Values are weighted percentages. There were no missing data. For this analysis, 2 respondents who reported that they experienced no cascades were excluded. The figure shows the distribution of responses to a series of questions: “In the past year, how often did you experience an incidental finding for your patient that led to each of the following?” “Physician harm” included anxiety, frustration, and wasted time and effort. “Patient harm” included physical or psychological harm, treatment burden, disrupted social relationships or status, financial burden, dissatisfaction with care, and death.15
Values are weighted percentages. For this analysis, we excluded 2 respondents who reported that they experienced no cascades and 3 respondents for whom it was not possible to assign responses about their most recent cascades to the given categories (eg, the initial event was reported to be a procedure or a panel of unspecified tests). The figure shows the distribution of responses to a series of questions after the following statement: “Thinking back to the last time you experienced any cascade from an incidental finding for your patient….” This “Sankey” diagram shows the percentage of physicians who gave each possible response to questions about their most recently experienced cascade. The heights of the question response boxes are proportional to the percentage of physicians who chose each response (shown in parentheses after the response); the heights of the connecting lines are proportional to the percentage of physicians who then chose the subsequent response. This diagram visualizes the “flow” of the reported cascades from how they started (Who ordered the initial test?) to their outcomes (Did the cascade cause patient or physician harm?). Initial test “Other” includes cardiac, urine, stool, microbiology, and pathology tests.
eAppendix 1. Survey Instrument
eAppendix 2. Response Rate Calculation
eTable 1. Sample Weights
eTable 2. Selected Survey Responses Stratified by Attending vs Trainee Physician Status
eFigure. Reported Events and Outcomes of Cascades Following an Incidental Finding, Stratified by Attending vs Trainee Physician Status
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Ganguli I, Simpkin AL, Lupo C, et al. Cascades of Care After Incidental Findings in a US National Survey of Physicians. JAMA Netw Open. 2019;2(10):e1913325. doi:10.1001/jamanetworkopen.2019.13325
How common are cascades of care after incidental findings, what occurs during the cascades, and what are the physician-reported consequences?
Among 376 practicing US internists in this nationally representative survey study, most physicians reported that they had experienced cascades after incidental findings that did not lead to clinically important and intervenable outcomes yet caused harm to patients and themselves.
This study’s findings suggest that policy makers and health care leaders should address cascades after incidental findings as part of efforts to improve health care value and reduce physician burnout.
Incidental findings on screening and diagnostic tests are common and may prompt cascades of testing and treatment that are of uncertain value. No study to date has examined physician perceptions and experiences of these cascades nationally.
To estimate the national frequency and consequences of cascades of care after incidental findings using a national survey of US physicians.
Design, Setting, and Participants
Population-based survey study using data from a 44-item cross-sectional, online survey among 991 practicing US internists in a research panel representative of American College of Physicians national membership. The survey was emailed to panel members on January 22, 2019, and analysis was performed from March 11 to May 27, 2019.
Main Outcomes and Measures
Physician report of prior experiences with cascades, features of their most recently experienced cascade, and perception of potential interventions to limit the negative consequences of cascades.
This study achieved a 44.7% response rate (376 completed surveys) and weighted responses to be nationally representative. The mean (SE) age of respondents was 43.4 (0.7) years, and 60.4% of respondents were male. Almost all respondents (99.4%; percentages were weighted) reported experiencing cascades, including cascades with clinically important and intervenable outcomes (90.9%) and cascades with no such outcome (94.4%). Physicians reported cascades caused their patients psychological harm (68.4%), physical harm (15.6%), and financial burden (57.5%) and personally caused the physicians wasted time and effort (69.1%), frustration (52.5%), and anxiety (45.4%). When asked about their most recent cascade, 33.7% of 371 respondents reported the test revealing the incidental finding may not have been clinically appropriate. During this most recent cascade, physicians reported that guidelines for follow-up testing were not followed (8.1%) or did not exist to their knowledge (53.2%). To lessen the negative consequences of cascades, 62.8% of 376 respondents chose accessible guidelines and 44.6% chose decision aids as potential solutions.
Conclusions and Relevance
The survey findings indicate that almost all respondents had experienced cascades after incidental findings that did not lead to clinically meaningful outcomes yet caused harm to patients and themselves. Policy makers and health care leaders should address cascades after incidental findings as part of efforts to improve health care value and reduce physician burnout.
