Doctors often want to study a medicine or device in their own way, and they ask the company that makes it for money, product or both. Investigator-initiated study review is how the company decides which of those studies to support: checking the science, the fit, the budget, and whether someone is already studying the same question.
This guide is for anyone who runs, reviews or pays for that decision: medical affairs leaders, program managers, medical directors who review proposals, committee members, and the compliance, legal and finance teams who sign off. No background needed.
Imagine you are a doctor. Dr. Elena Varga treats Crohn's disease at a university hospital. She wants to know whether a drug called Veltrimab helps patients after other treatments have failed.
She can't fund the study herself, so she asks the drug's maker, Northwind Bio, for $448,000 and the drug. Northwind's medical team now has to review her proposal.
Her proposal · VELOCITY-CD
Patients
Crohn's, after anti-TNF
Drug
Veltrimab
Measures
Endoscopic response
When
Week 12
Northwind's own phase 3b
Patients
Same
Drug
Same
Measures
Same
When
Week 12
Same questionscore 74 → 62once the overlap is found.
On paper her idea scored well. But Northwind's own trial already answers the same question in 2027. So the reviewer does not just say no: she asks Dr. Varga to make the one new part of her study, healing seen on an MRI scan at week 24, the main question.
That decision, made fairly, quickly and on the record, is investigator-initiated study review.
What investigator-initiated study review is, and why it is hard
The studies, the steps, the hard questions and the rules, explained without selling anything. If you are new to this, start here.
01 · In plain words
Investigator-initiated study review in five minutes
Companies that make medicines, medical devices and diagnostic tests run their own clinical trials. But doctors who use those products often have questions the company never studied: does it work for a different group of patients, at a different dose, alongside another treatment, or in everyday care?
When a doctor wants to answer one of those questions with a study of their own, they can ask the company for support: money, the product, or both. That is an investigator-initiated study. The doctor or their hospital designs the study, runs it and owns the results. The company only decides whether to support it, and then pays in stages under a written agreement.23
Deciding is the hard part. For every request, someone has to check the science, whether it fits the topics the company said it would support, whether the budget is a fair price, whether the hospital can recruit enough patients, and whether someone is already studying the same question. It must all be done with sales and marketing kept out, because research money given to win prescriptions can be a crime.1214Investigator-initiated study review is that decision process.
The life of one study, in five steps
A doctor has a questionDr. Varga wants to know if Veltrimab helps patients after other drugs failed.
She sends a proposalA short concept and a $448,000 budget go to Northwind Bio.
The company reviews itScience, fit, budget, recruiting and overlap. This is where the copy of Northwind's own trial is caught, or missed.
A committee decidesApprove, decline, wait for the next round, or ask for changes.
The study runs, and is reportedPaid in stages to the hospital, reported to the government, results published.
Step 3 is the subject of this guide. It is where most of the time and most of the risk sit.
Four things that surprise people
Most requests end in a no
In one 2025 study of cancer research proposals, companies approved 25% and declined 72%. Every no still needs a careful review and a clear reason.1
The wait is long
The same study found a median of 61 days from proposal to the company's decision, and 18 months before the study started. Doctors notice, and good ideas go elsewhere.1
A company can fund a copy of its own trial
A proposal has to be checked against the company's own studies, the studies it already funds and public registries. In October 2026, 6,563 active registered studies had a non-industry sponsor and an industry partner.224
Research money is watched like a gift
It must be a fair price, decided without sales, and reported. US companies reported $9.50 billion in research payments for 2025.121819
Four words you will see on this page
Concept
The short first proposal a doctor sends. A full protocol and budget follow if the company asks.
Fair market value
What the work would cost at an arm's-length price. Research payments must not be more than that.
Overlap
A proposal that asks a question another study already asks.
Open Payments
The US program where companies report what they pay doctors and teaching hospitals, research included.
What an investigator-initiated study is, and the names it goes by
In shortThe doctor or hospital designs the study, runs it and owns the results. The company gives money, product or both, and does not run it. Companies call the same thing by different names: IIS, IIT, ISS, ISR or ESR.
An independent investigator or institution has a research question about a medicine, device or diagnostic. They write the protocol, take on the sponsor's legal duties, and ask the company that makes the product for funding, product, or both. The company reviews the request, and if it fits, supports it under a written agreement. It does not run the study.
The definition, from the people who write the rules
Medical affairs practice
"Unsolicited, independent research where the investigator or the institution … serves as the sponsor and a pharmaceutical or medical device company provides support in the form of study drug/device and/or funding."2
Good Clinical Practice · ICH E6(R3)
A sponsor-investigator is "an individual who both initiates and conducts, alone or with others, a clinical trial". In an investigator-initiated trial, that person (or their institution) carries the sponsor's obligations, not the company.45
Investigator-sponsored research, externally sponsored research: names for the same model
Companies use different names for the same thing. Searching your own company's term matters, because investigators and colleagues will use it.
IIS investigator-initiated studyIIT investigator-initiated trialIIR investigator-initiated researchISS / IST investigator-sponsored study or trialISR investigator-sponsored researchESR externally sponsored research
Investigator-initiated study versus a company-sponsored trial
Question
Company-sponsored trial
Investigator-initiated study
Who has the idea and writes the protocol?
The company
The investigator; the company "generally should not participate in the scientific design"3
Who is the sponsor?
The company
The investigator or institution, with all sponsor duties: regulatory filings, ethics approval, registration, monitoring, data and reporting4
What does the company give?
Everything; it runs and pays for the trial
Funding, product (drug, device, reagents), or both, under a written agreement26
Who owns the data and the publication?
The company
The investigator, who publishes; the company may review and comment on drafts3
Which company team handles it?
Clinical development and operations
Medical affairs, kept apart from sales and marketing212
Why companies support investigator-initiated research
Evidence gaps
Investigator studies "may address important biological and clinical questions that may not be prioritized by pharmaceutical sponsors".1 They test products in new populations, doses, combinations and settings, and study biomarkers.2
Real-world questions
Clinicians ask practical questions: how a product performs in routine care, its economic value, its effect on a lab's or a ward's workflow.18
The research community
A fair, transparent program builds lasting relationships with academic investigators, who are busy clinicians with little patience for slow administration.2
Part of the medical plan
Supported studies sit inside the medical affairs evidence plan, next to the company's own studies, not in a commercial budget.2
Studies can be interventional, observational, retrospective or preclinical, and device and diagnostic makers run the same model.7811
03 · The lifecycle
The investigator-initiated study lifecycle: concept, proposal, committee, contract, milestones and publication
In shortTen steps, from a published list of topics to a published result. Most of the time is spent waiting: for missing documents, for the next committee, for contracts. The calculator at the end sizes your own review load.
Ten steps, eight parties. Most of the elapsed time is waiting: for missing documents, for the next committee meeting, for contracts and ethics approval. The blue row is the investigator's institution, which is the sponsor throughout.
