Diagnostic Errors: When Hours Decide Whether You Live

Most diagnostic errors are not catastrophic. A urinary tract infection misdiagnosed as a bladder spasm gets caught at the follow-up visit and treated. The harm is annoyance and a delayed cure.

The cases that destroy lives cluster in three categories: cancers, vascular events (heart attack, stroke, pulmonary embolism), and infections, especially sepsis. The Society to Improve Diagnosis in Medicine calls these the "Big Three." In each, the time between symptom onset and correct treatment is the variable that determines whether the patient recovers, is permanently disabled, or dies.

The medical record in these cases tells a remarkably consistent story when something has gone wrong: the patient presented with the classic warning signs, the chart shows them documented, and the team did not put the pieces together in time. Plaintiff-side malpractice review of these cases is largely a question of what should have been ordered, when, and what the standards of care required.

Sepsis kills more Americans each year than breast cancer, prostate cancer, and AIDS combined. The 2016 international consensus, known as Sepsis-3, defined it as "life-threatening organ dysfunction caused by a dysregulated host response to infection" (Singer et al., JAMA 2016). Operationally, sepsis is identified by an acute increase in the patient's SOFA score of 2 or more points in the setting of suspected or confirmed infection.

Septic shock is the subset with worse circulatory and metabolic dysfunction: persistent hypotension requiring vasopressors to keep mean arterial pressure at or above 65 mm Hg, plus a serum lactate above 2 mmol/L despite adequate fluid resuscitation. In-hospital mortality for septic shock approaches 40%.

The bedside screening tool from Sepsis-3 is qSOFA: one point each for respiratory rate of 22 or more, altered mental status, and systolic blood pressure at or below 100 mm Hg. A qSOFA of 2 or more in a patient with suspected infection should prompt a full sepsis workup. Important caveat for any current article: the 2021 Surviving Sepsis Campaign guidelines recommended against using qSOFA as a single screening tool, in favor of broader systems like NEWS or MEWS. The standard of care continues to evolve.

The Surviving Sepsis Campaign's 2018 update consolidated the prior 3-hour and 6-hour bundles into a single "Hour-1 Bundle." The five elements: measure serum lactate (and remeasure if the initial result is greater than 2); obtain blood cultures before starting antibiotics; administer broad-spectrum antibiotics; begin 30 mL/kg of intravenous crystalloid for hypotension or a lactate of 4 or more; and start vasopressors during or after fluid resuscitation if the patient remains hypotensive, targeting a mean arterial pressure of at least 65 mm Hg.

The 2021 Surviving Sepsis Campaign Guidelines kept the same general framework but downgraded the strength of the 30 mL/kg recommendation and clarified that the one-hour timeline is most critical for patients in septic shock or with a high probability of sepsis. For possible sepsis without shock, the guidelines accept a brief diagnostic investigation window before antibiotics, with a three-hour outer limit. Norepinephrine is now the first-line vasopressor, with peripheral initiation accepted rather than waiting for central access.

These are not bedside opinions. They are guideline-supported standards that a defense expert will be asked to address in front of a judge or jury.

The single most important piece of evidence in many sepsis malpractice cases is the time between recognition of sepsis and administration of effective antibiotics.

Kumar et al. (Critical Care Medicine, 2006) studied 2,154 septic shock patients and found that each hour of delay in effective antimicrobial administration from the onset of persistent hypotension was associated with an average 7.6% decrease in survival. Liu et al. (American Journal of Respiratory and Critical Care Medicine, 2017), looking at roughly 35,000 Kaiser Permanente patients, found smaller but consistent associations: each hour of delay was associated with absolute mortality increases of 0.3% in sepsis, 0.4% in severe sepsis, and 1.8% in septic shock. Seymour et al. (NEJM, 2017) analyzed New York State SEP-1 data and reported similar dose-response findings.

These three studies, taken together, are the backbone of almost every plaintiff expert's causation testimony in a delayed-sepsis case. The defense will quibble with the absolute numbers, but the dose-response relationship is well established.

