Public Roles of US Physicians Community Participation, Political Involvement, and Collective Advocacy

ORIGINAL CONTRIBUTION

Russell L. Gruen, MBBS, PhD Context Whether physicians have a professional responsibility to address health- Eric G. Campbell, PhD related issues beyond providing care to individual patients has been vigorously de- David Blumenthal, MD, MPP bated. Yet little is known about practicing physicians’ attitudes about or the extent to which they participate in public roles, which we defined as community participation, OR MANY YEARS, THERE HAS BEEN political involvement, and collective advocacy. debate about the degree to Objectives To determine the importance physicians assign to public roles, their par- which physicians should as- ticipation in related activities, and sociodemographic and practice factors related to sume public roles, that is, their physicians’ rated levels of importance and activity. degreeF of social responsibility for ad- Design, Setting, and Participants Mail survey conducted between November 2003 dressing health-related matters be- and June 2004 of 1662 US physicians engaged in direct patient care selected from yond providing care to individual pa- primary care specialties (family practice, internal medicine, pediatrics) and 3 non– tients.1-5 Physician leaders, social primary care specialties (anesthesiology, general surgery, cardiology). commentators, and professional orga- Main Outcome Measures Rated importance of community participation, politi- nizations’ mission statements have of- cal involvement, collective advocacy, and relevant self-reported activities encompass- ten supported the physicians’ assump- ing the previous 3 years; rated importance of physician action on different issues. 1-4, 6 tion of public roles. These are at Results Community participation, political involvement, and collective advocacy were odds, however, with some education rated as important by more than 90% of respondents, and a majority rated commu- and practice environments that do not nity participation and collective advocacy as very important. Nutrition, immunization, foster physicians’ engagement with substance abuse, and road safety issues were rated as very important by more phy- broader public health issues.7,8 sicians than were access-to-care issues, unemployment, or illiteracy. Two thirds of re- Currently, little is known about spondents had participated in at least 1 of the 3 types of activities in the previous 3 practicing physicians’ attitudes years. Factors independently related to high overall rating of importance (civic- mindedness) included age, female sex, underrepresented race/ethnicity, and gradu- about, or the extent to which they ation from a non-US or non-Canadian medical school. Civic mindedness, medical spe- participate in community, political, cialty, practice type, underrepresented race/ethnicity, preceptors of physicians in training, or advocacy activities. A previous rural practice, and graduation from a non-US or non-Canadian medical school were attempt to define and operationalize independently related to civic activity. public roles included a 1998 survey Conclusions Public roles are definable entities that have widespread support among of junior family practice physicians physicians. Civic-mindedness is associated primarily with sociodemographic factors, who regarded relevant components but civic action is associated with specialty and practice-based factors. to be assimilating into the commu- JAMA. 2006;296:2467-2475 www.jama.com nity and its organizations, identifying and intervening in community health problems, coordinating local com- communities that affect the health of Author Affiliations: Department of Surgery, University munity health resources, and individuals.”10 We also identified a of Melbourne, Royal Melbourne Hospital, Melbourne, Australia and Department of Health Policy and Man- responding to particular issues of range of office-based and out-of-office agement, Harvard School of Public Health, Boston, Mass local cultural representatives when tasks that may define physicians’ in- (Dr Gruen); and Institute for Health Policy, Massachu- 9 setts General Hospital and Harvard Medical School, Bos- caring for patients. volvement in public roles and that are ton (Drs Campbell and Blumenthal). Recently, we defined physicians’ pub- compatible with busy medical prac- Corresponding Author: Russell L. Gruen, MBBS, PhD, Department of Surgery, University of Melbourne, Royal lic roles to be “advocacy for and par- tices. These activities are health- Melbourne Hospital, Parkville, Victoria, 3050 Austra- ticipation in improving the aspects of related and may be undertaken by phy- lia ([email protected]).

Downloaded From: http://jama.jamanetwork.com/ by a University of California - San Francisco User on 04/29/2014 PUBLIC ROLES OF US PHYSICIANS

