Social media—Web-based applications and tools that facilitate information sharing, such as blogs, wikis, podcasts, media-sharing sites, and social networking sites (SNSs)—have flourished as a new paradigm for communication. Yet the potential blurring of personal and professional identities has created hazards for medical professionals and trainees.1–4 Unprofessional online content posted by medical students has resulted in disciplinary action by medical schools, including dismissal in some cases.1 Without existing precedent, institutions are finding their own ways through unique professionalism challenges involving trainees,5 and some have created social media policies in response.6
Until recently, physician organizations did not maintain national guidelines on social media use. The American Medical Association released a policy statement in November 2010 advising medical professionals to proceed cautiously with social media tools, to be vigilant about protecting patient privacy, and to consider having separate professional and personal social networking accounts.7 However, questions remain on how to maintain professional boundaries while using social media. Also, there is still no consensus about what exactly constitutes professional behavior online, aside from the most egregious of professionalism lapses involving unethical and illegal activity.4,8
The popularity of social networking Web sites (61% of all U.S. adult users of the Internet, ranging from 26% for adults 65 and older to 86% for adults 18–29) has implications for the boundaries and relationships between faculty, trainees, and patients.9 Although social networking use continues to rise in all age groups yearly, the rate of growth is most noticeable for older adults. From 2009 to 2010, social networking use increased by 88% for adults aged 50 to 64, and 100% for adults 65 and older.9 As more adults embrace social networking, medical school faculty, who may carry both evaluative and mentorship roles for medical students, may be faced with opportunities to join students' social networks. Faculty may also have opportunities to join social networks with trainees and with patients. However, there are few data on how faculty perceive social networking boundaries, to what extent these kinds of interactions are occurring, or what faculty interpret as unprofessional online behavior for trainees. Knowing these perceptions may help guide other clinician–educators, help build consensus on appropriate social networking relationships, and inform possible institutional policy development.
We examined responses to a national survey of clerkship directors in internal medicine to uncover faculty perspectives of professional boundaries and trainee-posted content on SNSs.
In June 2010, the Clerkship Directors in Internal Medicine (CDIM) conducted its annual, voluntary, and confidential survey of its 110 U.S. and Canadian member institutions. CDIM staff invited one institutional member per school (typically, the internal medicine core clerkship director) by e-mail to complete the survey via an included link (Checkbox Survey Software, version 4.7). They contacted nonresponders up to three additional times by e-mail and/or telephone. They did not offer any incentives to participate.
The survey contained seven sections, including one demographics section (age, gender, academic rank, primary academic role, percent time supported in clerkship director role) and one section on social networking use patterns. We developed the social networking survey items based on gaps in the literature that we identified through an informal review. The CDIM research committee pilot-tested the survey items for clarity and edited them accordingly, and the CDIM council approved them before they were included in the final survey instrument. The institutional review board of Case Western Reserve University (the survey's host institution) determined that the research protocol did not fit the definition of human subject research per 45 CFR 46.102, and therefore the protocol did not require exemption status, further IRB review, or IRB approval.
The 26 survey items in the social networking section contained categorical yes/no, multiple-choice, and constructed responses. We asked participants about their use of SNSs (if they use them, and, if so, how frequently), whether a current student or current resident had ever sent them a “friend request” (defined broadly as an invitation to share personal profiles and enter one's social network), and whether they accepted those requests. We also asked why they did or did not use SNSs and why they did or did not accept the friend requests. A series of questions asked participants to rate their opinions on the appropriateness of certain social networking behaviors in which faculty might engage, on a five-point Likert-type scale (never appropriate, rarely appropriate, neutral, usually appropriate, always appropriate). Another series of questions asked participants to rate their opinion on the acceptability of specific examples of students' online postings on SNSs, on a five-point Likert-type scale (never acceptable, rarely acceptable, neutral, usually acceptable, always acceptable).
Analysis included descriptive statistics and comparisons of age, gender, and academic rank for respondents' use of SNSs, acceptance of trainee friend requests, and perceptions of appropriateness of faculty and student social networking behaviors using the χ2 or Fisher exact test statistic (if expected cell counts were less than 5). We performed multivariate logistic regression to examine the independent effects of respondents' age, sex, rank, and SNS use for two series of questions: one series about perceptions of the appropriateness of faculty social networking behaviors and another series about perceptions of the acceptability of students' social networking behaviors. We used SAS statistical software, version 9.1 (SAS Institute Inc., Cary, North Carolina), to perform our analysis, with statistical significance defined as P < .05. For the qualitative analysis of the three free-text response questions, two authors (T.K., J.M.F.) reviewed all responses independently before performing line-by-line coding of discrete statements. Each author generated overall themes within and across each of the three questions, and these themes were discussed until agreement was reached. Coding was subsequently peer reviewed by the two additional authors (KC, SRG). We resolved any coding discrepancies by consensus and made clarifying changes accordingly.