Who ↓ · Step →
1Areas of interest
2Concept
3Full proposal
4Scientific review
5Committee
6Contract & FMV
7Ethics & registration
8Supply & milestones
9Publication & closeout
10Disclosure
Investigator / institution (sponsor)
Reads published areas
Submits concept via portal
Protocol, budget, CV
Answers questions
Waits for the decision
Negotiates and signs
IRB/EC, IND or CTA, registry
Enrolls; reports progress
Publishes; final report
Field medical (MSL)
Explains areas when asked
Receives unsolicited ideas
Scientific exchange
Program manager
Publishes areas
Checks for gaps
Chases missing items
Assigns reviewers
Builds the docket
Hands to contracts
Checks approvals in place
Tracks milestones
Chases publication
Feeds payment data
Scientific reviewers
Quick screen
Merit, fit, feasibility, overlap
Present and recommend
May comment on drafts
Review committee
Sets priorities
Approve, decline, defer, ask for changes
Compliance & legal
Approves wording
Red-flag screen
Attends; may hold
Contract terms
Open Payments, EFPIA
Finance / FMV
Annual envelope
Budget lines vs benchmark
FMV sign-off
Pays the institution
Reconciles funds
Payment records
Clinical supply
Can we supply it?
Confirms stock
Labels and ships product
Returns or destruction
Sales and marketing stay out of selection, amount and approval, at every step →Every decision documented for audit
Swipe sideways to see all ten steps.
Steps compiled from published sponsor processes and industry guidance.2367811
What waits on whom
Investigator → company
61 daysMedian from concept to a sponsor's decision (range of the middle half: 31–183 days), 68 oncology concepts, 2014–2024.1
Proposal → committee
Monthly or lessPublished cadences: monthly at one specialty pharma; every two months at another, which "cannot commit" to a review time.910
Protocol → reply
20 days to 3 months20 working days for a complete protocol at one diagnostics company; one to three months at one device maker's program.78
Submission → activation
18 monthsMedian from initial submission to trial activation for sponsor-approved concepts in the same study.1
Company → institution
~4 paymentsTypically contract execution, 50% enrolled, 100% enrolled, and the manuscript or final report (the last 10–15%).2
How support is paid: in stages, to the institution
Swipe sideways to see the full diagram.
The four-payment pattern is from a medical affairs best-practice deck; one large pharma's guide also ties 10% to the final report and 10% to publication. Holding back the last payment is how companies make sure results are published.236
Size your own review load
Defaults are illustrative, not benchmarks: no published benchmark exists for concepts per year or hours per review. Replace each with your own number.
Rough review load
13concepts per committee meeting
19reviews per reviewer a year
450reviewer hours a year
38reviewer hours before each meeting
Medical directors review on top of their main role. These are hours taken from patients and their own research.
04 · Why it matters now
More money is reported, more studies are registered, and enforcement is reaching research support.
In shortMore research money is reported every year, more studies are registered to check against, and enforcement now reaches research support.
Six numbers worth knowing, each with its source.
$9.50B
in research payments reported to US Open Payments for 2025, out of $14.67B in all reported payments and transfers of value.
CMS Open Payments, program year 2025, published 2026.18
61 days
median from concept submission to a sponsor's decision, with the middle half between 31 and 183 days. Investigators feel the slowest cases.
Gerber et al., Clin Cancer Res 2025; 68 thoracic oncology concepts.1
25%
of concepts sent to pharmaceutical sponsors were approved (15 of 60); 72% were declined. Most of the review effort ends in a no, and each no needs a reason.
active studies on ClinicalTrials.gov have a non-industry lead sponsor and an industry collaborator (34,012 ever registered). A rough proxy for supported investigator studies, and a lot to check for overlap.
civil penalty for knowingly failing to report payments: the largest Sunshine Act recovery since the law passed, part of a $46M resolution in August 2026.
ICH E6(R3), the new Good Clinical Practice guideline, is final: principles and Annex 1 adopted January 2025, Annex 2 in June 2026, consolidated text dated June 16, 2026.4
OIG is replacing its old industry guidances with new industry-specific ones. None for drug makers has been published yet; the 2003 pharmaceutical guidance sits in OIG's archive and remains its most specific text on research funding.1213
A five-year Corporate Integrity Agreement signed in August 2026 names "support of ISSs" among the research activities it covers, and requires researchers to be paid from a pre-set, independently benchmarked fair-market-value rate structure.23
A note on the numbers. The 61-day median and 25% approval rate come from one academic consortium's concepts in thoracic oncology; your program's numbers will differ, so measure your own (section 17). The registry count is a proxy, not a full count of company-supported studies.
05 · The hard questions
Fair market value, overlap and fit: the three hard questions in every investigator-initiated study review
In shortEvery review asks three hard questions: is the budget a fair price, is someone already studying this, and does it fit what the company said it would support? Whether the hospital can recruit enough patients is a fourth.
Published review criteria are similar across companies: scientific merit and a sound statistical plan, fit with the company's areas of interest, feasibility at the site, a budget at fair market value, a publication plan, and a check that the study is not unnecessarily duplicative.7811 Three of these take the most work and carry the most risk.
1
Is the budget at fair market value?
Research payments "should be fair market value for legitimate, reasonable, and necessary services."12 Payments should be "reasonable and based on work performed", under a written contract.16
No regulation prescribes how to calculate it. In practice, each budget line is compared with an external benchmark: one large pharma assesses every budget "against an externally benchmarked database prior to approval".6
2
Does it overlap a study already under way?
Practice guidance says to "limit overlap with company sponsored strategy/studies", avoid patient overlap, and that a "good tracking system [is] necessary to identify overlap".2 One device maker asks whether the proposal is "unnecessarily duplicative".7
The comparison runs across company studies, other funded investigator studies and public registries.
3
Does it fit what we said we would support?
Companies "should not solicit research projects" but "may publicly disclose the general areas or types of studies that it does or does not support".3
Published areas of interest are therefore both a scientific filter and a compliance control: support follows the published priorities, not a relationship.
Fair market value budgets, line by line (a fictional example)
Swipe sideways to see the full diagram.
A fictional case: a $448,000 request is brought to $397,750, inside the benchmark range of $360,000–$410,000, by pricing central reading at the benchmark rate and applying the program's 25% overhead cap. Overhead caps are a program setting: sponsors and institutions set them differently check your policy. Any budget approved above FMV should carry a written reason.
Overlap, element by element
Two studies overlap when they ask the same question. Evidence synthesis compares questions on the PICOTS elements; a systematic check compares six of them, each marked same, partly or different:28
PPopulation
ITreatment
CControl
OOutcome
TTiming
SRegion or setting
Overlap is not always a reason to decline. A concept that repeats a company study's primary question may still add something new, such as a different endpoint or population, and can be reshaped around it. Letters to investigators should never reveal a company's internal, unpublished studies.
Feasibility: will the site enroll in time?
The arithmetic
months to enroll = patients needed ÷ patients per month at this site
In the fictional example, 60 patients in 18 months needs 3.3 a month; the site's own history is 1.1 a month, so about 55 months. A second site, or a powered sample of 40, changes the answer.
