This piece starts with what residents actually posted online this week—125 real threads and comments from Reddit and the other corners where people speak freely. We use AI to sift through the noise and surface the patterns that keep showing up. Then a human editor makes the judgment call on what’s worth your time, and that’s what you’re about to read.
Two options, stated plainly: finish your residency and start earning attending salary this year, or add two to seven more years of fellowship training at $70,000 annually while your co-residents start making $350,000 or more. The published compensation tables make the subspecialty look like an obvious win—GI earns $150K more than general IM, cardiology adds another $100K on top of that. But those tables don’t show you the math that actually determines whether fellowship pays off, and they definitely don’t show you what happens when the demands of residency extend for another three years while the market shifts underneath you.
The Arithmetic Nobody Does
Start with the direct cost: lost attending earning years. A three-year GI fellowship means three years at roughly $70,000 instead of $280,000. That’s $630,000 in foregone income before you see your first attending paycheck. Add the opportunity cost of delayed retirement contributions—three years of not maxing a 401(k) at $23,000 annually, compounded over a 25-year career at 7% average returns, costs you roughly $180,000 in retirement wealth. Three more years of interest accruing on $300,000 in student loans at 7% adds another $63,000 to your debt burden.
The break-even math varies wildly by specialty pairing. For general IM versus GI, if the compensation gap holds at $150,000 annually, you break even around year 8 post-fellowship—meaning you’re 41 or 42 before you’re actually ahead. For general pediatrics versus peds hem/onc, the gap is smaller (often $80,000-$100,000) and the fellowship is three years, putting break-even around year 10-12. For general surgery versus a surgical subspecialty like vascular or CT surgery, you’re looking at an additional two to three years of training for a compensation gap that ranges from minimal to substantial depending on practice setting.
These calculations assume the gap stays constant. It won’t.
What the Compensation Tables Actually Contain
That $550,000 GI average in the latest compensation report is a statistical artifact, not a job offer. It blends academic physicians making $380,000 with partnership-track private practice gastroenterologists pulling $800,000. It includes employed W-2 positions with no call and private practice owners doing 1:3 call. It averages Manhattan salaries with rural Nebraska salaries—a geographic spread that often exceeds the inter-specialty gap you’re trying to capture.
More importantly, it’s a snapshot, not a trajectory. GI reimbursement has been flat to declining for a decade. Colonoscopy reimbursement dropped 8% in the last five years while volume requirements increased. The specialty that paid $600,000 in 2015 pays $550,000 in 2026 while requiring more procedures to hit the same RVU targets. Meanwhile, general IM compensation has been climbing as health systems compete for primary care capacity.
A subspecialty can be highly paid and simultaneously on a downward trajectory. The question isn’t "what does it pay now?" but "what will it pay in 2035 when I'm mid-career?"
Reimbursement Trends and Scope Shifts
Procedural reimbursement is under structural pressure across specialties. CMS conversion factors have been flat or declining in real terms for fifteen years. Private payers follow Medicare rates with a lag. The procedures that made interventional cardiology a $700,000 specialty in 2010 now face competition from structural heart programs, electrophysiology, and interventional radiology—all fighting over the same procedural revenue.
Scope shifts don’t eliminate revenue; they move it between specialties. Diagnostic radiology is losing reads to AI-assisted triage and to other specialists doing their own imaging. Interventional radiology is gaining procedures that used to go to vascular surgery. GI is losing screening colonoscopies to primary care-based alternatives while gaining complex therapeutic endoscopy. The subspecialty that looks dominant today may be fighting for market share in ten years.
This doesn’t mean fellowship is a bad bet. It means the bet is on clinical differentiation and market durability, not on a static compensation premium.
The Variables That Change the Answer
Market durability: Some subspecialties have structural demand that’s hard to disrupt. Pediatric hem/onc treats childhood cancers that aren’t going away and aren’t being absorbed by other specialties. Interventional cardiology faces more competitive pressure but still controls procedures that require years of training to perform safely.
Geographic flexibility: General IM can practice anywhere. Transplant hepatology needs a transplant center. The more subspecialized you become, the fewer places you can work—which means less bargaining power in salary talks and less ability to optimize for lifestyle or cost of living.
Exit optionality: Some fellowships open doors beyond clinical practice. Cardiology and oncology have strong pharma and device industry pathways. Hospitalist medicine has administrative and consulting exits. Surgical subspecialties tend to lock you into clinical work with fewer pivots available.
Lifestyle at the attending stage: Don’t evaluate fellowship lifestyle; evaluate attending lifestyle. GI fellowship is brutal, but GI attending life can be predictable outpatient work with reasonable call. Surgical subspecialty fellowship is brutal, and surgical subspecialty attending life is often still brutal. The fellowship is temporary; the career is thirty years.
How to Actually Decide
Pursue fellowship if the clinical work genuinely interests you and the market fundamentals are durable. Don’t pursue fellowship primarily for a compensation gap that may compress by the time you’re mid-career.
When evaluating specific programs, ignore name recognition and ask about job placement. Where did the last five graduating fellows go? What’s the mix of academic versus private practice versus employed positions? What’s the average starting salary, and how does it compare to the specialty median? A prestigious program that places fellows into academic jobs at $380,000 may be a worse financial decision than a community program that places into private practice at $550,000.
Ask about procedural volume and case complexity. Fellowship is skills training. A program that gives you 2,000 colonoscopies and 200 ERCPs produces a different gastroenterologist than one that gives you 800 colonoscopies and 50 ERCPs. The second fellow will have a harder time getting hired for the jobs that pay the premium.
Ask current fellows about protected time, moonlighting policies, and whether attendings actually teach or just supervise. The fellowship experience varies more within specialties than between them.
The Real Calculation
Fellowship is a bet on your future self’s preferences and on a market you can’t fully predict. The compensation premium exists, but it’s smaller than the tables suggest, it’s shrinking in many subspecialties, and it comes with real costs in flexibility, time, and compounding opportunities.
If you’d do the clinical work for the same salary as your generalist colleagues, fellowship is probably right for you. If you’re primarily motivated by the compensation gap, run the numbers honestly—including the trends, not just the snapshot—and make sure the gap is large enough and durable enough to justify what you’re giving up to get it.
Confession: I still don’t know which path I’ll pick when the next offer lands.
P.S. PhysEmp has job opportunities and salary reports by specialty — handy to know now, handier once the job hunt actually starts: physemp.com. And DocCommons is building a community for residents and attendings who want to talk to each other, not at each other; waitlist’s open at doccommons.com