U.S. research

Start where you are

The right first move is different for a final-year student, a graduate sitting at home with no U.S. status, and someone already here on a dependent visa. Find your row.

Start from home, now

Almost everyone treats research as something that begins after they arrive. It does not have to. Work you can do from your own country, with no visa and no money:

• Systematic reviews and meta-analyses. Needs a question, a librarian's search, and discipline — not a laboratory or a U.S. address.
• Retrospective chart reviews and registry analyses. The analysis travels even when you cannot; many are run entirely on de-identified data by collaborators in different countries.
• Case reports and case series from your own practice — modest currency on their own, but they demonstrate you can carry something to submission.
• A real contribution to someone else's project. Statistics, a language, data cleaning, a literature search, drafting the methods.

The point is not the paper. It is that when you email a principal investigator six months from now, you are not a stranger asking for a favor — you are someone with a finished product and a collaborator who will vouch for you. It also means that if a visa, a fee, or an entry restriction stops you from traveling at all, your application is still moving.

The positions that exist

"Research fellowship" covers arrangements that have almost nothing in common. What actually differs is who pays you, who sponsors you, and what you are allowed to do.

Salaried is the default you should be aiming at. A funded position means someone had to justify your cost, which usually means a mentor with a reason to care whether you produce. It also carries visa sponsorship and benefits. Everything below that bar should be judged against it.

Where to actually look

Method matters, but names save months. These are real starting points, with the details that decide whether you are eligible before you spend anything.

Federal research — where your passport is the qualification

The NIH intramural Visiting Fellow Program is the rare U.S. opportunity open only to people who are not U.S. citizens or permanent residents — NIH's own guidance routes citizens and green-card holders to a different mechanism instead. NIH sponsors the J-1 through its Division of International Services. Concretely:

• Pay is published, not negotiated. Postdoctoral stipends run from $68,544 at zero to one year of experience up to $117,083 at the top of the chart, on a unified table that has applied across all NIH institutes since October 2024 and is currently dated July 1, 2026. One catch worth knowing: the higher tiers that credit internship and residency years apply to U.S. credentials only — residency you completed at home does not move you up the scale.
• Length: up to five years, with initial appointments of one to three. A sixth year exists only as an exceptional terminal extension — written justification from your investigator, senior review, requested at least six months ahead, and it cannot roll into anything else.
• Eligibility: your doctoral degree should be within about five years of starting, and you should have no more than five years of relevant postdoctoral experience.
• How you apply: directly to individual investigators. There is no central selection and applications are reviewed year-round — NIH suggests starting about nine months before your preferred start date. Browse the roughly 1,200 intramural investigators at irp.nih.gov/our-research/principal-investigators, then write to the ones whose work fits.
• Visiting Fellows are trainees, not employees, and NIH does not sponsor them for H-1B.
• Note what NIH is not: the campus runs no core residency, so what you take away is the output and the letter, not a department that knows you and can speak to its own program director. Weigh that against proximity to a residency.

Beyond NIH, ORISE administers research appointments at CDC, FDA, EPA, USDA and Defense laboratories, posted on Zintellect (zintellect.com). These are training appointments rather than employment, and eligibility is set posting by posting: CDC's are generally open to foreign nationals who can participate as J-1 exchange visitors, but some offices are restricted to citizens, and a number of individual fellowships are citizen-or-green-card only. Read the citizenship line before anything else.

Fellowships built specifically for international graduates

A handful of programs exist expressly for IMGs. They are worth knowing because they are hard to find and each works differently:

Academic centers: the ordinary postdoctoral route

The programs named above are the ones that get passed around in forums, and between them they hold a small number of places a year. The route below is far larger — and it is not a program at all. It is the ordinary research-fellow or postdoctoral appointment that large academic medical centers make routinely, all year, in whichever laboratory has the money.

In September 2026 we read what ten large U.S. academic medical centers publish about postdoctoral and international-scholar appointments, to settle one question: does an international medical graduate qualify for a postdoctoral research appointment? Two findings.

First — the MD is a qualifying doctorate. Eight of the ten say so in their own published text: Duke, Emory, Johns Hopkins, Michigan, Northwestern, Pittsburgh, Stanford, and Vanderbilt. Johns Hopkins offers postdoctoral study "to a limited number of individuals holding MD, PhD, or equivalent doctoral degrees from accredited institutions." Vanderbilt requires "a terminal degree (i.e., Ph.D. or equivalent doctorate degree such as a Sc.D., M.D., PharmD, J.D.)." Michigan names "a professional doctorate (e.g., M.D., D.D.S, D.V.M.)." Stanford, simply, "a doctoral or MD degree." One wording is worth noticing: Northwestern's postdoctoral scholars are "individuals with recently granted degrees of PhD or MD." Some institutions attach a years-since-degree limit to the postdoctoral title, so a physician a decade out may be appointed as a research associate or visiting scholar instead — same work, same visa, different title. Ask rather than assume you are too late.