The federal government's quality measure for sepsis care is CMS SEP-1. It uses the older Sepsis-2 definitions (severe sepsis and septic shock), not the newer Sepsis-3 framework, which is a frequent source of confusion among clinicians. It requires hospitals to document compliance with a three-hour bundle (lactate, blood cultures before antibiotics, broad-spectrum antibiotics, and 30 mL/kg crystalloid for hypotension or lactate of 4 or more) and a six-hour bundle (repeat lactate if elevated, vasopressors for persistent hypotension, and reassessment of volume status).

SEP-1 is all-or-nothing: missing any single element fails the case for reporting purposes. As of fiscal year 2026, SEP-1 performance affects hospital payment through the Hospital Value-Based Purchasing Program, which means hospitals now have a direct financial incentive to comply.

In litigation, SEP-1 compliance data is sometimes discoverable and can be powerful. A hospital with a documented pattern of SEP-1 failures has a harder time arguing that the index case was an isolated lapse.

Cancer misdiagnosis cases turn on the difference between curable and incurable. A stage I colon cancer caught on a 2022 colonoscopy has a five-year survival above 90%. The same cancer first diagnosed at stage IV in 2026 has a five-year survival around 13%. The malpractice question is whether the cancer should have been diagnosed at the earlier stage. The chart usually shows the failure: an abnormal lab not followed up, an imaging study read as normal but with a finding visible in retrospect, a patient complaint dismissed as functional.

Vascular events are about minutes. For ischemic stroke, the door-to-needle target for IV thrombolytics is 60 minutes from arrival, and time lost is brain lost. For ST-elevation myocardial infarction (STEMI), the door-to-balloon time for primary PCI is 90 minutes. For pulmonary embolism, the question is whether the patient's presentation triggered a CT angiogram in time to anticoagulate before hemodynamic collapse.

In each of these, the standard of care is well established and well documented in guidelines. The litigation does not turn on whether the standard exists. It turns on whether the chart shows it was met.

These cases live in the medical record. The work, in roughly this order, is to obtain the full chart and the audit trail; identify the patient's complete presentation (vitals, symptoms, prior history) at each timestamp; identify what diagnostic workup was indicated by that presentation under the applicable guidelines; identify what was actually ordered and when results came back; and identify whether and when the abnormal results were acted on.

A frequent finding: the test was ordered, the result was abnormal, and the result was viewed in the EHR by a clinician but no clinical action was taken for hours or days. EHR audit trails make this kind of failure visible in a way that paper charts never did.

The standard of care expert (typically an emergency medicine, internal medicine, or specialty physician) then opines on whether the team's conduct deviated from what a reasonably competent practitioner would have done. The causation expert opines on whether the delay actually changed the outcome.

Diagnostic errors are the area where statutes of limitations cause the most heartbreak. Cancer cases are the classic example. A mammogram in 2021 was read as normal. The patient learned in 2025 that the cancer was visible in retrospect on the 2021 film. By then, the standard two-year SOL in many states has run, and the patient has to navigate the discovery rule.

DC's three-year SOL with the Bussineau discovery rule is the most plaintiff-favorable in our region. Virginia's two-year SOL with narrow discovery exceptions is the least. Maryland sits in the middle with a five-year hard cap. For FTCA cases involving VA medical centers, the federal Kubrick rule applies: the clock starts when the patient knows of the injury and its cause, not when they learn the conduct was negligent. See our separate article on the statute of limitations for the full breakdown.

The first call should be to a medical malpractice attorney who handles these cases regularly. Bring the full medical record if you have it, or be prepared to authorize release of records. The most important questions on first review are: when did the symptoms first appear, when was the alleged diagnostic failure, when was the correct diagnosis ultimately made, and what was the consequence of the delay.

Peter Anderson has resolved diagnostic error cases including a $1.5 million FTCA matter for failure to diagnose sepsis at a VA medical center and a $900,000 FTCA matter for failure to diagnose prostate cancer.