Diagnostic Errors: A New Chapter in Patient Safety Science, Policy, and Practice

by Hardeep Singh, MD, MPH

Perspective

For the field of diagnostic error, 2015 was a busy year. Those of us who have been trying for years to elevate the issue were particularly gratified to see the release of an Institute of Medicine (IOM) report on this challenging topic in September.(1) The IOM made several concrete recommendations to reduce diagnostic error, focusing not only on individual clinicians but also patients, health care delivery systems, and diagnostic services such as laboratory, pathology, and radiology. In addition to the IOM report, we have seen increased media coverage of diagnostic errors, landmark papers in leading journals (2,3), and the launch of a new Coalition to Improve Diagnosis under the auspices of the Society to Improve Diagnosis in Medicine.(4) Despite these laudable developments, it will take years of sustained action to ensure that recommended reforms see the light of day: adequate time and reimbursement for diagnostic thinking, reformed medical malpractice standards, technologies that support rather than hinder patient care, and a research funding climate conducive to accelerating our understanding of diagnostic errors and preventive interventions.

Despite abundant evidence regarding the substantial toll of diagnostic error, the issue has remained outside the mainstream conversation on patient safety. The reasons for this are several. First, diagnostic errors are often confused with errors of screening, prevention, and management and are overshadowed by contemporary issues of overdiagnosis, overtesting, overtreatment, and high health care costs. Second, diagnosis sits squarely at the intersection of science and art in medicine and involves understanding the physician's mind. There is a palpable tension between missing a diagnosis on one hand and ordering unnecessary tests or procedures on the other. Efforts to identify the right balance are often obscured by hindsight. Many system factors also affect diagnostic reasoning, and our understanding of this complexity is only now emerging. Often it's difficult to separate systems and cognitive factors that lead to errors; the latter tend to be harder to understand. Third, making a definitive diagnosis is not always a black-and-white proposition, especially when patients' symptoms and underlying pathophysiology unfold over time. Despite this well-known fact, physicians are generally not comfortable expressing uncertainty to patients (or documenting their ambivalence). Calibrating utilization of tests and procedures for optimal patient outcomes, and managing uncertainty more deliberately, are challenging but exciting areas for future work.

While researchers tackle and, ultimately, solve these complex issues, it will be a huge leap forward if the IOM report fosters greater openness to discussing and learning from diagnostic errors and partnering with patients on solutions. Little is known about optimal strategies for clinicians and health care systems to receive feedback on diagnostic performance. In our work, we have found that clinicians are reluctant to discuss diagnostic errors and often become defensive when talking about their decision-making processes, even in confidential research settings. Progress will require not only advances in the science but also changes to today's culture, which inhibits fact-finding and error acknowledgement.

Several measurement and definitional challenges will also need to be addressed to ensure progress. We cannot measure what we cannot define. The IOM definition of diagnostic error—"the failure to (a) establish an accurate and timely explanation of the patient's health problem(s) or (b) communicate that explanation to the patient"—is good for conceptual thinking.(1) But operationalizing it in real-world clinical settings will be hard.

To enable more rigorous measurement, we use a pragmatic definition of diagnostic errors that accounts for the evolving diagnostic process and other real-world challenges described above: missed opportunities to make a correct or timely diagnosis based on the available evidence, regardless of patient harm.(5) This description includes several types of overlapping situations. For example, diagnoses can be completely missed (cancer missed despite alarming symptoms), wrong (patients told they have one diagnosis but actually have evidence of another), or delayed (patient had an abnormal test result suggestive of cancer but no one has reviewed it or notified the patient). Looking for unequivocal evidence that some critical finding or abnormality was missed, or not investigated when it should have been, is essential if we are to understand how to reduce preventable diagnostic harm. Some experts have criticized the practice of evaluating errors in hindsight, but we maintain that systematic analysis of known errors is crucial to uncovering process breakdowns that put patients at risk.(6)

Despite some progress, we need a lot more research to improve diagnosis in the real world. Few institutions and practices are studying this issue. We need data to understand and reduce errors from more institutions than the handful that have contributed to the emerging evidence base thus far. Concurrently, we must expand the community of researchers who are working on the science of measurement.(7) In addition to increasing federal research funds for study of diagnostic errors, health care systems themselves could fully or partially fund improvement projects. This benefits them and the patients they serve. For example, the Department of Veterans Affairs has successfully implemented initiatives where researchers are supported by and work closely with clinical operations stakeholders on quality and safety issues, including those related to diagnosis.(8)

