How to Succeed at Failing, Part 2: Life and Death (Update)

Episode Summary

Executive Summary: This episode of Freakonomics Radio examines failure in healthcare through stories of preventable patient harm, a wrongful-medication case, and major system breakdowns. It argues that most medical failures are systemic rather than purely individual, and that transparent, blame-aware learning systems are essential to prevent repeat harm. Amy Edmondson’s six-part failure spectrum and Bob Langer’s research-driven persistence frame failure as a tool for improvement when handled well.

Main Topics: Patient harm, transparency, and advocacy (Priority: 5/5): Carol Hemmelgarn recounts her daughter’s death from leukemia complicated by hospital-acquired infection, mislabeling as anxiety, and alleged concealment by the hospital. She argues that honest disclosure and learning are necessary to prevent repeat harm and later co-founded a patient-safety advocacy organization. Blame versus systems in medical error (Priority: 5/5): The episode contrasts individual blame with systemic failure using the Vanderbilt nurse prosecution and Amy Edmondson’s view that medication errors usually arise from broader organizational conditions such as staffing, labeling, and policies. Amy Edmondson’s failure spectrum (Priority: 5/5): Edmondson presents six causes of failure—sabotage, inattention, inability, task challenge, uncertainty, and experimentation—arranged from blameworthy to praiseworthy, to help organizations respond more intelligently. Healthcare as a high-stakes failure domain (Priority: 4/5): The conversation emphasizes that medical failure is literally life-or-death, and that fear of litigation and silence from leadership prevent learning in hospitals and health systems. Public-sector technology failure: NHS Connecting for Health (Priority: 4/5): John Van Reenen describes the UK’s massive electronic patient records initiative as a top-down, rushed, centralized reform that collapsed under complexity, resistance, and poor implementation. Scientific experimentation and productive failure (Priority: 4/5): MIT scientist Bob Langer describes repeated grant rejections and hundreds of failed experiments as normal parts of discovery, arguing that persistence is justified when the work could produce major benefits. Grit versus quit (Priority: 3/5): The episode closes by setting up a broader question: when does persistence become wasteful, and when is quitting the smarter choice? This applies to science, startups, and personal projects.

Key Arguments: Most harmful failures in healthcare are not isolated mistakes; they are usually produced by systems, policies, staffing problems, design flaws, and organizational silence. Blame and criminalization can suppress reporting, making it harder to catch errors and improve safety. Hospitals must treat the actual condition of the patient, not a label; misclassifying fear as anxiety can block correct treatment. Patients and families harmed by medical errors often want accountability mainly so the same harm does not happen again. Edmondson’s failure spectrum helps distinguish blameworthy acts from valuable experimentation, making responses to failure more nuanced. Not all failures should be punished equally; the right response depends on whether the cause was sabotage, inattention, inability, task challenge, uncertainty, or experimentation. The NHS IT disaster shows that ambitious reforms fail when leaders ignore local realities, rush contracts, and fail to build buy-in. Scientific progress depends on repeated failure, and successful researchers often persist because they can see evidence that the underlying idea is promising. Failure can be productive when it is visible, analyzed, and shared; hidden failure causes repeated mistakes across organizations. The episode suggests that sustainable improvement requires a culture where people can speak up, learn, and adjust without fear.

Data Points: Age of Carol Hemmelgarn’s daughter: 9 years old - Her daughter Alyssa was diagnosed with ALL leukemia at age nine. Time from leukemia diagnosis to death: 10 days - Alyssa died 10 days after diagnosis. Hospital-acquired infection preventability: Known today can be prevented - Hemmelgarn says the infection contributing to her daughter’s decline was hospital-acquired and preventable. Time until honest conversation with family: 3 years, 7 months, and 28 days - Hemmelgarn says the organization took this long to have an honest conversation after her daughter’s death. Year of Vanderbilt nurse prosecution: 2019 - Redonda Vaught was prosecuted for medication error in 2019. Patient age in Vanderbilt case: 75-year-old - The patient who died after the medication error was 75. Wrong medication administered: Vecuronium instead of Versed - The nurse mistakenly pulled and administered a paralytic instead of a sedative. Estimated U.S. hospital preventable deaths (1999 IOM): 44,000 to 98,000 per year - Early Institute of Medicine estimate of medical error deaths. Estimated U.S. hospital preventable deaths (2013 study): 200,000 per year - A Journal of Patient Safety estimate cited in the episode. Estimated U.S. hospital preventable deaths (2020 Yale meta-analysis): 22,000 per year - A later meta-analysis revised the estimate downward but still found a large preventable death toll. Number of failure categories in Edmondson’s spectrum: 6 - Sabotage, inattention, inability, task challenge, uncertainty, experimentation. NHS IT project cost estimate: Up to $20 billion lost - The Connecting for Health project is described as costing up to this amount in losses. Number of failed research grants for Bob Langer: First 9 grants turned down - Langer described repeated early grant rejections. Number of failed experiment attempts for Bob Langer: Over 200 - He said he failed over 200 times before getting a result that worked. Time spent working days: 20-hour days - Langer said he worked extremely long hours during his experiments. Years Carol Hemmelgarn co-founded advocacy group: 2021 - She co-founded Patients for Patient Safety US in 2021.

Pivotal Quotes: "We don’t treat scared, afraid, and frightened. And that’s what my daughter was." — Carol Hemmelgarn: Explaining how mislabeling her daughter as anxious contributed to inadequate care. "Human error is almost never criminal." — Amy Edmondson: Her critique of criminalizing the Vanderbilt medication error and the chilling effect it has on reporting. "I failed over 200 times before I finally got something to work." — Bob Langer: Describing repeated experimental failure as part of scientific progress.

Implications: Hospitals and other complex institutions need transparent, blame-aware learning systems. Punishment alone won’t improve safety; reporting, experimentation, and leadership accountability are more likely to prevent future harm.