What does this launch intelligence case study show?

According to Within3, a top pharmaceutical company preparing a gastroenterology launch combined advisory board discussions, congress monitoring and internal field insights on its Launch Intelligence platform. A synthesis report linked diagnostic delay to anxiety and depression, and the company widened its education to mental health professionals. The page reports faster alignment and better access, but gives no figures, no condition name and no account of what the AI did.

Is this a rare disease launch case study?

The page does not say so. It describes a chronic gastrointestinal condition and a gastroenterology launch. The method transfers to rare disease, where delays are longer: a Rare Barometer survey of 6,507 people with rare diseases in Europe, published in 2024, found an average of 4.7 years to diagnosis, with misdiagnosis among the main drivers of delay.

Does FDA regulate AI used for launch intelligence?

Not directly. FDA's January 2025 draft guidance on AI covers models that produce information supporting regulatory decisions on safety, effectiveness or quality, and does not address AI used for operational efficiencies. It was still a draft on 12 October 2026. What binds is what the insights feed: promotional rules for any resulting materials, and adverse event reporting for safety information the platform takes in.

Launch Intelligence Case Study: Diagnostic Delays

Q: Do advisory board fees paid to psychiatrists and psychologists have to be reported?

Under the US Open Payments rule, fees to physicians are reportable, and psychiatrists are physicians. So are fees to physician assistants, nurse practitioners, clinical nurse specialists, certified registered nurse anesthetists and certified nurse-midwives. Psychologists and counsellors are not covered recipients under 42 CFR 403.902, so the federal duty does not reach them. Check state disclosure laws separately.

On 21 October 2025 Within3 published a case study on diagnostic delays in a chronic gastrointestinal condition. A top pharmaceutical company, unnamed, had convened an advisory board of gastroenterologists ahead of a key medical congress. The doctors said patients were waiting months or even years for an accurate workup. The company then used Within3's Launch Intelligence platform to work out why.

As a launch intelligence case study, it is unusually clear about the decision it changed. The company added mental health professionals to its list of healthcare professionals to reach, and widened its education programme. It is also unusually thin on evidence. The page reports better alignment and better patient access, and contains no numbers at all. This post sets out what the page shows, what it does not, what independent research says about the finding, and which rules bind a programme like this.

In short

According to Within3, the platform combined advisory board discussions, congress monitoring and internal field insights, and a Strategic Readout tied diagnostic delay to anxiety and depression.
The company named mental health professionals as an overlooked audience and widened its education. The page reports faster alignment and better access but gives no figures.
The finding is plausible. A UK study of 19,555 IBD cases, published in 2020, found that patients with prior depression were less likely to get timely specialist review.
The page does not name the condition or say it is rare. It says the platform is AI-powered but not which step used AI.
For a US launch, what binds is old law: the anti-kickback statute for advisory boards, Open Payments, adverse event reporting from any source, and the promotional rules for any new materials.

The situation: a gastroenterology launch and patients waiting for a diagnosis

The page, filed in our library as a top pharma using Launch Intelligence on diagnostic delays, names no company, product, condition or market. Its meta description calls the work a gastroenterology launch. Nothing on the page says the condition is rare. Rare disease teams start from a longer delay. A Rare Barometer survey of 6,507 people with rare diseases in Europe, published in 2024, put the average time to diagnosis at 4.7 years. Misdiagnosis was among the main drivers.

According to Within3, the company faced three problems:

Diagnostic delay. Patients met barriers to timely evaluation, which kept them from appropriate therapy.
Unrecognised comorbidities. Providers hesitated to pursue GI workups in patients who also had mental health concerns.
Fragmented evidence. Insights sat in separate places: advisory boards, congress activity and internal data.

The goal was to find the root causes of the delay and turn them into a cross-functional plan. That is a narrower brief than a general launch dashboard. It names one question.

What was deployed

Within3 says the team combined three inputs into a single view, then produced a Strategic Readout, its term for a synthesis report across data sets.

