Global Payment Contract Attitudes and Comprehension Among Internal Medicine Physicians

Joshua Allen-Dicker, MD, MPH; Shoshana J. Herzig, MD, MPH; and Russell Kerbel, MD, MBA

Recent health reform efforts center on developing market incentives to promote high-value care.1-4 In this environment, global payment contracts (GPCs) are increasingly common. GPCs rely on the sharing of responsibility for cost and quality of care between insurance payers and healthcare providers.5,6 This approach incorporates aspects of risk adjustment, capitation, and pay-for-performance.
Physicians have been identified as being a potential obstacle to successful implementation of such recent payment reforms.7-9 Proposed reasons for this challenge include perceived threat to physician autonomy, uncertain financial benefit, physician averseness to risk, and residual antagonism following the managed care organization disputes of the 1990s.5,10,11 We were not able to locate any publications on the role physicians are playing in the organizational change required for GPC adoption.
Through a survey of general medicine and internal medicine subspecialist physicians in a care organization engaged in multiple GPCs, we aimed to quantify internist support for GPCs, quantify internist comprehension of key information relevant to GPCs, and identify specific physician attributes that might predict GPC support.
Setting and Subjects
Beth Israel Deaconess Care Organization (BIDCO) is an independent physician and hospital network in Massachusetts. At the time of this survey, it consisted of several hospitals as well as physician organizations accounting for 1700 physician providers. Of these, the Department of Medicine at Beth Israel Deaconess Medical Center in Boston, Massachusetts, accounted for over 500 physician providers, the majority of whom were paid a salary not based on productivity.
Since 2009, BIDCO has formed GPCs with multiple payer organizations. Existing global payment partners now include Massachusetts Blue Cross Blue Shield via the Alternative Quality Contract (announced 2011), Tufts Health Plan (2011), CMS via the Pioneer Accountable Care Organization program (2012), and Harvard Pilgrim Health Care (2012). At the time of this survey, approximately 60% of patients cared for by BIDCO providers were covered by a GPC.
Prior to and immediately following GPC adoption, the hospital and care organization held global payment orientation sessions throughout the provider network. These educational sessions included formal presentations at divisionwide faculty meetings, and more informal presentations at local practice-based business meetings. Select primary care physicians were identified as champions responsible for communicating important global contract details to local primary care physician colleagues. At the time of our survey, no complementary structure existed for other physician specialties. As physician salary was determined by individual affiliated employers and not the care organization itself, GPC adoption did not directly alter physician compensation.
Ongoing organizational outreach efforts include regular distribution of data on utilization, cost, and quality, as well as medical management initiatives to assist providers in achieving the highest quality and efficiency in healthcare delivery within the system.
Survey Design
We designed a 3-part anonymous survey instrument (eAppendix A, available at The first section collected demographic information, including time since completion of residency, clinical area of practice, time spent on a clinical service, and sources previously consulted to learn about GPCs.
The second section consisted of 10 questions meant to assess comprehension of key concepts related to GPCs. Question topics included definitions and goals, mechanism for billing, role of budgets, role of quality metrics, and organizational-specific information. Question content was developed in conjunction with organizational leadership to ensure accuracy and appropriateness. Respondents were asked not to consult any outside sources while completing this section.
The third section quantified the agreement with the decision to participate in a GPC. The survey asked physicians to rate their level of agreement with the statement, “I support our physician organization’s decision to enter into global payment systems (eg, Medicare Accountable Care Organization, Blue Cross Blue Shield Alternative Quality Contract),” using a Likert scale (strongly agree, agree, disagree, strongly disagree). Depending on physician response, an additional question soliciting reasons for agreement or disagreement was included. Respondents could select multiple predefined answer choices that were sourced from the ongoing dialogue on GPCs in the lay and academic press. A free-text “other” response was allowed, as was no response.
Prior to formal survey administration, 2 focus groups were held that included a total of 8 internal medicine physicians who would not be included in the survey population. In a roundtable format, the survey was assessed for length and each individual question was assessed for reliability of physician understanding.