Why check
About one activated site in ten enrolls no patient, and 41% miss their target (2012 data).25 A study that never finishes is support spent with no evidence in return, and the final payment exists to tie money to delivery.2
06 · The rules
Separation from sales, anti-kickback risk and Open Payments research reporting
In shortResearch support is a payment to doctors and hospitals. So it must be a fair price, decided without sales or marketing, and reported. In the US, the Anti-Kickback Statute and Open Payments set the stakes.
Research support is a transfer of value to people who prescribe, use or buy the company's products. That is why it is reviewed on science, paid at fair market value, kept away from sales and marketing, and reported. Statements below are as of October 2026 and are a summary, not legal advice: confirm with your own counsel.
United States · law
The Anti-Kickback Statute
It is a federal crime to knowingly and willfully offer or pay anything of value to induce the purchase, order or recommendation of items paid for by federal health care programs (42 U.S.C. 1320a-7b(b)).14 A research grant offered to win prescriptions or a formulary listing can be a kickback.
United States · OIG guidance (2003)
Keep research grant-making away from sales and marketing
"To reduce risk, manufacturers should insulate research grant making from sales and marketing influences." "Research contracts that originate through the sales or marketing functions … are particularly suspect." "Post-marketing research activities should be especially scrutinized …"
The same guidance requires payments for research at fair market value. As of October 2026, it is OIG's most specific guidance for drug makers; OIG has archived it pending a new industry-specific version.1213
Industry codes
PhRMA's clinical-trial principles and 2014 position; AdvaMed for devices
The PhRMA Code on interactions with health care professionals (effective 2022) does not govern research: it points to PhRMA's separate Principles on Conduct of Clinical Trials.1516 PhRMA's 2014 position on investigator-initiated studies says support should not go preferentially to investigators who are customers, should not be conditioned on purchases, should be paid in milestones to the institution, and should be no more than fair market value.3 The AdvaMed Code for device makers lets sales staff give input, but they "should not control or unduly influence" who receives support or how much.17
Good Clinical Practice
The investigator is the sponsor
Under ICH E6(R3), the sponsor-investigator holds the sponsor's duties.4 In the US, the responsible party registers an applicable clinical trial on ClinicalTrials.gov within 21 days after enrollment begins;26 in the EU, trials of medicines run under the Clinical Trials Regulation, and all ongoing trials had to move to the public CTIS system by January 30, 2025.27 Companies typically check that ethics approval, registration and the agreement are in place before releasing money or product.67
Transparency reporting, stated precisely
United States · Open Payments (Physician Payments Sunshine Act)
What is reported: research payments made under a written agreement, a research protocol, or both.19
Payments to an institution count: a research payment to an entity with at least one physician covered recipient as principal investigator is reportable, with the physician listed as a principal investigator.19
Delayed publication: research payments for new product development or new uses of an existing product can be granted a delay in publication for up to four years, or until FDA approval. Reported is not the same as immediately published.19
Penalties (base amounts, adjusted annually): $1,000–$10,000 per payment not reported, up to $150,000 a year; $10,000–$100,000 per payment for knowing failures, up to $1,000,000 a year.20
Europe · EFPIA Code disclosure
Research and development transfers of value to health care professionals and organizations are disclosed in aggregate, not per named recipient.21
Other transfers of value are disclosed per recipient where possible.21
National laws and codes can add their own rules; device makers follow their own industry code check each country.
So a program needs, for each payment: the recipient, the principal investigators, the amount, the agreement, the country, and whether a delay in publication applies.
What enforcement looks like
2020 · $3.5M civil settlement
Research grants used as kickbacks (as alleged)
A drug maker settled allegations that its research grants were usually initiated by sales representatives or marketing executives and conditioned on the hospital adding its product to the formulary. It approved 28 grants with "little or no information regarding the proposed research", did little follow-up, and some recipients did no work at all. The company admitted no wrongdoing.22
2026 · $46M criminal and civil resolution
A record Sunshine Act penalty and a CIA covering research
A drug maker resolved kickback allegations centered on advisory boards, consulting and specialty-pharmacy payments, including a $1.55M penalty for knowingly failing to report payments. Its five-year Corporate Integrity Agreement covers "support of ISSs", requires written agreements, an annual research budgeting plan, a benchmarked FMV rate structure, and a monitoring program reviewing 5% or 10% of researcher arrangements each period.23
What an auditor will ask to see, for any supported study: the request as received, who reviewed it and against which criteria, the overlap and feasibility checks, the budget against fair market value, who approved it and when, the agreement, the milestones paid, the publication, and the transparency record. If that takes days across systems to assemble, the program is exposed.
07 · How teams review today
Four ways companies review investigator-initiated study proposals
In shortMost programs combine two or three of these. The real questions are when a proposal is first read closely, how overlap is found, and whether each decision can be rebuilt for an auditor.
1 · By hand
Email, spreadsheets and the committee
Proposals arrive by email or a portal. A program manager tracks them in a spreadsheet and chases what is missing. Reviewers read each one before the meeting.
Catches
What an experienced reviewer notices and remembers
Misses
Overlap nobody remembers; missing items found at the meeting
2 · Workflow
A grants management system
A portal, routing, reviewer assignment, contracts and milestone payments, all in one system.
Catches
Deadlines, approvals and payments
Misses
It moves documents but does not read them: fit, overlap and budget checks stay manual
3 · Benchmarks
Fair market value data
Budget lines compared with an external benchmark database for the country, as some large companies do before approval.6
Catches
Budget lines priced above the market
Misses
The science, the fit and the overlap
4 · Automated
Prepared, cited review packets
Software reads each proposal as it arrives, checks fit, overlap, recruiting, budget and compliance, and drafts a packet with sources. People still recommend and decide.
Catches
Overlap and missing items before the committee meets
Watch for
Whether every finding cites its source, who decides, and that internal studies stay internal
Review by hand and prepared review, side by side
Question
Typical review by hand
Prepared, cited review
When is a proposal read closely?
Often the week of the committee, sometimes in the room
On arrival, with missing items named at once
How is overlap found?
From memory, if someone knows the other study
A search of company, funded and public studies, element by element
How is the budget checked?
"Looks reasonable," or a separate benchmark exercise
Each line against your benchmark and overhead cap
Whose criteria?
Each therapeutic area's habits
The committee's own written rubric and weights
What is left behind?
Meeting minutes and an email trail
A cited packet, a signed recommendation and the decision record
Whichever way you review, the decision belongs to people: a scientific reviewer recommends, compliance clears red flags, and the committee decides.
Who is involved, and what each cares about
Roles in a typical program
HE
Head of medical affairs or evidence generation
Owns the program and its budget, often chairs the committee. Cares about evidence that answers the company's real gaps, a fair process and a clean audit.
PM
Investigator-study or grants program manager
Runs intake, completeness checks, reviewer assignment, the docket, contracts, milestones and letters. Cares about cycle time and chasing less.
MD
Medical directors as scientific reviewers
Review concepts in their area on top of their main role. Care about reading less and knowing quickly whether the question is already answered.
RC
Review committee
Medical affairs, clinical, biostatistics, regulatory, legal; sometimes others ad hoc. Care about a consistent basis for every decision.211
CO
Compliance and legal
Care about separation from sales, fair market value, contracts and transparency reporting, and about rebuilding any decision for an auditor.
