Recent changes in the delivery and financing of health

OUTCOMES IN PRACTICE Improving Physician Satisfaction on an Academic General Medical Service Robert C. Goldszer, MD, MBA, James S. Winshall, MD, Monte Brown, MD, Shelley Hurwitz, PhD, Nancy Lee Masaschi, MT, MBA, and Victor Dzau, MD Recent changes in the delivery and financing of health care have increased the pressures on attending physicians at academic medical centers in the United States. Traditionally, the salary of academic faculty has been based on the amount of revenue they generate through clinical activity and research. However, reimbursements for inpatient clinical activity are declining, and because reimbursement systems emphasize greater productivity, physicians are required to see more patients, leaving less time for teaching and increasing the pressure on attending physicians during both clinical and teaching activities [1,2]. Because attending physicians usually do not receive direct reimbursement for teaching, there is rarely sufficient payment to replace lost productivity during time spent teaching. Meanwhile, hospitals are looking for ways to decrease length of stay, while regulations are in place to control house staff and nursing workload [3]. Given these increased pressures on attending physicians, methods for monitoring and improving faculty satisfaction are essential. This article describes initiatives that were developed and implemented on the General Medical Service (GMS) at Brigham and Women s Hospital to meet a range of institutional goals, including improving attending physician satisfaction. Background Brigham and Women s Hospital is a 720-bed academic medical center located in Boston, MA, that provides primary, secondary, tertiary, and quartenary care to residents of Boston and the surrounding New England area. The GMS has a census of 70 to 90 patients. In 1999, the GMS was made up of 6 teams, each with 1 attending physician, 1 second- or thirdyear resident, 2 first-year residents, and 1 or 2 medical students. We were aware of attending physician evaluations at other academic medical centers and the changes these centers were trying to make to maintain quality of teaching and improve physician satisfaction. Wright and colleagues reported that the attributes of excellent attending physician role models included significant teaching experience, conducting rounds while teaching, stressing the importance of the doctor-patient relationship and psychosocial aspects of medicine, and being a chief resident [4]. Sheffield and colleagues described the need for allocating adequate time for scholarly pursuits [5]. New programs such as clinician-educator tracks have been established to fulfill the educational and clinical care missions of the academic medical center [6]. In light of these studies, we evaluated our own department. The evaluation was performed by department of medicine leadership, education leaders, and financial analysts. We reviewed opinions and feedback from house staff and attendings provided by residency directors. We reviewed our department budget, billings, and collections to assess sources of funding for paying teachers. This evaluation was done at a series of meetings with the department chair and clinical leaders. It was determined that the GMS attending physicians needed more orientation and feedback, specific goals, and a financial incentive for time spent teaching. Also, we determined that more focused leadership was needed to implement changes and be responsible for results. To address these needs and meet our institutional goals (eg, improving attending physician satisfaction, maintaining a high quality education program, recruiting the best house staff and faculty candidates, and improving patient satisfaction and efficiency), we implemented 8 new initiatives. A plan was formulated and reviewed by division chiefs and faculty, who provided feedback and suggestions. The initiatives were implemented between June and September 1999 by the vice chair for general medicine (RCG) and medical director for general medicine (JSW). Satisfaction and Quality Improvement Initiatives Hospitalist Program We placed 2 full-time hospitalists and 1 chief resident as the attending physicians on 3 of the 6 GMS teams. These attending physicians are responsible for patient care and teaching. The hospitalist teams care for the primary care physicians From the Department of Medicine and the Clinical Research Center, Brigham and Women s Hospital, Boston, MA. www.turner-white.com Vol. 9, No. 1 January 2002 JCOM 27

