Rural and Urban Primary Care Physician Professional Beliefs and Quality Improvement Behaviors

Authors

  • Anne C. Kirchhoff PhD, MPH,

    Corresponding author
    1. Department of Pediatrics, University of Utah School of Medicine, Salt Lake City, Utah
    2. Cancer Control and Population Sciences, Huntsman Cancer Institute, Salt Lake City, Utah
    • For further information, contact: Anne Kirchhoff, PhD, MPH, Assistant Professor, Pediatric Hematology/Oncology, 2000 Circle of Hope, Salt Lake City, UT 84103; e-mail: anne.kirchhoff@hci.utah.edu.

    Search for more papers by this author
  • Gary Hart PhD,

    1. Center for Rural Health, University of North Dakota School of Medicine and Health Sciences, Grand Forks, North Dakota
    Search for more papers by this author
  • Eric G. Campbell PhD

    1. Mongan Institute for Health Policy, Massachusetts General Hospital, Boston, Massachusetts
    2. Department of Medicine, Harvard Medical School, Boston, Massachusetts
    Search for more papers by this author

  • Disclosures: The authors report no conflicts of interest.

  • Funding: This study was supported by a grant from the Institute on Medicine as a Profession at Columbia University.

  • Acknowledgments: The authors thank Julia Bodson, Gina Nam, Echo Warner, and Daniel St. Hilaire for their assistance with data management and literature review during the writing of this manuscript.

Abstract

Purpose

We evaluated whether primary care physicians (PCPs) from urban and rural practices differ on attitudes and behaviors related to quality improvement (QI) activities, patient relationships, and professionalism/self-regulation.

Methods

Data from a national survey that assessed physician attitudes and behaviors based on the Physician Charter on Medical Professionalism were used. Of the 1,891 survey respondents, N = 840 were PCPs (n = 274 family medicine (response rate = 67.5%); n = 257 general internal medicine (60.8%); and n = 309 pediatricians (72.7%)). Using Rural-Urban Commuting Area (RUCA) codes, PCPs were classified as urban and rural according to their practice ZIP code.

Findings

A total of n = 691 physicians were urban and n = 127 rural. Attitudes regarding participating in QI did not differ by practice location; however, rural PCPs were more likely to have reviewed an other physician's records for QI than urban PCPs (65.6% vs 48.0%, P < .001). Rural physicians were more likely to agree that physicians should talk with their patients about the cost of care than urban PCPs (40.5% vs 29.2%, P = .02). While all PCPs endorsed attitudes regarding the importance of professional behaviors (eg, reporting impaired/incompetent colleagues, disclosing medical errors) at generally similar levels, their behaviors differed. More rural physicians had a personal knowledge of an impaired/incompetent physician than urban physicians (20.7% vs 12.7%, P = .02).

Conclusions

PCPs from rural and urban areas share similar attitudes regarding the importance of participating in QI and fulfilling professional responsibilities. However, certain behaviors (eg, knowledge of impaired colleagues) do differ. These results should be confirmed in larger studies of rural PCPs.

During the past decade, physician groups have increasingly focused on issues of physician professionalism as an avenue for improving patient care. In 2002, the seminal Physician Charter on Medical Professionalism developed core professional responsibilities for physicians that included commitment to professional competence and responsibilities, commitment to improving quality of and access to care, and dedication to maintaining patient confidentiality and appropriate relationships with patients.[1] In rural areas, however, patients may bypass local primary care providers (PCPs) due to perceptions that they provide a lower quality of care to their patients than do their urban counterparts.[2-4] Yet, studies that have analyzed quality of care indicators reveal that patients in rural areas fare as well—and sometimes better—than those in urban areas.[5]

Some of the negative perception is attributable to differences that exist between urban and rural PCPs regarding their practice settings and patient relationships. By definition, rural physicians practice in smaller communities where there are typically fewer economic and institutional resources, so they may have fewer opportunities to engage in training or quality improvement (QI) activities.[6, 7] Conversely, rural PCPs are often part of small practices, hospitals and communities rather than larger organizations and hospitals,[8, 9] meaning that opportunities for participating in QI activities may be greater in rural communities because the pool of local physicians is much smaller. As such, rural physicians’ attitudes and preparedness for engaging in QI may actually be better than their urban counterparts, although this has not been addressed in the literature to date.

