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

Learn More →

Burden of Difficult Encounters in Primary Care: Data From the Minimizing Error, Maximizing Outcomes Study

Burden of Difficult Encounters in Primary Care: Data From the Minimizing Error, Maximizing... Nearly 1 of 6 outpatient visits is considered difficult by physicians.1 Difficult encounters are more likely to occur with patients who have a mental disorder, present with more than 5 somatic symptoms, exhibit high use of health services, possess a list of complaints, or have threatening and abrasive personalities.2-7 Physicians report that they secretly hope that their challenging patients will not return and find that, in general, difficult encounters are time-consuming and personally and professionally unsatisfying.4 Although the attributes of challenging patients are well defined in the literature, the characteristics of physicians involved in high numbers of difficult encounters are less understood. For example, age and years in clinical practice have been inversely correlated with frequency of difficult encounters in some investigations, yet other studies have found no such relationship.1,6,8,9 Subspeciality physicians, compared with family physicians, are more frustrated by difficult encounters and feel ill-equipped to manage them,8,10 yet associations between difficult encounters and patient outcomes remain to be determined. We sought to compare levels of stress, burnout, satisfaction, time pressure, intent to leave the practice, and medical errors between primary care physicians who report having high numbers of difficult encounters with patients and those who do not. Methods Participants This is a nested analysis of physician survey data from the Minimizing Error, Maximizing Outcome (MEMO) study.11 The MEMO study is designed to assess the effect of the primary care work environment on quality of care and the role of physicians as mediators of this effect. General internists and family physicians were recruited from October 2002 to June 2003 from ambulatory care clinics in Chicago, Illinois; Madison, Wisconsin; Milwaukee, Wisconsin; New York, New York; and rural/small town Wisconsin. These 5 regions offered a diverse patient base. Physicians were recruited by on-site presentations or by mail and were asked to complete a 15-minute self-administered survey. The survey was derived from the Physician Worklife Survey12,13 and from comments made during focus groups at the inception of the MEMO study. Measures The attributes of challenging, difficult patients most consistently identified in the literature were used to develop an 8-item index, the Burden of Difficult Encounters measure (Table 1). Frequently cited characteristics of difficult patients and the Difficult Doctor-Patient Relationship Questionnaire,5-7,14,15 made up the Burden of Difficult Encounters measure, and, using this index, study physicians estimated how often they encountered patients with each attribute (1, never; 2, sometimes; 3, frequently; and 4, often). Based on these responses, using latent cluster analysis (LCA), physicians were clustered into 3 groups: those who perceived high, medium, and low numbers of difficult encounters. Table 1. View LargeDownload Burden of Difficult Encounters Measure Physicians also responded to several other measures. Job stress was rated on a 4-item scale,16 and physicians with a mean score of 3.5 or more (out of 5) were identified as stressed. Single questions measured burnout17 and intent to leave one's current job within 2 years. A 5-item scale queried global job satisfaction,12 with a mean score of 3.5 or more (out of 5) denoting high satisfaction. Physicians recorded the average time allotted for routine follow-up appointments and the time needed to provide quality care. A ratio greater than 1 of time needed to time allotted denoted perceived time pressure. Based on findings by Shanafelt et al,18 physicians rated their frequency of providing suboptimal care during the past year in 5 areas, such as committing treatment or medication errors not attributed to a lack of knowledge. We defined “error prone” as a mean score of 3 or more on the 5-item scale (1, never; 2, once; 3, several times; 4, monthly; 5, weekly). Finally, physicians predicted future error with the newly developed MEMO Occupational Stress and Preventable Error measure. This 9-item scale queried the likelihood of neglecting to provide common aspects of chronic disease management. A mean score of 4 (somewhat likely) or higher on a scale of 1 (very unlikely) to 6 (very likely) was used as the cutoff point. The scales for job stress, satisfaction, and burnout have been validated in previous studies.12,16,17 Data Analysis An LCA was used to classify physicians based on their responses to the Burden of Difficult Encounters measure (Table 1). Latent GOLD statistical software, version 4.0,19 was used to conduct this analysis. In essence, LCA allows for the detection of homogeneous subgroups in a heterogenous group through the minimization of associations among responses across a set of indicators. This tool uses a probabilistic approach: although each object is assumed to belong to 1 cluster, it is taken into