Second — no eligibility clause conditions the appointment on ECFMG certification or a USMLE score. In the eight policies that name the degree, the requirement is the degree itself; none adds an examination or a certification on top of it. ECFMG is named exactly once, by Duke, and not as a requirement to hold the appointment. (OHSU's appointment policy sits behind an internal login and could not be read; nothing on its public pages adds such a requirement either.)

Duke in fact writes down the answer most institutions leave unsaid. Its policy defines the candidate as someone who "holds a PhD or equivalent doctorate (e.g. ScD, MD, DVM)," then addresses ECFMG head-on: for appointments requiring a medical degree, equivalency determinations "have been made by the Educational Commission for Foreign Medical Graduates (ECFMG)" — but "if the Appointee does not have a valid ECFMG Certificate, Duke Visa Services will conduct an MD equivalency determination." A missing ECFMG certificate is treated there as a verification step the university performs itself, not as a bar.

Two qualifiers belong with all of that. Where a policy says how a foreign degree is handled, it is by an equivalency check rather than an exclusion — but Emory and Johns Hopkins require the degree to come from a "recognized" or "accredited" institution, which for a foreign school is a question to settle before you apply, not after. And the funding source still decides eligibility: as the federal traps above show, a laboratory's training grant can be closed to you while its ordinary research grant is not.

So the question worth asking is not "does this institution accept international medical graduates?" Among large academic medical centers the published answer is yes far more often than not, and none of these ten says no. The question is "does this investigator have money that can pay a non-citizen, and a reason to want me?" That turns the search from a short list of named IMG programs into most of the funded laboratories in the country — which is the argument for the method in the next section.

Two of the ten do not settle the degree question in public text at all: UAB's international-scholar page asks for a "Medical Diploma or certificate of highest degree earned" without naming a degree threshold, and OHSU's postdoctoral appointment policy sits behind an internal login. Neither says no; neither states a threshold.

Two centers where the numbers are public:

• MD Anderson Cancer Center publishes a postdoctoral stipend floor — $64,000, rising to $70,000 for biostatistics, bioinformatics, computational biology, therapeutics discovery and data science — and caps postdoctoral time at seven years, counting postdoctoral appointments held anywhere. It runs a dedicated visa office for trainees.
• Mayo Clinic posts a salary range on each individual listing rather than a single scale; research-fellow postings reviewed in September 2026 ranged from roughly $64,000 to $79,000. Research fellow roles require a doctoral degree and are aimed at people with no more than one prior postdoctoral fellowship, and Mayo sponsors J-1 Research Scholar status for them.

The approach is not to browse job boards hoping for a fit. Identify a funded investigator at the institution first — see finding a funded lab — and let HR and the visa office formalize a position the investigator already wants to create.

Departmental specialty research fellowships

In surgical and procedural fields especially, individual departments run one- and two-year research fellowships — two is common in surgical fields — that never go through any match and are often unadvertised. You apply by writing to the program or research director. These are conventional in general and subspecialty surgery, orthopedics, neurosurgery, urology, ophthalmology, plastic surgery, cardiology, oncology, and dermatology, and in several of those a research year is closer to an expectation for an international applicant than an advantage.

One distinction to keep straight: some programs with "international fellowship" in the name are clinical fellowships for people who have already completed residency, and require full ECFMG certification and USMLE Steps. Those are not pre-residency research bridges. Check what the program actually is before you apply.

The portals worth bookmarking

Finding a funded lab

Most people search for advertised jobs and send the same letter to everyone. The positions that suit IMGs are mostly not advertised — they are created when an investigator with money meets someone useful. So search for funded science, not for job postings.

Find investigators who have money

• NIH RePORTER (reporter.nih.gov) is the public record of who holds which federal grant. Use the advanced search: filter by activity code — an active R01, U01, or P01 is a funded project or program with a budget behind it, while a K-series award is mentored career-development support for the investigator's own salary rather than a budget to hire you — often excellent mentorship, rarely a paid slot. Filter by institution, or search the abstract text for your clinical topic. Each record gives you the contact investigator, the abstract, and the institution.
• PubMed — an author search tells you whether the lab is currently producing. A grant with no recent papers is a warning.
• ClinicalTrials.gov — for trial-running investigators, where coordinator and analyst roles live.
• Institutional postdoc offices and job boards for the positions that are advertised, and the NIH intramural investigator directory for the federal campus.