FN
Finance and FMV
Benchmarks, budgets, accruals and payments. Care about predictable spend against the annual envelope.
CS
Clinical supply
Forecasts, labels and ships product. Cares about knowing the quantity before approval.
MS
Medical science liaisons
Explain areas of interest and receive unsolicited ideas. Care about investigators getting a fair, timely answer.
The investigator is the customer of this process. A slow or opaque review costs the company more than time: good science goes to another sponsor, and relationships the medical team built over years wear thin.
Part B · See it working
Botminds Investigator-Initiated Study Review, one proposal at a time
Part A described the work without naming a product. This part shows how Botminds does it, for a fictional company, Northwind Bio. Every company, product, doctor, hospital and study name in the screens is sample data.
08 · One proposal, start to finish
Follow Dr. Varga's proposal through the review
In shortThis is the example from the top of the page, inside the product. Agents read, check and draft. A reviewer recommends and signs, and the committee decides.
1
The proposal is read, and a packet is built
Dr. Varga's concept form, budget sheet and cover email arrive. Seven agents each do one job: read the proposal, match it to Northwind's published research priorities, judge the study design, search for overlap, check recruiting and budget, screen for compliance, and draft the packet.
Each step is logged with what it found. The last step waits for the scientific reviewer.
The packet, built in seconds. Seven agents, each step logged with what it found.
2
The overlap is found before the committee meets
412 studies are searched: Northwind's own, the doctor-led studies it already funds, and a public registry. The closest is Northwind's phase 3b, VELTRIS-CD: same patients, same drug, same measure, same week 12. Overlap: 0.86.
The concept falls from 74 to 62 and from seventh to last. The check also names what is new: healing on an MRI scan at week 24, which the company study does not measure.
Overlap with ongoing studies. What is the same, what the proposal adds, and a suggestion to reshape it rather than decline.
3
The reviewer recommends, and signs
Dr. Maya Chen, a medical director, reads a drafted summary with seven citations. Recruiting 60 patients in 18 months is unlikely at the site's rate of 1.1 a month, and the budget's 30% overhead is above Northwind's 25% cap.
She recommends asking for a full protocol with four conditions, starting with making week-24 healing the main question. She edits the conditions and signs. The agents decide nothing.
Summary and recommendation. Every sentence cites its source; the reviewer chooses what goes to the committee.
4
The committee decides, on one page
The committee sees the signed recommendation, the scores and the sources on one page per proposal. Voting members vote; clinical supply and compliance attend without a vote; the chair signs the decision.
The letter to Dr. Varga explains the conditions, and never mentions Northwind's internal studies.
Committee. The signed recommendation, reviewer scores and each member's vote.
09 · Every screen
The rest of the work: the ranked round, budgets, recruiting, compliance and priorities
In shortOne proposal is the story. These screens show the whole round: every open proposal ranked, the budget against fair market value, recruiting, the compliance screen, and where the evidence gaps are.
Scorecard. Every open concept in the cycle, scored on four criteria with weights the committee sets (30 · 30 · 20 · 20 by default). A new concept climbed to 74, then the overlap check found the company's own phase 3b asking the same question at week 12, and it fell to 62. A compliance hold sits below the line, unranked.
Overlap finder. Company, funded investigator and public registry studies searched; the closest match compared on population, treatment, control, outcome, timing and region; the score's arithmetic shown.
Budget against fair market value. Each line against the benchmark range, the overhead cap applied, and the revised total shown against the FMV range. The reviewer chooses the adjustments.
Priority fit. The best-matching published priority, quoted, with runners-up and what the concept is missing.
Feasibility. Months to enroll at the site's own historical rate, with scenarios: as requested, a second site, a powered sample of 40.
Compliance screen. Sponsor status, promotional wording, sales involvement, recent payments to the investigator, and the transparency duty, each cited. Red flags go to compliance, which clears them.
Signed, with meaning. The e-signature records who signed, when, and what the signature means.
Research priorities. Each published priority with what the company runs, funds and receives, marked covered, crowded or open gap, plus the committee countdown and funding left by area.
Ask in plain words. "Why is this concept ranked last?" The answer shows each criterion, the reason and its source, and what would change the ranking.
Review rules. Weights, overlap alert, minimum fit, overhead cap, the commercial-involvement screen and the benchmark set. Every change is versioned in the audit trail.
Dashboard. Receipt-to-decision time by step, fit with published priorities, overlaps caught early. All figures are fictional sample data: the 27-day median is the solution's design target, not a measured result.
Select any screen to open it full size. Company, products, doctors, hospitals and study names in the screens are fictional.
10 · Agents, and who signs
Seven agents prepare the review. People recommend, clear and decide.
In shortEach agent has one job and one thing it never does. The reviewer recommends and signs, compliance clears red flags, and the committee decides.
The seven agents
Concept reader
Turns the concept form, budget sheet and cover email into a structured record, every field cited, and names what is missing.
Never guesses a missing field.
Priority matcher
Finds the best-fitting published priority, quotes it, scores the fit and lists the runners-up.
Never uses an unapproved priorities version.
Design appraiser
Scores scientific quality on the committee's rubric: a clear question, a validated endpoint, a design that answers it, a power calculation.
Never has the final word: the reviewer confirms every score.
Overlap finder
Searches company, funded investigator and public registry studies, compares six elements, scores the overlap and says what the concept adds.
Never shows internal company studies to investigators.
Feasibility and budget checker
Projects enrollment from the site's own history, compares each budget line with fair market value benchmarks and the overhead cap, and works out product quantity.
Never edits the investigator's budget.
Compliance screener
Looks for sales or marketing involvement, links to prescribing, promotional wording and recent payments, and notes the transparency duty.
Never clears a red flag: compliance does.
Packet and letter writer
Drafts the cited reviewer summary, the one-page committee packet and, after the decision, the letter to the investigator.
Never signs, and never names internal studies in a letter.
People decide; everything is cited
Recommends · e-signsScientific reviewerConfirms scores, edits conditions, signs the recommendation.
Decides · chair signsReview committeeVotes with a quorum; the chair signs the decision.
Clears holdsComplianceClears or keeps every red flag before scientific review continues.
Confirms · non-votingClinical supplyConfirms product quantity and timing before approval.
Every score and finding cites the passage it came from: concept, budget sheet, registry record, site history or priorities document. Budgets above fair market value need a written reason. Weight and threshold changes are versioned. Every step, by a person or an agent, lands on the audit trail.
The people in the screens
Fictional, from the working demo
MC
Dr. Maya Chen · medical director and scientific reviewer
Reviews three to five concepts a cycle on top of her medical role. Recommends and signs.
"Tell me early whether we already answer this question."
LO
Lena Ortiz · program manager
Screens concepts from the portal, assigns reviewers, builds the docket, sends letters.
"Half my week goes on chasing what's missing."
OH
Dr. Omar Haddad · committee chair, head of evidence generation
Signs committee decisions.
"Rank them, show me why, and show me the overlap."
DO
Dana Okafor · compliance
Clears or holds red flags.
"Support can never follow prescribing."