account that there is uncertainty about an object's cluster membership.20 The LCA begins with the assumption that there is only 1 cluster and subsequently estimates 2, 3, or more classes until an LCA model is found that statistically fits the data. We used the bayesian information criterion and consistent Akaike information criterion21 to compare models with 1 through 4 latent clusters (Table 2). The 3-cluster model was found to best fit our data (lowest fit indices indicate the best fit). Table 2. View LargeDownload Latent Cluster Analysis Cluster Solutions Once physician clusters were established based on perceived difficulty, this information was used in a logistic regression model as a k - 1 dummy variable to predict dichotomous physician outcome measures (ie, stress, burnout, job satisfaction, time pressure, intent to leave, suboptimal care in the past year, and likelihood of future errors), controlling for physician age, sex, and racial/ethnic minority status. Two more adjustments were applied to the results of our logistic regression models. First, because information was obtained from physicians who were recruited from clinics, these clustered physician responses can typically produce negatively biased standard errors and, subsequently, increase false positives. To adjust for this bias we used the Huber/White sandwich estimator.22,23 Second, the Sidak-Holms procedure24 was applied to adjust for type I errors derived from multiple testing of difficulty across various physician outcome measures. Results A total of 449 physicians from 118 clinics agreed to participate (59.8% of those invited to participate), and 422 (94.0%) completed the survey. Nonparticipants did not differ substantially from participants in specialty or sex. The physicians were evenly divided between general internists (51.9%) and family physicians (48.1%). The mean age was 43 years (age range, 29-89 years); 44.4% were women, 22.0% were of a racial or ethnic minority group, and 83.3% worked full-time. “Patients who insist on being prescribed an unnecessary drug” was the most frequently cited challenge, with 155 of 422 respondents (36.7%) claiming frequent encounters with such patients (Table 1). In addition, 16.1% of physicians claimed that they frequently saw patients who showed dissatisfaction with their care. Patients with unrealistic expectations for their care were frequently encountered by 13.7% of physicians. Physicians in the low (n = 41) and medium (n = 269) difficulty clusters were characterized by relatively low to medium scores on all difficulty items relative to physicians in the high difficulty cluster (n = 113). However, physicians in the low difficulty cluster were distinguished from those in the medium difficulty cluster by an almost complete and unanimous indication of no perceived difficulty with patients they saw. Those in the high cluster had an almost complete and unanimous indication of working with difficult patients (Table 1). High difficulty cluster physicians were significantly younger (mean age, 41 years) compared with medium (mean age, 43 years; P =.01) and low (mean age, 46 years; P =.02) difficulty cluster respondents. Physicians in the high and medium difficulty clusters were more likely to be women (50.4% and 44.6%, respectively) than their low difficulty cluster counterparts (26.8%; P =.005 and .03, respectively). However, high difficulty cluster physicians did not differ from the other 2 groups with regard to racial/ethnic minority or full-time work status (Table 3). Table 3. View LargeDownload Characteristics of Physicians by Perceived Difficulty Clustera For most tested end points, physicians in the high difficulty cluster reported more adverse outcomes than those in the low and medium difficulty clusters (Table 4). After adjusting for age, sex, and racial/ethnic minority status, as well as for negatively biased standard errors and type I errors arising from multiple comparisons, the following significant findings remained. High difficulty cluster physicians were 2.2 times more likely than medium difficulty cluster physicians (95% confidence interval [CI], 1.3-3.6; P = .02) to report burnout and were 12.2 times more likely to be burned out than low difficulty cluster physicians (95% CI, 2.7-55.6; P = .01). High difficulty cluster physicians were also 2.7 times less likely to indicate high job satisfaction compared with medium difficulty cluster physicians (95% CI, 1.8-4.0; P = .001) and were 3.8 times less likely to be highly satisfied with their jobs compared with low difficulty cluster respondents (95% CI, 1.6-9.1; P = .02). This “dose-like” response was found across all tested end points, including stress, time pressure, intent to leave one's practice, and perception of suboptimal care practices. Moreover, compared with low difficulty cluster physicians, high difficulty cluster physicians were more likely to report suboptimal care practices in the past year (odds ratio, 9.4; 95% CI, 1.5 to infinity) and to expect future errors in their practices (odds ratio, 2.8; 95% CI, 1.2-7.8); however, after adjusting for negatively biased standard errors and multiple comparisons, the data did not reach statistical significance (P = .09 and