sicians either individually or as part of Political Involvement. Participants unable to pay), and number of hours physician organizations.10 were asked, “How important is it for of clinical patient contact per week If calls for social responsibility are to physicians to be politically involved (which was subsequently divided into be promoted, it is important to under- (other than voting) in health-related 2 approximately equal groups of Ͻ50 stand the degree to which practicing matters at the local, state, or national hours and Ն50 hours). physicians are supportive of assuming level?” and “In the past 3 years have you Also requested was the ZIP code of public roles and the sociodemo- been politically active (other than vot- primary practice location. This infor- graphic and practice factors that influ- ing) on a local health care issue?” mation was used to link the following ence their attitudes and activity in this Collective Advocacy. Participants additional independent variables that regard. In this study, we addressed these were asked, “How important is it for were hypothesized to potentially influ- issues using data from the Institute on physicians to encourage medical orga- ence physicians’ public engagement: (1) Medicine as a Profession’s (IMAP) Sur- nizations to advocate for the public’s Location of practice, whether each phy- vey on Medical Professionalism. The health?” and “In the past 3 years, have sician’s practice was located in a met- IMAP survey collected data about atti- you encouraged your professional so- ropolitan statistical area or in a non– tudes toward and participation in ac- ciety to address a public health or policy metropolitan statistical area based on tivities related to physician profession- issue that is not primarily concerned the US Department of Agriculture Ru- alism from a nationally representative with physician welfare?” ral-Urban Continuum Codes12;(2) sample of physicians in 3 primary care The Importance of Issues for Phy- County mean life expectancy ob- specialties (general internal medicine, sician Advocacy. Participants were pre- tained from the US Burden of Disease family practice, and pediatrics) and 3 sented with a list of health-related is- and Injury Project13 (combined male nonprimary care specialties (general sues and asked to rate how important and female) in each physician’s county surgery, anesthesiology, and cardiol- it is that physicians, individually or col- of practice, as a surrogate measure of ogy). lectively,advocate for each of them. The local burden of illness; (3) Compre- issues included health insurance for the hensive Social Capital Index (CSCI), METHODS uninsured, reduction in obesity and bet- from the 2000 Social Capital Commu- Dependent Variables: ter nutrition, reduction in unemploy- nity Benchmark Survey, which was as- Measures of Public Roles ment, reduction in air pollution, pre- sociated with Robert Putnam’s book We operationalized the concept of so- vention of teenage substance abuse, Bowling Alone: The Collapse and Re- cial responsibility by explicating 3 types increased basic literacy levels, cultur- vival of American Community,14 we ob- of public roles in which physicians may ally sensitive responsiveness of health tained for each physician’s state the engage: community participation, po- services in ethnically diverse areas, to- CSCI, an index derived from mea- litical involvement as an individual, and bacco control, seat belt use, gun con- sures of the general public’s participa- collective advocacy through profes- trol, and complete immunization of eli- tion in community organizational life, sional organizations.10 Survey ques- gible populations. The response engagement in public affairs, commu- tions were developed that assessed phy- categories ranged from 1 to 4 (not at nity volunteerism, informal sociabil- sicians’ ratings of importance of each all important to very important). ity, and social trust. of these dimensions on a 4-point scale (1, not at all important; 2, not very im- Independent Variables Survey Sample portant; 3, somewhat important; and 4, We asked about personal characteris- This study used data from the IMAP very important) and assessed whether tics including sex and race/ethnicity, Survey on Medical Professionalism. The or not each physician had acted in each classified into underrepresented mi- sampling frame for the IMAP survey of the 3 dimensions in past 3 years (for norities according to the Association of was the list of physicians in the Ameri- which responses were yes or no). American Medical Colleges’ defini- can Medical Association (AMA) Mas- Community Participation. To assess tion.11 Professional characteristics on terfile in 2003. From that list, we se- attitudestowardcommunityparticipation, which we collected data included medi- lected all physicians in the 50 states in physicians were asked, “How important cal specialty, professional age (years primary care specialties (internal medi- isitforphysicianstoprovidehealth-related since graduation from medical school), cine, family practice, pediatrics), a expertise to local community organiza- non-US and non-Canadian medical medical specialty (cardiology), a sur- tions (eg, school boards, parent-teacher school graduate status, and whether or gical specialty (general surgery), and an organizations,athleticteams,andlocalme- not the physician was a preceptor of inpatient specialty (anesthesiology). dia)?” Toassess whether physicians had physicians in training. Practice char- From this group of physicians we ex- beeninvolvedinanyrelatedactivitiesthey acteristics about which we inquired included all doctors of osteopathy, resi- were asked, “In the past 3 years have you cluded primary practice type, patient dent physicians, physicians in feder- provided health-related expertise to local demographics (percentage of patients ally owned hospitals, physicians who community organizations?” insured by Medicaid or uninsured and had no address in the AMA Masterfile,

Downloaded From: http://jama.jamanetwork.com/ by a University of California - San Francisco User on 04/29/2014 PUBLIC ROLES OF US PHYSICIANS

