scale is thereby or lack of prevention, inadequate assessment (diagnoses) poorly validated with respect to impairment and work and treatment. Mangory et al. BMC Health Services Research (2021) 21:369 Page 3 of 8 This literature review therefore aims to identify studies Two of the studies included both nurses and physicians examining the association between physician burnout, in intensive care units, which are considered stressful measured by the MBI, and observed quality of care workplaces [16]. We included these studies because they among their patients (including number of medical comprised substantial samples of physicians. errors). Results Methods Eleven studies were included in the review (Table 1). We used the search engines MEDLINE, EMBASE, and There was a large variation in the size of the samples, in PsychINFO. The final search consisted of keywords and the observed outcomes reported in the studies and in Medical Subject Headings for predictor (e.g., burnout), how MBI burnout was measured and defined (i.e. sample (e.g., physician) and potential outcomes (e.g., whether the subscales were dichotomized, categorized or medical errors) using the following three groups of used as continuous variables); see Table 1 footnotes. search terms: Seven studies were of general practitioners (GPs) or from primary care settings, two studies were from inten- 1) [Burnout, Professional/]OR [Physician burnout] sive care units, one study was among paediatric resi- 2) [Practice Patterns, Physicians/]OR [Physicians, dents, and one study was about first-year residents in Primary Care/]OR [Physicians, Family/]OR internal medicine. Three of the studies were from the [Physicians/]OR [Primary Health Care/]OR United States, three from Denmark, and one study each [Internship and Residency/] from the Netherlands, Israel, Spain, France and 3) [Quality of Health Care/]OR [Medical Errors/]OR Switzerland. [Medication Errors/]OR [Physician-Patient Fahrenkopf et al. [22] studied cohorts of paediatric res- Relations/]OR [Patient Care/]OR [Referral and idents from three U.S. hospitals over one and a half Consultation/]OR [Attitude of Health Personnel/ month. Burnout was measured by EE and DP in, and ad- ]OR [Patient Outcome] verse patient outcome was the rate of errors made in or- dering medication. Residents with high combined score The search combined “OR” between the search terms for EE (> 27) and DP (> 10) were found to make similar in each group with “AND” between the groups, resulting rates of errors per resident month as residents with low in 357 articles. In addition, experts in the field (KIR and burnout scores. There was a significant association be- RT) provided three articles [19–21]. The articles were tween depression and medical errors (Harvard national subject to a manual review by two separate authors depression screening day scale); depressed residents (KYM and LYA). Inclusion criteria were as follows: (I) made 6.2 times more errors than non-depressed resi- the papers had abstracts written in English; (II) they dents. The medical errors were collected and rated by were published between 2007 and 2019; (III) they mea- trained observers (nurses and physicians). sured physician burnout with MBI as a predictor; and Zantinge et al. [23] observed videotaped consultations in (IV) they included observed measures of patient out- Dutch GPs. The findings were somewhat discrepant: GPs comes. The original search was for the student authors’ with low PA communicated less affectively, were less master thesis and had a 10-year frame (2007–2017). For patient-centred and had less eye contact, whereas those the preparation of this manuscript, we updated the with high EE and DP talked more about psychosocial search to include the past 2 years, resulting in the range problems in their consultations. The latter may have in- 2007–2019. The last year – 2020 – is heavily influenced creased possibility for adequate mental health care. by the COVID-19 pandemic at hospitals and other Kushnir et al. [24] found in a study from Israel among health services, and is not comparable to previous years primary care physicians that referral rates increased for with respect to possible confounders, such as, e.g., stress both diagnostic tests and to specialist clinics among phy- outside of work. First, we excluded articles that did not sicians who scored high on burnout. Being a specialist in have an abstract or had an irrelevant title and abstract. family medicine was more important than burnout for The remaining 29 articles were then read in full text by referring to a specialist clinic. Burnout was here mea- the two first authors (KYM and LYA) and excluded if sured as a common mean of items of EE and DP. they did not include objective measures of patient out- Yuguero Torres et al. [25] studied doctors from 22 pri- comes or did not investigate the association between mary care centres in a region in Spain (Catalonia). The physician burnout, measured by the MBI, and adverse doctors’ prescribing of sick leave was not associated with patient outcomes. The senior authors (KIR and RT) higher burnout among them. Burnout scores were di- reviewed the articles and took actively part in the last se- vided into low, moderate and high on EE, DP and PA. lection and writing. Finally, 11 articles seemed to meet There was an association between higher burnout and the inclusion criteria and were included in this review. lower empathy (Jefferson), but this did not influence the Mangory et al. BMC Health Services Research (2021) 21:369 Page 4 of 8 Table 1 Description of studies retained by the literature review Study Country Design Sample N Response Measured Observed measure on Association with burnout rate dimensions quality of care of burnout Fahrenkopf USA Observational Paediatric 246 50% EE and DP Missing or wrong Not associated with adverse et al. 2008 prospective residents prescription of drugs patient treatment [22] cohort study Zantinge Netherlands Observation GPs 142 89% EE, DP and Length of consultations, level PA associated with GPs et al. 2009 of video- PA of verbal communication, communicating less affectively, [23] recorded con- eye contact, and focus on being less patient-centred and less sultations and psychosocial issues eye contact. EE and DP not associ- questionnaire ated with adverse patient treatment Kushnir Israel Cross- GPs/Primary 136 99% EE, DP and Number of referrals for Associated with more referrals et al. 2014 sectional care PA diagnostic imaging, [24] observational specialized health services or study nurse sensitive treatments Yuguero Spain Prospective GPs 217 50% EE, DP and Number of prescribed sick Not associated with prescribing Torres et al. observational PA leaves more or longer sick leaves 2015 [25] study (1 year) Garrouste- France Prospective Doctors, 1988 77% EE, DP and Medical errors (i.e., error of Not associated with adverse Orgeas observational nurses and (330 PA execution or error of patient treatment et al. 2015 study (2-year) care doctors) planning). Adverse events [26] workers in were patient harms caused intensive by medical interventions. care unit Welp et al. Switzerland Observational Doctors 1425 Not EE, DP and Length of stay in hospital Associated with increased 2015 [19] study and nurses (243 specified PA and standardized mortality standardized mortality ratio, but in intensive doctors) ratio not length of stay care unit Pedersen Denmark Register study GPs 835 72% EE, DP and Number of requisitions for Not associated with increased et al. 2015 and PA PSA among healthy male requisitions for PSA [27] questionnaire patients Kwah et al. USA Prospective First-year 54 98% EE and DP Medication prescription Not associated with increase in 2016 [28] observational residents in (initial), errors with potential for medical errors study (1 year) internal 59% adverse drug effects medicine (cohort) Sun et al. USA Cross- Primary 102 Not EE, DP and Antibiotic prescriptions for Not associated with increase in 2017 [29] sectional care specified PA acute respiratory infections prescriptions observational study Noroxe Denmark Prospective GPs 781 50.2% EE, DP and Conditions not requiring Associated with increased et al. 2019 observational PA hospitalization in the case of frequency of hospitalizations of [20] study (6 appropriate intervention in ambulatory care sensitive months) primary care (ambulatory conditions care sensitive conditions) Noroxe Denmark Prospective GPs 409 50.2% EE, DP and Change of GP (unrelated to Associated with increased et al. 2019 observational PA change of address) likelihood of changing GPs [21] study (6 months) Abbreviations: EE emotional exhaustion, DP depersonalization, PA personal accomplishment (reduced), GP general practitioner, PSA prostate-specific antigen, MBI Maslach Burnout Inventory Defined burnout as EE > 27 and DP > 10 EE, DP and PA were dichotomized into low and high scores, and outcomes reported for each dimension Overall burnout was measured as an average of the responses to items EE and DP (continuous variable) Scores for each category were divided into low, moderate and high Burnout measured as (1) a combination of high EE and DP with low PA (dimensions dichotomized), and (2) defined as over a cut-off of a global MBI score of 9 Each dimension used as a continuous variable Burnout defined as high level of emotional exhaustion > 26 