Incidental Findings from Low-Value Screening and Resulting Cascades of Care in the United States

by yaleglobalhealthreview, posted in Uncategorized

BY FRANK HORRIGAN

Introduction

Before undergoing surgery for a painful hernia, Mr. X, prompted by his medical history, underwent a preoperative chest CT scan – an imaging practice of uncertain value. That scan revealed a small nodule on his left lung, prompting a delay in surgery to follow up on the potential cancer. His follow-up CT found another nodule, leading to further scans and delays. Only after months of nagging hernia pain and anxiety over the CT results was he able to proceed with his planned surgery (Niess and Prochazka, 2014). 

Mr. X’s experience was not an anomaly – incidental findings disconnected from a patient’s chief complaint following a routine scan are highly prevalent. Approximately 15-30% of all diagnostic imaging and 20-40% of CT scans contain such an “incidentaloma” (Lumbreras et al., 2010). While the percentage of scans with incidentalomas remains consistent, the percentage with clinically significant findings varies depending on the value of the screening, with low-value screening having signals almost exclusively from incidentalomas. Low-value screens and their subsequent barrages of follow-up tests, diagnoses, and procedures, dubbed a “cascade of care,” rarely catch clinically significant diagnoses. Instead, they often unnecessarily increase healthcare costs and subject patients to emotional and physical complications.

In this review, I argue that cascades of care are problematic because of their significant and unnecessary costs to the US healthcare system. Next, I will dissect the factors fueling low-value screening that results in cascades of care and explore why these practices persist. Finally, I propose several potential solutions, encompassing economic, policy, and technology-driven approaches.

Costs

Cascades of care increase healthcare spending across patient populations and health payer models. Several observational and retrospective studies have indicated the overall costs associated with cascades of care: from $35 million in additional Medicare spending from pre-cataract surgery electrocardiograms (EKGs) to $14 million in Veterans Health Administration spending because of preoperative EKGs and chest radiographs (Ganguli et al., 2019; Pickering et al., 2019). Commercially insured adults also generate higher spending from insurance and out-of-pocket payouts (Ganguli et al., 2023). In total, low-value screening, testing, and procedures are estimated to cost $17.2-27.9 billion annually (Shrank et al., 2019). 

Beyond their direct financial costs, cascades of care impose emotional tolls on patients and healthcare providers, which have knock-on effects on economic output. In a national survey of physicians, 86.7% reported that cascades of care had harmed their patients, including psychological harm (68.4%), treatment burden (65.4%), and financial toxicity (57.5%) (Ganguli et al., 2019). These burdens on patients can have effects on their ability to work, as being in poor health, both physical and mental, increases a patient’s risk of job loss or unemployment (Antonisse and Garfield, 2018). For the physicians themselves, 69.1% reported wasting time and effort on cascades of care. As a result, substantial numbers of physicians reported feeling frustration (52.5%) and anxiety (45.4%). When left unaddressed, these issues regarding physicians’ mental health and job satisfaction can lead to burnout (Ganguli et al., 2019). As demonstrated by the national survey of physicians, cascades of care create burdens on patients and physicians that can potentially lead to emotional distress, loss of employment, and burnout. 

Even when incidentalomas happen to lead to early diagnosis of a real condition, treatment at such early-stages can counter-intuitively lead to worse patient outcomes due to complications with overdiagnosis. Overdiagnosis occurs when incidence of a disease is inflated by relatively benign cases (Davenport, 2023). This means that in many cases, especially for older patients who are more likely to die from an unrelated condition before the diagnosed condition, patients are disproportionately exposed to the adverse effects of treatment rather than the benefit (Patz et al., 2000; Hollingsworth et al., 2007). Overdiagnosis exemplifies how cascades of care are more costly than beneficial when detecting insignificant disease. 

Driving Factors

Having demonstrated the sizable costs of cascades of care to the US healthcare system, I now identify the factors explaining why they remain prevalent. Cascades of care are initiated by low-value screening, which persist due to issues in medical practice and financial incentives.

