NIH Just Told Academic Medical Centers to Start Signing Partnership Agreements. The UDN Solicitation Is 18 Months Out.

Academic medical center PIs planning NIH consortium bids should read grants.gov opportunity 363834 now: NINDS forecast PAR-27-137 on September 10, 2026, with an estimated May 25, 2027 application deadline, and it quietly rewrites the Undiagnosed Diseases Network's clinical sites into regional networks.

The document is a forecast, and that is the most useful thing about it

Start with the honest framing, because it changes how you should spend the next six months. PAR-27-137 — "Undiagnosed Diseases Network (UDN) Clinical Networks of Diagnostic Excellence" is not an open solicitation. The text says so in one flat sentence: "Applications are not being solicited at this time."

What NINDS posted on September 10, 2026 is a forecast with a schedule attached. Estimated synopsis posting: March 25, 2027. Estimated application due date: May 25, 2027. Estimated award and project start: April 25, 2028, against fiscal year 2028. The funding instrument is a cooperative agreement under CFDA 93.853, and there is no cost-sharing requirement. Program contact is a shared mailbox, UDNProgramStaff@ninds.nih.gov, rather than a named officer.

NINDS is explicit about why it published eighteen months early: "This Forecast is being published to provide potential applicants time to plan, establish collaborations, formalize regional partnerships, and prepare responsive projects."

That word — formalize — is the tell. NIH does not usually tell you to go sign agreements. Here it does, because the thing being competed cannot be assembled in the eight weeks between a synopsis and a due date.

What changes: single-site centers become geographically distributed networks

The current UDN clinical sites are called Diagnostic Centers of Excellence. Each is one academic medical center with a multidisciplinary team, funded on its own U01. The forecast retires that shape:

"Unlike prior phases, clinical NODEs are expected to function as geographically distributed regional clinical networks rather than single-site centers. Each clinical NODE would establish and lead a regional network built on existing partnerships with academic medical centers, hospitals, clinics, federally qualified health centers, and other community-based clinical partners."

Read that partner list again. Federally qualified health centers are in it. So are community-based clinical partners. NINDS is not asking for a consortium of peer research hospitals — it is asking a research hospital to become the hub of a referral and evaluation network that reaches into settings where undiagnosed patients actually present and currently go nowhere.

The scope of work expands accordingly. A NODE recruits and enrolls participants, conducts medical record reviews, makes enrollment determinations, communicates outcomes to participants and referring providers, and coordinates clinical evaluation across local partner sites. On top of that clinical load: participating in network-wide development of new diagnostic tools for "both genomic and non-genomic conditions," supporting "systematic re-evaluation of unsolved UDN cases," running access-and-experience research, contributing to network governance, training partner sites on UDN standards, standing up "a Community Advisory Board or comparable mechanisms," and monitoring site performance.

That is a program office, a clinical service, a training function, and a research portfolio in one award.

The number NINDS did not publish, and the one you can look up

The forecast carries no award ceiling, no floor, and no expected number of awards. Applicants planning against it have to build a budget from the existing network, and that data is public.

In FY2026, NIH RePORTER shows ten UDN clinical site U01s totaling roughly $7.6 million, at these annual levels:

The band is tight: $690,000 to $843,000 per site per year, clustering near $760,000. Several sites carry administrative supplements on top — Duke $376,455, Vanderbilt $167,174, University of Washington $137,000, Baylor three separate $160,000 awards — which is itself a signal that the base award has been running lean against the actual work.

Now hold that against what a NODE is being asked to do. Roughly three-quarters of a million dollars a year funds a single-site diagnostic team. It does not fund that team plus subawards to federally qualified health centers, plus a community advisory board, plus a training function for partner sites, plus performance monitoring across a region.

So one of two things is true about the eventual NOFO, and applicants should plan for both branches. Either the per-award budget rises materially to fund a hub-and-spoke structure, or the number of awards falls and each surviving NODE absorbs the territory of sites that do not get renewed. A forecast that replaces ten single-site centers with an unstated number of regional networks is not neutral about consolidation.

Three forecasts posted the same day describe a whole architecture

PAR-27-137 did not arrive alone. NINDS posted three UDN forecasts on September 10, 2026, all sharing the same March 25, 2027 synopsis date and May 25, 2027 estimated due date:

• PAR-27-135 — Clinical and Data Coordinating Center (CDCC)
• PAR-27-136 — Diagnostic Research Innovation Center (DRIC)
• PAR-27-137 — Clinical Networks of Diagnostic Excellence (NODEs)

Together they lay out the next network: a CDCC for operations, protocol and regulatory management, data sharing, patient navigation, and "implementation science research to facilitate the transfer of successful UDN diagnostic approaches into mainstream medical care"; a DRIC providing advanced sequencing, multi-omics, and gene-function studies in animal and in vitro models; the intramural NIH Undiagnosed Diseases Program; the UDN Foundation; patient advocacy organizations; and individuals with lived experience named as network participants.

The coordinating center's scale is worth knowing before you budget a NODE. The current CDCC — Harvard Medical School's DUCAS award, U2C NS132415 — drew $4,184,192 in FY2026, plus a $2.7 million supplement, with an administrative core at $1.06 million and a clinical research support core at $1.84 million. The coordinating center is roughly half the size of all ten clinical sites combined. Centralized functions in this network are not an afterthought, and a NODE application that proposes duplicating CDCC infrastructure locally will read as not having understood the architecture.

The CDCC forecast adds one requirement with real teeth: applicants must show they can "assume CDCC responsibilities with minimal start-up time following award." Whoever is not the incumbent has to demonstrate a running operation before the money exists.

Why a mechanism forecast deserves eighteen months of your attention

Structural changes to NIH network mechanisms move more money than most headline increases, and they are consistently under-read because they look like plumbing. We made the same argument about NIGMS converting its IDeA clinical research program from a U54 cooperative agreement to a P30 center core grant — a single letter that reassigned who owns the scientific direction of a decade-old program.

PAR-27-137 is the same species of change, and the required response is not writing. It is partnership formation, and it has a real clock:

Between now and March 2027, identify the regional gap you can credibly claim. Geographic distribution is the organizing principle, which means two strong applicants in one metro are now competing for territory that used to hold two separate DCoEs.

Get executed agreements, not letters of support. NINDS asked applicants to formalize partnerships. A reviewer comparing a signed affiliation agreement with an FQHC network against a warm letter from a hospital system will not treat them as equivalent evidence.

Stand up community input early. A Community Advisory Board convened in April 2027 for a May 2027 submission is visible as exactly that. One with a year of minutes is infrastructure.

Email the program mailbox. UDNProgramStaff@ninds.nih.gov is published precisely so that pre-NOFO questions land before the terms are locked. The window in which program staff can still shape a forecast into a NOFO closes when the synopsis posts.

The one thing not to do is wait for March 25, 2027. A NOFO that requires a formalized regional network cannot be answered by an institution that starts building one the week it drops.

Track the live NIH consortium and network opportunities in your portfolio — search NIH cooperative agreement and clinical network solicitations on Granted — and set a watch on PAR-27-135, PAR-27-136, and PAR-27-137 so the synopsis posting reaches you the day it lands, not the week your competitors' agreements are already signed.