Burnout defined as a high EE or high DP subscore (dichotomized variables) Burnout measured 1) in four quartiles for each dimension 2) As a composite score by adding up points corresponding to the quartile of each subscale (reversed score for personal accomplishment); one point for scores in the first quartile, and two, three and four points for subsequent quartiles, respectively. The composite score was categorized into five groups, with increasing score indicating higher burnout Mangory et al. BMC Health Services Research (2021) 21:369 Page 5 of 8 number or duration of sick leave prescriptions among number of hospitalisations was obtained from national the doctors. register data. Garrouste-Orgeas et al. [26] conducted a large and im- Noroxe et al. [21] did a second study of the same eli- pressive 2-year prospective, observational study of doc- gible sample of Danish single-practice GPs, as men- tors (N = 330), nurses and care workers in 31 Intensive tioned above. Here the outcome was change of GP Care Units (ICUs) in France. There was no significant among the patients (unrelated to their change of ad- association between medical errors or adverse events dress). There was a dose-response like association be- and burnout measured either as (1) a combination of tween increase in DP and PA subscales (but not in EE) high EE and DP with low PA, or (2) defined as over a among the GPs and the likelihood that their patients left cut-off on a global MBI score. There was a significant them for a new GP. There was also a significant associ- association between depression symptoms (CES-Depres- ation with an increase in the composite burnout score. sion Scale) and medical errors. Medical errors were de- Data regarding change of GP were obtained from a na- fined by Delphi technique among 60 experts. tional register. Welp et al. [19] studied patient safety in ICUs in 48 Swiss hospitals. They assessed standardized mortality ra- Summary of the findings tios (SMR) and length of stay, and related these mea- Of the11studies,only4[19–21, 24] found a clear sures to burnout among medical doctors (N = 243) and significant association between physician burnout and nursing staff. The three burnout subscales were mea- adverse patient outcomes, whereas 6 [22, 25–29] sured continuously, with EE predicting a significant in- failed to find such an association. One study had di- crease in SMR, whereas DP and PA did not. Burnout did vergent findings, with some aspects of patient care not predict length of stay for patients in ICU. (observed communication skills) being reduced, Pedersen et al. [27] studied the association between whereas other aspects (time devoted to psychosocial burnout and other psychological measures in Danish issues) increased among burnout doctors [23]. There solo GP practices and testing of prostate specific antigen was no convincing association related to either posi- (PSA) in healthy male patients. Burnout measured as tions or practice venue. Among the 7 studies from high EE (> 26) was not associated with increased PSA primary care/general practice, 3 found a clear signifi- testing, and neither was empathy (Jefferson). On the cant association [20, 21, 24], while 4 did not [23, 25, other hand, high anxiety scores and bad outcome con- 27, 29]. This discrepancy was the same with respect cerns among the GPs predicted increased incident PSA to the two ICU studies [19, 26]. Twostudies (from testing. PSA tests were obtained from register data. Israel and Denmark) that investigated referral rates to Kwah et al. [28] studied first-year internal medicine specialist care and hospitalization from primary care/ residents in a smaller one-year prospective study in the GPs found a significant association with physician United States. Surprisingly, there were fewer cumulative burnout [20, 24]. Two studies that failed to show an medical errors in residents with high burnout, measured association with burnout found an association be- as high EE or high DP scores. Other professionalism tween adverse events and measures of depression, measures such as adequate time to complete discharge which highlights that burnout and depression are dif- summaries and time to review laboratory tests results ferent concepts with respect to impaired functioning were not affected by burnout status. [22, 26]. Sun et al. [29] studied 36 primary care practices in In studies where burnout has been measured by con- Cleveland, Ohio, the United States. They found no asso- tinuous variables (Kushnir et al., Welp et al.) [19, 24]or ciation between either high burnout or low empathy with variables divided into more than three categories (Jefferson) and antibiotic prescribing for acute respira- (the two studies by Noroxe et al.) [20, 21], relationships tory infections. Burnout was measured continuously on to adverse care tended to be found. In the study by Sun all three subscales: EE, DP and PA. Prescription data et al. [29] no associations were found, despite a continu- were obtained from the patients’ medical records. ous burnout dimension measure. However, in this study, Noroxe et al. [20] conducted a large 6-month pro- only 36 patients were included. Zantinge et al. [23] spective study among Danish single-practice GPs and found discrepant results for the different dimensions of the number of hospitalisations of their patients. The burnout measured above a cut-off. Associations between number of hospitalisations for so-called ambulatory care PA and observed poorer communication skills were sensitive conditions (that could potentially be managed found despite a dichotomous measure of burnout. How- by the GP) increased with increased level of burnout in ever, this study was manually rated by observers, and the GP. All three burnout subscales, EE, DP and PA, hence more subject to individual assessments than were were independent predictors of hospitalisation, and so many of the other studies. was a composite burnout score of all subscales. The Mangory et al. BMC Health Services Research (2021) 21:369 Page 6 of 8 Discussion Many studies indicate that physicians with burnout be- A major finding in our review was that physician burn- lieve they provide insufficient patient care [2, 4]. Physi- out as measured by conventional use of MBI did not cians’ perceptions of their own abilities often do not lead to observed adverse patient care in 6 of 11 studies. correspond with observable evaluations of their work Four of the studies in the review found a clear link be- [36]. Self-criticism is common among students who tween burnout and observed adverse patient care, and choose to study medicine, and often reinforced by the one found a link between one of the burnout dimensions studies and the work situation [37]. This makes self- with respect to adverse patient outcomes, but not with reported data on own performance a less valid measure the two other dimensions. Among the four studies with of observed performance. We know that some levels of a clear association with adverse outcomes, two were re- self-criticism and even stress may drive performance, garding unnecessary referrals from GPs with burnout to and the level of burnout among physicians that will re- specialist investigation or care, one was regarding the sult in observable adverse patient care remains to be likelihood of the patient changing from a GP with burn- identified [10]. out to a new GP, and one was about an outcome as ser- Our review found more unnecessary referrals to spe- ious as increased patient mortality when doctors and/or cialist care among burned-out physicians, and one of the nurses reported burnout in an ICU. Danish studies found that the likelihood of patients Several other studies suggest that burnout is not a vali- changing their GP was higher in the case of GPs with dated measure with respect to observed impaired func- burnout [21]. This may indicate that burnout affects as- tioning, unlike depression and other psychiatric pects of the physician–patient relationship not captured conditions [10, 11]. In the original definition of burnout by the measures of quality of patient care most com- by Maslach, the levels of all three dimensions were re- monly used presently [38], such as communication skills quired to be over a certain cut-off to denote burnout [39]. Zantinge et al.’s observation study of Dutch GPs [30], but this does not necessarily associate with the cut- and their patient-doctor relationship indicates this since off levels of any clinical symptom that impairs function- the burnout doctors showed physical signs of more de- ing. In later studies, often only one or two dimensions tached contact with their patients [23]. There is a need have been used to define and measure burnout and for further validation of the dependent measures of out- some studies have used fewer or even just one item to comes that are in use, to be able to understand this in a define a dimension of burnout [31]. In our review, all more nuanced way. On the other hand, one may argue four studies that were associated with adverse patient that it is doubtful whether referral to specialist care or a outcomes used the full three-dimension Maslach ver- patient choosing another primary care physician actually sion, and they also utilized continuous variable data, or reduces observed quality of patient care. While it cer- data divided into several (more