those who specifically requested that eligible because they were not provid- activity if they answered yes to activ- they not be contacted by mail sourced ing direct patient care, deceased, out of ity in any of the 3 dimensions in the pre- from the AMA Masterfile, and those the country, incorrectly classified or ceding 3 years. who were retired. This provided a sam- practicing a nonsampled specialty, non– Logistic regression models were es- pling frame of 271 148 physicians from English-speaking, or on leave. Of the timated to assess the independent as- which 3504 physicians were ran- remaining 3167 eligible physicians, sociations between civic mindedness domly selected with the same sample 1662 returned a completed question- and respondents’ personal, profes- size (584) in each specialty. The sample naire yielding a weighted response rate sional, and practice characteristics. size was calculated to provide 80% of 57.8%. The weighted response rate Again using a logistic regression model, power to detect specialty-related dif- was calculated as the weighted sum of we then examined which variables were ferences of 10% for estimates around the respondents (including the com- significantly related to civic activity. We 0.50 at a significance level of .05, and pleted interviews and ineligible cases) included civic-mindedness as a dichoto- assuming a response rate of 66.5%. Un- divided by the weighted sum of the full mous variable in the latter model. Pro- der these same specifications, the sample. The weighted response rate by fessional age and hours per week pro- achieved respondent counts can de- specialty was 42.6% among cardiolo- viding clinical care were dichotomized tect a difference of 12% to 13%. gists, 52.6% among internists, 55.3% in bivariate analyses to facilitate inter- among family practitioners, 57.7% pretation, but were included as con- Survey Development among general surgeons, 58.2% among tinuous variables in multivariate analy- and Administration anesthesiologists, and 65.8% among pe- ses. CSCI was treated as a continuous The survey instrument was developed diatricians. variable. Mean life expectancy was based on 4 individual interviews with omitted from multivariate models due physicians and a review of the extant lit- Analyses to its high correlation with CSCI erature. An initial draft of the survey was All analyses were conducted using Stata (r=0.385; PϽ.001). pretested using 4 additional cognitive in- version 8.15 and were weighted to ac- terviews conducted by professional in- count for the differential probability of RESULTS terviewers at Mathematica Policy Re- selection within the specialty we sur- Characteristics of Respondents search. A revised survey was again veyed. Presented is the percent of phy- The weighted characteristics of pretested using 4 cognitive interviews sicians rating each of the 3 dimen- respondents are presented in with physicians, and a final version was sions of public roles, and each topic for TABLE 1.Approximately80%ofthe developed using the results of those in- physician advocacy, as “important” (de- physicians were men, more than 70% terviews. The final survey instrument fined as somewhat important and very were white, and 74% were graduates was approved by the institutional re- important combined) or as “very im- of US or Canadian medical schools. view board of Massachusetts General portant” and the percent of respon- More than half had been in profes- Hospital. dents reporting activities in each of the sional practice for more than 20 The survey was mailed to physi- 3 dimensions in the previous 3 years. years. Almost half reported their pri- cians between November 2003 and June To capture the breadth of attitudes mary practice to be a group private 2004 with a cover letter, a fact sheet de- to and activities consistent with phy- practice, one quarter reported it to be scribing the study, a postage-paid re- sicians’ public roles as single vari- asoloor2-personpractice,a10tha turn envelope, a separate identifying ables, we combined the responses for hospital or public clinic, and a 10th a postage-paid postcard, and a prepaid in- the 3 dimensions of such roles. The sum university or medical school practice. centive check for $20. To track who re- of attitude scores to community par- Almost 89% practiced in metropoli- sponded to the survey and who did not ticipation, political involvement, and tan areas. The characteristics of while preserving anonymity of survey collective advocacy could range from respondents were representative of respondents, physicians were asked to 3 (if community participation, politi- all physicians in those specialties in mail back the postcard separately from cal involvement, and collective advo- the AMA Masterfile from which our the completed survey instrument, cacy were all rated as “not at all impor- sample was drawn. which contained no individual identi- tant”) to 12 (if all 3 dimensions were fiers. When the postcard with the re- rated as “very important”). Respon- Reported Importance spondent’s name was received, the phy- dents were defined as being civic of Public Roles sician was considered to have minded if the total was 10 or greater (ie, More than 90% of respondents over- responded to the survey. at least 1 dimension was rated to be very all, and more than 87% in each spe- Nonrespondents were contacted by important and, if any dimension was cialty, rated community participation, mail and telephone and encouraged to rated not very important, the other 2 political involvement, and collective participate. Of the 3504 sampled phy- were rated very important). Respon- advocacy to be important (Table 1). sicians, 337 (9.6%) were considered in- dents were defined as undertaking civic Furthermore, 51.9%, 38.6%, and 61.9%

Downloaded From: http://jama.jamanetwork.com/ by a University of California - San Francisco User on 04/29/2014 PUBLIC ROLES OF US PHYSICIANS

Table 1. Participant Response Characteristics, and Bivariate Analysis of Rated Importance and Activities Undertaken in the Dimensions of Public Roles* Community Political Collective Participation, %† Involvement, %† Advocacy, %†