Patient relationships are also different for rural providers. Often, a rural PCP may be the only provider in his or her community or one of a small number of providers.[10] This means there is potentially less anonymity for patients and physicians and thus a greater potential for dual relationships—where a physician engages in both personal and professional relationships with his or her patients.[11-14] While rural physicians may feel that their personal connections with their community and practice are rewarding,[12, 15] it may be more challenging for them to maintain appropriate professional boundaries. Patient expectations related to health care may differ as well because of personal relationships with their rural physicians. In rural primary care, the communication between physicians and patients has been described as “socioemotional” and care-oriented, rather than “instrumental” and cure-oriented in urban and suburban care.[11] This could affect not only patient-provider expectations and behaviors, but also health outcomes. Rural PCPs have the advantage of better understanding their patients because of the knowledge of these patients from outside the clinic walls and within the context of the community.

Despite these documented practice and patient-related differences between rural and urban physicians, it is unknown whether rural PCPs’ attitudes and behaviors differ on key components of medical practice, such as quality improvement activities, patient relationships and responsibilities, and professionalism/self-regulation. Using a national sample, we surveyed practicing PCPs regarding their attitudes and behaviors related to professional activities as endorsed by the Physician Charter on Medical Professionalism[1] to determine differences and similarities between rural and urban PCPs that may affect patient care.

Methods

Survey Development and Testing

The current survey was developed from the professionalism questionnaire that was administered in 2004.[16] The 2004 survey included questions addressing a variety of topics from the domains of the Physician Charter on Medical Professionalism.[1] For the 2009 survey, we revised the 2004 questions that did not perform adequately for respondent discrimination (ie, had ceiling effects whereby almost all physicians agreed with a given statement).[17, 18] We based our revisions on 4 focus groups with a total of 40 physicians, along with input from an interdisciplinary expert advisory board composed of experts on professionalism. The survey was rigorously pretested with 21 practicing physicians and results between the 2004 and 2009 survey were consistent, indicating response reliability. The final questionnaire was 8 pages long with 110 individual survey items. The survey is available by request. All study activities were approved by the Massachusetts General Hospital institutional review board.

Survey Administration

Using the AMA 2008 MasterFile, we identified all US physicians in primary care (family medicine, general internal medicine, and general pediatrics) and 4 nonprimary care specialties (anesthesiology, cardiology, general surgery, and psychiatry). From this group we excluded osteopathic physicians because of major differences in training that could impact professionalism and because of concerns about their representation in the AMA MasterFile. We also excluded resident physicians and physicians in federally owned hospitals; those with no address; those who requested not to be contacted; and those who were retired. From the remaining pool of eligible participants, we randomly selected 500 physicians from each of the 7 specialties (total sample n = 3,500).

The Center for Survey Research (CSR) at the University of Massachusetts, Boston, administered the survey in May 2009. CSR sent sampled physicians the survey packet by Priority Mail. The packet contained a cover letter; fact sheet; questionnaire, with a sticker on the back containing a random subject identification number; postage-paid return envelope; and a $20 incentive. The CSR telephoned nonrespondents to solicit participation and 2 additional mailings were sent to all nonrespondents.

Of the 3,500 sampled physicians, 562 were ineligible because they were deceased, out of the country, practicing a nonsampled specialty, on leave, or not actively practicing. Of the remaining 2,938 eligible physicians, 1,891 completed the survey (a 64.4% overall response rate). Because we were interested in primary care and because very few specialists practice in rural areas, for the current study we limited our analyses to PCPs (family medicine, general internal medicine, and general pediatrics). Of the 1,891 respondents to the survey, 840 were PCPs. Response rates by specialty were: family medicine n = 274/406 (67.5%); general internal medicine n = 257/423 (60.8%); and general pediatrics n = 309/425 (72.7%).