than three) categories at tainly increases costs of care and perhaps waiting time the expense of dichotomization and categorization into for the patients, a second opinion by another physician few categories. By using all the variance in the variables, may be good for the patients with respect to quality of there is a greater chance of capturing any effects on the care. There may also be indirect effects on patient care outcome. from physician burnout, such as physicians’ sick leave or More studies on this issue are thus needed. More recent early retirement and, thereby, lack of continuity of care. Dutch and Swedish measures and assessments of an What about the six studies of physician burnout “emotional exhaustion syndrome” define validated cut-off that showed no association with observed adverse pa- levels that are used for classification of a disease or dis- tient outcome? A relevant theory suggests that patient order that impairs functioning [11]. This can further the care is maintained at the expense of the physicians’ research on how burnout is related to patient care. well-being and mental health, which is known as the Some studies show a majority of participants simultan- Conservation of Resources theory [40]. The profession eously experience depressive symptoms and burnout might provide physicians with a work ethic that re- [22, 32]. Although depressive symptoms and burnout sults in a high threshold of patient care, regardless of share an appreciable amount of variance, the concepts the favourability of their working conditions. It is are different entities [33]. However, burnout in more ad- known that burnout among physicians can lead to a vanced stages has been found to lead to depressive corresponding increase in family conflicts, as the re- symptoms [34]. Depressive symptoms can influence sources devoted to their personal lives are rather well-being and thereby work ability, with increased risk spent in maintaining their quality of work [40, 41]. It of mistakes or suboptimal functioning, which can, in is likely that it is late in the process of burnout, only turn, increase the risk of development of burnout [35]. when personal and professional resources are de- The relationships between burnout, depressive symp- pleted, that observable patient treatment will be ad- toms and impaired functioning needs further study. versely affected [4]. Mangory et al. BMC Health Services Research (2021) 21:369 Page 7 of 8 Nevertheless, a study in our review found an alarming Authors’ contributions RT, KYM and LYA conceived of and designed the study. KYM and LYA effect of burnout on standardized mortality in Swiss searched the databases and did a manual review of the literature. RT and KIR ICUs [19]. The study investigated both physicians and provided the articles from other sources, and took part in the final selection nurses in these units (N = 1425), and the predictor of articles. All authors read and approved the final manuscript. model (where merely emotional exhaustion was signifi- Authors’ information cant) contributed with only 10% of the explained vari- KYM: Medical Student, Faculty of Medicine, University of Oslo, Norway. ance in the regression towards mortality. Other LYA: Medical Student, Faculty of Medicine, University of Oslo, Norway. KIR: M.D. PhD, MHA. Director for LEFO - Institute for Studies of the Medical important variables such as patient characteristics and Profession, Norway. various work-related psychosocial factors were not in- RT: Professor M.D. PhD. Department of Behavioural Medicine, Institute of cluded in the predictor model. Basic Medical Sciences, Faculty of Medicine, University of Oslo, Norway. The strengths of this study are the aggregation of a Funding large number of participants from the included studies Not applicable. and diversity of countries where the studies have been performed. Our search was limited to studies published Availability of data and materials since 2007, and among the 11 studies included in this Data sharing is not applicable to this article as no datasets were generated or analysed during the current study. review, two encompassed both physicians and nurses in the sample [19, 26]. By only including studies with ob- Declarations servable measures for adverse patient treatment, we did not consider patient satisfaction, though it is an import- Ethics approval and consent to participate Not applicable. ant indicator of quality in modern medicine. One limita- tion is that we did not include other burnout measures Consent for publication such as, e.g., OLBI and CBI. The time limits for the re- Not applicable. view could have been broader, since this literature is ra- Competing interests ther limited, and we may have missed some studies The authors declare that they have no conflicting interests. before 2007. The lack of studies with objective measures of quality of care might also be due to publication bias Author details 1 2 Faculty of Medicine, University of Oslo, Oslo, Norway. Institute for Studies because negative outcome studies may not have been of the Medical Profession, Oslo, Norway. Department of Behavioural published. Medicine, Institute of Basic Medical Sciences, Faculty of Medicine, University Future research should investigate possible links be- of Oslo, Oslo, Norway. tween burnout and impaired quality of patient care. Received: 6 April 2020 Accepted: 9 April 2021 Studies of this kind can be a challenge to design and costly to perform. Some might even be questionable from an ethical point of view, for example designing a References 1. Wallace JE, Lemaire JB, Ghali WA. Physician wellness: a missing quality randomized controlled trial with an expectation that one indicator. Lancet. 2009;374:1714–21. of the groups will deliver less sufficient medical treat- 2. Panagioti M, Geraghty K, Johnson J, Zhou A, Panagopoulou E, Chew- ment. We also need more non-Western studies on this Graham C, et al. Association between physician burnout and patient safety, professionalism, and patient satisfaction: a systematic review and meta- issue. analysis. JAMA Intern Med. 2018;178:1317–30. 3. West CP, Dyrbye LN, Shanafelt TD. Physician burnout: contributors, consequences and solutions. J Intern Med. 2018;283:516–29. Conclusion 4. Rathert C, Williams ES, Linhart H. Evidence for the quadruple aim: a In our review, only four, and in part a fifth, of 11 studies systematic review of the literature on physician burnout and patient found a relationship between physician burnout mea- outcomes. Med Care. 2018;56:976–84. 5. Angerer P, Weigl M. Physicians' psychosocial work conditions and quality of sured by MBI and observed adverse patient outcomes. care: a literature review. Prof Professionalism. 2015;5. https://doi.org/10.7577/ More studies should be performed to further under- pp.960. stand whether burnout among physicians leads to low- 6. Rotenstein LS, Torre M, Ramos MA, Rosales RC, Guille C, Sen S, et al. Prevalence of burnout among physicians: a systematic review. JAMA. 2018; ered patient care, and in particular to validate the 320:1131–50. specific level of burnout that impairs physician’ function- 7. Poghosyan L, Aiken LH, Sloane DM. Factor structure of the Maslach burnout ing with respect to observed adverse patient outcomes. inventory: an analysis of data from large scale cross-sectional surveys of nurses from eight countries. Int J Nurs Stud. 2009;46:894–902. Abbreviations 8. Maslach C, Jackson S, Leiter M. The Maslach burnout inventory manual. In: MBI: Maslach Burnout Inventory; DP: Depersonalization; EE: Emotional Zalaquett CP, Wood RJ, editors. Evaluating stress: a book of resources. USA: exhaustion; PA: Personal accomplishment (reduced); OLBI: Oldenburger The Scarecrow Press; 1997. Burnout Inventory; CBI: Copenhagen Burnout Inventory; GP: General 9. Ahola K, Hakanen J, Perhoniemi R, Mutanen P. Relationship between practitioner; ICU: Intensive care unit burnout and depressive symptoms: a study using the person-centred approach. Burn Res. 2014;1.https://doi.org/10.1016/j.burn.2014.03.003. Acknowledgements 10. Tyssen R. What is the level of burnout that impairs functioning? J Intern The final manuscript was reviewed by OnLine English. Med. 2018;283:594–6. Mangory et al. BMC Health Services Research (2021) 21:369 Page 8 of 8 11. Maslach C, Leiter MP. Understanding the burnout experience: recent research 35. West CP, Huschka MM, Novotny PJ, Sloan JA, Kolars JC, Habermann TM, and its implications for psychiatry. World Psychiatry. 2016;15:103–11. et al. Association of perceived medical errors with resident distress and 12. Schwenk TL, Gold KJ. Physician burnout-a serious symptom, but of what? empathy: a prospective longitudinal study. JAMA. 2006;296:1071–8. JAMA. 2018;320:1109–10. 36. Davis DA, Mazmanian PE, Fordis M, Van Harrison R, Thorpe KE, Perrier L. Accuracy of physician self-assessment compared with observed measures of 13. Demerouti E, Bakker AB, Nachreiner F, Schaufeli WB. The job demands- competence: a systematic review. JAMA. 2006;296:1094–102. resources model of burnout. J Appl Psychol. 2001;86:499–512. 37. Gabbard GO. The role of compulsiveness in the Normal physician. JAMA. 14. Crawford E, Lepine J, Rich B. Linking job demands and resources to 1985;254:2926–9. employee engagement and burnout: a theoretical extension and meta- 38. Linzer M, Manwell LB, Williams ES, Bobula JA, Brown R, Varkey AB, et al. analytic test. J Appl Psychol. 