Responses, Rated as Activity Rated as Activity Rated as Activity No. (%) Important in Past 3 y Important in Past 3 y Important in Past 3 y Overall 1662 (100.0) 94.9 54.2 91.6 25.6 97.0 24.3 Sex Men 1248 (79.7) 95.9 55.3 93.2 20.6‡ 97.7 25.4 Women 403 (20.2) 94.5 54.7 90.9 27.4‡ 96.7 23.9 Race/ethnicity Non-underrepresented minority White 1194 (71.8) 94.1 56.6‡ 91.5 27.6‡ 96.8 23.9 Asian 248 (15.4) 96.8 36.7‡ 90.1 15.0‡ 97.2 28.5 Other non-underrepresented minority 53 (3.6) 96.7 42.4‡ 95.9 22.4‡ 97.6 24.0 Underrepresented minority Black 63 (4.2) 97.0 71.8‡ 95.0 34.8‡ 100.0 29.7 Hispanic 59 (3.7) 96.3 59.8‡ 92.7 18.7‡ 100.0 23.5 Other underrepresented minority 19 (1.4) 100.0 70.8‡ 90.8 56.0‡ 97.7 35.2 Professional age, y Ͻ10 282 (12.5) 96.2‡ 59.3‡ 91.9 21.7 97.7 16.3‡ 10-19 516 (32.1) 96.1‡ 60.0‡ 92.0 24.6 97.2 23.6‡ 20-29 473 (30.9) 92.3‡ 53.6‡ 90.8 30.4 96.4 29.3‡ Ͼ29 381 (24.5) 95.2‡ 41.9‡ 91.6 23.9 96.5 25.8‡ Location of medical school United States/Canada 1252 (74.4) 94.2 58.6‡ 91.2 28.0‡ 96.9 24.0 Other 402 (25.6) 96.8 41.7‡ 92.7 18.3‡ 97.2 24.7 Specialty Family practice 298 (11.0) 95.3‡ 64.1‡ 90.9 28.2‡ 97.0 23.6 Internal medicine 256 (7.6) 93.3‡ 47.8‡ 90.2 19.1‡ 97.2 22.6 Pediatrics 323 (14.3) 97.8‡ 65.2‡ 93.5 27.2‡ 97.8 28.4 Anesthesiology 289 (20.7) 96.5‡ 32.8‡ 94.8 30.0‡ 96.2 20.5 Cardiology 229 (36.5) 92.6‡ 55.3‡ 87.3 26.1‡ 95.2 27.9 General surgery 267 (9.9) 93.2‡ 57.0‡ 94.0 32.4‡ 96.6 26.4 Primary practice organization Hospital/clinic 167 (11.3) 93.4‡ 46.5‡ 92.7 23.8 96.7 22.2‡ Staff-model health maintenance organization 68 (5.6) 87.2‡ 43.9‡ 84.7 13.1 95.6 8.4‡ Group 776 (43.6) 96.6‡ 59.3‡ 92.5 27.2 97.2 25.1‡ Solo or 2-person 385 (24.1) 94.0‡ 51.5‡ 91.2 24.6 97.4 24.4‡ Other 67 (4.7) 100.0‡ 45.0‡ 93.3 22.8 97.0 22.4‡ University/medical school 199 (10.7) 93.6‡ 57.1‡ 90.3 30.5 95.9 31.8‡ Hours in clinical patient care per week Ͻ50 711 (46.2) 94.5 52.4 90.6 25.2 96.1 24.0 Ն50 951 (53.8) 95.2 55.8 92.4 25.9 97.7 24.5 Preceptor of physicians in training No 619 (38.4) 94.4 41.5‡ 91.2 20.9‡ 96.1 18.8‡ Yes 1033 (61.6) 95.2 62.2‡ 91.8 28.5‡ 97.4 27.7‡ Location of practice Metropolitan statistical area 1382 (88.8) 94.8 52.4‡ 91.4 25.4 97.1 23.8 Non–metropolitan statistical area 176 (11.2) 96.8 67.7‡ 93.0 30.5 97.3 30.9 Percentage of patients in practice insured by Medicaid or uninsured Ͻ25 828 (56.4) 94.6 54.2 91.4 25.5 96.9 21.9‡ Ն25 657 (43.6) 95.0 55.9 91.0 27.6 96.5 28.8‡ *All percentages weighted for probability of selection in survey sample and nonresponse. †Community participation defined as providing health related expertise to local community organizations. Political involvement defined as being politically involved (other than voting) in health-related matters at the local, state, or national level. Collective advocacy defined as encouraging a medical organization to advocate on an issue of the public’s health that is not primarily concerned with physician welfare. ‡Statistically significant intragroup differences (PϽ.05).

Downloaded From: http://jama.jamanetwork.com/ by a University of California - San Francisco User on 04/29/2014 PUBLIC ROLES OF US PHYSICIANS