Defensive medicine is a practice in which physicians change their clinical behavior out of fear of malpractice liability, often leading to increased screening. Defensive medicine occurs in all specialities, but is particularly commonplace in emergency medicine, where the fast-paced environment pushes physicians to order excessive testing. A 2015 survey reported that 85% of emergency physicians (EPs) believed their emergency departments ordered too many tests, with 97% acknowledging personally ordering medically unnecessary imaging; 92% cited defensive medicine as sometimes, often, or always being a contributing factor (Kanzaria et al., 2015). Indeed, liability immunity has been shown to reduce inpatient spending by 5% without harming patient outcomes, although not all of these cost savings are from reducing cascades of care (Frakes and Gruber, 2019).

Another driver of incidentalomas is a poor understanding and communication of risk between physicians and patients. It is well-documented that people are generally not good at estimating risk, particularly around rare events like plane crashes or winning the lottery. The same applies to medical screening. In the above mentioned 2015 survey of EPs, 94.2% of respondents reported not wanting to miss a diagnosis, even with a low-likelihood of detection, as a reason for unnecessary advanced imaging (Kanzaria et al., 2015). The results of this survey illustrate how physicians can have difficulty estimating risk-to-benefit for scans, leading to unnecessary screening to err on the side of caution. Patients are also poor at estimating scan-risk, with interviews revealing that most do not see potential downsides including false positives or incidental findings (Mulligan et al., 2022). Combined with a failure of many physicians to communicate risks, the benefits of low-value tests and risks of not being tested can be overblown, leading to more low-value testing such as prostate-specific antigen screening in older males or mammograms in younger women (Kalavacherla et al., 2023, Keating and Pace, 2018). 

Congruent with poor risk perception, physicians often implement routine diagnostic screening too aggressively, as I will illustrate with a cancer-screening case study. Between 1975 and 2009, thyroid cancer incidence in the US nearly tripled without a corresponding improvement in mortality. Indeed, the vast majority of increased detection was for papillary thyroid cancer, a relatively benign tumor that often does not produce symptoms in one’s lifetime (Davies and Welch, 2014). Still more people were diagnosed and became patients, being subjected to further tests and treatments without clear benefits. This rise in incidence without a drop in mortality continues to be problematic today (Davies and Hoang, 2020). As evidenced by our continued challenges with excessive thyroid cancer screening, overly aggressive routines drive cascades of care. 

There are also several financial drivers of cascades of care, such as the fee-for-service (FFS) health insurance system that dominates in the US, which financially incentivizes cascades of care. Although there have been recent shifts towards alternative models of insurance including value-based care and capitated payments, FFS health reimbursement remains substantial with around 56% of Medicare beneficiaries as of 2020 and many commercially insured patients under similar volume-based systems (Clerveau et al., 2023). In FFS health insurance, providers are reimbursed based on services performed, thereby incentivizing excess healthcare utilization (Pearl, 2017). Although the evidence is not clear that FFS health insurance leads to more costly cascades of care than alternatives like Medicare Advantage plans (Park et al., 2021), it is clear the FFS health insurance fails to disincentivize them. The lack of insurance model effect on cascades of care could reflect that 88.1 percent of physicians reported receiving some compensation from FFS payment methods in 2020, demonstrating their predominance in physician compensation (Rama, 2021). 

A recent, particularly alarming driving factor behind cascades of care is for-profit businesses offering preventative full-body scans, which have the potential to massively increase incidental findings (Kwee and Kwee, 2019). Companies such as Prenuvo, Ezra, and simonONE offer preventative full-body MRI scans with the selling-point of allowing patients to take control of their health and treat ailments before they become serious. By now in this paper, it should be clear that scanning completely asymptomatic people for early-disease is begging for incidentalomas. Indeed, the American College of Radiology (ACR) issued a statement discouraging their use, citing risk for cascades of care (ACR, 2023). Still, these companies remain in action, with Prenuvo having raised $70M in series A funding in 2022. Preventative screening companies demonstrate the extreme of financial incentives driving low-value screens and cascades of care.

