Diagnosis

Interpreting the Shadows

Pat Croskerry, Karen Cosby, Mark L. Graber, Hardeep Singh

• Publisher: CRC Press
• Year: 2017
• ISBN:
  9781351650199
• Edition: 1
• Topic: Médecine
• Subtopic: Santé publique, administration et soins

Despite diagnosis being the key feature of a physician's clinical performance, this is the first book that deals specifically with the topic. In recent years, however, considerable interest has been shown in this area and significant developments have occurred in two main areas: a) an awareness and increasing understanding of the critical role of clinical decision making in the process of diagnosis, and of the multiple factors that impact it, and b) a similar appreciation of the role of the healthcare system in supporting clinicians in their efforts to make accurate diagnoses. Although medicine has seen major gains in knowledge and technology over the last few decades, there is a consensus that the diagnostic failure rate remains in the order of 10-15%. This book provides an overview of the major issues in this area, in particular focusing on where the diagnostic process fails, and where improvements might be made.

Section IV

Challenges and Controversies in Diagnosis

10

Diagnostic Error

Karen Cosby

CONTENTS

Introduction

Incidence of Diagnostic Error

Sources of Error in Diagnosis

Patient Interview

Physical Examination

Radiology

Clinical Laboratory

Anatomical Pathology

Communication and Coordination of Care

The State of Medical Knowledge

Strategies and Solutions to Address Diagnostic Error

Optimize the Performance of Clinicians

Improve Systems of Care

Engage and Empower Patients

Innovation and Technological Advancements

Conclusion

References

Introduction

Diagnostic work is quite amazing. There are myriad diagnoses, yet relatively few physical expressions of disease; there are nearly infinite possibilities, yet only one right answer. The diagnostician often seems to be part scientist, part shaman. The fact that our diagnoses are accurate most of the time is almost surprising considering the conditions of uncertainty under which we operate and the endless demands of a busy clinical practice. The gambler might well hedge his bets on diagnostic success! We would be impressed with ourselves if it were not for the remaining 10% or so of missed diagnoses— cases that reflect human lives that are diminished by our failures. While there are countless steps and processes in diagnostic evaluations, diagnostic errors are generally attributed to cognitive errors, system flaws, or both. This chapter examines processes common to many diagnostic workups, reviews contributing factors to diagnostic errors, and provides suggestions for clinicians, healthcare institutions, and patients to improve the odds of making the right diagnosis.

Incidence of Diagnostic Error

The Institute of Medicine (IOM) report To Err Is Human heralded the birth of the patient safety movement in 2000 and awakened the healthcare community’s consciousness to the reality that harm comes to many in the course of their medical care [1]. A quiet rumble of concern was raised early on by a few who felt that diagnostic delays and errors should be addressed as part of patient safety initiatives, but was largely ignored to attend to the “ low lying fruit” of treatment-related harm [2– 7]. Judgments about diagnostic errors were considered controversial, since they inevitably involve some degree of hindsight bias. Admittedly, some of the concern was rooted in the sense that we ought not to judge another person’s cognitive process, lest we too be judged. Doing so was almost a betrayal of the inviolable trust among our peers and profession. Some of the hesitancy in addressing diagnosis error is understandable; diagnostic error is hard to define, difficult to detect, and challenging to study [8– 12].

We are only just beginning to grapple with the definition and measurement of diagnostic error, and the early numbers reveal what many have suspected: diagnosis is a highly uncertain and imperfect process [13]. Most of us will experience a diagnostic error in our lifetimes [14]; 12 million Americans experience a diagnostic error each year [14,15]. The annual death toll attributable to diagnostic errors has been estimated to be between 40,000 and 80,000 lives each year in the United States [16]. Estimates from adult autopsies suggest that 71,400 of adults who die in hospitals each year (8.4% of all adult deaths) have a major diagnostic error, half of them significant enough to have likely impacted their outcome [17]. More than 34,000 patients die each year in intensive care units (ICUs) with Class I diagnostic errors— major conditions that might have been treatable and survivable had they been identified [18]. Data from pediatric ICUs reveal diagnostic errors in approximately 20% of autopsied cases [19]. Not all diagnostic errors occur in hospitals; half of missed diagnoses occur in ambulatory settings [14]. Five percent of outpatients experience a diagnostic failure each year, most commonly involving delays in the detection of lung, breast, and colon cancer [14,20]. Diagnostic errors occur across the spectrum of healthcare, including hospital wards, emergency departments (EDs), ICUs, and ambulatory settings; they also occur in all specialties. One recent U.S. survey found that 35% of adults recalled having personally experienced a medical error (in either themselves, a family member, or close friend); half of these were diagnostic errors [21,22].

Diagnosis error is the leading source of paid malpractice claims, accounting for the highest proportion of payments and the largest settlements, and involving the cases most likely to experience the worst patient outcomes (death or major disability). In a summary of 25 years of claims data, diagnostic error accounted for $38.8 billion U.S. dollars in settlements [23]. The biggest revelation is perhaps in finally admitting that healthcare providers, institutions, and processes are imperfect. Despite advanced training, commitment, passion, and high-tech care, we have simply failed to recognize and design for fallibility.

Sources of Error in Diagnosis

On the face of it, diagnostic activity would seem to be a mostly cognitive process that is largely dependent on expert clinical reasoning; we have explored the elements of reasoning and clinical decision making in the first half of this book. Diagnosis also relies on coordinated activity with different phases and processes of care provided by the healthcare system. The recent report from the National Academies of Sciences, Engineering, and Medicine, Improving Diagnosis in Health Care, describes diagnosis as a process, as illustrated in Figure 10.1 [24]. Attempts to address diagnostic error can be broken down into improvements made in each of these steps, beginning with patient entry into the healthcare system.

Patient Interview

Most diagnoses begin with a conversation— an interview with the patient. No one has yet established how accurate, reproducible, or valid the patient history is. However, clinicians will attest to how variable and difficult even a basic routine history can be. Patients may have difficulty describing their symptoms or use adjectives that differ from standard textbook descriptions. In academic settings, the information obtained from an inexperienced medical student may differ from that of other team members (sometimes better if they take the time to listen, but often not if they approach the history with closed-ended questions). The questions that are asked, how they are asked, and how much of a rapport the examiner has with the patient, can all influence the quality of information obtained. Sometimes patients don’t offer the most relevant history; a worried or frightened patient may sense that something is wrong but fixate on extraneous sensations, unable to discriminate between those that are relevant and those that may simply be benign or normal variants. In some cases, patients may be unable to contribute to a meaningful interview— too ill, medically illiterate, or simply disengaged.

FIGURE 10.1

Diagnosis as a process, and points in the process where failure can occur. (Reprinted from National Academies of Science, Engineering, and Medicine. Improving Diagnosis in Healthcare , National Academy Press, Washington, DC, 2015. With permission from the National Academies Press, Copyright 2016 [24].)

The history is probably one of the most important yet precarious aspects of diagnosis. The quality of information gained from the clinical history is essential to an accurate diagnosis, but is largely dependent on the expertise of the interviewer. There are two types of information obtained in the clinical history: (1) facts and (2) subjective descriptions. The facts include (in part) past medical events, current medications, timeline of symptoms, and results of previous testing. The subjective description includes the patient’s account of their illness as they experience it. Features of the history may have var...