Every MD/PhD student who successfully completes the program immediately confronts two issues: Will I do a residency? If so, which one should I choose?
Many students are surprised by the difficulty they experience in pondering these questions, but the reasons for their discomfiture are not hard to identify. The central problem is that even though both the PhD and MD are doctoral degrees, they mark very different levels of achievement in their respective careers. PhDs are nearing the end of their formal training--most PhDs have not only done significant research, but have also presented at national meetings and interacted at the highest level with national and international leaders in their fields. By their graduation, PhDs are living and functioning like mature scientists--indeed, there are numerous examples of wunderkind graduate students who have gone directly to faculty-level positions.
By contrast, newly minted MDs are nowhere close to functioning at the level of established physicians; most have never written an unsupervised order and never performed an unsupervised procedure, let alone directed others in such an effort. As such, senior medical students reside at the bottom of a steeply vertical hierarchy whose upper ranks they rarely even see, and in which they certainly do not live. (This last problem has been compounded for MD/PhDs by the changing nature of hospital care, which has greatly diminished the role of the physician-scientist on ward rounds.) And they do not encounter even the nether parts of that hierarchy until 12 to 18 months before they are called upon to decide whether they wish to join it. They therefore have to imagine the world of the academic physician-scientist by reconstructing it mentally from the fleeting glimpses of it they acquire during their long training. This is especially true for the many students who did their thesis research in labs not headed by physician-scientists.
So it's no surprise that their image of it resembles a Rohrshach blot more than a photograph, and that they are nervous about relying upon this image for a decision of such moment. Frankly, I see no simple solution to this problem. I do think, however, that we could do a better job in the early years of the MD/PhD pathway to help students understand that these two aspects of their training are by necessity out of register.
No formula for the answer
There is no unique formula for answering the question: Should I do a residency? But there are surely some ways not to approach the question. The poorest reason I have heard in favor of clinical residency is that "it will increase the number of jobs for which I am eligible." This is a seductive argument because, unlike most poor arguments, it happens to be true. But it's an ignorant position because it does not factor in the tremendous physical and emotional commitment required for successful residency training. These costs are acceptable if you genuinely like clinical problem-solving and patient contact, but are unbearable if you don't. If this is the only reason you can put forward for doing a residency, my strong advice would be: Don't do it.
A common mistake is to think "if I can just get through the residency, I'll be able to capture the lifetime benefit of more job availability with a one-time payment of time." Wrong. Most of the many research positions in academic medicine that are uniquely open to physician-scientists are in subspecialty fields that require considerable postresidency clinical training and a sustained commitment thereafter to at least some clinical time. In my experience, many of those who choose residency for this reason alone wind up dropping out--at great cost to themselves and their fellow interns.
The only reason to do a residency is if you genuinely enjoy at least some discipline(s) of clinical medicine. For those who do, residency can be a very satisfying--at times even exhilarating--experience.
Select residency field, not program
Assuming that you do enjoy clinical work, how do you select a residency? The important question is not "Which program?" (most programs are more similar than they care to acknowledge), but "Which field?" The latter requires careful consideration of one's longer-term goals. If your interest is in rigorous bench science, then procedure-oriented fields like surgical subspecialties present many postresidency challenges. The clinical demands of such fields (and the maintenance of procedural skills) require a greater commitment to them as a faculty member, which can erode the time available for laboratory work. This is not to say that procedure-oriented fields are incompatible with a laboratory research career, but only to point out that they pose additional issues that must be realistically confronted. These issues are of lesser moment if one's interest is clinical investigation, since the clinic forms a large part of the research enterprise in such careers.
Decisions about specific residencies are made simpler if you already know the subspecialty field in which you are interested. For example, if you're committed to the study of childhood cancer, then you know you'll need to do a pediatric residency followed by an oncology fellowship. In such a case, the decision should revolve around whether you're willing to stay in the same institution for both programs. If so, then your search will require not only vetting the pediatric residency but also (i) the quality of the oncology fellowship; and (ii) the nature of the research opportunities in that institution.
You may also wish to look into the availability of specialized programs offered by some residencies to foster the career development of physician-scientists. For example, at my institution (the University of California, San Francisco) a program called Molecular Medicine offers internship applicants in medicine, pediatrics, and pathology guaranteed admission to the subspecialty fellowship of the applicant's choice, together with research support and opportunities for "short-tracking" into that fellowship. (Short-tracking involves streamlining the time required to complete residency and fellowship requirements and requires active consultation with both residency and fellowship directors.) Several other institutions have or are developing such programs, although the specifics of each are different. It pays to inquire about these early during the application process, as they are highly competitive, and some require parallel application forms.
When evaluating research opportunities at a given institution, make certain that you inquire about any strictures your subspecialty division may place on those with whom you can do research. In the past, many divisions restricted their fellows to working with the divisional faculty only. Most such strictures have been relaxed nowadays, but it's important to inquire explicitly about this if you think you might wish to work outside that department.
Finally, remember that because clinical and research training are somewhat out of register, it is expected that your plans for your clinical future will be less secure than those for your research career. It is completely normal to have highly articulated interests in signal transduction or structural biology and only a vague feeling for what, if any, subspecialty clinical area one might elect. It's not necessary that by graduation day you have mapped out your entire career trajectory. If, in your final year of medical school, you can make a plan that you are comfortable with for the upcoming 2 years and that keeps your relevant options open, you're doing well. By the time you've done that much more clinical work, the answers about your clinical directions will very likely become clear.