Survey Administration
In June 2013, e-mails containing a unique link to our survey were sent to all physicians within the Department of Medicine at Beth Israel Deaconess Medical Center, which includes general medicine and internal medicine subspecialist physicians. Contact information was obtained from administrative rosters. Over the subsequent month, 2 reminder e-mails were sent to physicians who had not already completed the survey. All e-mails originated from study authors and not departmental or care organization leadership. The institution’s Committee on Clinical Investigations certified this research protocol as exempt.
Statistical Analysis
Summary statistics were calculated for all variables. General medicine physicians were defined as those who reported practicing primary care, hospital medicine, or geriatric medicine. Specialist physicians were defined as those who reported practicing any other specialty within internal medicine. Time since residency was dichotomized into less than and greater than or equal to 20 years (the time elapsed since the last attempt at national healthcare reform).12
For the 10 GPC knowledge questions, each respondent was given a score based on the percentage of questions they answered correctly.
We used the Fisher’s exact test to assess the bivariable associations between physician characteristics and level of support for GPC involvement as measured via our 4-category response variable (strongly agree, agree, disagree, strongly disagree). We used a multivariate ordinal logistic regression model to derive factors independently associated with level of support for GPC involvement, where our 4-category response variable (strongly disagree, disagree, agree, strongly agree) was the dependent variable, and all physician characteristics were simultaneously included as independent variables. We chose ordinal regression to make full use of our ordered data, maximizing statistical power. All statistical analyses were performed with the JMP Pro 10 statistical software package (SAS Institute, Cary, North Carolina).
Internist Demographics

Of the 569 Department of Medicine physicians contacted, 281 (49.4%) completed our survey (Table 1). Of all respondents, 52% identified themselves as general medicine physicians, while 48% identified themselves as specialist physicians. Respondents were representative of the total survey population with respect to the proportion of generalists (52% vs 47.4%; P = .22). Forty-two percent of respondents had completed their residency training more than 20 years prior, and 150 physicians (53.4%) reported spending more than half their time working on a clinical service.
Two hundred and thirty-one (82.3%) respondents reported being previously exposed to educational information sources on GPCs. The most common sources of information were informal discussions with other healthcare providers, news media, and academic journals (eAppendix B).
Internist Comprehension
Median score on the knowledge portion of the survey was 70% (25th quartile = 60%; 75th quartile = 80%; mean = 66.2%; SD = 13.7). Physicians who scored equal to or greater than the 50th percentile were significantly more likely to report previously consulting any educational source about GPCs than those who scored below the 50th percentile (58.4% vs 41.6%; P <.001).
Physician Attitudes
Two hundred and thirty-eight (84.7%) respondents reported agreeing or strongly agreeing with the care organization’s decision to enter into GPCs. Of those, 219 respondents (92%) selected at least 1 of the predefined reasons for this decision. Each respondent selected an average of 2.9 predefined reasons (25th quartile, 1 response; 75th quartile, 4 responses) The most popular reasons were a) perceived reduction in the cost of patient care and b) perceived increased competitiveness in the healthcare market (Table 2a). The least common reason was perceived personal financial benefit.
Of those who disagreed or strongly disagreed with the decision to enter into GPCs, 39 respondents (90.7%) selected at least 1 of the predefined options. Each respondent selected an average of 3.7 predefined responses (25th quartile = 2 responses; 75th quartile = 5 responses). The most popular reasons were a) perceived failure to improve the quality of patient health and b) failure to improve the quality of the physician experience (Table 2b). The least common reason was perceived failure to improve competitiveness in the healthcare market.
Associations Between Physician Characteristics and Level of GPC Support
See Table 1 for bivariable and multivariable associations between physician characteristics and level of GPC support. In our multivariate ordinal logistic regression model, variables independently associated with higher likelihood of supporting involvement in GPCs included scoring at or above the 50th percentile in the knowledge assessment (P = .01); reporting prior exposure to informational sources about GPCs (P = .01); working less than or equal to 50% clinical full-time equivalents (<.001); and completion of residency more than 20 years ago (<.01). There was no association between status as a generalist versus a specialist and likelihood of support for GPC involvement.