Analytic Group

Rural-Urban Commuting Area (RUCA) codes are a ZIP code-based geographic taxonomy that utilizes the standard Bureau of Census Urban Area and Urban Cluster definitions in combination with Census work commuting flows to characterize all of the nation's ZIP code areas regarding their rural and urban status, functional relationships, and focal settlement nodes. RUCA codes are based on population size and density, and primary and secondary worker commuter flow patterns to larger urban areas and clusters. The RUCAs were collaboratively developed and are disseminated with resources from the Department of Agriculture's Economic Research Service and the federal Office of Rural Health Policy.[19, 20]

To determine participants’ primary practice location type, we linked the most recent 2004 RUCA codes to our survey data using primary practice ZIP codes. The 33 RUCA codes were aggregated into a 4-level categorization: (1) Urban (codes 1.0, 1.1, 2.0, 2.1, 3.0, 4.1, 5.1, 7.1, 8.1, and 10.1 indicating metropolitan area core: primary commuter flow within an Urbanized Area [UA] of +50,000), (2) Large Rural (codes 4.0, 4.2, 5.0, 5.2, 6.0, and 6.1 indicating primary commuter flow within an Urban Cluster [UC] of 10,000-49,999), (3) Small Rural (codes 7.0, 7.2, 7.3, 7.4, 8.0, 8.2, 8.3, 8.4, 9.0, 9.1, and 9.2 indicating primary commuter flow within an UC of 2,500-9,999), and (4) Isolated (codes 10.0, 10.2, 10.3, 10.4, 10.5, and 10.6 indicating primary commuter flow to a tract outside UA or UC).[19, 20] Because of response number limitations, the 3 rural categories were combined into 1 category for our primary analyses (ie, urban vs rural).[19] Additionally, as there were certain differences in our outcomes within the rural category among the large rural and small/isolated rural groups, we also report relevant differences by 3-level practice location (ie, urban, large rural, and small/isolated rural). A total of 22 of the 840 PCP respondents (2.6%) did not supply usable ZIP code information for determination of RUCA codes and were excluded, leaving 818 participants for the analyses.

Dependent Variables

The survey included questions regarding physicians’ attitudes and behaviors related to professional activities, such as quality improvement efforts, and professionalism/self-regulation. As this was an exploratory analysis, we first examined the distribution of all measures of interest by rural/urban status. Then, we created binary outcome variables. As we have done in our past work, we dichotomized our responses to indicate complete endorsement of the attitude or behavior (eg, “1 = completely agree” and “0 = somewhat agree/somewhat disagree/completely disagree”), focusing on the extremes of the response categories as the attitudes and behaviors of interest are considered core norms of professionalism.[16-18]

Statistical Analyses

All analyses were weighted using Stata statistical software (StataCorp, College Station, Texas, USA), incorporating Stata's survey commands.[21] The data were weighted to reflect the inverse probability of sampling and response rates by survey strata (physician specialty). First, we examined univariate and bivariate relationships in the data by RUCA groups. To test for differences between groups, 2-sided t tests (continuous variables) or chi-squared tests (categorical variables) were applied as appropriate. Then, we examined differences in the association of outcomes by RUCA location (urban vs rural) by calculating unadjusted and adjusted percentages. Adjusted percentages were generated from logistic regressions using the svypxcat command in Stata. Covariates were limited to gender, physician specialty (ie, general internal medicine, family medicine, and general pediatrics) and medical school faculty status as these factors were thought to be associated with practice location rather than an outcome of practice location (eg, clinical hours worked per week). As our adjusted percentages were very similar to the unadjusted, we report the unadjusted percentages as our primary findings and indicate only where the adjusted estimates differ in statistical significance.