P = .16, respectively). Table 4. View LargeDownload 2-Level Logistic Regression Analysis Comparing Physicians Comment Difficult encounters are a readily recognized challenge in primary care, and the shared responsibility and contributions to such interactions by physicians and patients have been increasingly acknowledged during the past decade.6,9,14 Our data confirm that difficult encounters are common and that adverse outcomes are reported by physicians who perceive that they have high numbers of such visits. Physicians in our study who perceived high numbers of difficult visits were younger and more likely to be women than their counterparts, as demonstrated in previous studies.1,8 Older, more experienced practitioners may have developed coping mechanisms to mitigate the difficulty associated with such encounters. In addition, the process of self-selection (patients for whom encounters are difficult may seek other physicians to coordinate and provide care) may contribute to this observation. Others have previously reported on a clear relationship between female sex and professional burnout,26 possibly explaining our finding that burdened physicians tended to be women. Although significant differences in burnout rates remained after controlling for sex, our results suggest that difficult encounters are more prevalent within the practices of female physicians and that more time during patient visits may need to be allotted to address this discrepancy. Physicians who perceived a higher volume of difficult encounters were significantly more burned out and dissatisfied with their jobs than those reporting fewer difficult encounters. Others have reported similar findings,1,6,8 but the consistency of our results across numerous physician variables and in a graded fashion (more significant odds ratio when using low difficulty cluster respondents as the reference group compared with medium difficulty cluster respondents), as well as the use of validated tools for burnout and satisfaction, strengthen the hypothesized relationship between difficult encounters and adverse physician outcomes. Most salient of our findings was that high difficulty cluster physicians were 12 times more likely than low difficulty cluster physicians to report burnout. This has critical implications for the future of primary care because fewer trainees are choosing careers in primary care, perhaps in part owing to burned-out role models.27,28 Our study has several limitations. First, the data are self-reported. The simultaneous measurement of our primary end points and of the frequency of difficult encounters—all through the report of physicians—make causal relationships difficult to determine. In other words, whether difficult encounters lead to burnout or whether burned-out physicians consider more visits to be difficult cannot be determined by our study. Our goal was to highlight the coexistence of these physician experiences to better characterize difficult encounters as a whole, not to suggest cause and effect. We strongly believe that these associations are relevant in that, for example, a burned-out physician experiencing many difficult patient encounters is likely to need help in overcoming both perceived challenges. It is also likely that patients may feel the challenge the physician is feeling. As such, we believe this issue to be of considerable importance. In addition, the potential impact of physician burnout and perceived difficulty on patient satisfaction and perception of difficulty remains to be determined. A second limitation is that the most burned-out, stressed, and dissatisfied physicians may have declined to enroll in our study; physician outcome data from those who did not participate are not available. Finally, the present analysis did not assess actual patient outcomes. Our results indicate the potential value of strategies to help physicians manage difficult encounters more effectively. Previously suggested coping mechanisms include demonstrating more empathy, practicing nonjudgmental listening, and communicating more directly with patients involved in difficult encounters.2,29 Increased training on approaching difficult encounters is warranted, as is the provision of more support personnel (eg, social service) and perhaps the allotment of more time for difficult encounters. Because of the prevalence of difficult encounters and their strong association with physician burnout and dissatisfaction, explicitly addressing difficult encounters in primary care is of considerable importance. Correspondence: Dr An, Department of Medicine, Newton-Wellesley Hospital, 2014 Washington St, Newton, MA 02462 (perryan@post.harvard.edu). Author Contributions: All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: An, Rabatin, Manwell, Linzer, Brown, and Schwartz. Analysis and interpretation of data: An, Rabatin, Manwell, Linzer, Brown, and Schwartz. Drafting of the manuscript: An, Rabatin, Manwell, Linzer, Brown, and Schwartz. Critical revision of the manuscript for important intellectual content: An, Rabatin, Manwell, Linzer, Brown, and Schwartz. Administrative, technical, and material support: An, Rabatin, Manwell, Linzer, Brown, and Schwartz. Study supervision: