Thousands of Americans cycle through specialist after specialist for years without receiving a diagnosis. The federal government’s answer, the Undiagnosed Diseases Network, has shown that concentrated expert evaluation and advanced genomics can crack cases that standard care cannot. But the program reaches only a fraction of those who need it, and its funding model leaves most patients without access to these tools.
How Diagnostic Errors Drive the Problem
Misdiagnosis is not a fringe issue. A study using three large datasets of U.S. adult outpatient visits estimated how often diagnostic errors occur by reviewing medical records directly rather than relying on patient surveys. By applying standardized criteria to chart reviews, researchers could pinpoint missed and delayed diagnoses that might never be documented as such in billing codes or discharge summaries. The authors noted that their approach likely still underestimates the true burden, because many errors leave no obvious paper trail.
Separate research in BMJ Quality and Safety produced a national estimate of serious harms, including death and permanent disability, caused by diagnostic error across clinical settings. That analysis used a disease-based method to rank conditions by the frequency and severity of harm. Stroke, certain cancers, and infections accounted for a large share of catastrophic outcomes, illustrating that diagnostic failures often involve common symptoms (weakness, headache, fever) that are easy to dismiss in busy clinics and emergency departments.
Emergency departments face their own version of this challenge. A systematic review by the Agency for Healthcare Research and Quality examined diagnostic errors in the emergency department and found that evidence for measuring error rates and harms remains uneven. The review highlighted gaps in data, inconsistent definitions, and limited tracking of near-misses. The distance between what clinicians suspect they are missing and what actually goes unrecognized is wide, and closing it requires infrastructure (data systems, dedicated review time, and specialized expertise) that most hospitals do not have.
What the Undiagnosed Diseases Network Actually Does
The NIH launched the Undiagnosed Diseases Program at the NIH Clinical Center in 2008. By 2013, the effort expanded into the Undiagnosed Diseases Network, funded through the NIH Common Fund to build a national system for patients who remained undiagnosed despite extensive workups. The network was designed as a last resort for people who had already exhausted traditional care channels.
UDN clinical sites do not simply repeat what referring doctors already tried. According to a study in Genetics in Medicine, these centers perform deep phenotyping, advanced genomics, data integration, and targeted laboratory investigations that go far beyond standard specialty clinics. Evaluations can include exome or genome sequencing, RNA analyses, metabolomic profiling, and functional studies in cell or animal models. Each site requires multidisciplinary teams (neurologists, geneticists, radiologists, bioinformaticians) and specialized diagnostic equipment, which is precisely why they are expensive to operate and difficult to replicate.
Early performance data from the network show both promise and limits. A study in the New England Journal of Medicine reported the initial outcomes of UDN evaluations, including how many referrals were accepted, how many patients ultimately received a diagnosis, and how often those diagnoses changed care. Roughly one in three evaluated patients obtained an explanation for their condition, a striking yield given that nearly all had already been through years of negative tests. In a separate analysis, the NIH Common Fund summarized promising results from the network, noting that many diagnoses led to specific treatment changes, new surveillance plans, or refined genetic counseling for families.
For patients who had spent years in limbo, a confirmed diagnosis often meant the difference between aimless treatment and targeted care. Some could discontinue ineffective therapies; others became eligible for clinical trials or off-label use of drugs tailored to the newly identified mechanism. Even when no treatment existed, families gained clarity about prognosis and recurrence risk, information that standard care had failed to provide.
Too Few Sites, Too Many Barriers
The network’s results are real, but its reach is thin. Clinical sites operate at institutions such as Vanderbilt University Medical Center, Stanford, and Harvard-affiliated hospitals, each with its own intake rules and capacity constraints. Some locations accept referrals from across the country, while others limit access to patients already within their health systems, effectively closing the door to outside families.
Outside the UDN, a handful of institutions have built their own programs. Emory Special Diagnostic Services, for example, focuses on patients with undiagnosed symptoms by coordinating intensive record review and multi-specialty case conferences. Mayo Clinic’s Center for Individualized Medicine offers genomic testing and interpretation that can identify rare variants and atypical disease presentations. These programs help fill gaps, but they remain concentrated at elite academic medical centers, leaving wide geographic dead zones where patients have no realistic path to such evaluations.
The financial architecture of the UDN itself limits expansion. A 2023 NIH grant notice underscored that the network’s model depends on competitive federal awards, with research subcontracts covering sequencing, RNA studies, metabolomics, and model organism work. Because much of this testing is classified as research rather than routine clinical care, insurance plans generally do not pay for it. That means every new site needs dedicated federal funding before it can open, and every lapse in appropriations threatens existing operations. Hospitals cannot simply bill standard payers to sustain the program.
This grant-dependent structure also creates uncertainty for patients and clinicians. Referral pipelines, staffing, and long-term follow-up all hinge on multi-year awards that must be re-competed. When timelines are unclear, sites may hesitate to invest in new technologies or expand capacity, even as waiting lists grow.
The Cost of Doing Nothing
Leaving patients undiagnosed is not just a clinical failure; it carries steep financial and social consequences. A U.S. Government Accountability Office report documented systemic barriers to diagnosing rare diseases, including fragmented records, limited provider awareness, and uneven access to specialists. The report described how delayed or missed diagnoses lead to repeated hospitalizations, redundant imaging, and trial-and-error treatments that drive up spending without improving health. Families often shoulder additional costs for travel, lost wages, and out-of-pocket consultations that never resolve the underlying problem.
For children, the stakes are especially high. Years without a diagnosis can mean missed windows for early intervention, special education services, or disease-modifying therapies. Adults may lose employment, disability benefits, or insurance coverage while they search for answers. The cumulative effect is a slow erosion of financial stability and trust in the healthcare system.
From a population perspective, undiagnosed and misdiagnosed conditions also impede research. Without accurate labels, patients cannot be enrolled in relevant clinical trials or natural history studies. Scientists lose opportunities to discover new disease genes, pathways, and potential drug targets. Large databases such as the resources maintained by the National Center for Biotechnology Information depend on well-characterized cases to link genetic variants with clinical phenotypes; every missed diagnosis is a data point that never makes it into the scientific record.
What It Would Take to Scale Up
Expanding access to undiagnosed disease evaluations will require both policy and technical changes. On the policy side, federal agencies could treat advanced diagnostics as essential infrastructure, similar to trauma centers or transplant programs, with more stable funding streams. Pilot projects could test blended payment models in which insurers cover clinically actionable components of genomic and phenotypic workups, while research grants support experimental methods.
Technically, some elements of the UDN model could be decentralized. Regional hubs might handle in-person examinations and complex testing, while community clinicians contribute standardized data using shared templates. Telemedicine case conferences could bring UDN-level expertise to hospitals that lack their own genetics teams. Over time, tools developed in the network (phenotyping software, variant interpretation pipelines, and decision-support algorithms) could be adapted for broader use.
Yet even with better technology, the core challenge remains human: assembling and sustaining teams with the time and expertise to tackle the hardest cases. As long as this work is funded as short-term research rather than long-term care, many patients will remain stuck in diagnostic limbo. The experience of the Undiagnosed Diseases Network shows that answers are often possible. The question is whether the health system is willing to build a durable way to deliver them.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