Incidental findings on screening and diagnostic tests are common and may trigger cascades of further testing and treatment that are of uncertain value.1-10 By some estimates, up to 52% of radiology and laboratory tests produce incidental findings,3,5,11-13 and these rates are likely to increase with gains in technology.5 In some cases, further evaluation of these findings may reveal a clinically important and intervenable discovery, such as an early-stage cancer first detected on chest radiography that would have caused death if left untreated. More often, subsequent evaluations may find nothing significant, such as an electrocardiogram anomaly triggering a stress test and cardiac catheterization that ultimately shows no cardiovascular disease.9,10,14 Such cascades of care come with substantial potential for harms9,14-16: patients may experience anxiety and additional treatment risks in addition to monetary costs and inconvenience,1,7,14,17,18 and physicians may be distressed, conflicted, or burdened by additional work.14,16,19,20
Although this phenomenon has been described anecdotally and in specific clinical contexts, no study to our knowledge has examined the national scope of incidental findings or the cascades that may follow. We surveyed US internists to understand physician perceptions and experiences of these cascades at a national level. We aimed to estimate the frequency of cascades of care after incidental findings, identify the contents and characteristics of these cascades, understand their perceived consequences for patients and physicians, and pinpoint ways to minimize any negative consequences.
We conducted a web-based national survey of US internists between January 22 and March 3, 2019, in collaboration with the American College of Physicians (ACP), and analysis was performed from March 11 to May 27, 2019. The study followed the American Association for Public Opinion Research (AAPOR) reporting guidelines and was approved by Partners Institutional Review Board, which also waived the need for participant informed consent.
We used the Internal Medicine Insider Research Panel, a nationally representative panel of ACP member physicians. The ACP Research Center used stratified random sampling to create the panel in 2011 and regularly adjusts the panel to represent its membership across multiple demographic characteristics. Panelists agree to participate in approximately 2 projects per month and are rewarded for survey completion with points redeemable for gift cards.
Within this panel, we identified 991 physicians (57.4% of the member panel) who were residents, fellows, internists, or geriatricians practicing outpatient medicine. We further screened for respondents who were active in medicine, specialized in general internal medicine or geriatrics (among attendings), and practiced primarily or entirely in the outpatient setting.
Using literature review and clinical knowledge, we developed a survey instrument that included items on physicians’ prior experiences with cascades in general, items focused on their most recently experienced cascades (used to describe a sample of cascades), and an item on interventions to limit the negative consequences of cascades (based on the 2013 Presidential Commission on the Study of Bioethical Issues report on incidental findings).5,15,21 We defined incidental findings as “actionable results that are unrelated to why one ordered the tests” and described a cascade stemming from such a finding as “additional medical care, such as telephone calls, office visits, further testing, and treatment.”
We included questions on physician characteristics, including age, race/ethnicity, US vs foreign medical training, trainee vs attending status, clinical site, practice setting (urban, suburban, or rural), time in direct patient care, and malpractice history. We used a validated single-item discomfort with uncertainty scale22 ranging from 1 to 6, with 6 signifying the greatest discomfort.21,23 We also used a validated 6-item cost-consciousness scale.24-26 For this scale, we reversed items with negative wording to ensure that a higher score meant greater cost-consciousness, then calculated summary scores by summing the responses (6-36, with 36 denoting the most cost-consciousness). We conducted in-person, 30-minute cognitive interviews with 10 physician experts using thinking-out-loud and retrospective probing approaches27 between November 30 and December 10, 2018. These responses were then incorporated into the final 44-item cross-sectional, online survey instrument (eAppendix 1 in the Supplement).
We obtained physician sex from the ACP master file. We correlated physicians’ self-reported practice zip codes with US Census regions as well as with area-level education and income using data from the US Census Bureau American Community Survey.
We distributed the survey via email to 991 panel members on January 22, 2019, for a 41-day period. We paid a $10 incentive for survey completion and sent 6 reminders to improve the response rate. Respondents were required to answer all questions in the survey, so there were no missing response items.
The response rate was calculated using the American Association for Public Opinion’s Research Response Rate 3 definition (eAppendix 2 in the Supplement).28 To mitigate nonresponse bias, the ACP provided final survey weights to approximate national ACP membership on the basis of variables for which respondents were significantly different from the overall population, including respondent age category and experience level (eTable 1 in the Supplement).29,30 For relevant survey responses, we stratified analyses by trainee vs attending physician status (eTable 2 in the Supplement). To assess physician characteristics associated with physician harm, we built a multivariable logistic regression model in which the outcome was report of any physician harm (anxiety, frustration, and wasted time and effort) from a cascade in the past year, and covariates (chosen based on clinical plausibility) included age, sex, US vs foreign medical school training, trainee vs attending status, practice setting, time in direct patient care, presence of prior medical malpractice lawsuit, discomfort with uncertainty, and cost-consciousness. Two-sided P < .05 was considered statistically significant.31 All analyses used weighting commands and were performed with Stata statistical software, version 14.2 (StataCorp LLC).