PHYSICIAN SATISFACTION patients, who have the option to assign their patients for inpatient care. The hospitalists also care for many patients referred by physicians outside Brigham and Women s Hospital and for patients without primary care physicians. This reorganization has resulted in several improvements. Because we have 3 rather than 6 teams to fill with faculty attendings, we have been able to select physicians who are highly committed to teaching and patient care and who demonstrate the best practices. In addition, the nonhospitalist team attendings have fewer direct patient care responsibilities and more time for teaching because the hospitalist attendings receive many admissions directly referred from primary physicians and care for many patients who do not have a physician. The nonhospitalist teams may have approximately 10 total patients, yet the faculty attending is responsible for only 3 to 5 patients; the patients primary care physician or specialist cares for the remaining patients assigned to these teams. Leadership Reorganization The department of medicine reorganized its leadership, adding a vice chair for general medicine. The vice chair is responsible for the attendings on the GMS, with special focus on quality of teaching, quality of care, patient satisfaction, and efficiency. The addition of the vice chair aligns teaching initiatives with patient care and places leadership responsibility in the chair s office. Weekly Meetings We instituted a weekly meeting with attending physicians run by the vice chair for general medicine or the medical director of the GMS. At these meetings, we review billing procedures, team functioning, goals of the medical service, teaching and feedback techniques, and announcements. The weekly meeting helps to orient the attendings and allows them to ask questions and provide weekly feedback to the GMS leadership on the functioning of the medical service. Prior to 1999, there were no regularly attended scheduled meetings for GMS attendings. Evaluation and Feedback A new process for evaluating attending physicians was established; the physicians are evaluated on teaching, patient care, providing feedback to the house staff, team processes, and providing assistance in difficult situations. The attendings see the evaluation form in advance (Figure 1), so they are well aware of the GMS goals. At the end of the rotation, the form is given by hand to the house staff team members, who complete it anonymously and send it to the vice chair for collating. (In the future this process will be carried out via the Internet.) The attendings receive written and/or verbal feedback on the evaluation. They each receive a letter from the vice chair that summarizes the responses from the house staff, provides point totals and comments, and reports areas for improvement. Prior to this process, house staff would send evaluations to the director of the residency program, who did not routinely provide specific feedback to the attendings. Afternoon Rounds We incorporated afternoon contact with house staff or rounds into the required schedule for the attending physicians. Though this originally consisted of phone contact with the resident to review the patients, we now rely on afternoon rounds and contact with the house staff and nurses on the floor. This intervention was designed to provide a care plan and teaching every 12 hours and to improve patient, family, nurse, and house staff communication through the review of the care plan twice daily. Chairman s Letter The letter that all attendings receive from the chairman of the department before starting their rotation was rewritten to focus on teaching, patient satisfaction, and working with the nurses. The letter describes the schedule (including the implementation of afternoon rounds or contact with house staff), goals (including quality improvement, teamwork, and patient satisfaction), and patient care and education responsibilities. The prior letter had focused on only the education and patient care responsibilities. 2-Week Service Rotations We gave attending physicians the option of spending 2 weeks rather than 4 weeks on service. Attendings had made a request for this change to the vice chair of general medicine, pointing out that such a change would help them accommodate their other responsibilities. Attendings felt they could be most focused on the GMS if they removed themselves from their other responsibilities for a shorter period of time. Two (3.4%) of 58 attendings requested 4-week rotations in 1999, while only 1 (1.7%) did so in 2000. In the academic year 2000, 42 (72%) of the faculty did only 2 weeks of attending, 14 (24%) did 2 2-week rotations, and 2 (3.4%) did more than 4 weeks of attending spread throughout the academic year. Compensation for Teaching We developed a financial model whereby we could pay attendings who did not receive a significant portion (> 30%) of their salary from the department. Using funds from clinical collections, the attendings were paid $2000 for a 2-week rotation (not considered part of their salary). By improving our billing and collections, we have increased the percentage of collections, making more funds available; we also supplement the attending stipends with money from department of medicine education funds. Attendings received this payment directly along with a letter indicating that the bonus is 28 JCOM January 2002 Vol. 9, No. 1 www.turner-white.com