of physicians rated community partici- Rated Importance of Topics relevant to the health of individual pa- pation, political involvement, and col- for Public Advocacy tients. Other influences and behaviors lective advocacy to be very important, FIGURE 2 shows a conceptual model of that have less obvious or less direct ap- respectively. physicians’ public roles in which so- parent influence on individual health After summing the rated importance cioeconomic, political, environmen- care, labeled “broad socioeconomic is- of each of the 3 dimensions, 1162 re- tal, and behavioral issues are orga- sues in the model,” would be rated less spondents (70.3%) were defined as civic- nized into domains reflecting their important as topics for public involve- minded with a total score of greater than proximity to the concerns a physician ment by physicians.10 9outof12.Inlogisticregressionanaly- may have for an individual patients’ The percent of respondents in the sis, being of an underrepresented race/ health.10 The model hypothesized that IMAP survey who rated items within ethnicity, graduating from a non-US or physicians would rate as most impor- each domain as being very important non-Canadian medical school, being a tant those factors, such as access to for physician advocacy are overlaid on woman, and increasing professional age health care, that appear most directly this model (Figure 2). Overall, most were significantly related to civic- mindedness (TABLE 2). Table 2. Associations Between Physician or Practice Characteristics and Civic-Mindedness The adjusted impact of specialty on (Multivariate Analyses) rated importance of community par- Physician or Practice Characteristic Adjusted OR (95% CI) P Value ticipation, political involvement, and Female sex 1.54 (1.07-2.22) .02 collective advocacy is shown in Underrepresented minority 1.86 (1.05-3.28) .03 FIGURE 1.Withfamilypractitioners Professional age, per year* 1.03 (1.01-1.04) Ͻ.001 as the reference group, pediatricians Graduate of a non-US or non-Canadian medical school 1.46 (1.01-2.12) .046 and family practitioners were signifi- Specialty cantly more likely than internists and Family practice 1.00 nonprimary care physicians to rate General surgery 0.81 (0.53-1.23) .32 community participation as being Pediatrics 1.03 (0.67-1.58) .89 very important, indicated by 95% Cardiology 0.70 (0.45-1.10) .12 confidence intervals (CIs) that do Internal medicine 0.72 (0.47-1.09) .12 not overlap the point estimates. Anesthesiology 0.86 (0.57-1.31) .49 Anesthesiologists and general sur- Primary practice type Group practice 0.99 (0.58-1.68) .96 geons were significantly more likely University/medical school 1.06 (0.56-1.99) .86 to rate political involvement as being Solo/2-person practice 0.74 (0.42-1.30) .30 very important than were cardiolo- Staff-model health maintenance organization 0.58 (0.25-1.34) .20 gists and pediatricians. Pediatricians, Hospital/clinic 1.00 however, were significantly more Preceptor of physicians in training 1.31 (0.96-1.77) .08 likely to rate collective advocacy as Rural practice location 1.19 (0.78-1.82) .42 being very important than were Percentage of patients uninsured/unable to pay* 1.00 (0.99-1.01) .55 internists, anesthesiologists, and gen- Comprehensive social capital index* 0.91 (0.71-1.18) .49 eral surgeons. There was no signifi- Hours per week of clinical contact* 1.00 (0.99-1.01) .74 cant relationship between civic Abbreviations: CI, confidence interval; OR, odds ratio. mindedness and primary practice *Continuous variables (odds ratio based on a 1-unit increase). type, rural practice, or hours per week providing clinical care. Figure 1. Effect of Physician Specialty on Attitudes to Public Roles Being Very Important (Multivariate Analyses) Although a high CSCI was not independently related to civic- Community Participation Political Involvement Collective Advocacy mindedness (odds ratio [OR], 0.91; Specialty Family Practice 95% CI, 0.71-1.18; P=.49), or rated Internal Medicine importance of community participa- Pediatrics tion (OR, 0.92; 95% CI, 0.72-1.16; Anesthesiology P =.47), increasing CSCI was Cardiology General Surgery inversely related to rated importance of political involvement (OR, 0.77; 0.2 1.0 2.0 0.2 1.0 2.0 0.2 1.0 2.0 Odds Ratio Odds Ratio Odds Ratio 95% CI, 0.61-0.98; P=.04), and rated (95% Confidence Interval) (95% Confidence Interval) (95% Confidence Interval) importance of collective advocacy (OR, 0.79; 95% CI, 0.62-1.00; Odds ratios indicate likelihood, by specialty, of a rating of very important relative to family practitioners (reference group). P=.05).

Downloaded From: http://jama.jamanetwork.com/ by a University of California - San Francisco User on 04/29/2014 PUBLIC ROLES OF US PHYSICIANS

physicians rated issues related to risky is not primarily concerned with physi- resented race/ethnicity than those not behaviors and proximal causes of ill- cian welfare (24.3%) within the past 3 of underrepresented race/ethnicity ness and injury, such as obesity and years. There were 1077 (65%) who had tended to be active in each dimension poor nutrition, immunization, sub- been active in at least 1 of the 3 dimen- (statistically significant for commu- stance abuse, and road safety, which are sions. Civic-minded physicians were nity participation and political involve- labeled as direct socioeconomic influ- more likely to have been active in at least ment), and a significantly smaller per- ences in the model, as very important. 1dimensionthanthosewhowerenot centage of Asian physicians had been Approximately half rated access-to- (70.7% v 53.2%; PϽ.001). active in community organizations and care issues (universal health insur- Compared with other specialties, had been politically active. Older phy- ance or culturally appropriate care) as more family practitioners (64.1%), sicians were significantly less active in being very important, fewer than half more pediatricians (65.2%), and fewer community organizations, but signifi- rated air pollution or illiteracy as being anesthesiologists (32.8%) reported par- cantly more involved in collective ad- very important, and only 22.6% rated ticipating in community organiza- vocacy through their professional so- unemployment as being very impor- tions (PϽ.05). Anesthesiologists also re- cieties. Physicians who were graduates tant, each of which was hypothesized ported the lowest percentages active in from non-US or non-Canadian medi- to be more distal causes of illness or in- collective advocacy (20.5%). On the cal schools were significantly less ac- jury. other hand, anesthesiologists (30.0%) tive in community organizations and and general surgeons (32.4%) more of- politics. Physicians in staff model health Activities Consistent ten reported being politically in- maintenance organizations were sig- With Public Roles volved than other physicians (PϽ.05). nificantly less active in all 3 dimen- Table 1 shows the overall percentage of Variation in civic activity was corre- sions than were physicians working in respondents who reported providing lated with other physician characteris- other settings. Physicians who were pre- health-related expertise to local com- tics. Rural physicians tended to be more ceptors of physicians in training were munity organizations (54.2%), being po- active in pursuits in each of the 3 di- significantly more active than those who litically active (other than voting) on a mensions than physicians in metro- had not precepted physicians in train- local health care issue (25.6%), and en- politan areas, and significantly more so ing. Physicians with a greater propor- couraging a professional society to ad- in community participation. A higher tion of patients insured by Medicaid or dress a public health or policy issue that percentage of physicians of underrep- uninsured and unable to pay were sig-

Figure 2. Attitudes About the Extent of Physicians’ Public Responsibilities Showing the Percentage Rating as Very Important