We obtained 443 completed surveys, of which 67 were deemed ineligible based on screening criteria. Our final sample included 376 completed surveys, for a response rate of 44.7% (eAppendix 2 in the Supplement). We weighted responses to be nationally representative. Before and after weighting, respondents were predominantly male (weighted 60.4%), worked in urban and suburban settings, and practiced medicine at least 75% of their time (Table 1). The weighted mean (SE) age of respondents was 43.4 (0.7) years. Of the almost one-fifth who reported personal experience of a medical malpractice lawsuit, 11.4% had been sued for missed follow-up of an incidental finding. Respondents scored a mean of 3.8 (95% CI, 3.6-3.9) on the discomfort with uncertainty scale (minimum of 1 and maximum of 6) and 24.9 (95% CI, 24.3-25.5) on the cost-consciousness scale (minimum of 6 and maximum of 36).
Almost all respondents (99.4%) reported that they had experienced cascades resulting from incidental findings as a physician. Many physicians had also experienced cascades when they themselves were the patient (39.3%) or the patient’s family member (54.1%). Physicians reported that cascades for their patients commonly included telephone calls with patients (21.7% reported them at least weekly), new noninvasive tests (16.1% at least weekly), and repeated tests (14.7% at least weekly) (Figure 1). In addition, most physicians had experienced a cascade for their patient that led to a new invasive test (77.2%), emergency department visit (54.8%), or hospitalization (50.6%).
Most physicians (90.9%) had experienced cascades with clinically important and intervenable outcomes, such as discovery of a chronic condition (64.7%), cancer (59.0%), or an acute medical problem (36.6%) as well as cascades with no such outcome (94.4%). When asked about the frequency of experiencing these cascades, physicians were more likely to report that they experienced ones with no clinically important or intervenable outcome (31.1%) than those with meaningful outcomes (14.8%) on at least a monthly basis (Figure 1).
Most physicians reported that cascades had caused their patients harm (86.7%), including psychological harm (68.4%), treatment burden (65.4%), financial burden (57.5%), dissatisfaction with care (27.6%), physical harm (15.6%), disrupted social relationships (8.7%), and death (0.2%). They reported that they personally experienced wasted time and effort (69.1%), frustration (52.5%), and anxiety (45.4%). More than two-thirds (68.9%) of all respondents reported experiencing at least 1 of these harms in the past year. Physicians working in rural areas and those who had greater discomfort with uncertainty were more likely to report experiencing at least 1 of these harms in the past year (Table 2).
One-third (33.5%) of physicians reported that their experience of a cascade that uncovered something clinically important and intervenable made them more likely in the future to order the initial test that revealed the incidental finding. Of those who experienced a cascade resulting in no clinically important and intervenable outcome, 31.4% reported that this experience made them less likely to order that same initial test in the future. Finally, of those who experienced a harmful cascade, 30.1% reported that this made them less likely to order that initial test in the future.
When asked about their most recent cascade, physicians were most likely to report that the initial test was performed in an outpatient setting (64.9%), was an imaging test (54.4%), and was done for diagnostic purposes (48.7%) (Figure 2). For these results (ie, features of most recent cascades), we restricted our sample to 371 respondents who reported experiencing cascades and for whom their most recent cascades could be assigned to relevant categories. One-third (33.7%) reported that the initial test may not have been clinically appropriate. When faced with the incidental finding, 47.3% reported that they were concerned about an undiagnosed chronic disease, 44.2% were concerned about a new cancer, and 13.0% were concerned about the possibility of an acute event. Of those who pursued further evaluation (unweighted n = 361), the most commonly reported reasons for doing so were because the finding seemed clinically important (59.1%), they were following practice or community norms (43.7%), they were concerned about being sued (30.8%), the patient asked for it (20.1%), or another physician advised it (19.5%) (responses were not mutually exclusive). Most physicians reported that guidelines to inform these evaluations of their most recent cascade either did not exist to their knowledge (53.2%) or were not followed (8.1%). They reported that they attempted to lessen the cascade’s consequences by talking with their patient (62.5%), a specialist (28.9%), or a generalist colleague (18.3%); they also reported consulting educational references (46.8%), guidelines (39.3%), or the primary literature (17.1%). Notably, most cascades that did not uncover anything clinically important or intervenable (86.5%), as well as most cascades that did (81.8%), were reported to cause patient or physician harm (Figure 2).
