Mark Graber notes there is no course on diagnosis in medical school. A new field of diagnosis education, situativity, and a competency framework aim to fill that gap.
ImproveDx Blog
Patients bring visible risk factors—age, race/ethnicity, sex—that affect diagnostic accuracy, yet these effects on symptoms, test results, and diagnosis remain understudied, leaving no strategies to address the cognitive pitfalls tied to these overlooked patient factors.
Health professions education—like any system—produces the outcomes it's designed for, and current training yields roughly 90% diagnostic accuracy. Can better education raise that number?
Uncertainty is inherent to diagnosis—sometimes brief, sometimes chronic. Between these extremes, each case moves through a period of dynamic uncertainty as history, exam, and testing narrow the mystery toward resolution.
CIDM's Patient Stories showcase families harmed by missed or delayed diagnosis, spanning errors from delayed testing to cognitive bias across all patients and conditions—from sepsis to stroke. These narratives give voice to those affected, with many becoming advocates themselves, and inform CIDM's broader effort to include patients as active partners in improving diagnostic accuracy and outcomes.
ECRI Institute's 2018 Top 10 Patient Safety Concerns named diagnostic errors as the leading issue, noting they're hard to measure since they often surface only after patients leave care. ECRI recommended organizations capture data through event-reporting systems, malpractice claims, patient complaints, surveys, autopsies, and record reviews.
Pennsylvania's Act 112 requires imaging providers to notify patients directly of non-emergency abnormal results warranting physician review, within 20 days if follow-up is needed within 90 days. Effective December 2018 and inspired by cases of harm from lost results, the law is seen as a workaround for systemic gaps in test-result follow-up, with organizations like Penn State Health's Hershey Medical Center piloting notification programs while state groups clarify compliance.
The March 2019 issue of Diagnosis examines the pre-analytic stage—sample collection and handling—as a critical, under-recognized threat to diagnostic safety, since specimens are gathered outside the lab's controlled environment. This context highlights persistent gaps in problem analysis that have resisted systematic improvement.
The National Academy of Medicine estimates one in ten diagnoses is wrong, harming one in twenty patients annually and causing 40,000–80,000 U.S. deaths—from missed sepsis to biased dismissal of symptoms. Progress has been slow, hampered by low public awareness and a field historically driven by researchers and policymakers rather than patients themselves.
While quality improvement (QI) techniques have resulted in many changes in health care, using QI to drive improvements in diagnosis is a developing field. In partnership with the Institute for Healthcare Improvement (IHI) six leading medical centers apply the IHI Breakthrough Series methodology to develop specific interventions designed to reduce diagnostic error.
Pennsylvania's Act 112 requires imaging providers to notify patients directly—not just the ordering physician—when a "significant abnormality" is found, addressing the persistent problem of patients with serious undetected conditions never learning their results. Despite years of efforts to "close the loop" on critical findings, the problem persisted; Act 112 asserts patients' right to know and positions them as part of the solution.
Historian Ken Durr distinguishes chronology (recording events in order) from history (shaping facts into an engaging narrative). Physicians, trained to treat patient histories as narrative storytelling toward diagnosis, take histories some 100,000–200,000 times per career. Stanford clinicians instead favor chronological clarity over storyline, trading narrative appeal for protection against diagnostic pitfalls.
Diagnostic error is a global issue. Recognized as a problem in different cultures around the world, it is a concern in settings that vary dramatically in terms of economic, physical, and workforce resources. Despite that variety, diagnostic error (like medical error in general) is most often studied and addressed in settings that enjoy good healthcare resources, stable economies, and modern technologies.
Learning what causes patient harm and figuring out how to prevent it has been an essential patient safety tactic for nearly 20 years. When something bad happens that shouldn't have—preventable harm—or nearly happens—close call or near miss—people work hard to figure out why and take action to avoid recurrence. Learning in that way, from past experience, has value but should not be relied on alone to improve the safety of healthcare, including diagnosis.
The Agency for Healthcare Research and Quality (AHRQ), a government partner to the Coalition to Improve Diagnosis, is the lead federal agency investing in research to improve diagnostic safety. It has been an early supporter of, and provided conference grants to, the Society to Improve Diagnosis in Medicine. More broadly, the Agency invests in research to advance the knowledge of diagnostic accuracy and timeliness and to develop practical tools and resources to improve diagnostic safety.
Controversy erupted soon after Lyme disease was discovered more than 40 years ago. Debates focused on both diagnosis and treatment are often characterized by heated exchanges and accusations. The existence of conflicting sets of clinical guidelines—with their own societies and communities of physicians—reflects the way Lyme disease has travelled on two separate tracks since the 1980s.
MedStar Health, the largest healthcare provider in Maryland and greater Washington DC, has established a center devoted to diagnostic safety housed within the MedStar Institute for Quality and Safety (MIQS). The Center for Improving Diagnosis in Healthcare, one of five centers of expertise at MIQS, is focused on achieving the goals defined in the National Academy of Medicine's (NAM) 2015 report.
Transgender patients' diagnostic risk stems more from clinician bias and poor communication around gender identity than from medical issues tied to their transgender status—making individualized, patient-centered care essential. As a population, they face social and economic barriers leaving many uninsured and without routine care, while federal surveys and the census omit gender identity data, leaving their health trends poorly understood. Nurse and risk specialist Susan Boisvert calls this invisibility the single greatest risk transgender patients face.
Measurement is a perennial thorn in efforts to improve healthcare delivery, including diagnosis. Being a relatively new science, diagnostic safety lacks established ways to measure the quality of diagnostic performance, the incidence of errors, and the results of actions intended to prevent them. Developing useful, rational, and reliable systems for measurement is therefore a priority.
Internists consider diagnosis the hallmark of their practice. Accordingly, the American College of Physicians (ACP), the medical specialty society for internal medicine physicians, was one of the first organizations to join the Coalition to Improve Diagnosis (CID). Since then, ACP has been contributing to the coalition's collective actions, with involvement on the CID's subcommittees, and has taken individual action to create resources available to the public, as well as ACP members.