2010;95:834–48. Working conditions in primary care: physician reactions and care quality. 15. Parsa-Parsi RW. The revised declaration of Geneva: a modern-day physician's Ann Intern Med. 2009;151:28–36. pledge. JAMA. 2017;318:1971–2. 39. Passalacqua SA, Segrin C. The effect of resident physician stress, burnout, 16. Van Mol MM, Kompanje EJ, Benoit DD, Bakker J, Nijkamp MD. The and empathy on patient-centered communication during the long-call shift. prevalence of compassion fatigue and burnout among healthcare Health Commun. 2012;27:449–56. professionals in intensive care units: a systematic review. PLoS One. 2015;10: 40. Hobfoll S. Conservation of resources: a new attempt at conceptualizing e0136955. stress. Am Psychol. 1989;44:513–24. 17. Zarei E, Ahmadi F, Sial MS, Hwang J, Thu PA, Usman SM. Prevalence of 41. Shanafelt TD, Hasan O, Dyrbye LN, Sinsky C, Satele D, Sloan J, et al. Changes burnout among primary health care staff and its predictors: a study in Iran. in burnout and satisfaction with work-life balance in physicians and the Int J Environ Res Public Health. 2019;16:2249. general US working population between 2011 and 2014. Mayo Clin Proc. 18. Asante JO, Li MJ, Liao J, Huang YX, Hao YT. The relationship between 2015;90:1600–13. psychosocial risk factors, burnout and quality of life among primary healthcare workers in rural Guangdong province: a cross-sectional study. BMC Health Serv Res. 2019;19:447. Publisher’sNote 19. Welp A, Meier LL, Manser T. Emotional exhaustion and workload predict Springer Nature remains neutral with regard to jurisdictional claims in clinician-rated and objective patient safety. Front Psychol. 2014;5:1573. published maps and institutional affiliations. 20. Noroxe KB, Pedersen AF, Carlsen AH, Bro F, Vedsted P. Mental well-being, job satisfaction and self-rated workability in general practitioners and hospitalisations for ambulatory care sensitive conditions among listed patients: a cohort study combining survey data on GPs and register data on patients. BMJ Qual Saf. 2019;28:997–1006. 21. Noroxe KB, Vedsted P, Bro F, Carlsen AH, Pedersen AF. Mental well-being and job satisfaction in general practitioners in Denmark and their patients' change of general practitioner: a cohort study combining survey data and register data. BMJ Open. 2019;9:e030142. 22. Fahrenkopf AM, Sectish TC, Barger LK, Sharek PJ, Lewin D, Chiang VW, et al. Rates of medication errors among depressed and burnt out residents: prospective cohort study. BMJ. 2008;336:488–91. 23. Zantinge EM, Verhaak PFM, de Bakker DH, van der Meer K, Bensing JM. Does burnout among doctors affect their involvement in patients' mental health problems? A study of videotaped consultations. BMC Fam Pract. 2009;10:60. 24. Kushnir T, Greenberg D, Madjar N, Hadari I, Yermiahu Y, Bachner YG. Is burnout associated with referral rates among primary care physicians in community clinics? Fam Pract. 2014;31:44–50. 25. Yuguero Torres O, Esquerda Aresté M, Marsal Mora JR, Soler-González J. Association between sick leave prescribing practices and physician burnout and empathy. PLoS One. 2015;10:e0133379. 26. Garrouste-Orgeas M, Perrin M, Soufir L, Vesin A, Blot F, Maxime V, et al. The Iatroref study: medical errors are associated with symptoms of depression in ICU staff but not burnout or safety culture. Intensive Care Med. 2015;41: 273–84. 27. Pedersen AF, Carlsen AH, Vedsted P. Association of GPs' risk attitudes, level of empathy, and burnout status with PSA testing in primary care. Br J Gen Pract. 2015;65:e845–51. 28. Kwah J, Weintraub J, Fallar R, Ripp J. The effect of burnout on medical errors and professionalism in first-year internal medicine residents. J Grad Med Educ. 2016;8:597–600. 29. Sun BZ, Chaitoff A, Hu B, Neuendorf K, Manne M, Rothberg MB. Empathy, burnout, and antibiotic prescribing for acute respiratory infections: a cross- sectional primary care study in the US. Br J Gen Pract. 2017;67:e565–71. 30. Maslach C, Jackson SE. The measurement of experienced burnout. J Organ Behav. 1981;2:99–113. 31. Dyrbye LN, West CP, Shanafelt TD. Defining burnout as a dichotomous variable. J Gen Intern Med. 2009;24:440 author reply 441. 32. Ratanawongsa N, Roter D, Beach MC, Laird SL, Larson SM, Carson KA, et al. Physician burnout and patient-physician communication during primary care encounters. J Gen Intern Med. 2008;23:1581–8. 33. Glass DC, McKnight JD. Perceived control, depressive symptomatology, and professional burnout: a review of the evidence. Psychol Health. 1996;11:23–48. 34. Glass DC, McKnight JD, Valdimarsdottir H. Depression, burnout, and perceptions of control in hospital nurses. J Consult Clin Psychol. 1993;61: 147–55. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png BMC Health Services Research Springer Journals

Effect of burnout among physicians on observed adverse patient outcomes: a literature review

Loading next page...
 
/lp/springer-journals/effect-of-burnout-among-physicians-on-observed-adverse-patient-G2KycfkKQd

References (44)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
eISSN
1472-6963
DOI
10.1186/s12913-021-06371-x
Publisher site
See Article on Publisher Site

Abstract

Background: Physician burnout has potentially harmful effects for both physicians and their patients. Despite relationships between physician burnout and lowered patient satisfaction and clinician-rated adverse patient outcomes, there is scarce literature regarding effects on objective patient outcomes. This study aimed to examine the relationship between physician burnout and observed adverse patient outcomes via a review of the literature. Methods: A search was performed on the MEDLINE, EMBASE and PsychINFO databases, using keywords and Medical Subject Headings. The identified studies were in English, published from 2007 to 2019, measured burnout among physicians using the Maslach Burnout Inventory (MBI), and included observed adverse patient outcomes. In total, 360 eligible articles were identified, and 11 were included in the final review. All included studies measured patient outcomes by observed clinical measures (e.g. quality of care and medical errors). Results: Four studies found a clear significant relationship between physician burnout and observed adverse patient outcomes, while 6 did not. One study found a significant relationship with one of the MBI subscales. Burnout was, in contrast to depression, only partly associated with observed patient outcomes. Conclusions: This review illustrates the need for a validation of physician burnout measured by MBI with respect to observed patient outcomes. Further studies are required to investigate the effects of physician burnout on observed quality of their patient care. Keywords: Physician burnout, Patient outcome, Medical error Background from 2015 distinguished between different types of nega- Burnout among physicians is a widespread phenomenon tive treatment outcomes, and described one category of and has been found to be of importance for both physi- more subjectively perceived outcomes, and two categor- cians and their patients [1]. Recent reviews suggest a link ies of objectively measured outcomes [5]: (I) The sub- between physician burnout and negative impact on pa- jectively perceived outcomes were the physicians’ and/or tient care, including increased medical errors [2–4]. A patients’ perception of quality of care: These were out- review study on psychosocial work stress and burnout comes related to patient centeredness of clinical activ- ities and the physician–patient relationship as well as to physician empathy, accessibility, reliability and exchange * Correspondence: reidar.tyssen@medisin.uio.no of information. (II) One category of objectively measured The first two authors (Mangory and Ali) share the first authorship. Kashan Yasin Mangory and Lavin Yadgar Ali contributed equally to this outcomes included assessments of treatment success work. and absence of complications. These outcomes were Department of Behavioural Medicine, Institute of Basic Medical Sciences, evaluated by different observed indicators of treatment Faculty of Medicine, University of Oslo, Oslo, Norway Full list of author information is available at the end of the article © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Mangory et al. BMC Health Services Research (2021) 21:369 Page 2 of 8 success (such as diabetes or blood pressure control), performance [10, 11]. In an editorial from 2018 titled clinical quality or lack of treatment complications. (III) “Physician burnout—a serious symptom, but of what?” Another category of objectively measured outcomes was Schwenk and Gold discuss the high but varying preva- prevalence of medical errors: These outcomes included lence of burnout reported in many studies and they observed or recorded prescriptions, drug administration strongly emphasize the need for a deeper understanding or surgical errors. of its origins and possible consequences, which have not Most of the reviews so far build mainly on the subject- been sufficiently explained [12]. It is therefore important ively perceived data, from physicians and patients them- to relate burnout levels both to the physician’s individual selves. There have been surprisingly few reviews of suffering, but also to impact of burnout in the clinical observational studies on the relationship