OUTCOMES IN PRACTICE EVALUATION OF GENERAL MEDICINE ATTENDING Dear House Officer or Student, Attending on the General Medical Service has always been a privilege reserved for the most highly regarded teachers and clinicians in the Department of Medicine. We are committed to continuing this tradition of excellence, but need your assistance to identify those physicians who will best fulfill the role. Please take a few minutes to fill out this evaluation and return it to Bob Goldszer (BWH Interoffice) within one week. We will review your comments and share a summary with your attending. Specific feedback will be kept anonymous. Your evaluations will be considered as amongst the most important criteria for choosing the attending staff for next year. Thank you in advance for your assistance in this matter. ATTENDING NAME DATE TEAM Please circle the number that best reflects your assessment 1 = rarely, 2 = occasionally, 3 = consistently, NA = not assessed 1. Your attending was on time, reliable, and devoted adequate time to teaching and 1 2 3 NA clinical responsibilities 2. Your attending provided a positive teaching experience (in terms of preparation, 1 2 3 NA content, and style) 3. Your attending provided instruction in patient interviewing, physical diagnosis, and 1 2 3 NA data analysis 4. Your attending ensured that the highest quality care was provided to patients 1 2 3 NA 5. Your attending was supportive and available to help in difficult situations (with 1 2 3 NA patients, families, nursing staff, etc) 6. Your attending paid attention to the cultural diversity of the 1 2 3 NA staff and patients 7. Your attending communicated well with you, including one-on-one feedback 1 2 3 NA 8. Your attending communicated well with other members of the medical, nursing and 1 2 3 NA ancillary staff 9. Your attending contacted you twice a day to review your service 1 2 3 NA I recommend this physician to attend again next year on the General Medical Service Yes No Provisional (give details) What are your suggestions for improvements for this attending? Other comments or suggestions: Figure 1. Form used for evaluation of attending performance by the house staff. www.turner-white.com Vol. 9, No. 1 January 2002 JCOM 29

PHYSICIAN SATISFACTION GMS ATTENDING SURVEY NAME: Approximate date you were last on service: Please answer the following questions with regards to your most recent time as an attending on the GMS since July 1999. A = too little, B = just right, C = too much, NA = not applicable Time commitment per week A B C NA Number of weeks as Attending A B C NA Time spent on didactic teaching A B C NA Time spent on bedside teaching A B C NA Time spent with medical students A B C NA Time spent providing or overseeing patient care A B C NA Time spent on administrative duties A B C NA Please rate the statements below using the following scale: A = very satisfactory, B = satisfactory, C = unsatisfactory, NA = not applicable Your overall experience as a GMS Attending A B C NA Your interactions with house staff and students A B C NA Your interactions with nurses, care coordinators, etc. A B C NA Your ability to ensure quality care for patients A B C NA Your ability to address issues of patient satisfaction A B C NA Your ability to integrate your duties as GMS Attending with other A B C NA professional responsibilities Information (written and verbal) you received describing teaching A B C NA responsibilities, service goals, billing procedures, etc. Quality of feedback you received about your performance A B C NA Opportunities to give feedback to your trainees A B C NA Compensation you received for your work as GMS Attending A B C NA COMMENTS: PLEASE RETURN THIS FORM TO Robert C. Goldszer, MD, MBA; Brigham and Women s Hospital PB-4 Figure 2. Survey mailed to attending physicians to analyze their experience on the General Medical Service. compensation for their teaching time. Attendings who received significant salary support from the department (eg, vice chairs and program directors) did not receive this extra payment. Prior to initiating this program, there was no direct payment to the physician for time spent on the GMS. Impact of Changes We used a mailed survey to measure the attending physicians perception of the initiatives and satisfaction with teaching (Figure 2). The survey was sent and returned via regular mail. We surveyed 58 attending faculty physicians in 30 JCOM January 2002 Vol. 9, No. 1 www.turner-white.com

OUTCOMES IN PRACTICE July 1999 to gather baseline information and then repeated the survey in March 2000 to assess the impact of the changes. The attendings who were surveyed had worked throughout the year, so the physicians in each group had comparable experience with the teaching service. A team consisting of the vice chair for general medicine (RCG), the medical director for general medicine (JSW), and a representative of quality measurement (NM) performed the data compilation and review of the completed surveys. Statistical analyses were supported by the Biostatistical Consulting Service, Center for Clinical Investigation, Brigham and Women s Hospital. Survey results were analyzed using statistical methods for discrete data with correlated responses, since some physicians responded in both years [7]. The categorical response distributions for the time allocation items were dichotomized, and the analysis assumed the binomial distribution with the logit link function. The just right category was compared with the other categories combined. The response distribution for the satisfaction items was assumed to be ordinal categorical levels of amount of satisfaction, and the multinomial distribution with the cumulative logit link function was used. Survey Results Forty of 58 attendings responded to the survey in 1999, and 42 of 58 responded in 2000; 16 responded in both years. The proportion of physicians who felt that the time commitment per week