Domains of Physician Responsibility Issues for Rating as Physician Advocacy Very Important, % Access to Care Health insurance for uninsured 58.1 Cultural responsiveness of health services Individual in ethnically diverse areas 45.5 Patient Care Direct Socioeconomic Influences Reduced obesity and better nutrition 81.9 Immunization of eligible populations 79.8 Access to Care Tobacco control 76.9 Seat belt use 72.6 Direct Prevention of teenage substance abuse 72.2 Socioeconomic Gun control 49.9 Influences Broad Socioeconomic Influences Reduction in air pollution 42.7 Broad Socioeconomic Increased basic literacy 41.6 Influences Reduction in unemployment 22.6

Global Health Influences Domains of Professional Obligation Domains of Professional Aspiration

Socioeconomic influences, including their environmental, political, and behavioral associations, are shown as expanding domains depicting proximity to physicians’ core responsibility for individual patient care. Direct socioeconomic influences were more often rated as very important by physicians than were access-to-care issues, followed by broad socioeconomic influences. Adapted from Gruen et al.10

Downloaded From: http://jama.jamanetwork.com/ by a University of California - San Francisco User on 04/29/2014 PUBLIC ROLES OF US PHYSICIANS

nificantly more active in collective ad- Table 3. Associations Between Physician or Practice Characteristics and Civic Activity vocacy, but not more individually ac- (Multivariate Analyses) tive in community participation or Physician or Practice Characteristic Adjusted OR (95% CI) P Value politics. Number of hours spent in Civic-mindedness 2.37 (1.75-3.30) Ͻ.001 clinical patient contact was unrelated Female sex 0.86 (0.61-1.23) .41 to levels of activity in each of the 3 Underrepresented minority 2.34 (1.35-4.08) .003 dimensions. Professional age, per year* 0.99 (0.98-1.01) .33 The results of multivariate analyses Graduate of a non-US or non-Canadian medical school 0.71 (0.51-0.99) .049 are shown in TABLE 3.Civicactivitywas Specialty significantly associated with civic- Family practice 1.00 mindedness (OR, 2.37; 95% CI, 1.75- General surgery 0.97 (0.61-1.53) .89 3.30; PϽ.001). Other characteristics in- Pediatrics 0.83 (0.54-1.27) .39 dependently associated with civic Cardiology 0.70 (0.43-1.13) .14 activity included being a physician from Internal medicine 0.69 (0.45-1.07) .10 an underrepresented race/ethnicity, Anesthesiology 0.52 (0.34-0.80) .003 being a preceptor of physicians in train- Primary practice type Group practice 2.13 (1.30-3.48) .003 ing, having rural practice location, and University/medical school 1.94 (1.06-3.55) .03 being a graduate of a US or Canadian Solo/2-person practice 1.77 (1.03-3.02) .04 medical school. Staff-model health maintenance organization 1.15 (0.51-2.62) .73 In multivariate analysis, the odds of Hospital/clinic 1.00 family practitioners and general sur- Preceptor of physicians in training 2.10 (1.55-2.85) Ͻ.001 geons being active were significantly Rural practice location 1.81 (1.15-2.87) .01 greater than they were for anesthesi- Percentage of patients uninsured/unable to pay* 1.00 (0.99-1.01) .85 ologists. The odds of activity were sig- Comprehensive social capital index* 1.03 (0.78-1.38) .82 nificantly greater for physicians Hours per week of clinical contact* 1.00 (0.99-1.01) .76 whose primary practice was in a Abbreviations: CI, confidence interval; OR, odds ratio. group, a solo or 2-person practice, or *Continuous variables (odds ratio based on a one unit increase). an academic practice than for physicians whose primary practice was a non-university hospital or public These results indicate a high degree of tinction lies between advocacy on is- clinic. The number of clinical contact consensus and previously undocu- sues that are perceived to be closely hours per week, the percentage of mented willingness of physicians to connected to individual patients’ health patients insured by Medicaid or unin- engage in addressing US public health (such as tobacco control, nutrition, im- sured and unable to pay, and the concerns. munization, substance abuse, and seat CSCI had no significant association Our findings suggest that many phy- belt use), and broader population health with activity, nor did sex or profes- sicians believe they have professional concerns that physicians may per- sional age after including civic mind- responsibilities to health-related is- ceive to be less clearly linked to the edness in the model. sues outside their direct clinical prac- health of individual patients (such as tice, and that these are considered to unemployment, illiteracy, and air pol- COMMENT be responsibilities of individual physi- lution).10 The finding that 72% to 82% Physician participation in public cians, expressed through community of physicians regarded involvement in roles—defined as community partici- participation and individual political in- issues closely connected to individual pation and individual and collective volvement, as well as of their profes- patients’ health to be very important but health advocacy—was strongly sional associations. These findings are only 22% to 43% regarded involve- supported by US physicians in our consistent with a view of professional- ment in issues less clearly linked to in- sample. More than 90% regarded pub- ism in which physicians are respon- dividual patient health to be very im- lic roles as important. Furthermore, sible in their areas of expertise to con- portant, was consistent with our more than half regarded community tribute to helping the society that grants previous hypothesis.10 participation and collective advocacy them professional status.3,8,16 If physi- Given estimates of over 60 million to be very important, and more than cians subscribe to this view, then they uninsured and underinsured Ameri- one third regarded individual political likely perceive a point where their re- cans,17 and the impact that access to involvement to be very important. sponsibilities as experts end, and their care has on receipt of quality health Approximately two thirds of all physi- civic responsibilities are no greater than care,18 it is worth considering why cians reported participating in at least those of other members of society. We only 58% and 46% of our respondents 1publicroleintheprevious3years. had previously suggested that this dis- regarded universal health insurance