Input	What the page says it found	What a reviewer would ask
Advisory board of gastroenterologists	Mental health comorbidities complicate diagnosis and cause hesitancy around endoscopy and biomarker use	How many advisers, how chosen, how paid
Congress monitoring	Evidence that patients with comorbid anxiety or depression were more likely to have delayed GI workups	Which abstracts, what study designs, peer-reviewed or not
Internal field insights	Combined with the other two, but no specific finding is reported	Who recorded them, and how they were coded
Strategic Readout	Confirmed, in the page's words, that "clinicians were under-diagnosing GI conditions in patients with psychological comorbidities"	How the synthesis was produced, and who checked it

The page title calls the platform AI-powered. The body never says which step used AI, what kind of model, or how its output was checked. The site's platform menu advertises cutting reporting time from months to days, with more feedback and less workload. Those are general marketing claims, not results from this case.

Three actions followed. The company identified mental health professionals as an overlooked but critical audience. It expanded education on diagnosing GI disease in patients with mental health concerns. And it gave clinicians guidance on bringing psychological and physical health together in diagnostic pathways.

What it achieved, and what the numbers measure

There are no numbers. The page lists three results, each stated as an outcome.

Claim on the page	What would measure it	Reported?
Medical and commercial teams aligned quickly	Time from first signal to an agreed plan	No
Earlier, more accurate diagnosis improved patient access	Time from first GI symptom to specialist review or diagnosis, before and after, against a comparison group	No
Outreach to mental health professionals closed a gap in the patient journey	Referrals or workups initiated after mental health contact	No

The second claim is the strongest one and the least supported. The page offers no evidence that any patient was diagnosed sooner. Most mental health professionals do not order GI investigations. Their influence runs through referral, so referral is what should have been counted.

Within3 published a second case study the same day, filed in our library as a top-five pharma finding an overlooked prescriber segment. It describes a newly acquired chronic disease therapy in an unfamiliar therapy area, where mental health professionals again turned out to matter. Its one quantified claim is that insights arrived in real time, "avoiding the 6+ month delay typical of ATU studies". That compares the platform with a slow alternative, awareness, trial and usage tracking. It does not measure what the platform changed. The two pages do not say whether they describe the same company.

Independent research can test whether the finding is plausible. It is.

A UK case-control study in the Journal of Crohn's and Colitis (online July 2020) compared 19,555 people with inflammatory bowel disease (IBD) with 78,114 controls. Patients with a previous diagnosis of depression were less likely to get timely specialist review (hazard ratio 0.77). So were patients previously diagnosed with irritable bowel syndrome.
A 2023 meta-analysis of 101 studies found typical delays of 8.0 months in Crohn's disease and 3.7 months in ulcerative colitis (median of study medians). Delay in Crohn's disease was linked to higher odds of stricturing disease (1.88) and intestinal surgery (2.24).
A 2021 meta-analysis in Lancet Gastroenterology & Hepatology found pooled rates of anxiety symptoms of 32.1% and depression symptoms of 25.2% in adults with IBD.

These studies are about IBD. The Within3 page does not name its condition, so they show the pattern is real in one well-studied GI disease, not that it holds in this one.

What made it work, and what limits it

Three things went right. The question was specific. The sources pointed the same way: an expert view, published conference data and the company's own field reports. And the action was concrete, a new audience with its own education, rather than a slide. Our library holds another Within3 case on unifying launch intelligence where the claimed benefit was mostly organisational. This one at least names the decision it changed.

The limits are just as clear.

Three weak sources are not one strong one. Advisory boards are small, selected and paid. Congress abstracts are often preliminary. Field insights reflect what representatives and medical science liaisons chose to write down. Agreement between them raises confidence but does not replace a claims or records analysis.
A similar finding was already published. For IBD, the link between prior depression and slower specialist review was in a peer-reviewed journal by 2020. On the page's own account, the platform's value lay in getting medical and commercial teams to act together, not in discovery. A literature search is the cheapest first test of any launch insight.
No outcome was measured. Without a baseline for referral timing or diagnosis, nobody can say whether the mental health outreach worked, or by how much.
The AI is a black box on the page. If a synthesis drives a change in audience and spend, someone should be able to trace each conclusion to its sources. The page does not say that was possible.