Results

Of the PCP respondents, 84.5% practiced in urban locations, 9.3% in large rural, and 6.2% in small/isolated rural areas. Table 1 displays respondent characteristics by practice location. Fewer large rural and small/isolated rural PCPs were female (31.1% and 26.0%, respectively) compared to urban (41.1%; P = .04). Family medicine was more common in large rural (45.1%) and small/isolated rural practices (64.1%) than among urban practices (29.1%; P < .001). A total of 32.2% of urban and 23.0% of large rural respondents held medical school appointments compared to 16.2% of their small/isolated rural counterparts (P = .02). Large rural and small/isolated rural physicians also differed from urban physicians in self-reported patient mix and work hours. Urban physicians had fewer patients on Medicare (27.5% of their patients on average) compared to large rural (33.1%) and small/isolated rural (36.1%; P < .01). Similarly, PCPs in rural areas had more Medicaid patients (26.6% of large rural and 23.9% of small/isolated rural vs 18.7% of urban; P < .01). Small/isolated rural physicians tended to work more hours per week, on average, at 51.0 hours per week compared to urban (45.3 hours) and large rural (45.8 hours; P = .01).

Table 1. Demographics and Clinical Characteristics of Primary Care Physicians by Urban/Rural Practice Location (N = 818)
 UrbanLarge ruralSmall/Isolated rural 
 n = 691n = 76n = 51 
DemographicsN%aN%aN%aP value
  1. a

    Weighted for Sampling Strata.

  2. b

    Due to multiple reimbursement strategies, percentages add to more than 100%.

Gender       
Male38858.95068.93774.1.04
Female30341.12631.11226.0 
Clinical Specialty       
Family Medicine19929.13545.13564.1<.001
Internal Medicine21945.52138.51231.3 
Pediatrics27325.42016.444.7 
Race       
White46165.15067.04182.4.15
Asian12019.31420.035.8 
Other10715.61013.0511.8 
Clinical CharacteristicsN%aN%aN%aP value
Yearly Clinical Compensation ($)       
<100,00016024.21418.31223.3.10
100,000-149,99923033.41621.61531.3 
150,000-199,99912919.52334.5817.8 
≥200,00015622.91925.71327.6 
US/Canadian Medical School       
No18228.41622.01327.8.56
Yes50871.75978.03872.2 
Medical School Faculty       
No45467.85677.04383.8.02
Yes23532.22023.0816.2 
Practice Organization       
Hospital/University/Med School14421.91523.31019.2.96
Group31944.93847.22243.6 
Solo/2 person15322.41722.31223.6 
Other7410.867.3713.6 
Primary Reimbursementb       
Fee for service31445.64052.42753.5.58
Capitated, full/partial7610.444.848.9.31
Salary27941.32939.11835.5.77
Other527.733.4713.6.23
 Meana(SD)Meana(SD)Meana(SD)P value
Percentage of patients:       
Medicare27.5(18.3)33.1(18.7)36.1(13.0)<.01
Medicaid18.7(18.1)26.6(19.3)23.9(14.4)<.01
Uninsured9.6(12.1)10.9(9.5)11.7(7.4).29
Hours week direct patient care45.3(14.4)45.8(12.8)51.0(10.2).01
Years in medical practice21.9(9.3)22.9(8.5)24.5(9.0).24

Attitudes, Preparedness and Participation in Quality Improvement

PCPs’ attitudes about the need to engage in QI activities (eg, peer review and recertification) did not differ significantly by practice location (Table 2). The extent physicians felt prepared to contribute to formal QI efforts did not significantly differ by the 2-level practice location variable. However, physicians in small/isolated rural areas reported feeling significantly more prepared to contribute to QI activities (52.5%) than their urban (35.9%) or large rural (22.1%; P < .01) counterparts (not shown in table). Yet, PCPs in rural locations indicated feeling less prepared to critically evaluate new clinical knowledge (21.4% vs 34.8% of urban PCPs; P < .01). Almost 61% of rural PCPs reported engaging in formal medical error reduction initiatives compared to 51.2% of urban physicians (P = .06). Rural PCPs reviewed an other physician's records for QI more often (65.6% vs 48.0% of urban; P < .001, while among the small/isolated rural group this was higher at 71.6%). Yet, urban PCPs were marginally more likely to report looking for data on disparities at their practice, clinic, hospital, or other health care setting (12.8% vs 6.7% of rural; P = .06).