Downloaded From: http://jama.jamanetwork.com/ by a University of California - San Francisco User on 04/29/2014 PUBLIC ROLES OF US PHYSICIANS

and culturally sensitive care to be very to civic activity including professional group settings are also comfortable important concerns, respectively. It and practice characteristics such as spe- with the relationships that mediate may be that physicians don’t perceive cialty, practice setting, being a precep- public activity. Further research the link between access to care and tor of physicians in training, rural prac- might investigate if there are measur- health as strongly as other issues, that tice location, and being a physician of able characteristics that mark the they don’t believe they have a role in an underrepresented race/ethnicity or civic-minded physician and, if so, addressing these issues, or that they a graduate of a US or Canadian medi- whether they should be selected for don’t perceive themselves to be able to cal school. and cultivated. exert as much influence on access We can only speculate on the rea- Results of this study should be in- concerns as they can on smoking sons for these associations. Physicians terpreted with caution given their re- policy, for example. Another possibil- from underrepresented race/ethnicity liance on self-reported beliefs and be- ity is that the question of universal groups and those employed in aca- haviors, and if nonresponders regard insurance has become linked to parti- demic environments may have been public roles differently than respond- san debates and ideological positions more civic minded and more active in ers. Nonresponders may be less civic- in ways that make physicians feel public roles because of their greater di- minded and, if so, these results over- uncomfortable. Gun control is simi- rect experience with social influences estimate physician civic mindedness in larly linked to ideology and partisan- on health or greater awareness of the general. Our findings may also not be ship, which may explain why only evidence of the health effects of some generalizable to specialties that were not 50% of physicians regarded it to be a socioeconomic characteristics. Teach- surveyed. Furthermore, there may have serious public health issue. ing students (independent of aca- been associations with civic minded- We determined respondents’ civic demic practice setting) may be an ex- ness or civic activity that were not mea- mindedness based on the extent to tra factor that motivates physicians to sured but that might be influential, such which they rated particular public roles be active in public roles, or it may be as socially oriented medical school pro- as being very important. By this stan- an indicator of a civic-oriented prac- grams or personal experiences of hard- dard, more than 7 out of every 10 re- tice. Rural physicians may be more ac- ship. spondents were civic-minded. Our mul- tive because, historically at least, mem- Nonetheless, this study provides tivariate analysis showed that civic bers of rural communities were more some important evidence for profes- mindedness varied with sociodemo- often involved in community projects sional leaders and organizations, graphic characteristics, such as being and organizations. policy makers, and educators who of an underrepresented race/ethnic- The potential reasons why organi- may want to engage more physicians ity, a graduate of a non-US or non- zational setting was associated with in public health and health policy Canadian medical school, sex, and pro- civic activity are less clear. Group concerns. First, the majority of physi- fessional age, and was independent of practice environments seem to be cians are supportive of such engage- specialty and other professional and independently associated with much ment through community participa- practice factors that we took into ac- greater levels of civic activity than tion, political involvement, and count. This suggests that civic mind- hospital or public clinic environ- collective advocacy. Second, the per- edness is heavily influenced by physi- ments. The reasons for this observa- ceived importance physicians assign cians’ personal characteristics, and tion, if confirmed in future research, to involvement in any particular issue perhaps less by professional and prac- seem worth exploring in more detail. appears to be, at least in part, related tice attributes. Nevertheless, other un- The phenomenon does not appear to how directly the issue concerns measured professional and practice fac- related to physicians in groups having individual patients’ health. Clarifying tors could have influenced civic more time to pursue outside roles the evidence of such causative links mindedness. since group practice physicians who may therefore be critical in influenc- We found that physicians’ civic worked in clinical care more than 50 ing physician participation in public mindedness was an important influ- hours per week were just as active in roles. Third, a variety of personal, ence on civic engagement, but that at- public roles as those who worked less professional, and practice characteris- titude did not always predict action. For than 50 hours. Furthermore, there tics may influence physicians’ civic example, 28% of those identified as was no significant association mindedness and civic activity. Con- being civic minded had not been ac- between civic activity and the propor- firming and understanding these tive in public roles in the past 3 years, tion of all patients receiving care from potential influences could provide and 53% of those considered not to be the practice who are poor, the life important guidance to leaders and civic-minded had participated in civic expectancy of local residents, or mea- policy makers who want to enlist the activities. In multivariate analysis, a sures of statewide social capital. One positive energy of physicians in pro- number of factors other than civic possibility is that physicians who are moting public health at a societal mindedness were significantly related temperamentally comfortable in level.