Eighty-two of 110 institutional members completed the survey (a response rate of 75%). Because responses were not mandatory, not all respondents answered every question. Throughout this section, reported N values vary because of nonresponse. Fifty of the 81 (62%) respondents were male. Thirty-eight of the 82 respondents (46%) were associate professors, 20 (24%) were assistant professors, and 24 (29%) were full professors. Sixteen of 77 (21%) respondents who reported their age were between 31 and 40 years old, 32 (42%) were 41 to 50, 25 (32%) were 51 to 60, and 4 (5%) were 61 to 70. The majority of respondents (69 of 79; 87%) held the role of core internal medicine clerkship director, 1 (1%) was the internal medicine subinternship director, and 9 (11%) held other roles involving medical student education.
Social networking use patterns
Twenty-nine of the 82 (35%) respondents reported that they currently use SNSs. Twenty-three (28%) reported using SNSs for personal purposes only, 1 (1%) for professional purposes only, and 5 (6%) for both personal and professional purposes. Over half (42 of 82; 51%) had never used SNSs, and 11 (13%) had used SNSs in the past but were not currently.
Of the 40 respondents who currently use or had used SNSs in the past, most (17 of 40; 43%) reported using SNSs less than once a month, 5 (13%) reported using them one or two times per month, 12 (30%) reported using them weekly, and 6 (15%) reported using them daily.
Clerkship directors 45 years old and younger were more likely to currently use SNSs than those older than 45 years (19 of 34 [56%] versus 9 of 43 [21%]; P = .002).
In response to the open-ended prompt to describe why respondents did or did not use SNSs, answers fell into three major domains: Users were categorized in the positive domain, and nonusers were categorized in either negative or neutral domains. Domains, subthemes, and representative quotes are shown in Figure 1.
Friend requests by trainees
Over half (21/40; 53%) of respondents who currently use or had used SNSs in the past had received a friend request from a current student at their own institution. Of these, 4 (19%) accepted the request. Being sent a friend request by a current resident at one's institution was more common; 25 of 40 (63%) respondents reported having received such a request. Twelve of these 25 (48%) accepted the friend request from the current resident.
Qualitative analysis of the open-ended prompt asking respondents to describe why they had or hadn't accepted a friend request from a current student revealed the domains, subthemes, and representative quotes in Figure 2.
Qualitative analysis of the open-ended prompt asking respondents to describe why they had or hadn't accepted a friend request from a current resident paralleled the domains and subthemes found with students, with a few minor exceptions. We used two additional subtheme descriptors for accepting friend requests from current residents: collegial (“I feel residents are more like colleagues”) and professional reasons (“The current resident is also involved as a member of the professional society I joined the social networking site for and is the one who maintains that society's Facebook page”).
Clerkship directors who were associate professors were more likely to have received a friend request by a student than assistant professors or full professors (28 of 38 [74%] versus 10 of 20 [50%] versus 0 of 24 [0%]; P = .004). Whether a clerkship director accepted a friend request from a current student or current resident did not differ by clerkship director age, rank, or sex.
Perceptions of the appropriateness of faculty behaviors on SNSs
Table 1 shows respondents' perceptions of the appropriateness of certain faculty behaviors on SNSs. The majority of respondents felt that faculty members sending a friend request to a current student (63 of 80 [79%]) or accepting a friend request from a current student (61 of 80 [76%]) was never or rarely appropriate. However, becoming social networking friends with a former student was deemed more appropriate (41 of 80 [51%] responded with usually or always appropriate). This same pattern was seen with regard to becoming social networking friends with current versus former residents. The majority of respondents felt it was inappropriate to become social networking friends with current patients (65 of 80 [81%] responded with never or rarely appropriate).
Perceptions of the appropriateness of student postings on SNSs
Table 2 shows responding clerkship directors' perceptions of the appropriateness of specific student postings on SNSs.
Notably, for all faculty and student social networking behaviors deemed inappropriate by most respondents, clerkship directors aged 45 years and younger were less likely than those older than age 45 to consider these behaviors inappropriate, with odds ratios (ORs) ranging from 0.18 to 0.79 and adjusted ORs (controlling for respondents' sex, rank, and SNS use) of 0.12 to 0.86, although none reached statistical significance (see Table 3). Faculty perceptions of the appropriateness of faculty and student social networking behaviors did not differ based on respondents' SNS use, frequency of use, rank, or gender.