Table 2. Attitudes, Preparedness and Participation in Quality Improvement Activities for Primary Care Physicians by Rural/Urban Practice Location
 UrbanRural 
 N = 691N = 127 
 %%P value
Attitudes   
“Completely Agree” that physicians should   
Participate in peer review of quality of care53.757.0.51
Undergo recertification periodically56.256.2.99
Preparedness   
Extent you feel “Very Prepared” to   
Contribute to formal QI efforts36.034.8.81
Evaluate colleagues’ clinical performance26.822.2.29
Critically evaluate new clinical knowledge34.821.4<.01
Participation   
Percentage in the last 3 years   
Participated in formal medical error reduction initiative at office51.260.8.06
Reviewed another physician's medical records for QI48.065.6<.001
Participated in the development of clinical practice guidelines48.656.8.10
Looked for data on disparities due to race, gender, or income12.86.7.06

Attitudes and Behaviors Related to Patients

Forty percent of rural PCPs completely agreed that physicians should talk with patients about costs of care compared to 29.2% of urban physicians (P = .02; Table 3). Urban physicians endorsed the need to provide care regardless of patient ability to pay more often (76.5% vs 65.5%; P = .01). Rural PCPs were more likely than urban physicians to have asked patients for donations in the past 3 years (14.6% vs 7.1% of urban; P < .01). Rural physicians had provided patient care to someone with whom they had a financial relationship more often than urban PCPs (13.9% vs 7.5% of urban; P = .02).

Table 3. Attitudes and Behaviors Related to Patient Relationships, Responsibilities and Access to Care for Primary Care Physicians by Rural/Urban Practice Location
 UrbanRural 
 N = 691N = 127 
 %%P value
  1. a

    No longer statistically significant at α = 0.05 when adjusted for gender, physician specialty (general internal medicine, family medicine, general pediatrics) and faculty at medical school.

  2. b

    For this item, we report the percent of participants who indicated that they rarely, sometimes or often asked patients for donations.

Attitudes   
Personal Relationships With Patients   
“Completely Agree” that physicians should   
Talk to patients about costs of care29.240.5.02
Fully inform all patients about benefits/risk of treatment89.883.6.06
Never tell a patient something that is not true83.382.8.88
Patient Responsibilities and Access to Care   
“Completely Agree” that physicians should   
Provide necessary care regardless of patient ability to pay76.565.5.01a
Minimize disparities in care due to patient race, gender or income86.184.0.55
Know the overall cost of care they provide52.056.8.36
Behaviors   
Personal Relationships With Patients   
Answered “Yes”   
Asked a patient to make a donation to a charitable organization you're involved with in the past 3 yearsb7.114.6<.01
Provided patient care in the last year for a person with whom you have a financial relationship7.513.9.02a
Provided patient care in the last year for a person (other than spouse) with whom you have a sexual relationship0.53.3<.01
Patient Responsibilities and Access to Care   
Rarely/sometimes/often have   
Prescribed a brand name drug when an equal generic was available because a patient asked for it in the past year77.390.1<.01
Added one or more Medicaid/uninsured patients to your panel in the past year84.895.7<.01
Provided care to an uninsured patient without charging in the past year80.691.4<.01

When asked about their own behaviors regarding patient responsibilities and access to care, 90.1% of rural physicians reported prescribing a brand name drug because a patient asked for it compared to 77.3% of urban PCPs (P < .01). Rural physicians were more likely to have added Medicaid and uninsured patients to their panel than urban physicians (95.7% vs 84.8% urban; P < .01). A total of 91.4% of rural PCPs reported providing care to an uninsured patient without charging during the past year, whereas 80.6% of urban had done so (P < .01). When adjusted for gender, physician specialty, and medical school faculty status the differences in providing care regardless of ability to pay and care for a patient with whom the physician had a financial relationship were no longer statistically significant.