between phys- setting and for patient treatment. ician burnout and objectively measured outcomes of The relationships between job demands, burnout, work sub-optimal care of patients or errors [4]. We therefore engagement, physician impairment and quality of patient need to study whether physician burnout is actually as- care are complex. High job demands have been found to sociated with observed adverse patient care and if it in- relate to emotional exhaustion, whereas lack of resources creases the risk of medical errors. has been related to disengagement [13]. However, the doc- The distinction between objectively observed out- tor’s perception of the type of demand can influence how comes and subjective perception of quality of care is it impacts burnout. Demands perceived as challenges were crucial for patients, physicians and their employers. We found to be positively associated with work engagement, have therefore reviewed relevant observational studies to while demands perceived as hindrances could increase provide insights into this issue. burnout. Engagement in work tasks could be developed Burnout is usually measured using the Maslach Burn- despite a certain level of exhaustion [14], and could thus out Inventory (MBI). In a review of burnout prevalence, mitigate consequences of burnout. Additionally, there is a covering the period from 1991 to 2018, including 185 strong moral imperative in the profession to “put patients studies from 45 countries, 86% of the studies had used a first”, articulated in the Declaration of Geneva’sPhysi- version of the MBI [6]. The MBI has been found to have cians´ Pledge [15], and this can urge doctors to preserve a replicable factor structure within and between profes- the quality of care even during heavy workload periods for sional groups and across countries [7]. Although there the individual physicians. are several other valid measures of burnout (such as In addition, a stressful working climate in some hos- OLBI—Oldenburger Burnout Inventory, Burnout Inven- pital units may lead to burnout also among health pro- tory, CBI—Copenhagen Burnout Inventory), we have, on fessionals other than physicians [16]. This applies also to this background, limited the present review to include primary health care settings in non-Western countries, only studies using the MBI. The MBI has, however, been as shown in recent studies [17, 18]. Burnout in whole modified in several places in order to accommodate a hospital units and departments may also impact quality variety of professionals and it has been shortened to in- of patient care [5]. clude fewer items per dimension, or even only a single Several studies have shown that burned-out physicians item [3]. themselves report that their condition affects their pa- The MBI includes three main components or subscales: tient treatment and care negatively [2, 4]. A recent emotional exhaustion (EE), which has been thought to be meta-analysis on the link between physician burnout the initial reaction, followed by depersonalization/disen- and patient safety found that “physician burnout may gagement/cynicism (DP) which represents an emotionally jeopardize patient care”, but that most studies were detached attitude towards work (understood as a way of cross-sectional and relied on outcomes self-reported by coping with exhaustion) and, finally, an experience of re- physicians [2]. The authors “failed to show significant duced personal accomplishment (or performance) (PA) links between physician burnout and patient safety out- [8]. Both cynicism and reduced accomplishment could in- comes recorded in the health-care systems (e.g., the dicate impaired physician functioning and subsequent health records of patients, monitoring etc.).” In line with subordinate patient care. this, another recent review found no significant link be- However, critique of this burnout model points to un- tween physician burnout and observed clinical patient certainties about these possible relationships. Firstly, the outcomes in any of the five studies identified; however, MBI is not a diagnostic measure. Maslach surveyed per- there were associations between physician burnout and sons “at work”, defining the third with highest score as outcomes reported by physicians and patients [4]. “high burnout”. This has later been used as a cut-off In this literature review, the definition of “adverse pa- score. Those who were off sick or not able to be at work tient outcomes” includes observed measures for failure were excluded from definition [9]. The scale is thereby or lack of prevention, inadequate assessment (diagnoses) poorly validated with respect to impairment and work and treatment. Mangory et al. BMC Health Services Research (2021) 21:369 Page 3 of 8 This literature review therefore aims to identify studies Two of the studies included both nurses and physicians examining the association between physician burnout, in intensive care units, which are considered stressful measured by the MBI, and observed quality of care workplaces [16]. We included these studies because they among their patients (including number of medical comprised substantial samples of physicians. errors). Results Methods Eleven studies were included in the review (Table 1). We used the search engines MEDLINE, EMBASE, and There was a large variation in the size of the samples, in PsychINFO. The final search consisted of keywords and the observed outcomes reported in the studies and in Medical Subject Headings for predictor (e.g., burnout), how MBI burnout was measured and defined (i.e. sample (e.g., physician) and potential outcomes (e.g., whether the subscales were dichotomized, categorized or medical errors) using the following three groups of used as continuous variables); see Table 1 footnotes. search terms: Seven studies were of general practitioners (GPs) or from primary care settings, two studies were from inten- 1) [Burnout, Professional/]OR [Physician burnout] sive care units, one study was among paediatric resi- 2) [Practice Patterns, Physicians/]OR [Physicians, dents, and one study was about first-year residents in Primary Care/]OR [Physicians, Family/]OR internal medicine. Three of the studies were from the [Physicians/]OR [Primary Health Care/]OR United States, three from Denmark, and one study each [Internship and Residency/] from the Netherlands, Israel, Spain, France and 3) [Quality of Health Care/]OR [Medical Errors/]OR Switzerland. [Medication Errors/]OR [Physician-Patient Fahrenkopf et al. [22] studied cohorts of paediatric res- Relations/]OR [Patient Care/]OR [Referral and idents from three U.S. hospitals over one and a half Consultation/]OR [Attitude of Health Personnel/ month. Burnout was measured by EE and DP in, and ad- ]OR [Patient Outcome] verse patient outcome was the rate of errors made in or- dering medication. Residents with high combined score The search combined “OR” between the search terms for EE (> 27) and DP (> 10) were found to make similar in each group with “AND” between the groups, resulting rates of errors per resident month as residents with low in 357 articles. In addition, experts in the field (KIR and burnout scores. There was a significant association be- RT) provided three articles [19–21]. The articles were tween depression and medical errors (Harvard national subject to a manual review by two separate authors depression screening day scale); depressed residents (KYM and LYA). Inclusion criteria were as follows: (I) made 6.2 times more errors than non-depressed resi- the papers had abstracts written in English; (II) they dents. The medical errors were collected and rated by were published between 2007 and 2019; (III) they mea- trained observers (nurses and physicians). sured physician burnout with MBI as a predictor; and Zantinge et al. [23] observed videotaped consultations in (IV) they included observed measures of patient out- Dutch GPs. The findings were somewhat discrepant: GPs comes. The original search was for the student authors’ with low PA communicated less affectively, were less master thesis and had a 10-year frame (2007–2017). For patient-centred and had less eye contact, whereas those the preparation of this manuscript, we updated the with high EE and DP talked more about psychosocial search to include the past 2 years, resulting in the range problems in their consultations. The latter may have in- 2007–2019. The last year – 2020 – is heavily influenced creased possibility for adequate mental health care. by the COVID-19 pandemic at hospitals and other Kushnir et al. [24] found in a study from Israel among health services, and is not comparable to previous years primary care physicians that referral rates increased for with respect to possible confounders, such as, e.g., stress both diagnostic tests and to specialist clinics among phy- outside of work. First, we excluded articles that did not sicians who scored high on burnout. Being a specialist in have an abstract or had an irrelevant title and abstract. family medicine was more important than burnout for The remaining 29 articles were then read in full text by referring to a specialist clinic. Burnout was here mea- the two first authors (KYM and LYA) and excluded if sured as a common mean of items of EE and DP. they did not include objective measures of patient out- Yuguero Torres et al. [25] studied doctors from 22 