SP
Sam Patel · clinical supply (non-voting)
Confirms product and timing.
"Tell me the quantity before you approve."
11 · What goes in and out
Documents in, decisions out, with a person and a citation in between
In shortIt reads the proposal and your own lists, priorities and benchmarks. What comes out is a ranked round, a cited packet for each proposal, the decision record and the data for transparency reporting.
Inputs
Concept or letter of intent, protocol
Budget sheet and cover email
Investigator CV and site enrollment history
Published research priorities
Company study list and funded investigator studies
Public registries (ClinicalTrials.gov, CTIS)
Fair market value benchmarks
Prior payments to the investigator
Agents read, score, check, draftAgainst the committee's own criteria, with every finding linked to its source
A person recommends and decides
Audit trail on every step
Outputs
Structured, cited proposal record
The ranked cycle scorecard
Review packet and committee docket
Decision letter with conditions
Data for transparency reporting
Audit trail, exportable
What changes for the team
Reviewers read a prepared packet, not a raw submission
Overlap is a standing check, not a memory test
The committee spends its time deciding, not reading
Design targets, not results. The solution is designed to bring the median from receipt to decision to 27 days (against a 52-day by-hand baseline in its sample data), to catch overlaps before the committee meets, and to have four of five committee packets ready two weeks ahead. These are targets for the design; your own baseline and a measured pilot decide what it does for you.
12 · With what you already run
It works with the systems your program already runs
In shortKeep your grants system, trial management, contracts and transparency reporting. The new piece reads proposals and returns scored, cited packets and decision records to them.
Grants and investigator-study management systems
The portal, routing, contracts and milestones stay where they are. Agents read each submission as it lands and return a scored, cited packet and a decision record.
Clinical trial management system (CTMS)
The company's own studies, with populations, endpoints, sites and readout dates, become the list the overlap finder searches. Site enrollment history feeds feasibility.
Contracts and contract management
Approved conditions, budget lines and milestones pass to the agreement without re-keying; the letter and the contract say the same thing.
Transparency and aggregate-spend reporting
Recipient, principal investigators, amounts, country and delayed-publication flags are captured at approval, ready for Open Payments and EFPIA reporting.
Medical information or CRM system
Field medical interactions show whether a request was unsolicited and who discussed it, and open-gap priorities can go back to medical science liaisons.
Documents, registries and benchmarks
Your document store, public registries such as ClinicalTrials.gov and CTIS, and your fair market value benchmark set are read as sources, each finding linked to the passage behind it.
Three principles of an agentic review program
1
Prepare everything; decide nothing
Agents read, score, search, check and draft. Reviewers, compliance and the committee make every recommendation and decision, and sign.
2
Cite every finding
Every score, overlap, budget flag and sentence in a packet links to the passage it came from, so anyone can check it and an auditor can follow it.
3
Your rules, versioned
Weights, thresholds, caps and benchmarks are the committee's settings, changed by named people and kept on the audit trail.
Built on Botminds
One platform for agents, data and screens
Builder: describe the work, get a spec, build it, test it.
Runtime: agents, data, integrations, screens and models, in one place.
SecureIsolatedGovernedTestedScalableHighly available
Enterprise-ready
Roles and separation of duties: reviewer, chair, compliance and supply each sign their part
Electronic signatures that record meaning, name and time
Audit export an auditor or monitor can read
Integrations with grants, trial management, contracts, transparency and CRM systems
Service levels and support; human approval and citations everywhere
Part C · Are you ready to automate?
Should you automate investigator-initiated study review? A buyer's view
You run the program, chair the committee, or answer for it to compliance. This part is written for you: what slow review costs today, what automating the preparation could return, whether your team is ready, and what to ask any vendor, including us.
13 · What slow review costs
What a slow or inconsistent review costs, before you buy anything
In shortStart from your own costs: reviewer hours, days investigators wait, support spent on duplicates, and the time it takes to rebuild a decision for an auditor. Once you trust those numbers, judge any solution against them, ours included.
61 days
median from a proposal to the company's decision in one 2025 study, with the middle half between 31 and 183 days. Slow answers send good science to another sponsor.1
One academic consortium, cancer studies
~450 h
of reviewer time a year if 150 proposals take three hours each. Medical directors give these hours on top of their main role.
Illustrative arithmetic: 150 × 3, the defaults in section 3
$1.55M
civil penalty for knowingly failing to report payments, the largest Sunshine Act recovery so far, in an August 2026 settlement whose integrity agreement covers research support.23
Section 6 has the details
Find your own numbers
Cost
Where it hides
How to find your number in a week
Reviewer time
Medical directors reading raw submissions on top of their day job
Time five reviews; multiply by the proposals you receive a year
Chasing
Program managers chasing missing budgets, CVs and power calculations
Count open proposals waiting on missing items today
Waiting
Proposals carried over to the next committee
Median days from receipt to decision, and the slowest tenth
Duplicates
Support for a question an ongoing study already answers
Proposals declined or reshaped for overlap last year, times your average support
Audit time
Rebuilding one study's request, review, approval and payments from several systems
Time how long it takes for three supported studies
Costs no calculator counts
RelationshipsA slow or unexplained no wears thin the trust your medical team built with investigators over years.
Studies that stallAbout one trial site in ten enrolls no patient. Recruiting taken on trust is support with no evidence back.25
Compliance exposureSupport that cannot be shown to follow science, not prescribing, is the risk regulators look for.1222
Open evidence gapsPriorities nobody is studying stay invisible until the round has closed.
Your reviewersExperienced medical directors spend their time reading forms instead of judging the science.
14 · Your return
What automating the preparation could return for you
In shortPut in your own numbers. The calculator asks how many proposals you receive, how long each takes, and how many duplicates and audits you handle. Every default is an assumption: replace it with yours, or set a line to 0.
The value levers
Days out
Faster answers to investigators
Concepts arrive complete and prepared, so more are decided at the next meeting.
days saved = today's median − new median
Reviewer hours
Reviewers read a packet
Agents read, score, search and draft; reviewers check, edit and sign.
concepts × (hours by hand − hours with packet) × hourly cost
Funding redirected
Duplicates caught before the committee
Support goes to new questions, not ones an ongoing study already answers, and patients aren't split between competing studies.
overlaps caught × share otherwise funded × average support
Budgets
Fair market value, documented
Every line compared with the benchmark, every exception explained, every time.
Σ (requested − approved within FMV)
Audit
One record per study
Request, review, scores, approvals and payments in one place, ready for a monitoring sample.
arrangements sampled × hours saved each × hourly cost
Portfolio
Gaps seen in time
Open gaps against your priorities are visible while the cycle is still open, so field medical can explain them.
priorities with no study or concept this cycle
The calculator
Every default below is illustrative, not a benchmark or a promise. Replace each with your own number, and set a lever to 0 if it doesn't apply.