Professionalism and Self-Regulation Attitudes and Behaviors

The extent to which physicians completely agreed with several normative statements regarding professionalism and self-regulation, such as reporting impaired/incompetent colleagues and disclosing medical errors to patients, did not differ by practice location (Table 4). However, when behaviors related to this domain were examined, rural PCPs were more likely to have personal knowledge of an impaired or incompetent physician (20.7% vs 12.7% of urban; P = .02). Urban physicians were more likely to report they had never intentionally or unintentionally revealed to an unauthorized person health information about a patient in the past year compared to rural physicians (71.8% vs 60.8% of rural; P = .02).

Table 4. Attitudes and Behaviors Related to Professionalism and Self-Regulation for Primary Care Physicians by Rural/Urban Practice Location
 UrbanRural 
 N = 691N = 127 
 %%P value
Attitudes   
“Completely Agree” that physicians should   
Report impaired or incompetent colleagues63.765.9.64
Never disclose confidential patient information91.688.8.31
Disclose all significant medical errors to affected patients66.561.3.28
Disclose their financial relationship with drug/device to patients65.566.5.84
Extent “Very Prepared”   
Deal with colleagues who practice medicine while impaired16.817.8.80
Deal with colleagues who are incompetent in their medical practice16.220.3.29
Behaviors   
Answered “Yes”   
Personal knowledge of an impaired/incompetent physician12.720.7.02
Never intentionally/unintentionally revealed health information about one of your patients in the past year71.860.8.02

Discussion

We found that rural and urban PCPs are more alike than different on many important professional activity domains. There were no significant differences by practice location regarding PCPs’ attitudes related to several key components of professional behavior, such as the need to report medical errors to patients, inform patients about risks/benefits, be honest with patients, and disclose financial relationships with drug/device makers. Likewise, urban and rural PCPs’ attitudes about the importance of engaging in quality improvement activities are similar. These similarities suggest a deep and broad agreement on many of the core aspects of physician professionalism as articulated in the Charter on Medical Professionalism.

Differences do exist by practice location in participation in quality improvement activities. Despite endorsing similar attitudes about the importance of QI activities, a greater proportion of PCPs in small/isolated rural areas felt prepared to contribute to QI efforts than urban and large rural physicians. More rural physicians reported engaging in reviewing other physicians’ records for QI. Since 2001, the Centers for Medicare and Medicaid Services (CMS) has required quality reporting among acute care hospitals paid under the Medicare Prospective Payment System.[22] While critical access hospitals are exempt, many still choose to voluntarily report. Rural facilities also participate in other initiatives that mandate quality reporting and encourage provider participation, such as the Medicare Beneficiary Quality Improvement Program.[22] While both rural and urban hospitals are required by various entities to report on quality measures, rural PCPs in smaller practices or hospitals may have greater exposure to QI activities and are more likely to be involved in a quality initiative than physicians in larger urban hospitals where only a small percentage of physicians lead the QI efforts.

Rural and urban physicians also share similar beliefs about the need to report impaired colleagues and disclose medical errors. However, rural physicians are significantly more likely to report encountering impaired or incompetent colleagues. Based on this survey, we cannot ascertain whether impaired or incompetent physicians in rural communities are more likely to be known, or if the incidence of compromised physicians is higher in rural areas. Regardless of the explanation, few of the PCPs—irrespective of practice location—felt prepared to deal with impaired or incompetent colleagues, suggesting the challenge of reporting incompetent colleagues may be a problem across diverse practice settings.[18] Peer monitoring and reporting are the prime mechanisms for identifying impaired or incompetent physicians. As the fear of retaliation for reporting colleagues may be greater in smaller communities, rural physicians, in particular, may need to be connected with medical societies and accrediting organizations to provide support in such circumstances.[18]

