Salaries, incentives, fairness: rethinking the structure of primary care physician compensation Jennifer Lochner, MD Beth Potter, MD Brian Arndt, MD

Disclosures We have no conflicts of interest to disclose (except that we are paid via the compensation plan we are about to discuss)

Learning Objectives 1.Delineate the pros and cons of including different measures as part of a physician compensation plan 2.Consider local forces such as payer mix and culture when modifying a physician compensation plan 3.Describe theories of motivation for professionals to help guide the creation of physician compensation plans

A story of change in our community, how we modified our compensation plan in response to this change and some outcomes to date.

This report includes 14 DFMCH clinics in Dane County, WI: 4 residency training sites (32 faculty physicians) and 10 community clinics (52 faculty physicians)

Dane County, WI in 2010 Significant attrition of DFMCH physicians as local competitors build primary care systems in response to the passage of the ACA Our department’s physician compensation suddenly found to be lower than local market Survey data showed DFMCH physician satisfaction with amount and structure of compensation was found to be low Clinical compensation was almost entirely based on RVUs

Revenue sources for our clinics CapitationFee for service %65.3% %59.5%

UW Health Primary Care Job Description Formalized across primary care disciplines Emphasis on team based care including panel management and population health as well as caring for individual patients Defines work week as 27 hours face to face care, 13 hours non face to face care

Process DFMCH Compensation committee, working with UW Health, crafted two new compensation plans, one for residency faculty and one for community faculty – representatives from each group participated Conversations focused on mission and culture of the two groups, where they aligned and where they were different

Key changes to the clinical compensation plan Primarily based on panel size rather than RVU productivity – for 1.0 clinical FTE target panel size was set at 1800 Link to national benchmarks for FM clinical compensation Maintain some aspect of RVU productivity in the calculation For residency faculty minimize or remove any incentive to move patients onto faculty panels

80% panel-based compensation Individual clinical FTE * Total clinical FTE for all residency faculty 20% RVU- based compensation Individual RVU productivity Total RVU productivity for all residency faculty Residency Faculty Clinical Compensation National benchmark family medicine salary Target panel size Total panel size for all residency clinics × $220,000 1,800 patients Residency Compensation Pool = $3,911, FTE patient care FTE precepting 18 FTE across 4 clinics × 80% of compensation pool 32,000 patients across 4 clinics Total compensation for residency physician with 0.6 clinical FTE = $133,629 3,000 personal RVUs 80,000 RVUs across 4 clinics × 20% of compensation pool 50% panel-based compensation Individual panel size Total panel size at clinic 50% RVU-based compensation Individual RVU productivity Total RVU productivity at clinic Community Faculty Clinical Compensation National benchmark family medicine salary Target panel size Individual clinic panel size × $220,000 1,800 patients Community Compensation Pool = $855,556 2,000 patients 7,000 patients at clinic × 50% of compensation pool 7,000 patients at clinic Total compensation for community physician with 1.0 FTE and a panel size of 2,000 patients = $217,284 4,000 personal RVUs 18,000 RVUs at clinic × 50% of compensation pool

Compensation Plan Component ResidencyCommunityRationale Pool of dollars available for clinical compensation Based on panel size of 4 clinics pooled together Based on panel size for each individual clinic Residency clinics view themselves as one group practice, sharing responsibility for caring for the entire panel together; community clinics view themselves as separate practices Panel size and allocation 80% of salary is based on panel size Panel is allocated based on clinical FTE rather than actual individual faculty panel size 50% of salary is based on panel size Panel is allocated based on individual physician panel size Using clinical FTE (which includes precepting time) to assign panel size to residency faculty allows residents’ patients to be allocated across the faculty based on the amount of time a physician spends in the clinic caring for those patients. In the community clinics, physicians are credited for their personal panel. RVU production20% of salary is based on RVU productivity 50% of salary is based on RVU productivity Residency faculty wanted to de- emphasize RVU productivity and instead focus on panel management as the key determinant of compensation while community physicians wanted to value panel size and RVU productivity equally.

Outcomes - RVUs - Panel size - Physician satisfaction - Cost and quality

Trends in RVU per clinical FTE 2011 RVUs per FTE 2012 RVUs per FTE 2013 RVUs per FTE 2014 RVUs per FTE RVU variance National FM RVU benchmark b % Residency clinics % Residency clinics percentage of FM benchmark Community clinics % Community clinics percentage of FM benchmark a New compensation plan started January 2013, retroactive to July 2012 b National family medicine Relative Value Unit benchmarks were obtained from data from the Medical Group Management Association, the American Medical Group Association, and McGladrey & Pullen

Trends in panel size January 2012 January 2013 January 2014 January 2015 Community Panel size72,68175,01278,26182,632 Clinical FTE Panel size per 1.0 FTE Residency Panel size33,47533,58833,60834,321 Clinical FTE Panel size per 1.0 FTE a New compensation plan started January 2013, retroactive to July 2012 b This includes physician clinical FTE, not including resident, fellow and APP FTE. For faculty this number reflected both direct patient care and resident precepting.

Physician satisfaction - follow up survey on compensation plan

Satisfaction with salary received Group (Year) Very dissatisfied (%) Dissatis- fied (%) Neutral (%) Satisfied (%) Very satis- fied (%) Community (pre) n=42 5 (12)17 (41)9 (21)10 (24)1 (2) Community (post) n=31 0 (0)2 (7) 12 (39)15 (48)p<0.001 Residency (pre) n=33 2 (6)15 (45)5 (16)11 (33)0 (0) Residency (post) n=19 1 (5) 4 (21)8 (42)5 (26)p=0.001

Satisfaction with plan structure Group (Year) Very dissatisfied (%) Dissatis- fied (%) Neutral (%) Satisfied (%) Very satisfied (%) p-value Community (pre) n=42 7 (17)13 (31)8 (19)11 (26)3 (7) Community (post) n=31 4 (13)3 (10)1 (3)14 (45)9 (29)p=0.003 Residency (pre) n=33 4 (12)17 (51)6 (18) 0 (0) Residency (post) n=19 1 (5)3 (16)6 (32)7 ( 37)2 (11)p=0.004

Perception of salary equity Group (Year) Very unfair (%) Unfair(%) Neutral (%) Fair(%) Very unfair(%) p-value Community (pre) n=42 9 (27)18 (55)2 (6)3 (9)1 (3) Community (post) n=31 0 (0)3 (16)7 (37)4 (21)5 (26)p<0.001 Residency (pre) n=33 8 (19)15 (36)7 (17)10 (24)2 (5) Residency (post) n=19 3 (10)4 (13)13 (42)3 (10)8 (26)p=0.022

Cost and quality

Considerations for the future Are we incentivizing large panels to whom we cannot offer needed access? Will patient satisfaction and quality be maintained? Will physicians narrow their scope of practice to allow them to serve a larger panel? How to structure physician work day – even more time for non face to face care? Adequate panel weighting is needed – current system includes only age, sex and insurance status Organizational support for this care model and compensation model

Also… This does not address other important compensation issues such as inpatient call coverage, maternity care coverage, teaching and academic work

Please evaluate this presentation using the conference mobile app! Simply click on the "clipboard" icon on the presentation page.