Rabatin, Manwell, Linzer, Brown, and Schwartz. Financial Disclosure: None reported. Funding/Support: This study was sponsored by grant 053253 from the Robert Wood Johnson Foundation and grant R01 HS011955 from the Agency for Healthcare Research and Quality. Previous Presentations: A preliminary report of these results appeared as an abstract at the 2007 annual meeting of the Society of General Internal Medicine; April 25-28, 2007; Toronto, Ontario, Canada. Additional Information: A list of the MEMO investigators is available on request from the corresponding author. References 1. Mathers NJones NHannay D Heartsink patients: a study of their general practitioners. Br J Gen Pract 1995;45 (395) 293- 296PubMedGoogle Scholar 2. Steinmetz DTabenkin H The “difficult patient” as perceived by family physicians. Fam Pract 2001;18 (5) 495- 500PubMedGoogle Scholar 3. McDonald PSO’Dowd TC The heartsink patient: a preliminary study. Fam Pract 1991;8 (2) 112- 116PubMedGoogle Scholar 4. Hahn SR Physical symptoms and physician-experienced difficulty in the physician-patient relationship. Ann Intern Med 2001;134 (9, pt 2) 897- 904PubMedGoogle Scholar 5. Hahn SRKroenke NSpitzer RL et al. The difficult patient: prevalence, psychopathology, and functional impairment. J Gen Intern Med 1996;11 (1) 1- 8PubMedGoogle Scholar 6. Jackson JLKroenke K Difficult patient encounters in the ambulatory clinic. Arch Intern Med 1999;159 (10) 1069- 1075PubMedGoogle Scholar 7. Sharpe MMayou RSeacroatt V et al. Why do doctors find some patients difficult to help? Q J Med 1994;87 (3) 187- 193PubMedGoogle Scholar 8. Krebs EEGarrett JMKonrad TR The difficult doctor? characteristics of physicians who report frustration with patients: an analysis of survey data [published online October 6, 2006]. BMC Health Serv Res 2006;6128- 135PubMed10.1186/1472-6963-6-128Google Scholar 9. Haas LJLeiser JPMagill MKSanyer ON Management of the difficult patient. Am Fam Physician 2005;72 (10) 2063- 2068PubMedGoogle Scholar 10. Wetterneck TBLinzer M McMurray JE et al. Worklife and satisfaction of general internists. Arch Intern Med 2002;162 (6) 649- 656PubMedGoogle Scholar 11. Linzer MManwell LBMundt M et al. Organizational climate, stress, and error in primary care: the MEMO study. Advances in Patient Safety From Research to Implementation. Rockville, MD Agency for Healthcare Research and Quality2005;65- 77AHRQ publication 050021 (1). http://www.ahrq.gov/qual/advances/. Accessed November 13, 2007Google Scholar 12. Williams ESKonrad TRLinzer M et al. Refining the measurement of physician job satisfaction: results from the Physician Worklife Study. Med Care 1999;37 (11) 1140- 1154PubMedGoogle Scholar 13. Konrad TRWilliams ESLinzer M et al. Measuring physician job satisfaction in a changing workplace and a challenging environment. Med Care 1999;37 (11) 1174- 1182PubMedGoogle Scholar 14. Elder NRicer RTobias B How respected family physicians manage difficult patient encounters. J Am Board Fam Med 2006;19 (6) 533- 541PubMedGoogle Scholar 15. Groves JE Taking care of the hateful patient. N Engl J Med 1978;298 (16) 883- 887PubMedGoogle Scholar 16. Motowidlo SJPackard JSManning MR Occupational stress: its causes and consequences for job performance. J Appl Psychol 1986;71 (4) 618- 629PubMedGoogle Scholar 17. Freeborn DK Physician satisfaction in a prepaid group practice HMO. Group Health J 1985;6 (1) 3- 12PubMedGoogle Scholar 18. Shanafelt TDBradley KAWipf JEBack AL Burnout and self-reported patient care in an internal medicine residency program. Ann Intern Med 2002;136 (5) 358- 367PubMedGoogle Scholar 19. Vermunt JKMagidson J Latent GOLD User's Guide. Boston, MA Statistical Innovations Inc2004; 20. Magidson JVermunt JK Latent class models. Kaplan D The Sage Handbook for Quantitative Methodology Thousand Oaks, CA Sage Publications2004;175- 198Google Scholar 21. Fraley CRaftery AE How Many Clusters? Which Clustering Method? Answers via Model-Based Cluster Analysis. Seattle, WA Dept of Statistics, University of Washington1998;Technical Report 329 22. Huber PJ Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability. Berkeley University of California Press1967;221- 233 23. White H Maximum likelihood estimation of misspecified models. Econometrica 1982;501- 25Google Scholar 24. Hochberg YTamhane AC Multiple Comparison Procedures. New York, NY John Wiley & Sons1987; 25. Mehta CRPatel NRJajoo B Exact Logistic Regression: Theory, Methods, and Software. Cambridge, MA Cytel Software Corp1993;Cytel Technical Report 26. Linzer M McMurray JEVisser MROort FJSmets Ede Haes HC Sex differences in physician burnout in the United States and the Netherlands. J Am Med Womens Assoc 2002;57 (4) 191- 193PubMedGoogle Scholar 27. Moore GShowstack J Primary care medicine in crisis: toward reconstruction and renewal. Ann Intern Med 2003;138 (3) 244- 247PubMedGoogle Scholar 28. Whitcomb MECohen JJ The future of primary care medicine. N Engl J Med 2004;351 (7) 710- 712PubMedGoogle Scholar 29. Adams JMurray R The general approach to the difficult patient. Emerg Med Clin North Am 1998;16 (4) 689- 700PubMedGoogle Scholar http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Archives of Internal Medicine American Medical Association