Potential Solutions

To best address the costs of cascades of care, we need to stop them at their roots by preventing low-value screening and follow-ups to resulting incidentalomas. Firstly, we can address some of the medical system causes of cascades of care by supporting research into and development of appropriate care guidelines. The national survey of physicians that revealed the emotional tolls of cascades of care also reported that 53.2% physicians felt cascades of care were due to lack of guidelines for follow-up testing. As for potential solutions, 62.8% and 44.6% suggested accessible guidelines and decision aids respectively (Ganguli et al., 2019). Easily accessible and well-researched decision-guides will also address defensive medicine by giving physicians a clear list of necessary steps, leading to fewer cascades of care and less worry about being sued. 

Similarly, screening programs should be carefully crafted, considering the risks of incidental findings. If existing screening programs are found to be harmful, they should be publicly discouraged. For example, South Korea faced a similar problem to the US regarding overdiagnosis of thyroid cancer, with an even greater 15-fold increase in cases over approximately 20 years (Ahn et al., 2014). However, upon raising awareness of the overdiagnosis problem through outreach by a physician consortium and a government anti-screening campaign, they experienced a 35% drop in screening within a year, which has been sustained to this day (Ahn and Welch., 2015; Oh et al., 2021). When changing screening guidelines, we cannot forget the importance of communication. While the US Preventive Services Task Force has recommended against routine mammograms for women under 50 since 2009, as of 2018, there were few changes in screening practices, demonstrating ineffective communication and implementation of guidelines (Keating and Pace, 2018).

We can also take advantage of rapid growth in artificial intelligence (AI) technologies to improve risk evaluation and facilitate physician-patient communication. Assistive technologies have been around for a long time, helping radiologists with making reads (Hosny et al., 2018). While not all systems directly detect or point out potential incidental findings, by the nature of a blob not being picked up by AI, radiologists are more likely to ignore them in favor of more significant findings. Indeed, AI models for analyzing breast imaging have been shown to be effective when inserted into existing workflows in Sweden, with the potential for reduced false-positives (Dembrower et al., 2023). Similarly, AI could provide an objective, data-based opinion to inform the use of certain tests and enforce care guidelines. Indeed, clinical decision support systems (CDSSs), computerized systems that employ clinical and patient data to improve healthcare delivery, have been around since the 1970s, although they will need some modifications to be truly effective. Relevant to cascades of care, many CDSSs already alert physicians when they are ordering low-value care, but physicians reject these alerts up to 90% of the time because of a distrust in computer recommendations and alert fatigue, a phenomenon when physicians become desensitized to alerts because many of them are insignificant (Levi and Gorenstein, 2022; Sutton et al., 2020). To address distrust of computer recommendations, CDSSs could highlight low-value care in relation to the physician’s peers. For example, along with citing data of how unlikely the test they are ordering will be useful, also include data highlighting their overuse of a test relative to their peer physicians. If there is continued overuse, have the CDSS trigger a utilization review. Beyond improving physicians’ adherence to evidence-based medicine with data-driven opinions, AI could also help to facilitate communication between physicians and patients by generating reports and translating them into terms that are understandable for patients (Amin et al., 2023, Yu et al., 2023). 

To financially disincentivize cascades of care, the ideal method would be shifting to a value-based care approach, rather than the currently prevalent FFS or volume-driven system. Although high-deductible health plans discourage unnecessary screening by decreasing patients’ moral hazard when seeking care, they do not effectively limit the occurrence of unnecessary screening once the patient has already sought medical attention. (Chou et al., 2021). Meanwhile, accountable-care organizations have been shown to moderately reduce low-value screening through their risk-sharing payment structure (Schwartz et al., 2015). Accountable-care organizations, and similar value-based health systems, receive a capitated, risk-adjusted lump sum payment in exchange for complete responsibility over a patient’s care. This risk-sharing financially motivates them to cut out low-value screening and their subsequent cascades of care, as more tests and procedures will eat into profits rather than generate them. While this solution sounds promising on paper, it can be extremely challenging to implement by requiring a complete overhaul of the healthcare system, enormous upfront investment, and a change in mindset for providers (Goodrich, 2023).

