The Important but Rarely Studied Cascade of Care

While technology has greatly improved health care, its advances are not free of downsides. A potential harm is the clinical cascade following an incidental finding on a medical test. Many reports have described cascades, but none have attempted to systematically examine the national scope of the problem and its consequences on patients and clinicians. Ganguli et al¹ present the results of a 44-item online survey that asked US internists about their experiences with cascades and potential solutions. In a sample of 376 physicians (44.7% response rate), the authors reported that nearly all respondents had experienced a cascade—with both positive and negative outcomes.¹ The frequency of cascades with no meaningful outcome surpassed those with meaningful outcomes 31.1% to 14.8%.¹ Nearly 90% of physicians reported that a cascade had caused patient harm, including psychological harms, treatment burden, financial burden, and physical harm.¹ Cascades also negatively affected physicians in terms of wasted time and effort, frustration, and anxiety.¹ Notably, approximately half of physicians experienced a cascade when they or family members were patients.¹

Consider first that cascades are core to the practice of medicine. The whole point of mammography and prostate-specific antigen testing, for instance, is to trigger more imaging, a biopsy, and potentially beneficial therapy. Similarly, electrocardiography and a troponin assay, when positive in patients with chest pain, could lead to eventual cardiac catheterization and revascularization. The problem is that no clinical test has perfect sensitivity and specificity. Thus, if you practice medicine today, you have likely seen a patient who has fallen ill from cascades.

We are drawn to the work of Ganguli et al¹ because cascades following incidental findings will surely worsen as tests become more sensitive and relied on without being specific for a disease. The advent of the high-sensitivity troponin assay is a classic example of how a more technologically advanced test can be more difficult to use at the bedside. In the past, any detectable troponin in the blood indicated an acute cardiac problem. This made for binary decisions: positive was a rule-in and negative, a rule-out. However, higher-sensitivity troponin assays can detect low levels of troponin in patients without heart disease. Finding incidental troponin in the blood is akin to seeing unknown shadows on an ultrasound, white spots on a computed tomography scan, or distal subsegmental emboli in the pulmonary vasculature. New tests require new paradigms, but new paradigms require time to develop and be accepted.

One finding from Ganguli et al¹ identified the core problem clinicians face today: while the survey reveals that physicians often experience cascades from incidental findings that cause harm, nearly as many physicians report good outcomes from the workup of an incidental finding. A nonspecific finding on an electrocardiogram is more likely to lead to unnecessary, costly, and potentially harmful cardiac testing, but it sometimes finds a left main coronary lesion.

Probabilistic thinking and innumeracy are challenges for both patients and phyisicians.² What clinicians do not understand about numbers may hurt patients.³ The first step in dealing with uncertainty is recognizing it. Here we note that the survey results in the study by Ganguli et al¹ stratified by trainee vs attending status. Namely, trainees were more likely than attendings to consult a generalist, guidelines, or the primary literature to lessen the consequences of a cascade. Also, trainees more often believed that value-based payment models would help to mitigate cascades, suggesting that younger clinicians may place more weight on the role of incentives in human behavior. These observations tempt us to be optimistic that increased medical education in probabilistic thinking may help mitigate unwanted cascades.

Another potential cause of cascades turns on cultural expectations. Harm from overzealous screening is underrecognized. The general public is inundated with disease-of-the-month campaigns, but there are no overdiagnosis awareness months.⁴ A person who undergoes an unnecessary test or procedure today may feel satisfied or relieved that they don’t have cancer or heart disease. Ivan Illich, PhD, called this the sickening of society or social iatrogenesis,⁵ and it relates to what H. Gilbert Welch, MD, calls anticipatory medicine. In other words, the medical system can easily convince people with no symptoms that they need our help to ward off impending doom.⁶ Hindsight, omission and availability biases, plus the illusion of control among patients and clinicians bolster the cultural expectation that healthy people need health care.

Distraction also plays a role in harmful cascades. While there is great emphasis in the academic and lay press on medical errors, little emphasis is given to the role of clinical decisions in cascade iatrogenesis.⁷ Obvious medical errors, such as wrong-site surgical procedures and medication administration errors, are serious examples of medical harm, but more important may be the seemingly banal everyday decisions to order a test or perform a procedure. An echocardiogram is a painless noninvasive scan that instills little fear in the tester or tested. But an aberrant shadow that cascades all the way to needless heart surgery would not be characterized as a medical error. Looking upstream in a cascade to the initial inappropriate test done in the name of safety is a type of medical error that needs further exploration. Clinical studies to define the frequency of cascade medical errors and the tests most prone to harmful cascades would be a start.

We agree with many of the physicians surveyed who felt that decision support holds promise for reducing unwanted cascades. The problem with decision support is that it must be designed to add value and be easily accessible without increasing burden for clinicians. Decision support needs to better provide relevant information at the point of care to make decision-making easier for clinicians. To address clinical cascades, we need research on the actual meanings of incidental findings to be presented with reports. For many tests, providing the sensitivity and specificity with the test result could remove a barrier to accurate interpretation. For other tests, a report could contain information on the frequency of incidental findings, what results from those findings, and what options are available for workup. The movement in radiology to include evidence-based recommendations when reporting incidental nodules is a good first step.

Ganguli et al¹ have added greatly to the public good. In medical practice, the first step is always to identify the problem. Their survey study¹ brings needed attention to clinical cascades—a problem that will surely worsen as medical technology redefines normal and increasingly digitizes the human condition.

Published: October 16, 2019. doi:10.1001/jamanetworkopen.2019.13315

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Mandrola J et al. JAMA Network Open.

Corresponding Author: Daniel J. Morgan, MD, MS, University of Maryland School of Medicine, 10 S Pine St, MSTF 334, Baltimore, MD 21201 (dmorgan@som.umaryland.edu).

Conflict of Interest Disclosures: Dr Morgan reported receiving support to travel to meetings from IDWeek, the Lown Institute, and the Society for Healthcare Epidemiology of America; receiving honoraria for journal editing from Springer Nature; and receiving grants from the National Institutes of Health, the US Centers for Disease Control and Prevention, the Agency for Healthcare Research and Quality, and the US Department of Veterans Affairs. No other disclosures were reported.