Rural PCPs likely have different relationships with their patients than urban physicians—differences that may reflect the smaller communities in which they practice and the likelihood that they are more familiar with each other or have common acquaintances.[11-14] In this study we found that rural PCPs were more likely to completely agree that physicians should talk to patients about costs of care than urban physicians. While the overall endorsement of the acceptability of personal relationships with patients was low, rural physicians had asked patients for charitable donations more often than urban PCPs. Although we did not ask more detailed questions about rural PCPs’ personal relationships with patients, there are probably trade-offs with these closer patient-physician ties. While ethical issues of patient confidentiality and trust could affect care,[12-14], [23] rural physicians may have a more personal contact, connection, and familiarity with patients and have a better feel for their overall quality of life. Physicians in general could benefit from increased education in recognizing and managing personal relationships with patients.

Rural PCPs appear strongly committed to treating underserved patients. Most rural (95.7%) physicians reported adding Medicaid and uninsured patients to their panels compared to almost 85% of urban physicians (P < .01). Also, while urban physicians were more likely to agree physicians should provide care regardless of patient ability to pay than rural PCPs, they were less likely to say they had provided care to an uninsured patient without charging during the past year. These findings may reflect the practice communities of these physicians, as rural patients are more likely to report cost-related barriers to health care.[24, 25] Future research should explore whether physicians in rural communities may be reticent to forgo treating patients who cannot afford care because of their dual relationships with patients, even if it has a negative influence on their practice.

As the Patient Protection and Affordable Care Act (ACA) rolls out, rural primary care will face many opportunities and challenges. Recruiting and maintaining physicians in rural areas will likely continue to be problematic in terms of the numbers and geographic maldistribution of PCPs.[8, 26] As the ACA extends insurance coverage, PCP shortages are projected as patient panel sizes are exhausted.[27, 28] Patient-physician relationships in rural areas, in particular, may change as more previously uninsured patients come into the system. Also, as many of the rural PCPs in our sample reported currently providing care for uninsured patients, greater insurance coverage may mean rural PCPs will receive payments for care they previously provided free of charge. Unfortunately, the optional expansion of Medicaid in half of states means that many rural physicians may continue to disproportionately bear the burden of uninsured patients.

Our study has limitations that should be considered. Our measures reflect only certain aspects of physician attitudes and behavior related to QI, patient relationships, and professionalism. In particular, future research should determine whether different issues (eg, interacting with patients in personal settings such as community groups, ie, Rotary Club) may be relevant for rural physicians to better capture how their behaviors may be influenced. We also relied on self-report of attitudes and behaviors. Social desirability, especially regarding issues such as reporting medical errors or incompetent/impaired colleagues, may have influenced our findings. There may also be response bias; although the response rate was strong for a physician survey and we corrected for nonresponse using weighting, such methods are limited.

Additionally, as this was a secondary data analysis, the sample sizes for rural practice location are relatively small, meaning that our findings may be somewhat limited in detecting urban-rural differences and eliminating our ability to investigate differences by primary care physician specialty. We also had no information on physician response rate differences by urban or rural practice location; thus, we are unable to determine whether there may be reporting bias by practice location. Finally, our sample does not include osteopathic doctors, who are more likely to practice in rural areas. However, as osteopathic physicians comprise only 10% of rural physicians,[29] our results reflect the vast majority of rural primary care physicians.

Our study provides the first empirical portrait of PCPs’ professionalism and dedication to their patients and quality improvement. Rural physicians believe in important norms such as disclosing medical errors and being honest with patients at a similar level as urban PCPs. At the same time, they are engaging in QI activities, while caring for uninsured and publically insured patients more often than urban PCPs. Future research is needed to better understand what is driving the focus on QI in rural primary care practices and the supports PCPs may need to implement selected changes in patient-physician relationships. Furthermore, both urban and rural physicians need training in important areas of professionalism, such as dealing with incompetent colleagues, which should be designed to meet unique practice setting differences between urban and rural communities.

Ancillary