Reviewer and program timeDuplicate support avoidedAudit evidence
Illustrative yearly value
Reviewer and program time$0
Support redirected from duplicates$0
Audit evidence$0
Total$0
Hours freed a year0
Formulas, per year: Time = concepts × (hours by hand − hours with packet) × hourly cost Duplicates = overlaps caught × share funded × average support Audit = arrangements × hours saved × hourly cost
Where the defaults come from: 25% is the approval rate in one 2025 oncology study1; $100,000 echoes a 2015 industry survey headline that average spending per investigator-initiated trial topped $100,000 in 201429. All other defaults are placeholders. Not counted: faster answers to investigators, publications delivered, budgets brought to FMV and compliance risk avoided. Weigh the result against the cost of the solution and your rollout.
Start here15 · Readiness check
Are you ready to automate? Nine questions, five minutes
In shortPick the answer closest to your program today. Eight questions score 0 to 2; the last is for reflection. The meter adds them up as you go. Each question says why it matters and what a strong answer looks like.
1
Volume
How many investigator-initiated study concepts and proposals does your program receive in a year?
Why it mattersVolume turns hours saved per concept into reviewer capacity and faster answers. At low volume, a careful manual process may be enough.
A strong answer"About 200 a year across five therapeutic areas, and we know which months are busiest."
2
Research priorities
Do you have written, versioned research priorities or areas of interest that every concept is judged against?
Why it mattersFit has to be scored against something written down. Published areas are also a compliance control: support follows priorities, not relationships.
A strong answer"Each product has published areas of interest, approved and versioned each year."
3
Review criteria
Does the committee score proposals on a written rubric, with the same criteria for every concept?
Why it mattersAgents apply the committee's own criteria. A written rubric makes scores consistent, comparable across a cycle and defensible in an audit.
A strong answer"Merit, fit, feasibility, budget and overlap, each scored 1 to 5, with weights the committee agreed."
4
Study lists
Could a reviewer search one list of the company's own studies and the investigator studies you already fund?
Why it mattersOverlap checks need something to search. Public registries are open; your internal lists decide how complete the check is.
A strong answer"Company studies sit in our trial management system and funded investigator studies in our grants system; both can be exported."
5
Baseline
Do you know your median time from receipt to decision, and how many concepts are waiting today?
Why it mattersA pain with a number gets a budget. Without a baseline, no improvement can be shown.
A strong answer"Median 70 days last year, 40 open concepts, a third waiting on missing documents."
6
Systems and data
Can your grants or investigator-study system, trial management system and document store be read through APIs or exports?
Why it mattersIntegration sets effort and risk. Systems reachable through screens alone can still work, but take longer.
A strong answer"Our grants system has an API, and submissions land in one document store."
7
Owner and budget
Is there a named medical affairs leader who owns review cycle time and quality, with a budget to change it?
Why it mattersReview crosses medical, compliance, legal, finance and supply. It stalls without one owner.
A strong answer"Our head of evidence generation owns it and has budget this year."
8
Governance
Do you have a way to approve AI for use in medical affairs: compliance review, human sign-off on every decision, an audit trail?
Why it mattersAgents should prepare and people should decide, with every step on the record. A clear approval path shortens the road to production.
A strong answer"Compliance and IT have an AI policy, and any decision on support needs a named signer."
9
Vision · for reflection
If a strong concept arrived today, what would have to be true for the investigator to have a reasoned answer after the next committee meeting, with every finding cited?
Why it mattersYour answer lists the blockers. Each blocker is a candidate job for an agent: reading, checking, searching or drafting.
A strong answerA short, specific list, such as "completeness checked on arrival, overlap searched before review, and a packet the reviewer can sign in an hour."
What your score means
0–6
Explore
Build the foundations: write down and version your priorities and review rubric, count concepts by stage, and time a handful of reviews. Start with a narrow job, such as completeness checks on arrival.
7–11
Ready to pilot
Pick one therapeutic area and one cycle. Run agentic preparation next to your current review, on your own concepts, with your reviewers deciding, and measure days, hours and overlaps found.
12–16
Ready to implement
Plan the full program: integrations with your grants, trial management and transparency systems, roles and signatures, audit export and service levels. Roll out one area at a time against your baseline.
16 · Questions for any vendor
Sixteen questions to ask before you buy software for investigator-initiated study review
In shortAsk every vendor the same questions, us included. The answers show how the work is really done: what gets checked, what evidence you keep, and who is allowed to decide.
Ask
A good answer
A warning sign
What gets checked
Does it check fit, design, recruiting, budget, overlap and compliance on every proposal?
All six, on arrival, with what is missing named
Only a summary of the submission
Which studies does the overlap search cover?
Your own studies, the studies you fund, and public registries such as ClinicalTrials.gov and CTIS
Public registries only
Can it read what investigators actually send?
Concept forms, protocols, budget sheets, CVs and cover emails
Only data typed into its own form
What happens when something is missing?
It names the missing item and a person asks for it
It fills the gap with a likely value
Evidence
Does every score and finding cite its source?
One click from the finding to the passage in the proposal, registry record or priority
A score with no trail
Could an auditor rebuild one decision without you in the room?
One record per study: request, review, scores, approvals, conditions, payments
Evidence lives only inside the vendor's screens
Rules
Does it score on our rubric and our weights?
Your criteria and weights, versioned, changed by named people
One built-in scoring model for every customer
Does it use our fair market value benchmark and overhead cap?
Your benchmark set per country, your cap, and a written reason for any exception
Its own idea of a reasonable budget
Control
Can it approve or decline a proposal on its own?
No. A reviewer recommends and signs; the committee decides
"It auto-declines the weak ones"
Who clears a compliance red flag?
Compliance, before scientific review continues
The reviewer, or nobody
Can a letter to an investigator mention our internal studies?
Never; internal studies stay internal
Letters drafted straight from the overlap report
Your data
Where do proposals and our internal study lists live, and who at the vendor can see them?
Isolated to your company, in a region you choose, with access logged
Shared environments, vague answers
Is our data used to train anything?
No, in writing
"Only to improve the service"
Fit and proof
Does it work with our grants system, trial management and transparency reporting?
It reads from them and writes findings back; it replaces none of them
A migration comes first
Will you run a pilot on our own past proposals?
Yes: 20 to 30 past proposals scored against our own past decisions, then one live round
Demo data only
What does the price depend on?
A clear unit, such as proposals a year, and what is included
A number that appears only after the contract
17 · How to start
How to start: five steps to a decision you can defend
In shortWrite down your rules, measure where you are, then pilot on your own past proposals and one live round. None of these steps commits you to a vendor.
Write down your priorities and rubric. Published research priorities and scored criteria are what any review, by hand or automated, is judged against.
Measure your baseline: median days from receipt to decision, how many proposals are open today, and how long five reviews really take.
Gather 20 to 30 past proposals with their decisions, with an approval path to share them.
Pilot one therapeutic area for one round, next to your current review. Your reviewers recommend and your committee decides, as today.
Decide with numbers: days to decision, reviewer hours, overlaps found before the committee, and decisions you could hand an auditor.
The full working-session questions: 26 questions in 9 groups
A
Program and scope
A1
What does the program support: interventional, observational, retrospective or preclinical studies; funding, product or both; drugs, devices or diagnostics?
Why it mattersScope sets the criteria, the documents and the regulatory checks a review must cover.
Strong answer · warning sign
A written scope per product.