pri- comes or did not investigate the association between mary care centres in a region in Spain (Catalonia). The physician burnout, measured by the MBI, and adverse doctors’ prescribing of sick leave was not associated with patient outcomes. The senior authors (KIR and RT) higher burnout among them. Burnout scores were di- reviewed the articles and took actively part in the last se- vided into low, moderate and high on EE, DP and PA. lection and writing. Finally, 11 articles seemed to meet There was an association between higher burnout and the inclusion criteria and were included in this review. lower empathy (Jefferson), but this did not influence the Mangory et al. BMC Health Services Research (2021) 21:369 Page 4 of 8 Table 1 Description of studies retained by the literature review Study Country Design Sample N Response Measured Observed measure on Association with burnout rate dimensions quality of care of burnout Fahrenkopf USA Observational Paediatric 246 50% EE and DP Missing or wrong Not associated with adverse et al. 2008 prospective residents prescription of drugs patient treatment [22] cohort study Zantinge Netherlands Observation GPs 142 89% EE, DP and Length of consultations, level PA associated with GPs et al. 2009 of video- PA of verbal communication, communicating less affectively, [23] recorded con- eye contact, and focus on being less patient-centred and less sultations and psychosocial issues eye contact. EE and DP not associ- questionnaire ated with adverse patient treatment Kushnir Israel Cross- GPs/Primary 136 99% EE, DP and Number of referrals for Associated with more referrals et al. 2014 sectional care PA diagnostic imaging, [24] observational specialized health services or study nurse sensitive treatments Yuguero Spain Prospective GPs 217 50% EE, DP and Number of prescribed sick Not associated with prescribing Torres et al. observational PA leaves more or longer sick leaves 2015 [25] study (1 year) Garrouste- France Prospective Doctors, 1988 77% EE, DP and Medical errors (i.e., error of Not associated with adverse Orgeas observational nurses and (330 PA execution or error of patient treatment et al. 2015 study (2-year) care doctors) planning). Adverse events [26] workers in were patient harms caused intensive by medical interventions. care unit Welp et al. Switzerland Observational Doctors 1425 Not EE, DP and Length of stay in hospital Associated with increased 2015 [19] study and nurses (243 specified PA and standardized mortality standardized mortality ratio, but in intensive doctors) ratio not length of stay care unit Pedersen Denmark Register study GPs 835 72% EE, DP and Number of requisitions for Not associated with increased et al. 2015 and PA PSA among healthy male requisitions for PSA [27] questionnaire patients Kwah et al. USA Prospective First-year 54 98% EE and DP Medication prescription Not associated with increase in 2016 [28] observational residents in (initial), errors with potential for medical errors study (1 year) internal 59% adverse drug effects medicine (cohort) Sun et al. USA Cross- Primary 102 Not EE, DP and Antibiotic prescriptions for Not associated with increase in 2017 [29] sectional care specified PA acute respiratory infections prescriptions observational study Noroxe Denmark Prospective GPs 781 50.2% EE, DP and Conditions not requiring Associated with increased et al. 2019 observational PA hospitalization in the case of frequency of hospitalizations of [20] study (6 appropriate intervention in ambulatory care sensitive months) primary care (ambulatory conditions care sensitive conditions) Noroxe Denmark Prospective GPs 409 50.2% EE, DP and Change of GP (unrelated to Associated with increased et al. 2019 observational PA change of address) likelihood of changing GPs [21] study (6 months) Abbreviations: EE emotional exhaustion, DP depersonalization, PA personal accomplishment (reduced), GP general practitioner, PSA prostate-specific antigen, MBI Maslach Burnout Inventory Defined burnout as EE > 27 and DP > 10 EE, DP and PA were dichotomized into low and high scores, and outcomes reported for each dimension Overall burnout was measured as an average of the responses to items EE and DP (continuous variable) Scores for each category were divided into low, moderate and high Burnout measured as (1) a combination of high EE and DP with low PA (dimensions dichotomized), and (2) defined as over a cut-off of a global MBI score of 9 Each dimension used as a continuous variable Burnout defined as high level of emotional exhaustion > 26 Burnout defined as a high EE or high DP subscore (dichotomized variables) Burnout measured 1) in four quartiles for each dimension 2) As a composite score by adding up points corresponding to the quartile of each subscale (reversed score for personal accomplishment); one point for scores in the first quartile, and two, three and four points for subsequent quartiles, respectively. The composite score was categorized into five groups, with increasing score indicating higher burnout Mangory et al. BMC Health Services Research (2021) 21:369 Page 5 of 8 number or duration of sick leave prescriptions among number of hospitalisations was obtained from national the doctors. register data. Garrouste-Orgeas et al. [26] conducted a large and im- Noroxe et al. [21] did a second study of the same eli- pressive 2-year prospective, observational study of doc- gible sample of Danish single-practice GPs, as men- tors (N = 330), nurses and care workers in 31 Intensive tioned above. Here the outcome was change of GP Care Units (ICUs) in France. There was no significant among the patients (unrelated to their change of ad- association between medical errors or adverse events dress). There was a dose-response like association be- and burnout measured either as (1) a combination of tween increase in DP and PA subscales (but not in EE) high EE and DP with low PA, or (2) defined as over a among the GPs and the likelihood that their patients left cut-off on a global MBI score. There was a significant them for a new GP. There was also a significant associ- association between depression symptoms (CES-Depres- ation with an increase in the composite burnout score. sion Scale) and medical errors. Medical errors were de- Data regarding change of GP were obtained from a na- fined by Delphi technique among 60 experts. tional register. Welp et al. [19] studied patient safety in ICUs in 48 Swiss hospitals. They assessed standardized mortality ra- Summary of the findings tios (SMR) and length of stay, and related these mea- Of the11studies,only4[19–21, 24] found a clear sures to burnout among medical doctors (N = 243) and significant association between physician burnout and nursing staff. The three burnout subscales were mea- adverse patient outcomes, whereas 6 [22, 25–29] sured continuously, with EE predicting a significant in- failed to find such an association. One study had di- crease in SMR, whereas DP and PA did not. Burnout did vergent findings, with some aspects of patient care not predict length of stay for patients in ICU. (observed communication skills) being reduced, Pedersen et al. [27] studied the association between whereas other aspects (time devoted to psychosocial burnout and other psychological measures in Danish issues) increased among burnout doctors [23]. There solo GP practices and testing of prostate specific antigen was no convincing association related to either posi- (PSA) in healthy male patients. Burnout measured as tions or practice venue. Among the 7 studies from high EE (> 26) was not associated with increased PSA primary care/general practice, 3 found a clear signifi- testing, and neither was empathy (Jefferson). On the cant association [20, 21, 24], while 4 did not [23, 25, other hand, high anxiety scores and bad outcome con- 27, 29]. This discrepancy was the same with respect cerns among the GPs predicted increased incident PSA to the two ICU studies [19, 26]. Twostudies (from testing. PSA tests were obtained from register data. Israel and Denmark) that investigated referral rates to Kwah et al. [28] studied first-year internal medicine specialist care and hospitalization from primary care/ residents in a smaller one-year prospective study in the GPs found a significant association with physician United States. Surprisingly, there were fewer cumulative burnout [20, 24]. Two studies that failed to show an medical errors in residents with high burnout, measured association with burnout found an association be- as high EE or high DP scores. Other professionalism tween adverse events and measures of depression, measures such as adequate time to complete discharge which highlights that burnout and depression are dif- summaries and time to review laboratory tests results ferent concepts with respect to impaired functioning were not affected by burnout status. [22, 26]. Sun et al. [29] studied 36 primary care practices in In studies where burnout has been measured by con- Cleveland, Ohio, the United States. They found no asso- tinuous variables (Kushnir et al., Welp et al.) [19, 24]or ciation between either high burnout or low empathy with variables divided into more than three categories (Jefferson) and antibiotic prescribing for acute respira- (the two studies by Noroxe et al.) [20, 21], relationships tory infections. Burnout was measured continuously on to adverse care