The email

Expect to write to a great many investigators — this is a volume exercise as much as a targeting one, and a low response rate is normal rather than a verdict on you. What separates the answered emails is that they are specific — and that they are about the investigator's science rather than about you. (Emailing a residency program about yourself is a different act, and rarely read.)

• A subject line that names the science, not your need. Not "IMG seeking research opportunity."
• One short paragraph on who you are — degree, Step status if you have it, when you are available, and plainly what sponsorship you would need. Do not hide the visa question; discovering it at week three ends conversations.
• One paragraph proving you read their work — a specific recent paper or the aims of a specific grant — and stating concretely what you could contribute to it. Chart review, a systematic review, statistics in R or Python, REDCap, recruitment, a language, an imaging skill, drafting.
• A one-to-two-page research CV, attached. Reverse-chronological, and built out of evidence rather than adjectives. "Research assistant" says nothing; "built and validated a 4,000-patient retrospective cohort, ran the survival analysis in R, drafted methods and results, presented the abstract" says everything.
• Follow up once, after a week or two, then move on. Track every message in a spreadsheet — who, which grant, when sent, what came back.

On timing. None of this is fast, and you should write in batches over months rather than in one push. NIH suggests approaching investigators about nine months before your preferred start date; the MGH trauma fellowship quotes three to four months for visa processing alone, after selection. Plan on that order of time for any position, and longer where an investigator has to create the role — the appointment paperwork, the institutional visa office, and a consular appointment each take their own time, and they run in sequence, not in parallel.

Judging an offer before you accept

An offer is easy to accept when you have been waiting for one. Score it first — a bad year is expensive in a way that is hard to recover from, because it consumes the time you needed and produces nothing to show for it.

Before you accept, get answers in writing to: the title and exact dates; what pays your salary; which office handles your visa and which category; the specific projects you will join and what data already exists; the authorship plan; how often you will meet the investigator; what previous fellows in this role went on to do; how many fellows they have trained in the last few years (a lab that trains people regularly has a working pipeline; one that has not may be improvising); what happens if the project stalls; and exactly what patient contact, if any, is permitted. An investigator who is glad to answer all of that is the one you want.

What it pays, what it costs

What these positions pay

Pay varies more than the job titles suggest, and much of what circulates online is unreliable. Four things are worth knowing before you negotiate or accept:

• Published scales are your benchmark. NIH's stipend table and MD Anderson's floor — both in where to look — tell you what a funded position pays before you ever negotiate.
• Many postings now carry a range, because several states require pay transparency in job advertisements. That range is the institution's own figure — unlike a salary-aggregator estimate, which is crowd-sourced guesswork and should never be treated as reliable.

As a rule: a salaried academic research-fellow position is a real job with a real salary and benefits, not a stipend gesture. If the number being discussed would not support living in that city, treat that as information about the position.

What it costs you

Even a fully funded position carries costs that people routinely fail to plan for:

• Visa and exchange-program fees — the visa application itself, plus the separate SEVIS fee for exchange-visitor status, and a further set of fees later if you change from research to clinical status.
• Health insurance meeting the exchange-program's required minimums, for you and any dependents — this is a regulatory floor, not an optional benefit, and it is a real monthly number.
• Proof of funding. If the position is unpaid or self-funded, the host institution will require documented evidence that you can support yourself before it will issue your immigration paperwork at all. The threshold is set by the institution, not nationally, so ask early — this is a common reason a "confirmed" unpaid position never actually materializes.
• Relocation and the first two months — flights, deposits, and living costs before a first paycheck arrives.

The exam, certification, and application fees are itemized in ECFMG & the Pathways and Where to Apply.

The decisions that are hard to undo

Most of this page is recoverable if you get it wrong. This section is not. A handful of immigration rules around a research year close doors for years, and every one of them is avoidable if you know it before you accept a position. None of this is legal advice. Confirm your own case with the host institution's international office — and an immigration attorney where it is complicated — against the primary sources at the time you act: Intealth for the change of category, the State Department's exchange-visitor pages for J-1 rules and fees, and USCIS for H-1B. Re-check in the week before you file or travel.