Downloaded From: http://jama.jamanetwork.com/ by a University of California - San Francisco User on 04/29/2014 PUBLIC ROLES OF US PHYSICIANS

Author Contributions: Dr Gruen had full access to all of Funding/Support: This research was funded by a grant cine, Massachusetts General Hospital and Harvard the data in the study and takes responsibility for the in- from the Medicine as a Profession Project of the Open Medical School for assistance with data processing and tegrity of the data and the accuracy of the data analysis. Society Institute. Dr Gruen reports being a 2002- analysis, and Janice Ballou and Frank Potter at Math- Study concept and design: Gruen, Campbell, 2003 Harkness Fellow in Health Policy at the Har- ematica Policy Research Inc. Use of the US Burden of Blumenthal. vard School of Public Health, receiving support from Disease and Injury Project data were provided by Acquisition of data: Campbell, Blumenthal. The Commonwealth Fund, and a being a Fellow in Catherine Michaud, MD, PhD, Center for Population Analysis and interpretation of data: Gruen, Campbell, Medical Ethics at Harvard Medical School. and Development Studies, Harvard School of Public Blumenthal. Role of the Sponsors: The sponsors had no role in Health. The Social Capital Community Benchmark Sur- Drafting of the manuscript: Gruen, Campbell, the design and conduct of the study; collection, vey data used in this study were collected by Robert Blumenthal. management, analysis, and interpretation of the D. Putnam of the Saguaro Seminar Civic Engage- Critical revision of the manuscript for important in- data; or preparation, review, or approval of the ment in America, a project of the John F. Kennedy tellectual content: Gruen, Campbell, Blumenthal. manuscript. School of Government at Harvard University and nu- Statistical analysis: Gruen. Acknowledgment: We thank Paul D. Cleary PhD, Yale merous community foundations nationwide, and made Obtained funding: Campbell, Blumenthal. School of Public Health, for advice on earlier drafts of available through the Roper Center for Public Opin- Administrative, technical, or material support: Campbell. the manuscript, Manju Ghokale, formerly of the In- ion Research. None of the individuals listed in this ac- Study supervision: Campbell, Blumenthal. stitute of Health Policy, Massachusetts General Hos- knowledgment received additional compensation for Financial Disclosures: None reported. pital, Susan Regan PhD, Division of General Medi- their work on this article.

REFERENCES 1. Stevens RA. Public roles for the medical profes- and Soros Foundations Network Web site. 2002. http: 1965-1994. Cambridge, Mass: Harvard Center for sion in the United States: beyond theories of decline //www.soros.org/resources/articles_publications Population and Development Studies; 1998. http: and fall. Milbank Q.2001;79:327-353. /publications/ideas_physicians_20020101 //www.hsph.harvard.edu/organizations/bdu/papers 2. Greenlick MR. Educating physicians for population- /ideas_physicians.pdf Accessed October 30, 2006 . /usbodi/index.html. Accessed July 25, 2005. based clinical practice. JAMA.1992;267:1645-1648. 9. Pathman DE, Steiner BD, Williams E, Riggins T. The 14. Putnam RD. Bowling Alone: The Collapse and 3. May WF. Beleaguered Rulers: The Public Obliga- four community dimensions of primary care practice. Revival of American Community.NewYork,NY: tion of the Professional. Louisville, Ky: Westminster JFamPract.1998;46:293-303. Simon & Schuster; 2000. John Knox Press; 2001. 10. Gruen RL, Pearson SD, Brennan TA. Physician- 15. Stata Statistical Software Release 8.0. College Sta- 4. Sullivan WM. Medicine under threat: profession- citizens–public roles and professional obligations. tion, Tex: Stata Corp; 2003. alism and professional identity. CMAJ.2000;162:673- JAMA.2004;291:94-98. 16. Sullivan WM. Work and Integrity: The Crisis and 675. 11. American Association of Medical Colleges. Un- Promise of Professionalism in America. 2nd ed. San 5. Benson JA Jr. The burdens of professionalism: pa- derrepresented in medicine definition. 2006. http: Francisco, Calif: Jossey-Bass; 2004. tients’ rights and social justice. Pharos.2000;63:4-9. //www.aamc.org/meded/urm/start.htm. Accessed July 17. Schoen C, Doty MM, Collins SR, Holmgren AL. 6. Medical Professionalism Project. Medical profes- 25, 2005. (Web exclusive) Insured but not protected: how many sionalism in the new millenium: a physician charter. 12. United States Department of Agriculture Eco- adults are underinsured? Health Aff (Millwood). 2005; Ann Intern Med.2002;136:243-246. nomic Research Service. Measuring rurality: rural- (suppl):W5-289–W5-302. doi:10.1377/hlth- 7. Rothman DJ. Medical professionalism: focusing on urban continuity codes. 2003. http://www.ers.usda aff.W5.289. Accessed October 25, 2006. 15956055. the real issues. NEnglJMed.2000;342:1284-1286. .gov/briefing/Rurality/RuralUrbCon/. Accessed July 18. Ayanian JZ, Weissman JS, Schneider EC, Gins- 8. Rothman DJ, O’Toole T. Ideas for an open soci- 25, 2005. burg JA, Zaslavsky AM. Unmet health needs of un- ety: physicians and the body politic. Occasional Pa- 13. Murray CJ, Michaud CM, McKenna MT, Marks insured adults in the United States. JAMA.2000;284: pers from OSI-US Programs, Open Society Institute JS. US Patterns of Mortality by County and Race: 2061-2069.

Downloaded From: http://jama.jamanetwork.com/ by a University of California - San Francisco User on 04/29/2014