Burden of Difficult Encounters in Primary Care: Data From the Minimizing Error, Maximizing Outcomes Study

Loading next page...
 
/lp/american-medical-association/burden-of-difficult-encounters-in-primary-care-data-from-the-ilWzHSMNpv
Publisher
American Medical Association
Copyright
Copyright © 2009 American Medical Association. All Rights Reserved.
ISSN
0003-9926
eISSN
1538-3679
DOI
10.1001/archinternmed.2008.549
Publisher site
See Article on Publisher Site

Abstract

Nearly 1 of 6 outpatient visits is considered difficult by physicians.1 Difficult encounters are more likely to occur with patients who have a mental disorder, present with more than 5 somatic symptoms, exhibit high use of health services, possess a list of complaints, or have threatening and abrasive personalities.2-7 Physicians report that they secretly hope that their challenging patients will not return and find that, in general, difficult encounters are time-consuming and personally and professionally unsatisfying.4 Although the attributes of challenging patients are well defined in the literature, the characteristics of physicians involved in high numbers of difficult encounters are less understood. For example, age and years in clinical practice have been inversely correlated with frequency of difficult encounters in some investigations, yet other studies have found no such relationship.1,6,8,9 Subspeciality physicians, compared with family physicians, are more frustrated by difficult encounters and feel ill-equipped to manage them,8,10 yet associations between difficult encounters and patient outcomes remain to be determined. We sought to compare levels of stress, burnout, satisfaction, time pressure, intent to leave the practice, and medical errors between primary care physicians who report having high numbers of difficult encounters with patients and those who do not. Methods Participants This is a nested analysis of physician survey data from the Minimizing Error, Maximizing Outcome (MEMO) study.11 The MEMO study is designed to assess the effect of the primary care work environment on quality of care and the role of physicians as mediators of this effect. General internists and family physicians were recruited from October 2002 to June 2003 from ambulatory care clinics in Chicago, Illinois; Madison, Wisconsin; Milwaukee, Wisconsin; New York, New York; and rural/small town Wisconsin. These 5 regions offered a diverse patient base. Physicians were recruited by on-site presentations or by mail and were asked to complete a 15-minute self-administered survey. The survey was derived from the Physician Worklife Survey12,13 and from comments made during focus groups at the inception of the MEMO study. Measures The attributes of challenging, difficult patients most consistently identified in the literature were used to develop an 8-item index, the Burden of Difficult Encounters measure (Table 1). Frequently cited characteristics of difficult patients and the Difficult Doctor-Patient Relationship Questionnaire,5-7,14,15 made up the Burden of Difficult Encounters measure, and, using this index, study physicians estimated how often they encountered patients with each attribute (1, never; 2, sometimes; 3, frequently; and 4, often). Based on these responses, using latent cluster analysis (LCA), physicians were clustered into 3 groups: those who perceived high, medium, and low numbers of difficult encounters. Table 1. View LargeDownload Burden of Difficult Encounters Measure Physicians also responded to several other measures. Job stress was rated on a 4-item scale,16 and physicians with a mean score of 3.5 or more (out of 5) were identified as stressed. Single questions measured burnout17 and intent to leave one's current job within 2 years. A 5-item scale queried global job satisfaction,12 with a mean score of 3.5 or more (out of 5) denoting high satisfaction. Physicians recorded the average time allotted for routine follow-up appointments and the time needed to provide quality care. A ratio greater than 1 of time needed to time allotted denoted perceived time pressure. Based on findings by Shanafelt et al,18 physicians rated their frequency of providing suboptimal care during the past year in 5 areas, such as committing treatment or medication errors not attributed to a lack of knowledge. We defined “error prone” as a mean score of 3 or more on the 5-item scale (1, never; 2, once; 3, several times; 4, monthly; 5, weekly). Finally, physicians predicted future error with the newly developed MEMO Occupational Stress and Preventable Error measure. This 9-item scale queried the likelihood of neglecting to provide common aspects of chronic disease management. A mean score of 4 (somewhat likely) or higher on a scale of 1 (very unlikely) to 6 (very likely) was used as the cutoff point. The scales for job stress, satisfaction, and burnout have been validated in previous studies.12,16,17 Data Analysis An LCA was used to classify physicians based on their responses to the Burden of Difficult Encounters measure (Table 1). Latent GOLD statistical software, version 4.0,19 was used to conduct this analysis. In essence, LCA allows for the detection of homogeneous subgroups in a heterogenous group through the minimization of associations among responses across a set of indicators. This tool uses a