Physicians identified several potential options to address these cascades: 62.8% believed that accessible guidelines on how to manage incidental findings would help limit the negative consequences of cascades, 48.1% cited patient and clinician education on potential harms from unnecessary medical care as potentially beneficial, 44.6% identified decision aids (ie, shared decision-making tools), and 42.0% chose malpractice reform. Fewer physicians thought that patient cost-sharing (18.1%) or value-based payment models (16.2%) would help (Table 3).
When we stratified our results by trainee vs attending status, we found that attendings generally reported higher frequencies of cascade events than trainees (eFigure in the Supplement). Trainees were more likely than attendings to report consulting a generalist colleague (27.6% vs 9.9%) or reading guidelines (49.2% vs 30.5%) or the primary literature (24.6% vs 10.3%) to shorten or lessen the consequences of their most recent cascade (eTable 2 in the Supplement). Compared with attendings, more trainees believed that value-based payment models would help mitigate cascades (23.9% vs 9.4%), and fewer believed that malpractice reform would (36.6% vs 46.8%, a nonsignificant difference) (Table 3).
In this national survey of US internists, we found that almost all had experienced cascades after incidental findings. We quantified the scope and contents of these cascades. Most physicians reported that incidental findings frequently prompted telephone calls with patients and repeated tests; most physicians had also seen their patients undergo new invasive tests, emergency department visits, and hospitalizations after an incidental finding. Many reported that they had experienced cascades as patients themselves. Physicians reported experiencing cascades for their patients that led to no clinically important or intervenable outcome more often than those that led to meaningful outcomes. Most physicians reported that the cascades had caused psychological or physical harms to patients and to themselves.
This work suggests that cascades are widely prevalent and may seem inevitable—even if not clinically significant—once an incidental finding is discovered.4,9 Indeed, the term cascade was coined to convey a sequence of events set irrevocably into motion,8,9 and a qualitative study4 found that primary care physicians reported feeling “compelled but frustrated” to pursue the “quagmire” of costly follow-up evaluations for incidental findings that were unlikely to be significant. Therefore, a key intervention may be to avoid that initial test whenever possible.9,32 One-third of physicians in our survey reported that the initial test in their most recently experienced cascade may not have been clinically appropriate: harms are unlikely to be offset by any benefits from testing in such cases.2,10
Most physicians reported negative consequences from cascades, including anxiety, frustration, and wasted time and effort. This was particularly true for those physicians reporting higher levels of discomfort with uncertainty, consistent with prior work linking discomfort with uncertainty and measures of burnout and depression.33 At a time of great concern about physician burnout and its potential drivers,34,35 our findings add to research suggesting that clinician involvement with what they perceive as futile or nonbeneficial care is linked to burnout and its components.36-38 Professional burnout has numerous negative consequences, with profound implications for physician well-being,39 patient satisfaction and safety,40,41 and quality of health care delivery.42,43 Understanding factors associated with burnout is an important step to enhancing physician well-being, with downstream consequences for high-quality patient care. Physicians who reported negative repercussions from cascades were also more likely to work in rural settings. This may reflect the effect that working in social and supportive environments has on well-being and may suggest that interventions to increase relational connection and community could help.44
The frequency of perceived harms from cascades—even from cascades with clinically important and intervenable outcomes—also shows that we need better ways to navigate incidental findings once they are found. Physicians in our survey believed that point-of-care tools might mitigate cascade burden, in line with suggestions by the 2013 Presidential Commission for the Study of Bioethical Issues report on incidental findings.5 Specifically, most respondents pointed to evidence-based recommendations written into radiology and laboratory result reports.4,14,45,46 Recommendations that quantify the likelihood of various differential diagnoses and are stratified by patient phenotypes (eg, level of anxiety and low vs high disease risk factors), in particular, may allow internists to make more nuanced evidence-based decisions with their patients.4 Respondents also cited the need for accessible guidelines on managing incidental findings. Such guidelines are in short supply,5 although organizations, such as the American College of Radiology Incidental Findings Committee, are working to fill this gap.5,14 Unfortunately, there is no guarantee that guidelines will lead to fewer harms from cascades47; indeed, in some cases, guidelines may lead radiologists to recommend, and ordering physicians to pursue, follow-up more often or more intensively than they otherwise would.48 Therefore, all of these efforts require more robust cost-effectiveness, decision analysis, and outcome studies of incidental findings to inform thoughtful recommendations and guidelines.11