In this national sample of internal medicine educators, nearly half of respondents reported firsthand experience using SNSs, either currently or in the past. Of these, the majority have received a friend request from a trainee, with some accepting these requests. The majority of responding clerkship directors believed that it was inappropriate to initiate or to accept a friend request from a current student but felt that this was more acceptable if he or she was a former student, that is, a student with whom the faculty member no longer had an evaluative role or who had graduated from the institution. To a lesser degree, we observed the same pattern when clerkship directors considered becoming friends with current versus former residents. This makes sense in terms of the idea of the progression of medical training from student to resident to professional colleague and the potential for forming problematic “dual relationships” with current versus former trainees. The vast majority of respondents felt it was inappropriate to engage in social networking friendships with current patients.
The American Medical Association's recent policy on social media use and professionalism recommends that physicians consider separating personal and professional information online and to preserve professional boundaries when interacting with patients online, but it does not provide guidance when it comes to faculty–trainee interactions on such sites.7 This study is the first that we are aware of to provide some consensus among faculty as to appropriate faculty behavior when interacting with trainees online.
Regarding respondents' perceptions of the appropriateness of certain student posting behaviors, it was surprising to learn that a photograph of a student simply holding an alcoholic beverage was deemed unacceptable by the majority of respondents. Although we did not specify the context, this could include a photograph taken during off-duty hours of an adult of legal drinking age without a depiction of intoxication. We were also surprised that the majority of respondents felt descriptions of patient encounters, even if properly deidentified and respectfully told, were unacceptable. This may reflect a generational difference of opinion; we did find a difference between younger and older clerkship directors, with older faculty more likely to perceive this behavior as unacceptable, although this difference was not statistically significant. The appreciation of narrative medicine is relatively new,10 and a prior study of clerkship directors in internal medicine found that younger clerkship directors were more likely to have a reflective writing program or to consider having a program in the near future compared with older clerkship directors.11 It is possible that we are seeing a paradigm shift in professional attitudes on sharing and reflecting on patient stories, both in traditional and social media (e.g., blog) formats, that could be explored further in studies using methodologies such as focus groups or content analysis of student versus faculty blogs.2,8
Indeed, the trend for older clerkship directors to hold more conservative views regarding the acceptability of a multitude of social networking behaviors may reflect a more general evolution of what faculty consider professional conduct online in this digital age in which more personal information is accessible about individuals than before. An awareness and appreciation of generational differences should be part of discussions of online professionalism within medical school curricula, residency training programs, and institutional policy setting.
Faculty consensus on what constitutes professional online behaviors can assist medical students who are in the process of forming their professional identities. Some medical trainees do not understand what their medical schools would classify as unacceptable online behavior.12 Also, medical students hold varying opinions about what is appropriate to post online, with the exception of patient privacy violations and illegal activity.8 Although students may be generally opposed to policies that restrict their online behaviors,8,13 they desire guidance from their institutions so that they can take personal responsibility for their actions.8
This study has several limitations. First, not all of the MD-granting medical schools in the United States and Canada are represented in CDIM; approximately 77% of medical schools accredited by the Liaison Committee on Medical Education had institutional representation in the 2010 CDIM survey. Because this study was limited to clerkship directors in internal medicine, we cannot generalize to all medical school faculty who may work with students or residents. Finally, we did not ascertain whether friend requests, received or accepted, were generated through primarily personal SNSs (e.g., Facebook) or professional ones (e.g., LinkedIn). Although the term “friend request” is more commonly associated with personal SNSs like Facebook and MySpace, we defined this term broadly to participants as an invitation to share profiles and enter one's social network. Because professional SNSs generally convey less personal information (i.e., less risk of blurring personal and professional identities), this may be considered an appropriate option for communicating and networking with trainees by respondents familiar with this type of site. Future studies might explore differences in faculty and trainee interactions on professional versus personal SNSs.
SNSs, and social media in general, have the potential to fundamentally change our professional relationships as well as the way the public views our profession in an increasingly digital world. Future studies are needed to better assess the possible generational differences in defining online professionalism and to explore the bidirectionality of opinion. For example, how would trainees react to seeing faculty engage in these online behaviors? Other questions to consider include, how does the lay public view these behaviors, and would seeing these behaviors lead to decreased trust in their physicians or in the medical profession in general? Further research and consensus gathering in this area is critical to better understand, define, and model “online professionalism” and set best practices for medical schools and the profession at large.
The data used in this survey are the property of the Clerkship Directors in Internal Medicine and have been used with permission. The authors acknowledge the Alliance for Academic Internal Medicine staff for their help in administering the online survey as well as the CDIM Research Committee, in particular Drs. Klara Papp and Dario Torre, for their instrumental assistance.
The institutional review board (IRB) of Case Western Reserve University determined that the research protocol did not require exemption status, further IRB review, or IRB approval.