tended to be found. In the study by Sun all three subscales: EE, DP and PA. Prescription data et al. [29] no associations were found, despite a continu- were obtained from the patients’ medical records. ous burnout dimension measure. However, in this study, Noroxe et al. [20] conducted a large 6-month pro- only 36 patients were included. Zantinge et al. [23] spective study among Danish single-practice GPs and found discrepant results for the different dimensions of the number of hospitalisations of their patients. The burnout measured above a cut-off. Associations between number of hospitalisations for so-called ambulatory care PA and observed poorer communication skills were sensitive conditions (that could potentially be managed found despite a dichotomous measure of burnout. How- by the GP) increased with increased level of burnout in ever, this study was manually rated by observers, and the GP. All three burnout subscales, EE, DP and PA, hence more subject to individual assessments than were were independent predictors of hospitalisation, and so many of the other studies. was a composite burnout score of all subscales. The Mangory et al. BMC Health Services Research (2021) 21:369 Page 6 of 8 Discussion Many studies indicate that physicians with burnout be- A major finding in our review was that physician burn- lieve they provide insufficient patient care [2, 4]. Physi- out as measured by conventional use of MBI did not cians’ perceptions of their own abilities often do not lead to observed adverse patient care in 6 of 11 studies. correspond with observable evaluations of their work Four of the studies in the review found a clear link be- [36]. Self-criticism is common among students who tween burnout and observed adverse patient care, and choose to study medicine, and often reinforced by the one found a link between one of the burnout dimensions studies and the work situation [37]. This makes self- with respect to adverse patient outcomes, but not with reported data on own performance a less valid measure the two other dimensions. Among the four studies with of observed performance. We know that some levels of a clear association with adverse outcomes, two were re- self-criticism and even stress may drive performance, garding unnecessary referrals from GPs with burnout to and the level of burnout among physicians that will re- specialist investigation or care, one was regarding the sult in observable adverse patient care remains to be likelihood of the patient changing from a GP with burn- identified [10]. out to a new GP, and one was about an outcome as ser- Our review found more unnecessary referrals to spe- ious as increased patient mortality when doctors and/or cialist care among burned-out physicians, and one of the nurses reported burnout in an ICU. Danish studies found that the likelihood of patients Several other studies suggest that burnout is not a vali- changing their GP was higher in the case of GPs with dated measure with respect to observed impaired func- burnout [21]. This may indicate that burnout affects as- tioning, unlike depression and other psychiatric pects of the physician–patient relationship not captured conditions [10, 11]. In the original definition of burnout by the measures of quality of patient care most com- by Maslach, the levels of all three dimensions were re- monly used presently [38], such as communication skills quired to be over a certain cut-off to denote burnout [39]. Zantinge et al.’s observation study of Dutch GPs [30], but this does not necessarily associate with the cut- and their patient-doctor relationship indicates this since off levels of any clinical symptom that impairs function- the burnout doctors showed physical signs of more de- ing. In later studies, often only one or two dimensions tached contact with their patients [23]. There is a need have been used to define and measure burnout and for further validation of the dependent measures of out- some studies have used fewer or even just one item to comes that are in use, to be able to understand this in a define a dimension of burnout [31]. In our review, all more nuanced way. On the other hand, one may argue four studies that were associated with adverse patient that it is doubtful whether referral to specialist care or a outcomes used the full three-dimension Maslach ver- patient choosing another primary care physician actually sion, and they also utilized continuous variable data, or reduces observed quality of patient care. While it cer- data divided into several (more than three) categories at tainly increases costs of care and perhaps waiting time the expense of dichotomization and categorization into for the patients, a second opinion by another physician few categories. By using all the variance in the variables, may be good for the patients with respect to quality of there is a greater chance of capturing any effects on the care. There may also be indirect effects on patient care outcome. from physician burnout, such as physicians’ sick leave or More studies on this issue are thus needed. More recent early retirement and, thereby, lack of continuity of care. Dutch and Swedish measures and assessments of an What about the six studies of physician burnout “emotional exhaustion syndrome” define validated cut-off that showed no association with observed adverse pa- levels that are used for classification of a disease or dis- tient outcome? A relevant theory suggests that patient order that impairs functioning [11]. This can further the care is maintained at the expense of the physicians’ research on how burnout is related to patient care. well-being and mental health, which is known as the Some studies show a majority of participants simultan- Conservation of Resources theory [40]. The profession eously experience depressive symptoms and burnout might provide physicians with a work ethic that re- [22, 32]. Although depressive symptoms and burnout sults in a high threshold of patient care, regardless of share an appreciable amount of variance, the concepts the favourability of their working conditions. It is are different entities [33]. However, burnout in more ad- known that burnout among physicians can lead to a vanced stages has been found to lead to depressive corresponding increase in family conflicts, as the re- symptoms [34]. Depressive symptoms can influence sources devoted to their personal lives are rather well-being and thereby work ability, with increased risk spent in maintaining their quality of work [40, 41]. It of mistakes or suboptimal functioning, which can, in is likely that it is late in the process of burnout, only turn, increase the risk of development of burnout [35]. when personal and professional resources are de- The relationships between burnout, depressive symp- pleted, that observable patient treatment will be ad- toms and impaired functioning needs further study. versely affected [4]. Mangory et al. BMC Health Services Research (2021) 21:369 Page 7 of 8 Nevertheless, a study in our review found an alarming Authors’ contributions RT, KYM and LYA conceived of and designed the study. KYM and LYA effect of burnout on standardized mortality in Swiss searched the databases and did a manual review of the literature. RT and KIR ICUs [19]. The study investigated both physicians and provided the articles from other sources, and took part in the final selection nurses in these units (N = 1425), and the predictor of articles. All authors read and approved the final manuscript. model (where merely emotional exhaustion was signifi- Authors’ information cant) contributed with only 10% of the explained vari- KYM: Medical Student, Faculty of Medicine, University of Oslo, Norway. ance in the regression towards mortality. Other LYA: Medical Student, Faculty of Medicine, University of Oslo, Norway. KIR: M.D. PhD, MHA. Director for LEFO - Institute for Studies of the Medical important variables such as patient characteristics and Profession, Norway. various work-related psychosocial factors were not in- RT: Professor M.D. PhD. Department of Behavioural Medicine, Institute of cluded in the predictor model. Basic Medical Sciences, Faculty of Medicine, University of Oslo, Norway. The strengths of this study are the aggregation of a Funding large number of participants from the included studies Not applicable. and diversity of countries where the studies have been performed. Our search was limited to studies published Availability of data and materials since 2007, and among the 11 studies included in this Data sharing is not applicable to this article as no datasets were generated or analysed during the current study. review, two encompassed both physicians and nurses in the sample [19, 26]. By only including studies with ob- Declarations servable measures for adverse patient treatment, we did not consider patient satisfaction, though it is an import- Ethics approval and consent to participate Not applicable. ant indicator of quality in modern medicine. One limita- tion is that we did not include other burnout measures Consent for publication such as, e.g., OLBI and CBI. The time limits for the re- Not applicable. view could have been broader, since this literature is ra- Competing interests ther limited, and we may have missed some studies The authors declare that they have no conflicting interests. before 2007. The lack of studies with objective measures of quality of care might also be due to publication bias Author details 1 2 Faculty of Medicine, University of Oslo, Oslo, Norway. Institute for Studies because negative outcome studies may not have been of the Medical Profession, Oslo, Norway. Department of Behavioural published. Medicine, Institute of Basic Medical Sciences, Faculty of Medicine, University Future research should investigate possible links be- of Oslo, Oslo, Norway. tween burnout and impaired quality of patient care. Received: 6 April 2020 Accepted: 9 April 2021 Studies of this kind can be a challenge to design and costly to perform. Some might even be questionable from an ethical point of view, for example designing a References 1. Wallace JE, Lemaire JB, Ghali WA. Physician wellness: a missing quality randomized controlled trial with an expectation that one indicator. Lancet. 