Conclusion

Cascades of care are a financially and emotionally costly inefficiency within our healthcare system, driven by a multifaceted set of practice and financial reasons. While we must recognize there is no easy fix to the issue, which is why it has existed for so long, there are a vast array of potential solutions from improving the dissemination of existing guidelines to completely changing the delivery of healthcare to a value and data-based system. By working on reducing cascades of care, we are not only preventing needless suffering of patients, but also building a healthcare system that will be more efficient and sustainable in the long-run.

Frank Horrigan is a sophomore in Pauli Murray College. He can be contacted at frank.horrigan@yale.edu.

——————————

References

  1. ACR. (2023, April 17). ACR statement on screening total body MRI. American College of Radiology. https://www.acr.org/Media-Center/ACR-News-Releases/2023/ACR-Statement-on-Screening-Total-Body-MRI 
  2. Ahn, H. S., Kim, H. J., & Welch, H. G. (2014). Korea’s thyroid-cancer “epidemic”–screening and overdiagnosis. The New England journal of medicine, 371(19), 1765–1767. https://doi.org/10.1056/NEJMp1409841
  3. Ahn, H. S., & Welch, H. G. (2015). South Korea’s Thyroid-Cancer “Epidemic”–Turning the Tide. The New England journal of medicine, 373(24), 2389–2390. https://doi.org/10.1056/NEJMc1507622
  4. Amin, K., Khosla, P., Doshi, R., Chheang, S., & Forman, H. P. (2023). Artificial Intelligence to Improve Patient Understanding of Radiology Reports. The Yale journal of biology and medicine, 96(3), 407–417. https://doi.org/10.59249/NKOY5498
  5. Antonisse L., Garfield, R. (2018, August 7). The relationship between work and Health: Findings from a literature review. KFF. https://www.kff.org/medicaid/issue-brief/the-relationship-between-work-and-health-findings-from-a-literature-review/
  6. Chou, S. C., Hong, A. S., Weiner, S. G., & Wharam, J. F. (2021). High-deductible health plans and low-value imaging in the emergency department. Health services research, 56(4), 709–720. https://doi.org/10.1111/1475-6773.13569
  7. Clerveau G., Ochieng, N., Cubanski, J., Neuman, T. (2023, August 14). A Snapshot of Sources of Coverage Among Medicare Beneficiaries. KFF. https://www.kff.org/medicare/issue-brief/a-snapshot-of-sources-of-coverage-among-medicare-beneficiaries/
  8. Davenport M. S. (2023). Incidental Findings and Low-Value Care. AJR. American journal of roentgenology, 221(1), 117–123. https://doi.org/10.2214/AJR.22.28926
  9. Davies, L., & Hoang, J. K. (2021). Thyroid cancer in the USA: current trends and outstanding questions. The lancet. Diabetes & endocrinology, 9(1), 11–12. https://doi.org/10.1016/S2213-8587(20)30372-7
  10. Davies, L., & Welch, H. G. (2014). Current thyroid cancer trends in the United States. JAMA otolaryngology– head & neck surgery, 140(4), 317–322. https://doi.org/10.1001/jamaoto.2014.1
  11. Dembrower, K., Crippa, A., Colón, E., Eklund, M., Strand, F., & ScreenTrustCAD Trial Consortium (2023). Artificial intelligence for breast cancer detection in screening mammography in Sweden: a prospective, population-based, paired-reader, non-inferiority study. The Lancet. Digital health, 5(10), e703–e711. https://doi.org/10.1016/S2589-7500(23)00153-X