Four years since first engaging in a GPC, a majority of Department of Medicine physician survey respondents within our care organization were supportive of this decision. We identified several physician attributes independently associated with GPC support, including greater knowledge of GPCs. To our knowledge, our study represents the first published survey of academic physician attitudes on this topic.
Prior industry surveys and analyses have identified physicians as potential barriers to implementation of organizational reform.5,7 In contrast, our study of academic internists demonstrates notable support for a model that requires significant change on an organizational level. This could be partially due to the presence of financial and nonfinancial incentives in academic medical centers that can insulate practitioners from market forces.13,14 While this has previously been depicted as a challenge, based on our results we propose that it may also serve as a strength in facilitating organizational change.15 Salaried physicians with academic appointments may worry less about risk-sharing agreements, as a majority of the immediate risk is carried by the physician organization or hospital. As a result, academic physicians may be more open to the idea of organizational restructuring required by GPCs, as well as issues such as cost control and quality of care (cited by 69.8% and 47.1% of respondents in support of GPC involvement). The degree to which our findings are generalizable to nonacademic settings with productivity-based compensation should be investigated.
Despite public views of physicians as being focused on financial concerns, physicians in our survey infrequently cited (11.3%) personal monetary benefit as playing a role in their decision to support or oppose involvement in a GPC.16,17 Furthermore, despite documented “wage gaps” between generalist and specialist physicians, there was no association between generalist versus specialist status and GPC support.18
Understanding predictors of internist support for GPCs within our care organization may help other academic health systems approach organizational change. We found that demonstrated knowledge about GPCs and a history of educational exposure to GPCs, regardless of format, was associated with an increased likelihood of supporting involvement in GPCs. These findings suggest that physician education efforts, whether formal or informal, may be effective tools for shaping physician opinion on an organizational level.
The likelihood of supporting GPCs was positively associated with increased time since completion of residency and decreased clinical time commitment. We hypothesize that older academic internists and those who spend less time working in clinical activities may be more involved in administrative positions that allow them a better understanding of the organizational issues that motivate GPC involvement.
Our study has several limitations. While our study sample was large enough to allow for identification of associations within our organization, the single-center nature of our study may limit generalizations beyond our organization. In particular, the academic nature and high proportion of salaried physicians within the Department of Medicine limits generalizability to settings with a higher proportion of nonsalaried physicians, where GPCs may have greater financial ramifications for individual physicians. Additionally, although physician beliefs regarding certain recent health reform efforts were similar across regions, support for GPCs may depend on geography and local political leanings.19,20 Within our organization, located in urban Massachusetts, political support for health reform was a frequently cited reason for support of GPCs. Although our survey was sent by study authors and not departmental or care organization leadership, and participants were informed that responses were anonymous, participants may have overestimated their support for GPC involvement due to perceived organizational norms. Additionally, it is possible that survey respondents differed from nonrespondents in ways that could have affected the results. Regarding our attempt to identify potentially modifiable predictors of opinion, the cross-sectional nature of our investigation limits our ability to demonstrate causation. While we found that physicians with exposure to informational sources on GPCs are most likely to support GPCs, it may be that physicians who support GPCs are more likely to seek out informational sessions. Additionally, despite our use of focus groups as part of the survey design process, it is possible that our questions were not interpreted as intended, or were interpreted differently between respondents. As we did not survey physicians in other departments, we are unable to draw conclusions about physicians outside the Department of Medicine.
In conclusion, in a survey of internal medicine physicians within an organization with more than 4 years of experience with GPCs, there was overwhelming support for this involvement. Greater knowledge regarding GPCs and exposure to informational sources about GPCs correlated with higher levels of support for GPC involvement. These findings suggest potential targets for increasing physician support for GPC involvement.

The authors would like to acknowledge the leadership of the Beth Israel Deaconess Medical Center Department of Medicine for their support. The statements contained in this document are solely those of the authors. The authors assume responsibility for the accuracy and completeness of the information contained in this document.
Print | AJMC Printing...