The rules it had to satisfy: what binds a programme like this

The page does not name a market. For a US launch, these are the rules that apply, position current at 12 October 2026. None is specific to AI.

Instrument	Status on 12 Oct 2026	What it means here
Anti-kickback statute, with OIG compliance guidance for manufacturers (May 2003)	Statute binding; guidance not	Advisory board arrangements should be written, meet a real need, pay fair market value and document the services
Open Payments, 42 CFR 403.902	Binding	Fees to physicians, including psychiatrists, are reportable; psychologists and counsellors are not covered
21 CFR 314.80(b)	Binding for approved drugs	Adverse event information must be reviewed whatever its source
21 CFR 202.1 and 314.81(b)(3)(i)	Binding	Branded material for a new audience is promotional labeling, filed on Form FDA 2253 at first use
Disease awareness definition, draft guidance announced 10 February 2004	Still draft	Education that names or suggests no product sits outside the promotional rules
FDA draft guidance on AI, January 2025	Draft	Does not address AI used for operational efficiencies

Advisory boards are where the insight starts, and where the risk sits. OIG's 2003 guidance, which its compliance guidance page still offers as a resource, says fair-market-value payments to small numbers of physicians for genuine advisory work are unlikely to raise significant concern. Manufacturers should check, among other things, that there is "a legitimate need for the services". OIG has issued newer industry-specific guidance for other sectors, but none yet for manufacturers. A launch intelligence platform helps here. A record showing what each board produced and how it was used is evidence of need.

Widening the audience changes the reporting. The federal Open Payments definition of a covered recipient lists physicians, physician assistants, nurse practitioners, clinical nurse specialists, certified registered nurse anesthetists and certified nurse-midwives. A psychiatrist on a future advisory board is reportable. A psychologist is not, under the federal rule.

Safety information arrives through insight channels. For an approved drug, 21 CFR 314.80(b) requires the applicant to review adverse drug experience information obtained "from any source, foreign or domestic". Advisory board transcripts, field notes and monitored discussions are sources. Any platform that pools them needs a route to pharmacovigilance intake, tested before go-live.

Branded or unbranded is the first decision for new audiences. Education for psychiatrists that names the product is promotional labeling. It goes through MLR and onto Form FDA 2253, and the net impression standard applies to the whole piece. Education about recognising GI disease that names or suggests no product is disease awareness. FDA's definition of that category still comes from a draft, which its CBER advertising Q&A continues to cite. Our post on agentic MLR review covers how the branded route is checked.

What does not bind. FDA's January 2025 draft on AI says it does not address AI "when used for operational efficiencies" that do not affect patient safety, drug quality or study results. Synthesising launch insights sits outside it on its face, as our post on the FDA AI credibility framework's scope explains. A Federal Register search on 12 October 2026 found no notice of a final version.

What this means in practice

For a launch team that wants this kind of result, in lead-time order:

Write down the decision the intelligence must change. Audience, message, sequencing or resourcing. Then name the measure that would show the change worked, such as referrals from a new specialty or time from first symptom to specialist review in claims data. Take the baseline now, before launch.
Run the literature and claims check first. Before convening anyone, search the published evidence on diagnostic pathways in your condition. It is cheap, fast and sometimes decisive, as the IBD studies show here.
Set up advisory boards to the OIG standard. Written agreements, a stated need, fair market value and a record of what each board produced. Map Open Payments status for every new professional group you plan to engage.
Connect safety before you connect sources. Agree with pharmacovigilance how adverse events in transcripts and field notes are flagged and sent on. Test it on real material before any platform goes live.
Decide branded or unbranded for each new audience. Route branded education through MLR and Form FDA 2253. Keep unbranded disease education free of any product suggestion.
Ask the vendor to show its working. For each conclusion in a synthesis report, you should see the sources behind it. If the platform cannot show that, treat its output as a lead to verify, not a finding.

Within3's case shows that pooling scattered insight can move an organisation to act. It does not show that patients were diagnosed sooner. That second question is the one worth paying to answer.

Case study: What Launch Intelligence Found Behind a GI Diagnostic Delay