  12. Frakes, M., & Gruber, J. (2019). Defensive Medicine: Evidence from Military Immunity. American economic journal. Economic policy, 11(3), 197–231. https://doi.org/10.1257/pol.20180167
  13. Ganguli, I., Lupo, C., Mainor, A. J., Raymond, S., Wang, Q., Orav, E. J., Chang, C. H., Morden, N. E., Rosenthal, M. B., Colla, C. H., & Sequist, T. D. (2019). Prevalence and Cost of Care Cascades After Low-Value Preoperative Electrocardiogram for Cataract Surgery in Fee-for-Service Medicare Beneficiaries. JAMA internal medicine, 179(9), 1211–1219. https://doi.org/10.1001/jamainternmed.2019.1739
  14. Ganguli, I., Simpkin, A. L., Lupo, C., Weissman, A., Mainor, A. J., Orav, E. J., Rosenthal, M. B., Colla, C. H., & Sequist, T. D. (2019). Cascades of Care After Incidental Findings in a US National Survey of Physicians. JAMA network open, 2(10), e1913325. https://doi.org/10.1001/jamanetworkopen.2019.13325
  15. Ganguli, I., Ying, W., Shakley, T., Colbert, J. A., Mulligan, K. L., & Friedberg, M. W. (2023). Cascade Services and Spending Following Low-Value Imaging for Uncomplicated Low Back Pain among Commercially Insured Adults. Journal of general internal medicine, 38(4), 1102–1105. https://doi.org/10.1007/s11606-022-07829-2
  16. Goodrich, K. (2023, Nov. 29). Clinical Experience in Value Based Care [Powerpoint slides]. Humana. https://yale.instructure.com/courses/88306/files/folder/Slides?preview=8644831
  17. Hollingsworth, J. M., Miller, D. C., Daignault, S., & Hollenbeck, B. K. (2007). Five-year survival after surgical treatment for kidney cancer: a population-based competing risk analysis. Cancer, 109(9), 1763–1768. https://doi.org/10.1002/cncr.22600
  18. Hosny, A., Parmar, C., Quackenbush, J., Schwartz, L. H., & Aerts, H. J. W. L. (2018). Artificial intelligence in radiology. Nature reviews. Cancer, 18(8), 500–510. https://doi.org/10.1038/s41568-018-0016-5
  19. Kalavacherla, S., Riviere, P., Javier-DesLoges, J., Banegas, M. P., McKay, R. R., Murphy, J. D., & Rose, B. S. (2023). Low-Value Prostate-Specific Antigen Screening in Older Males. JAMA network open, 6(4), e237504. https://doi.org/10.1001/jamanetworkopen.2023.7504
  20. Kanzaria, H. K., Hoffman, J. R., Probst, M. A., Caloyeras, J. P., Berry, S. H., & Brook, R. H. (2015). Emergency physician perceptions of medically unnecessary advanced diagnostic imaging. Academic emergency medicine : official journal of the Society for Academic Emergency Medicine, 22(4), 390–398. https://doi.org/10.1111/acem.12625
  21. Keating, N. L., & Pace, L. E. (2018). Breast Cancer Screening in 2018: Time for Shared Decision Making. JAMA, 319(17), 1814–1815. https://doi.org/10.1001/jama.2018.3388
  22. Kwee, R. M., & Kwee, T. C. (2019). Whole-body MRI for preventive health screening: A systematic review of the literature. Journal of magnetic resonance imaging : JMRI, 50(5), 1489–1503. https://doi.org/10.1002/jmri.26736
  23. Levi, R., Gorenstein, D. (2022, June 14). When routine medical tests trigger a cascade of costly, Unnecessary Care. NPR. https://www.npr.org/sections/health-shots/2022/06/13/1104141886/cascade-of-care 