2009;374:1714–21. of the groups will deliver less sufficient medical treat- 2. Panagioti M, Geraghty K, Johnson J, Zhou A, Panagopoulou E, Chew- ment. We also need more non-Western studies on this Graham C, et al. Association between physician burnout and patient safety, professionalism, and patient satisfaction: a systematic review and meta- issue. analysis. JAMA Intern Med. 2018;178:1317–30. 3. West CP, Dyrbye LN, Shanafelt TD. Physician burnout: contributors, consequences and solutions. J Intern Med. 2018;283:516–29. Conclusion 4. Rathert C, Williams ES, Linhart H. Evidence for the quadruple aim: a In our review, only four, and in part a fifth, of 11 studies systematic review of the literature on physician burnout and patient found a relationship between physician burnout mea- outcomes. Med Care. 2018;56:976–84. 5. Angerer P, Weigl M. Physicians' psychosocial work conditions and quality of sured by MBI and observed adverse patient outcomes. care: a literature review. Prof Professionalism. 2015;5. https://doi.org/10.7577/ More studies should be performed to further under- pp.960. stand whether burnout among physicians leads to low- 6. Rotenstein LS, Torre M, Ramos MA, Rosales RC, Guille C, Sen S, et al. Prevalence of burnout among physicians: a systematic review. JAMA. 2018; ered patient care, and in particular to validate the 320:1131–50. specific level of burnout that impairs physician’ function- 7. Poghosyan L, Aiken LH, Sloane DM. Factor structure of the Maslach burnout ing with respect to observed adverse patient outcomes. inventory: an analysis of data from large scale cross-sectional surveys of nurses from eight countries. Int J Nurs Stud. 2009;46:894–902. Abbreviations 8. Maslach C, Jackson S, Leiter M. The Maslach burnout inventory manual. In: MBI: Maslach Burnout Inventory; DP: Depersonalization; EE: Emotional Zalaquett CP, Wood RJ, editors. Evaluating stress: a book of resources. USA: exhaustion; PA: Personal accomplishment (reduced); OLBI: Oldenburger The Scarecrow Press; 1997. Burnout Inventory; CBI: Copenhagen Burnout Inventory; GP: General 9. Ahola K, Hakanen J, Perhoniemi R, Mutanen P. Relationship between practitioner; ICU: Intensive care unit burnout and depressive symptoms: a study using the person-centred approach. Burn Res. 2014;1.https://doi.org/10.1016/j.burn.2014.03.003. Acknowledgements 10. Tyssen R. What is the level of burnout that impairs functioning? J Intern The final manuscript was reviewed by OnLine English. Med. 2018;283:594–6. Mangory et al. BMC Health Services Research (2021) 21:369 Page 8 of 8 11. Maslach C, Leiter MP. Understanding the burnout experience: recent research 35. West CP, Huschka MM, Novotny PJ, Sloan JA, Kolars JC, Habermann TM, and its implications for psychiatry. World Psychiatry. 2016;15:103–11. et al. Association of perceived medical errors with resident distress and 12. Schwenk TL, Gold KJ. Physician burnout-a serious symptom, but of what? empathy: a prospective longitudinal study. JAMA. 2006;296:1071–8. JAMA. 2018;320:1109–10. 36. Davis DA, Mazmanian PE, Fordis M, Van Harrison R, Thorpe KE, Perrier L. Accuracy of physician self-assessment compared with observed measures of 13. Demerouti E, Bakker AB, Nachreiner F, Schaufeli WB. The job demands- competence: a systematic review. JAMA. 2006;296:1094–102. resources model of burnout. J Appl Psychol. 2001;86:499–512. 37. Gabbard GO. The role of compulsiveness in the Normal physician. JAMA. 14. Crawford E, Lepine J, Rich B. Linking job demands and resources to 1985;254:2926–9. employee engagement and burnout: a theoretical extension and meta- 38. Linzer M, Manwell LB, Williams ES, Bobula JA, Brown R, Varkey AB, et al. analytic test. J Appl Psychol. 2010;95:834–48. Working conditions in primary care: physician reactions and care quality. 15. Parsa-Parsi RW. The revised declaration of Geneva: a modern-day physician's Ann Intern Med. 2009;151:28–36. pledge. JAMA. 2017;318:1971–2. 39. Passalacqua SA, Segrin C. The effect of resident physician stress, burnout, 16. Van Mol MM, Kompanje EJ, Benoit DD, Bakker J, Nijkamp MD. The and empathy on patient-centered communication during the long-call shift. prevalence of compassion fatigue and burnout among healthcare Health Commun. 2012;27:449–56. professionals in intensive care units: a systematic review. PLoS One. 2015;10: 40. Hobfoll S. Conservation of resources: a new attempt at conceptualizing e0136955. stress. Am Psychol. 1989;44:513–24. 17. Zarei E, Ahmadi F, Sial MS, Hwang J, Thu PA, Usman SM. Prevalence of 41. Shanafelt TD, Hasan O, Dyrbye LN, Sinsky C, Satele D, Sloan J, et al. Changes burnout among primary health care staff and its predictors: a study in Iran. in burnout and satisfaction with work-life balance in physicians and the Int J Environ Res Public Health. 2019;16:2249. general US working population between 2011 and 2014. Mayo Clin Proc. 18. Asante JO, Li MJ, Liao J, Huang YX, Hao YT. The relationship between 2015;90:1600–13. psychosocial risk factors, burnout and quality of life among primary healthcare workers in rural Guangdong province: a cross-sectional study. BMC Health Serv Res. 2019;19:447. Publisher’sNote 19. Welp A, Meier LL, Manser T. Emotional exhaustion and workload predict Springer Nature remains neutral with regard to jurisdictional claims in clinician-rated and objective patient safety. Front Psychol. 2014;5:1573. published maps and institutional affiliations. 20. Noroxe KB, Pedersen AF, Carlsen AH, Bro F, Vedsted P. Mental well-being, job satisfaction and self-rated workability in general practitioners and hospitalisations for ambulatory care sensitive conditions among listed patients: a cohort study combining survey data on GPs and register data on patients. BMJ Qual Saf. 2019;28:997–1006. 21. Noroxe KB, Vedsted P, Bro F, Carlsen AH, Pedersen AF. Mental well-being and job satisfaction in general practitioners in Denmark and their patients' change of general practitioner: a cohort study combining survey data and register data. BMJ Open. 2019;9:e030142. 22. Fahrenkopf AM, Sectish TC, Barger LK, Sharek PJ, Lewin D, Chiang VW, et al. Rates of medication errors among depressed and burnt out residents: prospective cohort study. BMJ. 2008;336:488–91. 23. Zantinge EM, Verhaak PFM, de Bakker DH, van der Meer K, Bensing JM. Does burnout among doctors affect their involvement in patients' mental health problems? A study of videotaped consultations. BMC Fam Pract. 2009;10:60. 24. Kushnir T, Greenberg D, Madjar N, Hadari I, Yermiahu Y, Bachner YG. Is burnout associated with referral rates among primary care physicians in community clinics? Fam Pract. 2014;31:44–50. 25. Yuguero Torres O, Esquerda Aresté M, Marsal Mora JR, Soler-González J. Association between sick leave prescribing practices and physician burnout and empathy. PLoS One. 2015;10:e0133379. 26. Garrouste-Orgeas M, Perrin M, Soufir L, Vesin A, Blot F, Maxime V, et al. The Iatroref study: medical errors are associated with symptoms of depression in ICU staff but not burnout or safety culture. Intensive Care Med. 2015;41: 273–84. 27. Pedersen AF, Carlsen AH, Vedsted P. Association of GPs' risk attitudes, level of empathy, and burnout status with PSA testing in primary care. Br J Gen Pract. 2015;65:e845–51. 28. Kwah J, Weintraub J, Fallar R, Ripp J. The effect of burnout on medical errors and professionalism in first-year internal medicine residents. J Grad Med Educ. 2016;8:597–600. 29. Sun BZ, Chaitoff A, Hu B, Neuendorf K, Manne M, Rothberg MB. Empathy, burnout, and antibiotic prescribing for acute respiratory infections: a cross- sectional primary care study in the US. Br J Gen Pract. 2017;67:e565–71. 30. Maslach C, Jackson SE. The measurement of experienced burnout. J Organ Behav. 1981;2:99–113. 31. Dyrbye LN, West CP, Shanafelt TD. Defining burnout as a dichotomous variable. J Gen Intern Med. 2009;24:440 author reply 441. 32. Ratanawongsa N, Roter D, Beach MC, Laird SL, Larson SM, Carson KA, et al. Physician burnout and patient-physician communication during primary care encounters. J Gen Intern Med. 2008;23:1581–8. 33. Glass DC, McKnight JD. Perceived control, depressive symptomatology, and professional burnout: a review of the evidence. Psychol Health. 1996;11:23–48. 34. Glass DC, McKnight JD, Valdimarsdottir H. Depression, burnout, and perceptions of control in hospital nurses. J Consult Clin Psychol. 1993;61: 147–55.

Journal

BMC Health Services ResearchSpringer Journals

Published: Apr 21, 2021

There are no references for this article.