probabilistic approach: although each object is assumed to belong to 1 cluster, it is taken into account that there is uncertainty about an object's cluster membership.20 The LCA begins with the assumption that there is only 1 cluster and subsequently estimates 2, 3, or more classes until an LCA model is found that statistically fits the data. We used the bayesian information criterion and consistent Akaike information criterion21 to compare models with 1 through 4 latent clusters (Table 2). The 3-cluster model was found to best fit our data (lowest fit indices indicate the best fit). Table 2. View LargeDownload Latent Cluster Analysis Cluster Solutions Once physician clusters were established based on perceived difficulty, this information was used in a logistic regression model as a k - 1 dummy variable to predict dichotomous physician outcome measures (ie, stress, burnout, job satisfaction, time pressure, intent to leave, suboptimal care in the past year, and likelihood of future errors), controlling for physician age, sex, and racial/ethnic minority status. Two more adjustments were applied to the results of our logistic regression models. First, because information was obtained from physicians who were recruited from clinics, these clustered physician responses can typically produce negatively biased standard errors and, subsequently, increase false positives. To adjust for this bias we used the Huber/White sandwich estimator.22,23 Second, the Sidak-Holms procedure24 was applied to adjust for type I errors derived from multiple testing of difficulty across various physician outcome measures. Results A total of 449 physicians from 118 clinics agreed to participate (59.8% of those invited to participate), and 422 (94.0%) completed the survey. Nonparticipants did not differ substantially from participants in specialty or sex. The physicians were evenly divided between general internists (51.9%) and family physicians (48.1%). The mean age was 43 years (age range, 29-89 years); 44.4% were women, 22.0% were of a racial or ethnic minority group, and 83.3% worked full-time. “Patients who insist on being prescribed an unnecessary drug” was the most frequently cited challenge, with 155 of 422 respondents (36.7%) claiming frequent encounters with such patients (Table 1). In addition, 16.1% of physicians claimed that they frequently saw patients who showed dissatisfaction with their care. Patients with unrealistic expectations for their care were frequently encountered by 13.7% of physicians. Physicians in the low (n = 41) and medium (n = 269) difficulty clusters were characterized by relatively low to medium scores on all difficulty items relative to physicians in the high difficulty cluster (n = 113). However, physicians in the low difficulty cluster were distinguished from those in the medium difficulty cluster by an almost complete and unanimous indication of no perceived difficulty with patients they saw. Those in the high cluster had an almost complete and unanimous indication of working with difficult patients (Table 1). High difficulty cluster physicians were significantly younger (mean age, 41 years) compared with medium (mean age, 43 years; P =.01) and low (mean age, 46 years; P =.02) difficulty cluster respondents. Physicians in the high and medium difficulty clusters were more likely to be women (50.4% and 44.6%, respectively) than their low difficulty cluster counterparts (26.8%; P =.005 and .03, respectively). However, high difficulty cluster physicians did not differ from the other 2 groups with regard to racial/ethnic minority or full-time work status (Table 3). Table 3. View LargeDownload Characteristics of Physicians by Perceived Difficulty Clustera For most tested end points, physicians in the high difficulty cluster reported more adverse outcomes than those in the low and medium difficulty clusters (Table 4). After adjusting for age, sex, and racial/ethnic minority status, as well as for negatively biased standard errors and type I errors arising from multiple comparisons, the following significant findings remained. High difficulty cluster physicians were 2.2 times more likely than medium difficulty cluster physicians (95% confidence interval [CI], 1.3-3.6; P = .02) to report burnout and were 12.2 times more likely to be burned out than low difficulty cluster physicians (95% CI, 2.7-55.6; P = .01). High difficulty cluster physicians were also 2.7 times less likely to indicate high job satisfaction compared with medium difficulty cluster physicians (95% CI, 1.8-4.0; P = .001) and were 3.8 times less likely to be highly satisfied with their jobs compared with low difficulty cluster respondents (95% CI, 1.6-9.1; P = .02). This “dose-like” response was found across all tested end points, including stress, time pressure, intent to leave one's practice, and perception of suboptimal care practices. Moreover, compared with low difficulty cluster physicians, high difficulty cluster physicians were more likely to report suboptimal care practices in the past year (odds ratio, 9.4; 95% CI, 1.5 to infinity) and to expect future errors in their practices (odds ratio, 2.8; 95% CI, 1.2-7.8); however, after adjusting for negatively biased standard errors and multiple comparisons, the data did not reach statistical significance (P = .09 and P = .16, respectively). Table 