Each area decides for itself.
A2
Which countries do requests come from, and does each region run its own committee?
Why it mattersRegions bring different benchmarks, contracts and disclosure rules (Open Payments in the US, the EFPIA Code and national rules in Europe).
Strong answer · warning sign
A map of regions, committees and rules.
"Global" with no view of local rules.
A3
How are areas of interest published, and how often are they updated?
Why it mattersThey are the yardstick for fit, and the evidence that support was not solicited.
Strong answer · warning sign
Versioned, approved, on the website.
Shared verbally by the field team.
B
Volume and cadence
B1
How many concepts and full proposals arrive a year, by therapeutic area, and what share is invited, approved or declined?
Why it mattersThe base for every other number, including the ROI sketch below.
Strong answer · warning sign
Counts by stage and area.
Needs a special report to answer.
B2
How often does the committee meet, and how many concepts wait more than one meeting for a decision?
Why it mattersA missed meeting can add a month or more to the investigator's wait.
Strong answer · warning sign
Known cadence and carry-over count.
Packets finished the night before.
B3
What is the median and the slowest decile from receipt to decision, and from decision to signed contract?
Why it mattersInvestigators remember the slowest cases; the median hides them.
Strong answer · warning sign
Both measured by step.
Measured from the committee date, not receipt.
C
Intake and screening
C1
Through which channels do concepts arrive: a portal, email, field medical, post? Who checks they are complete?
Why it mattersMixed channels and incomplete submissions are where days disappear before review even starts.
Strong answer · warning sign
One portal and a completeness checklist.
Concepts handed over in hallways.
C2
How long does it take to get missing items (budget, CV, power calculation) back from investigators?
Why it mattersEarly, specific requests for what is missing shorten the cycle more than faster reviews do.
Strong answer · warning sign
Tracked, with reminders.
Found missing at the committee.
C3
How do you make sure a concept was unsolicited, and that no sales or marketing colleague is involved?
Why it mattersResearch that originates in sales is "particularly suspect" to regulators.
Strong answer · warning sign
A screen on every submission and its correspondence.
Relies on someone reading the cover email.
D
Scientific and strategic review
D1
How many hours does a reviewer spend on one concept, and on one full protocol?
Why it mattersReviewer time is the scarcest resource, and no published benchmark exists. Measure yours.
Strong answer · warning sign
A sample of timed reviews.
"It depends."
D2
Would two reviewers score the same concept the same way? Have you checked?
Why it mattersConsistency is fairness to investigators and defensibility to auditors.
Strong answer · warning sign
Calibration sessions and a shared rubric.
Each area has its own habits.
D3
Who checks the statistics and the enrollment plan against the site's own history?
Why it mattersFeasibility taken as written is a common reason studies stall after approval.
Strong answer · warning sign
A biostatistician and site data.
Nobody, unless asked.
E
Overlap and portfolio
E1
How do you find out today that a concept repeats a company study, another funded study or a registered trial?
Why it mattersFound late, an overlap costs a committee slot and a reviewer's afternoon; missed, it can fund a duplicate of your own study.
Strong answer · warning sign
A systematic search on every concept.
"Someone usually remembers."
E2
How many concepts last year were declined or reshaped because of overlap?
Why it mattersA number turns a vague worry into a value lever.
Strong answer · warning sign
Reasons recorded per decision.
Decline reasons not recorded.
E3
Can you see, by priority, what the company runs, funds and has received this cycle, and where there are open gaps?
Why it mattersKnowing where evidence is thin before the cycle closes is how the program serves the medical plan.
Strong answer · warning sign
A coverage view per priority.
Assembled once a year for a slide.
F
Budgets and fair market value
F1
Which benchmark do you compare budget lines with, by country, and how is the comparison documented?
Why it mattersFair market value is required, but the method is yours to define and defend.
Strong answer · warning sign
A named benchmark set and a recorded comparison.
"Looks reasonable."
F2
What is your overhead cap, and how are exceptions approved?
Why it mattersOverhead is a common budget flag, and exceptions need a written reason.
Strong answer · warning sign
A cap, with documented exceptions.
Negotiated case by case, unrecorded.
F3
Who confirms product quantity and timing before approval?
Why it mattersAn approved study without product is a promise the company can't keep.
Strong answer · warning sign
Clinical supply confirms before the vote.
Found out at contracting.
G
Compliance and transparency
G1
If an auditor sampled ten supported studies, could you show for each the request, review, scores, FMV check, approval, contract, payments and publication, in an hour?
Why it mattersRecent integrity agreements require monitoring a share of researcher arrangements every period.
Strong answer · warning sign
Yes, from one record per study.
It would take days across systems.
G2
How do research payment details reach your Open Payments and EFPIA reporting, including principal investigators and delayed-publication flags?
Why it mattersUnreported or misreported payments carry per-payment penalties.
Strong answer · warning sign
Captured at approval, reconciled at payment.
Rebuilt at year end from invoices.
G3
What does your AI policy require: compliance review, data residency, human approval, validation?
Why it mattersKnowing the approval path early avoids a late stop.
Strong answer · warning sign
A framework and a named reviewer.
"AI isn't allowed," with no path to change it.
H
Contracts, milestones and publication
H1
How are milestones tracked and payments released, and who notices a study that has stopped enrolling?
Why it mattersMilestones tie money to work actually done.
Strong answer · warning sign
Milestones tracked with alerts.
Payments on request.
H2
What share of completed studies were published within the agreed time?
Why it mattersPublication is the evidence the program exists to create.
Strong answer · warning sign
Tracked per study.
Unknown.
I
Decision, success measures and vision
I1
What would a pilot have to show for you to call it a success? Which numbers, measured how?
Why it mattersA measured pilot settles the decision.
Strong answer · warning sign
Days to decision, reviewer hours, overlaps found, audit-ready records.
"Let's see what it does."
I2
Could you gather 20 to 30 past concepts, with their decisions, for a working session within two weeks?
Why it mattersReal files, scored against your own past decisions, are the fastest way to see what agentic review would do.
Strong answer · warning sign
Yes, with an approval path.
No way to share anything.
I3
Which steps would you never let software do without a person deciding?
Why it mattersTells you where the signatures belong.
Strong answer · warning sign
Recommendations, decisions, letters, budget exceptions, clearing red flags.
"All of it": then start with reading and checking.
An investigator-initiated study is research that an independent investigator or institution designs and sponsors, with funding, product or both from a pharmaceutical, biotech or medical device company. The investigator writes the protocol and holds the sponsor's duties; the company reviews the request and supports it under a written agreement, but does not run the study.
Is an investigator-sponsored study the same as an investigator-initiated study?
Yes. Companies use different names for the same model: investigator-initiated study or trial, investigator-sponsored study or trial, investigator-sponsored research, investigator-initiated research and externally sponsored research. In each, the investigator or institution is the sponsor.
How are investigator-initiated study proposals reviewed?
Most companies check a short concept for completeness, invite a full protocol and budget, and have medical reviewers assess scientific merit, fit with published areas of interest, feasibility, budget and overlap with other studies. A cross-functional committee then approves, declines, defers or asks for changes. One 2025 study of oncology concepts found a median of 61 days from submission to a sponsor's decision.