Research status and clinical status are different things

A research appointment normally runs on a J-1 Research Scholar visa sponsored by the host institution. Residency runs on a J-1 Alien Physician visa, for which ECFMG/Intealth is the only sponsor. Moving between them is a formal change of category that Intealth files and the U.S. Department of State decides. Intealth's memo on it (Request for a Change of J-1 Visa Category, dated February 23, 2026) sets out three ways to lose that option.

Budget for the fees as separate line items rather than one number: an Intealth application fee of $370; $567 in case-processing fees if the request is determined to be a Department of State case (a $200 Intealth review fee plus the $367 DOS processing fee, which the State Department also publishes on its own exchange-visitor pages); and the $220 SEVIS fee, payable again for the new category.

There is a second way across, worth knowing because the rules above can close the first one: end the research appointment, leave the United States, and re-enter on a new clinical DS-2019 issued through ECFMG. That sidesteps the change-of-category restrictions entirely — at the cost of consular processing time and a gap in status you have to plan for and fund. It is a months-ahead decision, not a June one.

The two-year home-residence requirement

Known by its statutory section, 212(e), this sends some J-1 physicians home for two years before they can hold an H or L visa, adjust to permanent residence, or in most cases change status inside the United States at all. It attaches for any one of three independent reasons:

• Government funding — from your home country, from the U.S. government, or from an international organization. Direct or indirect.
• Your field appears on your country's Exchange Visitor Skills List. The list was revised on December 9, 2024, and the revision is retroactive: someone subject only because of an earlier list is no longer subject if their country is not on the current one. Several large source countries — including Brazil, China, India, and South Korea — are not on the 2024 list.
• Graduate medical education as an ECFMG-sponsored physician. A J-1 residency is essentially always subject, whatever the Skills List says.

A home-government scholarship attached to a research year can carry a consequence that outlives the position by years — which is why the funding source belongs on the list of things you get answered before you accept.

Patient contact on a research visa

A physician doing research may have only incidental patient contact, and it is documented in writing. The certification the State Department requires — five specific points under 22 CFR 62.27(c)(1)(ii), signed by the dean of the involved U.S. medical school or a designee and appended to the DS-2019 — states that the program is predominantly observation, consultation, teaching, or research; that any incidental contact is under the direct supervision of a physician who is a U.S. citizen or permanent resident licensed in that state; that you will not be given final responsibility for diagnosis and treatment; that activities conform to state licensing rules; and that the experience is not creditable toward specialty board certification.

Two things follow. Never accept an informal suggestion that you can "help out" clinically — the consequences land on you, not on the person who offered. And note the last point: this exposure is not U.S. clinical experience in the sense a program director means, and cannot be counted as such. For that, see U.S. Clinical Experience.

What a visitor visa cannot be used for

The Foreign Affairs Manual is explicit that a physician may enter in B status to observe U.S. medical practice and consult with colleagues, provided there is no remuneration from a U.S. source and no patient care. That is the whole permitted scope, and visa-free entry under the Visa Waiver Program carries the same limits. What B status does not cover is productive work — the test is the nature of the activity, not whether you are paid for it, so declining a salary does not convert a job into a permitted visit.

A separate narrow provision lets a student at a foreign medical school take an unpaid elective clerkship — and it is written for students in their normal third or fourth year, as part of the foreign degree. The manual says plainly that it does not apply to graduate medical training. If you have graduated, that route is closed to you.

Two bars on coming back as a research scholar

These restrict repeat research-scholar participation. Neither blocks a clinical J-1, F-1, or H-1B.

• The 24-month bar. Having completed a J-1 program as a professor or research scholar, you cannot begin a new professor or research scholar program for two years. The regulation ties this to having completed such a program; State Department guidance as relayed by university sponsors applies it regardless of how short the program was, and to J-2 dependents as well — so treat even a brief research-scholar stint as triggering it.
• The 12-month bar. If you were in the U.S. in any J status for six months or more in the preceding year, you cannot begin a new professor or research scholar program for twelve months. The exceptions are a bona fide transfer, a prior presence of less than six months, and time in the Short-Term Scholar category.

That last exception is the useful one: a Short-Term Scholar appointment of six months or less triggers neither bar, which makes it a clean way to do a defined project or a scouting visit without spending your research-scholar eligibility.