4. View LargeDownload 2-Level Logistic Regression Analysis Comparing Physicians Comment Difficult encounters are a readily recognized challenge in primary care, and the shared responsibility and contributions to such interactions by physicians and patients have been increasingly acknowledged during the past decade.6,9,14 Our data confirm that difficult encounters are common and that adverse outcomes are reported by physicians who perceive that they have high numbers of such visits. Physicians in our study who perceived high numbers of difficult visits were younger and more likely to be women than their counterparts, as demonstrated in previous studies.1,8 Older, more experienced practitioners may have developed coping mechanisms to mitigate the difficulty associated with such encounters. In addition, the process of self-selection (patients for whom encounters are difficult may seek other physicians to coordinate and provide care) may contribute to this observation. Others have previously reported on a clear relationship between female sex and professional burnout,26 possibly explaining our finding that burdened physicians tended to be women. Although significant differences in burnout rates remained after controlling for sex, our results suggest that difficult encounters are more prevalent within the practices of female physicians and that more time during patient visits may need to be allotted to address this discrepancy. Physicians who perceived a higher volume of difficult encounters were significantly more burned out and dissatisfied with their jobs than those reporting fewer difficult encounters. Others have reported similar findings,1,6,8 but the consistency of our results across numerous physician variables and in a graded fashion (more significant odds ratio when using low difficulty cluster respondents as the reference group compared with medium difficulty cluster respondents), as well as the use of validated tools for burnout and satisfaction, strengthen the hypothesized relationship between difficult encounters and adverse physician outcomes. Most salient of our findings was that high difficulty cluster physicians were 12 times more likely than low difficulty cluster physicians to report burnout. This has critical implications for the future of primary care because fewer trainees are choosing careers in primary care, perhaps in part owing to burned-out role models.27,28 Our study has several limitations. First, the data are self-reported. The simultaneous measurement of our primary end points and of the frequency of difficult encounters—all through the report of physicians—make causal relationships difficult to determine. In other words, whether difficult encounters lead to burnout or whether burned-out physicians consider more visits to be difficult cannot be determined by our study. Our goal was to highlight the coexistence of these physician experiences to better characterize difficult encounters as a whole, not to suggest cause and effect. We strongly believe that these associations are relevant in that, for example, a burned-out physician experiencing many difficult patient encounters is likely to need help in overcoming both perceived challenges. It is also likely that patients may feel the challenge the physician is feeling. As such, we believe this issue to be of considerable importance. In addition, the potential impact of physician burnout and perceived difficulty on patient satisfaction and perception of difficulty remains to be determined. A second limitation is that the most burned-out, stressed, and dissatisfied physicians may have declined to enroll in our study; physician outcome data from those who did not participate are not available. Finally, the present analysis did not assess actual patient outcomes. Our results indicate the potential value of strategies to help physicians manage difficult encounters more effectively. Previously suggested coping mechanisms include demonstrating more empathy, practicing nonjudgmental listening, and communicating more directly with patients involved in difficult encounters.2,29 Increased training on approaching difficult encounters is warranted, as is the provision of more support personnel (eg, social service) and perhaps the allotment of more time for difficult encounters. Because of the prevalence of difficult encounters and their strong association with physician burnout and dissatisfaction, explicitly addressing difficult encounters in primary care is of considerable importance. Correspondence: Dr An, Department of Medicine, Newton-Wellesley Hospital, 2014 Washington St, Newton, MA 02462 (perryan@post.harvard.edu). Author Contributions: All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: An, Rabatin, Manwell, Linzer, Brown, and Schwartz. Analysis and interpretation of data: An, Rabatin, Manwell, Linzer, Brown, and Schwartz. Drafting of the manuscript: An, Rabatin, Manwell, Linzer, Brown, and Schwartz. Critical revision of the manuscript for important intellectual content: An, Rabatin, Manwell, Linzer, Brown, and Schwartz. Administrative, technical, and material support: An, Rabatin, Manwell, Linzer, Brown, and Schwartz. Study supervision: Rabatin, Manwell, Linzer, Brown, and Schwartz. Financial Disclosure: None reported. Funding/Support: This study was sponsored