  24. Lumbreras, B., Donat, L., & Hernández-Aguado, I. (2010). Incidental findings in imaging diagnostic tests: a systematic review. The British journal of radiology, 83(988), 276–289. https://doi.org/10.1259/bjr/98067945
  25. Mulligan, K. L., Ganguli I., Rudin R. S.. (2022, April 27). Promoting Patient-Clinician Conversations to Reduce Cascades of Care. RAND. https://www.rand.org/pubs/commentary/2022/04/promoting-patient-clinician-conversations-to-reduce.html
  26. Niess, M. A., & Prochazka, A. (2014). Preoperative chest x-rays: a teachable moment. JAMA internal medicine, 174(1), 12. https://doi.org/10.1001/jamainternmed.2013.10531
  27. Oh, C. M., Lim, J., Jung, Y. S., Kim, Y., Jung, K. W., Hong, S., & Won, Y. J. (2021). Decreasing trends in thyroid cancer incidence in South Korea: What happened in South Korea?. Cancer medicine, 10(12), 4087–4096. https://doi.org/10.1002/cam4.3926
  28. Park, S., Jung, J., Burke, R. E., & Larson, E. B. (2021). Trends in Use of Low-Value Care in Traditional Fee-for-Service Medicare and Medicare Advantage. JAMA network open, 4(3), e211762. https://doi.org/10.1001/jamanetworkopen.2021.1762
  29. Patz, E. F., Jr, Goodman, P. C., & Bepler, G. (2000). Screening for lung cancer. The New England journal of medicine, 343(22), 1627–1633. https://doi.org/10.1056/NEJM200011303432208
  30. Pearl, R. (2017, September 25). Healthcare’s dangerous fee-for-service addiction. Forbes. https://www.forbes.com/sites/robertpearl/2017/09/25/fee-for-service-addiction/?sh=1293505c8adb 
  31. Pickering, A. N., Zhao, X., Sileanu, F. E., Lovelace, E. Z., Rose, L., Schwartz, A. L., Oakes, A. H., Hale, J. A., Schleiden, L. J., Gellad, W. F., Fine, M. J., Thorpe, C. T., & Radomski, T. R. (2023). Prevalence and Cost of Care Cascades Following Low-Value Preoperative Electrocardiogram and Chest Radiograph Within the Veterans Health Administration. Journal of general internal medicine, 38(2), 285–293. https://doi.org/10.1007/s11606-022-07561-x
  32. Rama, A. (2021). (rep.). Payment and Delivery in 2020: Fee-for-Service Revenue Remains Stable While Participation Shifts in Accountable Care Organizations During the Pandemic. Retrieved December 3, 2023, from https://www.ama-assn.org/system/files/2020-prp-payment-and-delivery.pdf. 
  33. Schwartz, A. L., Chernew, M. E., Landon, B. E., & McWilliams, J. M. (2015). Changes in Low-Value Services in Year 1 of the Medicare Pioneer Accountable Care Organization Program. JAMA internal medicine, 175(11), 1815–1825. https://doi.org/10.1001/jamainternmed.2015.4525
  34. Shrank, W. H., Rogstad, T. L., & Parekh, N. (2019). Waste in the US Health Care System: Estimated Costs and Potential for Savings. JAMA, 322(15), 1501–1509. https://doi.org/10.1001/jama.2019.13978
  35. Sutton, R. T., Pincock, D., Baumgart, D. C., Sadowski, D. C., Fedorak, R. N., & Kroeker, K. I. (2020). An overview of clinical decision support systems: benefits, risks, and strategies for success. NPJ digital medicine, 3, 17. https://doi.org/10.1038/s41746-020-0221-y
  36. Yu, F., Endo, M., Krishnan, R., Pan, I., Tsai, A., Reis, E. P., Fonseca, E. K., Lee, H. M., Abad, Z. S., Ng, A. Y., Langlotz, C. P., Venugopal, V. K., Rajpurkar, P. (2023). Evaluating progress in automatic chest X-ray radiology report generation. Patterns, 4(9), 100802. https://doi.org/10.1016/j.patter.2023.100802

Leave a comment