Online Identities of Physicians

ployed compared with those who did not deploy, this difference was not statistically significant (relative risk, 1.3; 95% CI, 1.0-1.7). The authors concluded that the increase in suicides paralleled an increase in mental disorder prevalence, mirroring our findings. Furthermore, the study did not control for covariates, particularly mental disorders. Suicidal behavior is extremely complex, and we agree with Drs Bryan and Clemans that other unmeasured factors may influence risk. It is important not to misinterpret our findings as suggesting that the conflicts in Iraq and Afghanistan have had no cumulative effect. Since 2005, suicide incidence in the Army increased just as sharply among soldiers who did not deploy as those who did. This corresponded to a steep increase in mental disorders in the military population overall, likely a product of intense wartime demands affecting soldiers throughout all phases of the deployment cycle. The aggregate research indicates that individuals who develop mental disorders following deployment or other military stressors unrelated to deployment will be at higher risk for suicide than individuals relatively less affected by such stressors. We agree that TBI is an important health concern requiring further study, along with other serious injuries and chronic illnesses associated with suicidal behaviors. However, the research on deployment-related TBI has been inconsistent due to definitional problems, failure to differentiate mild TBI (concussions) from more debilitating moderate or severe TBI, lack of adequate control groups (eg, non–head-injured controls), and the life-threatening context in which combat-related concussions occur, which contributes to depression, posttraumatic stress disorder, substance abuse, complex grief, and other comorbid mental and physical health problems.4-6 With regard to potential moderators, stratification by age (<35 years and ≥35 years) did not change the results; deployment was not associated with increased risk for suicide in either age group. We took extensive steps to ensure valid results and address potential limitations of time-dependent covariates in the Cox models mentioned by Dr Larson and colleagues. This included adjusting for separation from the military and assessing deployment as both non–time-varying and timevarying covariates (in which individuals changed to a deployed status on the first day of deployment and remained in that category for the remainder of follow-up). The linkage of military and postmilitary data directly addresses potential biases of attrition from service. Most importantly, we conducted an entirely separate analysis using a nested casecontrol design with deployment as a static variable and found identical results.


Online Identities of Physicians
To The Editor Dr DeCamp and colleagues 1 recently wrote about the online identities of physicians, stating that it is "operationally impossible" and "potentially harmful" to separate personal and professional identities. Rather than identity separation, they proposed physicians decide whether content is appropriate for a physician to be posting, regardless of the forum. We believe this latter point to be necessary, but not sufficient, for physicians to manage their online identity.
Most physicians will be best served by proactively managing their online identity, including separation of professional and personal profiles where appropriate. 2,3 Powerful search engines merge identities and juxtapose social networking profiles next to practice websites. 4 Even though it may be impossible to completely separate identities online, it does not mean that one should not try.
Separating identities means doing 2 things: proactively creating and managing a professional online identity, and hiding personal social networking profiles and sites from public view. This approach has several advantages, including making it easy for patients to find relevant contact information and allowing physicians to share personal pictures and experiences with friends and family while retaining some sense of privacy. Without privacy boundaries, some patients, and even state boards, may take issue with photos that physicians or their friends share, such as one of a physician holding a glass of wine at a party. 5 Sharing personal information with patients should be an active decision, not one made by search engines. When physicians disclose personal information or experiences, place a family portrait on a desk, or wear a lapel pin supporting a cause, they are making an active decision to share with their patients. Without privacy settings on personal websites, physicians lose control over whether or not they wish to share information. If a physician actively wishes to share pictures of In Reply We agree with Dr Crotty and colleagues that physicians should be proactive in how they create and manage their online presence. This includes regularly reviewing and monitoring their online presence, as guidelines suggest. 1,2 We also agree that the choice to share personal elements, such as family portraits or hobbies, rests with the physician alone.
To clarify, our intent was not to provide a necessary and sufficient condition for determining the appropriateness of a specific social media contribution. We sought primarily to cast doubt on a dominant assumption in ongoing discussions about professionalism and social media (ie, the separation of personal and professional identities online). Instead we recommended that physicians first ask whether a potential social media contribution is appropriate in a public or potentially public space.
This approach recognizes, as do Crotty and colleagues, that completely separating identities is impossible (due to the power of search engines) and that the inability to separate them is potentially damaging (because personal or private behaviors can be objectionable to patients, state medical boards, and the general public). Privacy settings offer no guarantee of privacy. They reduce but cannot eliminate disclosure-related risks.
Importantly, our approach allows other factors to affect a physician's decision regarding participation in social media. A physician's specialty might be one such factor. For example, psychiatrists may be less comfortable posting about certain personal activities because of the nature of the illnesses of some patients and the strict boundaries sometimes necessary for successful therapeutic relationships. Emphasizing the inherently public or potentially public nature of social media is a fundamentally different starting point from which to begin the ethics discussion than a description of a logically necessary and sufficient condition.
We also remain concerned about whether existing recommendations to separate personal and professional identities provide truly usable advice. One guideline recommends physicians separate personal and professional spheres and "comport themselves professionally in both." 2 This raises more questions than answers, partly for the reasons discussed in our Viewpoint and also because what it means to be professional in the 2 spheres remains unclear. Does one professional standard apply to both, thereby collapsing the personal and professional spheres and aligning with our approach? Or do multiple professional standards exist, one for each sphere plus a potential third standard applicable only to a deeper personal space? Practicing physicians and medical students struggling to manage their professional identities 3 might find this confusing.
In addition, we believe our approach is more consonant with the experience of many physician social media users. 4 This was reinforced in a public Twitter conversation hosted by the Johns Hopkins Berman Institute of Bioethics the week the article was published. 5 Further work is necessary to define appropriate behavior in the public space of social media. We believe our approach lays a foundation for doing so.