by grant 053253 from the Robert Wood Johnson Foundation and grant R01 HS011955 from the Agency for Healthcare Research and Quality. Previous Presentations: A preliminary report of these results appeared as an abstract at the 2007 annual meeting of the Society of General Internal Medicine; April 25-28, 2007; Toronto, Ontario, Canada. Additional Information: A list of the MEMO investigators is available on request from the corresponding author. References 1. Mathers NJones NHannay D Heartsink patients: a study of their general practitioners. Br J Gen Pract 1995;45 (395) 293- 296PubMedGoogle Scholar 2. Steinmetz DTabenkin H The “difficult patient” as perceived by family physicians. Fam Pract 2001;18 (5) 495- 500PubMedGoogle Scholar 3. McDonald PSO’Dowd TC The heartsink patient: a preliminary study. Fam Pract 1991;8 (2) 112- 116PubMedGoogle Scholar 4. Hahn SR Physical symptoms and physician-experienced difficulty in the physician-patient relationship. Ann Intern Med 2001;134 (9, pt 2) 897- 904PubMedGoogle Scholar 5. Hahn SRKroenke NSpitzer RL et al. The difficult patient: prevalence, psychopathology, and functional impairment. J Gen Intern Med 1996;11 (1) 1- 8PubMedGoogle Scholar 6. Jackson JLKroenke K Difficult patient encounters in the ambulatory clinic. Arch Intern Med 1999;159 (10) 1069- 1075PubMedGoogle Scholar 7. Sharpe MMayou RSeacroatt V et al. Why do doctors find some patients difficult to help? Q J Med 1994;87 (3) 187- 193PubMedGoogle Scholar 8. Krebs EEGarrett JMKonrad TR The difficult doctor? characteristics of physicians who report frustration with patients: an analysis of survey data [published online October 6, 2006]. BMC Health Serv Res 2006;6128- 135PubMed10.1186/1472-6963-6-128Google Scholar 9. Haas LJLeiser JPMagill MKSanyer ON Management of the difficult patient. Am Fam Physician 2005;72 (10) 2063- 2068PubMedGoogle Scholar 10. Wetterneck TBLinzer M McMurray JE et al. Worklife and satisfaction of general internists. Arch Intern Med 2002;162 (6) 649- 656PubMedGoogle Scholar 11. Linzer MManwell LBMundt M et al. Organizational climate, stress, and error in primary care: the MEMO study. Advances in Patient Safety From Research to Implementation. Rockville, MD Agency for Healthcare Research and Quality2005;65- 77AHRQ publication 050021 (1). http://www.ahrq.gov/qual/advances/. Accessed November 13, 2007Google Scholar 12. Williams ESKonrad TRLinzer M et al. Refining the measurement of physician job satisfaction: results from the Physician Worklife Study. Med Care 1999;37 (11) 1140- 1154PubMedGoogle Scholar 13. Konrad TRWilliams ESLinzer M et al. Measuring physician job satisfaction in a changing workplace and a challenging environment. Med Care 1999;37 (11) 1174- 1182PubMedGoogle Scholar 14. Elder NRicer RTobias B How respected family physicians manage difficult patient encounters. J Am Board Fam Med 2006;19 (6) 533- 541PubMedGoogle Scholar 15. Groves JE Taking care of the hateful patient. N Engl J Med 1978;298 (16) 883- 887PubMedGoogle Scholar 16. Motowidlo SJPackard JSManning MR Occupational stress: its causes and consequences for job performance. J Appl Psychol 1986;71 (4) 618- 629PubMedGoogle Scholar 17. Freeborn DK Physician satisfaction in a prepaid group practice HMO. Group Health J 1985;6 (1) 3- 12PubMedGoogle Scholar 18. Shanafelt TDBradley KAWipf JEBack AL Burnout and self-reported patient care in an internal medicine residency program. Ann Intern Med 2002;136 (5) 358- 367PubMedGoogle Scholar 19. Vermunt JKMagidson J Latent GOLD User's Guide. Boston, MA Statistical Innovations Inc2004; 20. Magidson JVermunt JK Latent class models. Kaplan D The Sage Handbook for Quantitative Methodology Thousand Oaks, CA Sage Publications2004;175- 198Google Scholar 21. Fraley CRaftery AE How Many Clusters? Which Clustering Method? Answers via Model-Based Cluster Analysis. Seattle, WA Dept of Statistics, University of Washington1998;Technical Report 329 22. Huber PJ Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability. Berkeley University of California Press1967;221- 233 23. White H Maximum likelihood estimation of misspecified models. Econometrica 1982;501- 25Google Scholar 24. Hochberg YTamhane AC Multiple Comparison Procedures. New York, NY John Wiley & Sons1987; 25. Mehta CRPatel NRJajoo B Exact Logistic Regression: Theory, Methods, and Software. Cambridge, MA Cytel Software Corp1993;Cytel Technical Report 26. Linzer M McMurray JEVisser MROort FJSmets Ede Haes HC Sex differences in physician burnout in the United States and the Netherlands. J Am Med Womens Assoc 2002;57 (4) 191- 193PubMedGoogle Scholar 27. Moore GShowstack J Primary care medicine in crisis: toward reconstruction and renewal. Ann Intern Med 2003;138 (3) 244- 247PubMedGoogle Scholar 28. Whitcomb MECohen JJ The future of primary care medicine. N Engl J Med 2004;351 (7) 710- 712PubMedGoogle Scholar 29. Adams JMurray R The general approach to the difficult patient. Emerg Med Clin North Am 1998;16 (4) 689- 700PubMedGoogle Scholar

Journal

Archives of Internal MedicineAmerican Medical Association

Published: Feb 23, 2009

Keywords: primary health care,burnout

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$499/year

Save searches from
Google Scholar,
PubMed

Create folders to
organize your research

Export folders, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month