Finally, our results highlight the importance of engaging patients in these efforts (eg, talking to patients about the possibility of incidental findings, even before ordering a test). Recognizing the uncertainty inherent in many of these conversations, this also requires strategies to help physicians with this communication.49,50 Almost half of respondents cited patient and clinician education on potential harms from unnecessary medical care as a solution.5,9 If (or when) incidental findings arise, clinicians and patients could use decision aids incorporating factors like clinical significance and actionability of the finding, patient time, and cost to decide next steps.5,6,51-53 Clinicians might also use scripts to help frame an incidental finding and the need for evaluation. Although only 19.9% of physicians reported that they ordered follow-up testing because their patient asked for it, we expect patients to have a bigger role as they gain increased access to laboratory and radiology results through electronic portals. Early, open, and evidence-driven communication with patients may also reassure those physicians who pursue cascades not because they are clinically needed but out of obligation to community norms or fear of getting sued. Almost half of physicians reported that malpractice reform would help reduce cascades, although few believed that payment reforms like patient cost-sharing and value-based payment models would be effective solutions, in keeping with the notion that the decision to pursue cascades is not driven by financial incentives.54
This study has several limitations. We cannot determine if physicians would act in practice as they reported in hypothetical scenarios. We achieved a 44.7% response rate, raising the possibility of response bias. However, respondents were found to be different from nonrespondents on just 2 observable demographic characteristics (experience level and age category), by which we weighted our results. We also acknowledge the possibility of recall bias. For example, physicians may better remember emotionally salient events, leading to overestimation of cascades resulting in clinically important and intervenable outcomes or harm and underestimation of cascades resulting in no clinically important or intervenable outcome. Future work might examine the role of the health literacy, patient-clinician relationships, and patient and clinician anxiety.
The survey findings indicated high national rates of internists experiencing cascades of care after incidental findings that did not lead to clinically important outcomes yet caused harm to patients and physicians. Policy makers and health care leaders can address cascades after incidental findings in their efforts to improve the value of health care and to reduce physician burnout. Initiatives could include the development and effective dissemination of point-of-care guidelines and shared decision-making tools, along with other strategies to embrace and communicate uncertainty.
Accepted for Publication: August 25, 2019.
Published: October 16, 2019. doi:10.1001/jamanetworkopen.2019.13325
Correction: This article was corrected on November 8, 2019, to correct the number of participants given in the Findings portion of the Key Points.
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Ganguli I et al. JAMA Network Open.
Corresponding Author: Ishani Ganguli, MD, MPH, Division of General Internal Medicine and Primary Care, Department of Medicine, Brigham and Women’s Hospital, 1620 Tremont St, Third Floor, Boston, MA 02120 (email@example.com).
Author Contributions: Dr Ganguli had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Ganguli, Simpkin, Lupo, Weissman, Rosenthal, Colla, Sequist.
Acquisition, analysis, or interpretation of data: Ganguli, Simpkin, Lupo, Mainor, Orav, Colla, Sequist.
Drafting of the manuscript: Ganguli, Simpkin, Lupo.
Critical revision of the manuscript for important intellectual content: Ganguli, Simpkin, Weissman, Mainor, Orav, Rosenthal, Colla, Sequist.
Statistical analysis: Ganguli, Simpkin, Lupo, Orav, Colla.
Obtained funding: Colla, Sequist.
Administrative, technical, or material support: Lupo, Weissman, Mainor, Colla, Sequist.
Conflict of Interest Disclosures: Dr Ganguli reported receiving grants from the Agency for Healthcare Research and Quality (AHRQ) and International Business Machines (IBM) Corporation and receiving personal consultancy fees from Haven. Ms Lupo, Mr Mainor, and Drs Rosenthal, Colla, and Sequist reported receiving grants from AHRQ. No other disclosures were reported.
Funding/Support: This study was funded by grant 1R01HS023812 from AHRQ.
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: The following expert physicians contributed to the development of the survey instrument: David W. Bates, MD, MSc, Asaf Bitton, MD, MPH, Aaron F. Mann, MD, Stuart M. Pollack, MD, Katherine D. Rose, MD, and Gordon D. Schiff, MD (all of Brigham and Women’s Hospital, Boston, Massachusetts); Mark W. Dickinson, MD, MPH, Susan R. Hata, MD, and Joshua C. Ziperstein, MD (all of Massachusetts General Hospital, Boston, Massachusetts); and Paul Simpkin, MBBS (of King Edward VII’s Hospital, London, United Kingdom). They were not compensated for their contributions.