Targets for improvement efforts have become clearer in the past few years. In addition to improved measurement of diagnostic errors, these areas include clarifying responsibilities for follow-up of abnormal clinical findings, identifying at-risk patients for better tracking or follow-up, enhancing doctor–patient communication and relationships, and tracking follow-up on certain types of high-risk abnormal test results (such as those suspicious for cancer). To bolster measurement (9), patient safety teams can tap into a treasure trove of data from incident reports, patient complaints, root cause analyses, peer reviews, and malpractice claims. Additionally, institutional leaders should consider using electronic triggers (10,11) and developing local clinician champions who can motivate other clinicians to discuss and learn from error.(12) Another avenue for future work is the inclusion of smaller practices that lack the traditional quality improvement infrastructures of larger systems. For example, with further refinement, a single clinician could use the recently developed Safer Dx instrument for identifying diagnostic error through record reviews.(13)

There are numerous challenges and opportunities ahead. These include how to leverage teamwork in diagnosis, improve transparency of the diagnostic process, and break down silos among clinicians as well as between frontline clinicians and diagnostic specialties (such as lab and radiology). Patient-facing interventions, including apps, online second opinions, symptom checkers, and crowdsourcing to obtain a diagnosis are on the horizon. While these are exciting, we need to ensure that these interventions are rigorously evaluated.(14,15) Our work ahead should also define how to maintain our critical thinking skills, particularly in an environment of increasingly complex and automated technologies.(16)

Despite all the exciting innovations, one of the most important pathways to reduce error is to preserve the fundamentals of clinical practice amid the time pressures and the overwhelming electronic data. We must work to ensure that clinicians have the time and skills they need to communicate effectively with patients, and with one another, to formulate a correct diagnosis.(17) Upholding those fundamentals will likely remain key to timely and accurate diagnosis for decades to come.

Hardeep Singh, MD, MPH
Chief, Health Policy, Quality and Informatics Program
Veterans Affairs Center for Innovations in Quality, Effectiveness and Safety
Michael E. DeBakey Veterans Affairs Medical Center and Baylor College of Medicine, Houston, Texas

References

1. Improving Diagnosis in Health Care. Committee on Diagnostic Error in Health Care. National Academies of Science, Engineering, and Medicine. Washington, DC: National Academies Press; 2015. [Available at]

2. Singh H, Graber ML. Improving diagnosis in health care—the next imperative for patient safety. N Engl J Med. 2015;373:2493-2495. [go to PubMed]

3. McGlynn EA, McDonald KM, Cassel CK. Measurement is essential for improving diagnosis and reducing diagnostic error: a report from the Institute of Medicine. JAMA. 2015;314:2501-2502. [go to PubMed]

4. Society to Improve Diagnosis in Medicine. [Available at]

5. Singh H. Editorial: Helping health care organizations to define diagnostic errors as missed opportunities in diagnosis. Jt Comm J Qual Patient Saf. 2014;40:99-101. [go to PubMed]

6. Landro L. A medical detective story: why doctors make diagnostic errors. Wall Street Journal. September 26, 2015. [Available at]

7. In conversation with Hardeep Singh, MD, MPH [interview]. AHRQ WebM&M [serial online]. December 2013. [Available at]

8. Patient Safety Centers of Inquiry (PSCI). Washington, DC: VA National Center for Patient Safety, US Department of Veterans Affairs; 2015. [Available at]

9. Singh H, Sittig DF. Advancing the science of measurement of diagnostic errors in healthcare: the Safer Dx framework. BMJ Qual Saf. 2015;24:103-110. [go to PubMed]

10. Murphy DR, Wu L, Thomas EJ, Forjuoh SN, Meyer AND, Singh H. Electronic trigger-based intervention to reduce delays in diagnostic evaluation for cancer: a cluster randomized controlled trial. J Clin Oncol. 2015;33:3560-3567. [go to PubMed]

11. Singh H, Giardina TD, Forjuoh SN, et al. Electronic health record-based surveillance of diagnostic errors in primary care. BMJ Qual Saf. 2012;22:93-100. [go to PubMed]

12. Okafor N, Payne VL, Chathampally Y, Miller S, Doshi P, Singh H. Using voluntary reports from physicians to learn from diagnostic errors in emergency medicine. Emerg Med J. 2015 Nov 3; [Epub ahead of print]. [go to PubMed]

13. Al-Mutairi A, Meyer AND, Thomas EJ, et al. Accuracy of the Safer Dx measurement instrument to identify diagnostic errors in primary care. J Gen Intern Med. [In press].

14. Semigran HL, Linder JA, Gidengil C, Mehrotra A. Evaluation of symptom checkers for self diagnosis and triage: audit study. BMJ. 2015;351:h3480. [go to PubMed]

15. Meyer AND, Longhurst CA, Singh H. Crowdsourcing diagnosis for patients with undiagnosed illnesses: an evaluation of CrowdMed. J Med Internet Res. [In press].

16. Upadhyay DK, Sittig DF, Singh H. Ebola US Patient Zero: lessons on misdiagnosis and effective use of electronic health record. Diagnosis. 2014;1:283-288. [Available at]

17. Singh H, Giardina TD, Meyer AND, Forjuoh SN, Reis MD, Thomas EJ. Types and origins of diagnostic errors in primary care settings. JAMA Intern Med. 2013;173:418-425. [go to PubMed]