H-1B, honestly

H-1B carries no home-residence requirement of its own and is the cleaner long-run status — but if you are already subject to 212(e) from a J-1, it is closed to you until that is satisfied or waived. Its cost picture has also been in open dispute for a year, so do not plan around a headline. As of September 2026, in sequence: a September 2025 proclamation imposed a $100,000 payment on certain new H-1B petitions for beneficiaries abroad; a federal court in Massachusetts vacated the guidance implementing it in June 2026 and an appeals court declined to restore it in July; the underlying restriction was then extended on paper in September 2026, through September 21, 2027; and on August 25, 2026 DHS proposed a separate $103,265 fee on every cap-subject petition, excluding cap-exempt employers such as universities; comments closed September 24, 2026, and it is not final. The practical position today is that the payment is not being collected while the vacatur stands, and that this could change.

The other vehicles, briefly

• O-1 for genuinely accomplished researchers — 212(e) does not bar an O-1 the way it bars H and L.
• TN for Canadian and Mexican citizens: the physician category is teaching or research only. Patient care incidental to that is permissible; a recurring or scheduled patient load is not.
• E-3 for Australian citizens in a specialty occupation, outside the H-1B lottery entirely.
• Dependent work authorization — a J-2 spouse who obtains an employment authorization document — and an H-4 spouse in the narrower circumstances where one is available — can take a research job on the open market, which quietly solves sponsorship for a household where one partner already holds status. A J-2's income may not be used to support the J-1.

A trap common to O-1, E-3, and TN: if you are subject to 212(e), you generally cannot change status to them from inside the United States — you would need to apply at a consulate abroad.

Making the year produce

Treat the appointment as a production project with a deadline, because that is what it is.

• Run more than one project. A sensible pipeline is one substantial original study, one shorter database or secondary analysis, and one review. A single project with one delayed approval or one slow co-author leaves you with nothing to submit.
• Budget for onboarding before you can touch anything. Expect human-subjects and research-integrity training, HIPAA and data-security modules, conflict-of-interest disclosure, background screening, and occupational-health clearance if you will be anywhere near patient areas — and the project itself may be waiting on IRB approval or a reliance agreement. Ask in your first week what you can work on while that clears, or a month disappears.
• Settle authorship at the start. Ask what would count as excellent work in this role, and what authorship position the work would earn. Investigators respect the question.
• Favor finished over numerous. One credible first-author piece of work carries more than a pile of thin case reports, and a long list of thin ones reads as padding. See Red Flags & Filtered Metrics.
• Go to the meeting with something. A national meeting in your specialty is worth the cost when you have an abstract accepted and a mentor who will introduce you. Follow up within days with a specific reason to continue the conversation.
• Cultivate the letter deliberately — ask early what excellent looks like, and a few months before you apply ask directly whether they can write a strong one. A detailed letter from a year-long supervisor is a different instrument from a polite one. See Earning a Strong Letter.
• Work where the residency is, if you can. The real prize is often not the publication — it is that faculty in a department that runs a residency know you, and can say so to their own program director. That internal advocacy cannot be manufactured in an application.

When it is time to write it all down, the ERAS application guide covers how to describe research in the Experiences section — what you did and what came of it, not what you learned from it.

Sequencing it

A research year helps when it is finished before you apply, and hurts when it competes with the things that actually gate your application.

The most useful way to hold all of it in your head is as five tracks running at once, not five boxes in a row: the research itself (mentor → data → analysis → abstract → manuscript → an advocate), your credentials (Steps → English → a Pathway → ECFMG certification), clinical exposure (recent, lawful, and specialty-relevant), immigration (hold status → protect the route into clinical status → start sponsorship the week you match), and the application (program list → signals → interviews → rank list). People who run these in parallel get far more out of a research year than people who finish one before starting the next — and the credential track is the one that must never wait.

• Credentials come first. Your Steps, the English requirement, and ECFMG certification are gates; research is a differentiator. Never let the research year be the reason a Step slips.
• Work backward from September. Applications go to programs on September 2, 2026, and programs begin reading on September 23, 2026. Abstracts, submissions, and letters need to exist before that, which means the work needs to be well underway by the spring.
• A spring start suits a one-year position. Starting in spring gives output time to mature before the following application season, rather than leaving you applying in the first weeks of a new job with nothing finished.
• Keep the visa runway in view. If the plan is research now and residency after, keep the research stint short and deliberate, and arrange any extension that bridges the spring-to-July gap well in advance.
• Build the list around sponsorship and cutoffs first — research does not move a program that cannot take you. See Where to Apply, and remember that a signal spent on a program that cannot sponsor you is wasted (Preference Signaling).