was just right rose significantly from 75% to 95% (P = 0.01) (Table 1). Responses to the remaining questionnaire items in Table 1 did not change significantly from baseline. The proportion of attendings who responded just right was approximately constant for the time spent with medical students (P = 0.40), bedside teaching (P = 0.83), and providing patient care (P = 0.58). There were a number of areas in which physicians satisfaction with the medical service increased, including quality of information received, quality of feedback, and compensation received (Table 2). More importantly, satisfaction with overall experience as an attending increased significantly (P = 0.01). Although nearly all of the respondents were at least satisfied with their overall experience in both years, the proportion who were very satisfied rose from 47% to 76%. The proportion of those at least satisfied with compensation tripled from 31% to 94%. The level of satisfaction did not change significantly for interactions with house staff, interactions with nurses, ability to address patient satisfaction issues, ability to integrate duties, and opportunities to give feedback. Table 1. Percentage of Just Right Responses to Time Allocation Items 1999 2000 Group Group Item (n = 40) (n = 42) P Value Time commitment per week 75 95 0.01 Number of weeks 60 74 0.17 Time spent on didactic teaching 90 92 0.72 Time spent on bedside teaching 35 38 0.83 Time spent with medical students 62 54 0.40 Time spent on patient care 83 81 0.58 Time spent on administrative duties 86 90 0.63 Discussion We introduced 8 major initiatives to the GMS designed to achieve improved patient care, higher patient and staff satisfaction, and continued excellence in teaching. Attending physicians responses to these initiatives, as measured by questionnaire, were generally positive, with the greatest improvements in satisfaction occurring in regard to availability of information and compensation. These programs are ongoing at Brigham and Women s Hospital and will continue to be evaluated for needed improvements. Evaluations from house staff and students reveal general satisfaction with the current programs, although some house staff and students have expressed the need for more time with the attending physician. We take all house staff criticisms very seriously and discuss them with the attending physician. If no improvement is demonstrated, the physician is not asked to return as attending. The applicability of our findings is limited by several factors. First, the initiatives were implemented at only 1 academic medical institution, so the results may not be similar at other institutions or in other settings. Second, the number of physicians involved in the program and surveyed was small. Only 16 physicians responded in both years; thus, this study is not strictly longitudinal but rather it describes the level of satisfaction in each year s group of physicians. Finally, we implemented the 8 initiatives as a block, so we cannot determine if any had a greater impact than others. We conclude that the interventions implemented have had a positive impact on attendings satisfaction with their teaching and clinical activities. It remains to be demonstrated whether this higher provider satisfaction may be associated with improved patient outcomes. www.turner-white.com Vol. 9, No. 1 January 2002 JCOM 31

PHYSICIAN SATISFACTION Table 2. Percentage of Satisfied and Very Satisfied Responses in 1999 and 2000 1999 2000 Item Very Satisfied Satisfied Very Satisfied Satisfied P Value* Overall experience 47 53 76 21 0.0100 Interactions with house staff and students 72 28 81 19 0.3700 Interactions with nurses and care coordinators 33 59 41 54 0.3600 Ability to ensure quality care 30 65 57 38 0.0200 Ability to address patient satisfaction 21 61 26 63 0.3900 Ability to integrate duties 13 46 12 62 0.2600 Quality of information received 18 57 54 44 0.0001 Quality of feedback received 09 06 26 37 0.0010 Opportunities to give feedback 40 53 37 63 0.9000 Compensation received 07 24 35 59 0.0001 *Compares amount of satisfaction (very satisfied, satisfied, unsatisfied) between 1999 and 2000 attendings. Corresponding author: Robert C. Goldszer, MD, MBA, Dept. of Internal Medicine, Brigham and Women s Hospital, 75 Francis St. PB 4, Boston, MA 02115; rgoldszer@partners.org. References 1. Shea S, Nickerson KG, Tenenbaum J, et al. Compensation to a department of medicine and its faculty members for the teaching of medical students and house staff. N Engl J Med 1996;334:162 7. 2. Kuttner R. Managed care and medical education. N Engl J Med 1999;341:1092 6. 3. Iglehart JK. The American health care system. Teaching hospitals. N Engl J Med 1993;329:1052 6. 4. Wright SM, Kern DE, Kolodner K, et al. Attributes of excellent attending physician role models. N Engl J Med 1998; 339:1986 93. 5. Sheffield JV, Wipf JE, Buchwald D. Work activities of clinician-educators. J Gen Intern Med 1998;13:406 9. 6. Levinson W, Rubenstein A. Mission critical integrating clinician-educators into academic medical centers. N Engl J Med 1999;341:840 3. 7. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika 1986;73:13 22. Copyright 2002 by Turner White Communications Inc., Wayne, PA. All rights reserved. 32 JCOM January 2002 Vol. 9, No. 1 www.turner-white.com