How is fair market value set for an investigator-initiated study budget?
US government guidance requires research payments to be at fair market value for legitimate, reasonable and necessary services, but does not prescribe a method. In practice, companies compare each budget line with an external benchmark for the country, apply an overhead cap, and document any exception with a written reason.
Are investigator-initiated study payments reported under the Sunshine Act?
Yes. As of October 2026, research payments made under a written agreement or protocol are reported to CMS Open Payments, including payments to an institution when a physician covered recipient is a principal investigator. Payments for research on new products or new uses can be granted a delay in publication of up to four years, or until FDA approval.
How are research payments disclosed in Europe?
Under the EFPIA Code's disclosure rules, research and development transfers of value to health care professionals and organizations are disclosed in aggregate rather than per named recipient, as of October 2026. National laws and codes can add requirements, so check each country.
Why must investigator-initiated study support be kept apart from sales and marketing?
Because support given to win prescriptions or formulary listings can be a kickback under the US Anti-Kickback Statute. Government guidance tells manufacturers to insulate research grant-making from sales and marketing, and enforcement cases have involved grants initiated by sales staff and approved with little information about the research.
Can AI agents approve investigator-initiated study proposals?
They shouldn't. In a well-governed design, agents read, score, search for overlap and draft a cited packet, while the scientific reviewer recommends and signs, compliance clears any red flag, and the committee decides. Every step is kept on an audit trail.
Does automated review replace our grants or investigator-study management system?
No. It works alongside it: it reads submissions, company study lists, registries and benchmarks, and returns scores, overlap findings, cited packets and decision records to the systems you already run.
19 · One-page summary
Investigator-initiated study review: one-page summary
Print this page
The readiness questions
How many concepts and proposals do we receive a year?
Are our research priorities written, published and versioned?
Does the committee score on a written rubric?
Can a reviewer search one list of our own and funded studies?
Do we know our median time from receipt to decision?
Can our grants, trial management and document systems be read through APIs or exports?
Is there a named owner with budget?
Do we have a way to approve AI, with human sign-off and an audit trail?
What would have to be true to give a strong concept a reasoned, cited answer after the next meeting?
Score 0–2 on 1–8: 0–6 explore · 7–11 pilot · 12–16 implement.
10 terms
Investigator-initiated study
Research the investigator designs and sponsors, with company support.
Sponsor-investigator
Holds both the sponsor's and the investigator's duties.
Areas of interest
Published topics a company will consider supporting.
Scientific review committee
Cross-functional group that decides support.
Fair market value
Arm's-length cost; research payments must not exceed it.
Overlap
A concept repeating a question already being studied.
US reporting of research payments; publication can be delayed.
EFPIA Code
European disclosure; R&D payments in aggregate.
20 · Sources
Sources
Accessed October 26, 2026. Regulatory statements are as of October 2026. Company programs are described generically in the text; their published pages are linked here.
Gerber DE et al. Development, review, and activation of thoracic oncology investigator-initiated trials. Clinical Cancer Research 2025;31(6):1103–1108 (PubMed 39820484). pubmed.ncbi.nlm.nih.gov
Medical Affairs Professional Society. Best practices for managing the life-cycle of an investigator initiated trial (2018). medicalaffairs.org
PhRMA. Position statement: principles to govern pharmaceutical company support of investigator initiated studies (March 11, 2014). www.phrma-jp.org
ICH E6(R3) Guideline for Good Clinical Practice, final version adopted June 16, 2026 (glossary: sponsor-investigator; document history). database.ich.org
A large pharmaceutical company's guide for prospective investigators of investigator-initiated trials (FMV assessment, milestone schedule). published page
A medical device maker's investigator-initiated study program page for CT (review criteria; one to three months to review). published page
A diagnostics company's investigator-initiated studies program page (20 working days; publication within 18 months). published page
A specialty pharmaceutical company's research grants FAQs (2022): committee meets every two months. published page
A specialty pharmaceutical company's investigator-initiated studies page: review meetings held monthly. published page
Investigator Initiated Sponsored Research Association (IISRA). Position statement: setting up an investigator initiated-sponsored research program (2010), hosted by ACRP. acrpnet.org
HHS Office of Inspector General. Compliance Program Guidance for Pharmaceutical Manufacturers, 68 Fed. Reg. 23731 (May 5, 2003). www.govinfo.gov
HHS Office of Inspector General. Compliance guidance page: General Compliance Program Guidance (2023), industry-specific guidances, and the archive of earlier guidances. oig.hhs.gov
PhRMA Code on Interactions with Health Care Professionals (effective January 1, 2022). phrma.org
PhRMA Principles on Conduct of Clinical Trials and Communication of Clinical Trial Results (copy hosted by Seton Hall Law). law.shu.edu
AdvaMed Code of Ethics on Interactions with U.S. Health Care Professionals, revised and restated, effective June 1, 2023, section on supporting independent third-party research. www.advamed.org
CMS Open Payments, data overview: Program Year 2025 update (published 2026). www.cms.gov
CMS Open Payments Frequently Asked Questions (current as of July 1, 2024): research payments, principal investigators, delay in publication. www.cms.gov
42 CFR 403.912, penalties for failure to report (Open Payments). www.ecfr.gov
A pharmaceutical company's EFPIA Code methodological note for 2025 disclosure (Norway): research and development transfers of value disclosed in aggregate. published page
US Attorney's Office, District of New Jersey: settlement agreement resolving allegations that research grants were used as kickbacks (July 2020). www.justice.gov
US Department of Justice settlement agreement (August 2026) and the HHS-OIG Corporate Integrity Agreement signed August 2026 (definition of Research; research-related activities; researcher monitoring program). www.justice.gov · oig.hhs.gov
ClinicalTrials.gov API v2, queried October 26, 2026: studies with lead sponsor class OTHER and collaborator class INDUSTRY, all statuses and active statuses (recruiting, not yet recruiting, active not recruiting, enrolling by invitation). clinicaltrials.gov
Getz K. Enrollment performance: weighing the facts. Applied Clinical Trials, May 1, 2012 (Tufts CSDD study of 151 phase II–III trials, 2008–2010). www.appliedclinicaltrialsonline.com
42 CFR Part 11, clinical trials registration and results information submission (final rule, 2016). www.govinfo.gov
European Medicines Agency: transition of clinical trials to the Clinical Trials Information System (CTIS) under Regulation (EU) No 536/2014. www.ema.europa.eu
PICOTS (population, intervention, comparator, outcomes, timing, setting), eCQI Resource Center glossary, after the AHRQ methods guide. ecqi.healthit.gov
Industry survey press release: average spending per investigator-initiated trial tops $100,000 in 2014 (March 18, 2015). www.globenewswire.com
Next step
See it on your own past proposals.
Explore Botminds Investigator-Initiated Study Review, or talk to us about a working session on 20 to 30 of your own past proposals and decisions.
We use essential cookies to make this site work, and cookieless, anonymised analytics to understand how it's used. "Accept all" also turns on Google Analytics, which sets cookies. Privacy policy