Abstract
Purpose:
To retrospectively analyze the nature and extent of oncology-related errors accounting for malpractice allegations in diagnostic radiology.
Methods:
The Comparative Benchmarking System of the Controlled Risk Insurance Company, a database containing roughly 30% of medical malpractice claims in the United States, was searched retrospectively for the period 2008–2017. Claims naming radiology as a primary service were identified and were stratified and compared by oncologic vs. non-oncologic status, allegation type (diagnostic vs. non-diagnostic) and imaging modality.
Results:
Over the 10-year period, radiology was the primary responsible service for 3.9% of all malpractice claims (2,582/66,061) and 12.8% of claims with diagnostic allegations (1,756/13,695). Oncology (neoplasms) accounted for 44.0% of radiology cases with diagnostic allegations, a larger share than any other category of medical condition. Among radiology cases with diagnostic allegations, high-severity harm occurred in 79% of oncologic but just 42% of non-oncologic cases. Of all oncologic radiology cases, 97.4% had diagnostic allegations, while just 55.0% of non-oncologic radiology cases had diagnostic allegations. Imaging misinterpretation was a contributing factor for a large majority (80.7% [623/772]) of oncologic radiology cases with diagnostic allegations. The modalities most commonly used in oncologic radiology cases with diagnostic allegations involving misinterpretation were mammography, CT, and MRI.
Conclusion:
Oncology represents the most common source of radiology malpractice cases with diagnostic allegations. Oncologic radiology malpractice cases are more likely than non-oncologic radiology cases to be due to diagnostic errors and be associated with high-severity harm. Efforts are warranted to reduce misinterpretations of oncologic imaging.
Summary Sentence:
Over a ten-year period, oncology (i.e., neoplasms) accounted for nearly half of radiology cases with diagnostic allegations, making it by far the most common category of such cases.
Introduction
Medical errors have long been recognized as an unacceptably frequent cause of preventable harm in the United States, leading to considerable morbidity, mortality and economic impact (1, 2). A 2015 Institute of Medicine (IOM) report highlighted diagnostic errors as an increasingly critical component of medical errors, given the growing complexity of the diagnostic process and multiple opportunities for failure (3). While the exact scope of diagnostic errors is difficult to estimate, the IOM study concluded that most people will experience at least one diagnostic error in their lifetime (3). Estimates of the numbers of individuals who suffer serious, misdiagnosis-related harms in the United States annually range from 40,000 to as high as 4,000,000 (2).
Though not all medical errors lead to malpractice claims, such claims provide a useful proxy for studying trends in medical errors. The Comparative Benchmarking System (CBS) of the Controlled Risk Insurance Company (CRICO) is a large, prospectively curated database of well-characterized, medical malpractice claims from captive and commercial malpractice insurers across the United States (2). It contains roughly 30% of all malpractice claims in the country, based on a comparison with paid claims contained in the National Practitioner Data Bank (NPDB) (4). An analysis of closed claims from the CBS for the years 2006–2015 found that cancers accounted for the greatest share of high-severity cases involving diagnostic error (37.8%), followed by vascular events (22.8% ) and infections (13.5%) (2). Furthermore, in an analysis of claims in the CBS for the years 2010–2014 that named radiology as the primary responsible service, 60% concerned diagnosis, and in cases involving misinterpretation of a diagnostic study, primary malignancies and/or metastases were the most commonly missed diagnoses (5).
Additional insights into the extent and nature of errors involving diagnostic oncologic imaging could be useful for raising awareness among radiologists, referring physicians, patients, and other stakeholders and directing appropriate actions to reduce such errors. Therefore, we conducted this study to analyze oncology-related errors in diagnostic radiology as reflected in medical malpractice claims in the CRICO CBS database.
Methods
This retrospective study did not represent human subjects research and thus did not require review by an Institutional Review Board.
The Comparative Benchmarking System (CBS, described above) is a national database of medical malpractice cases maintained for the purposes of research into medical error; it is supported by a partnership of U.S. captive and commercial insurers led by one such insurer, CRICO Strategies (2, 5). The database contains malpractice claims for over 500 healthcare entities and 185,000 healthcare providers, from both academic and community settings (2, 5). A claim represents an allegation of medical malpractice with a request for payment for damages. The database is organized at the level of individual medical events associated with a single patient outcome; thus, multiple claims with different defendants relating to the same underlying event are captured as a single event, or case, in the database. A team of clinical specialists, using a highly structured taxonomy, enters comprehensive information about each case into the database, based on a rigorous review of medical and legal records, in order to provide far more complete information relating to malpractice cases than is available in other national claims databases (2, 5). The taxonomy provides for coding cases to multiple main categories (e.g., Final Diagnosis, Allegation, Service), with various levels of sub-category detail, such that “like cases” can be analyzed collectively.
The database was searched over a 10-year period from 2008 to 2017 that included a total of 66,061 deeply coded medical malpractice cases. For the purposes of classifying allegations as diagnostic, the CBS utilizes the following definition of diagnostic errors provided by the Institute of Medicine Committee on Diagnostic Error in Healthcare: “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” (3). The primary responsible service was determined for all allegations and for all diagnostic allegations.
The percentages of all cases, and of all cases with diagnostic allegations, for which radiology was the primary responsible service (also referred to henceforth as “radiology cases” and “radiology cases with diagnostic allegations,” respectively) were computed. Interventional radiology (IR) and nuclear medicine cases were upfront not counted when radiology cases were identified from all cases. The final diagnoses of radiology cases with diagnostic allegations were extracted and classified as oncologic (i.e., concerning neoplasms, both malignant and benign) vs. non-oncologic. The most common diagnoses in each of the two groups were determined. Oncologic and non-oncologic radiology cases with diagnostic allegations were compared in terms of severity of harm (low, medium, and high) and incidence of death. The former was determined using the severity of harm scale of the National Association of Insurance Commissioners (NAIC) (6).
The percentages of cases with oncologic final diagnoses were determined from among all radiology cases and all radiology cases with diagnostic allegations. The percentages of oncologic and non-oncologic radiology cases designated in the database as diagnostic were compared. In addition, oncologic and non-oncologic radiology cases with diagnostic allegations involving misinterpretations were compared in terms of the associated imaging modalities. These assessments were performed in terms of both individual claims and dollar amounts when applicable. The percentage of radiology cases with diagnostic allegations that involved oncology was determined for each year of the analysis. The descriptive analyses were performed based on counts and percentages using Microsoft Excel for Windows (Microsoft Corporation, Redmond, WA). Statistical comparisons were performed by chi-square tests using MedCalc for Windows (MedCalc Software bvba; Ostend, Belgium).
Results
Comparison of diagnostic vs. non-diagnostic allegations in radiology
Of 66,061 coded cases in the CBS database for the 10-year period studied, radiology was the primary responsible service for 3.9% (2,582), and these cases represented 4.3% ($511,758,882/$11,951,434,718) of total paid amounts. However, of a total of 13,695 cases with diagnostic allegations, radiology was the primary responsible service for 12.8% (1,756) and 13.0% ($435,800,510/$3,351,585,949) of total paid amounts. Among all 2,582 cases for which radiology was the primary responsible service, 30.7% (793) were oncologic. However, among the 1,756 cases with diagnostic allegations for which radiology was the primary responsible service, oncology (i.e., neoplasms) accounted for 44% (772)--a larger share than any other category of medical condition. Among the remaining 56.0% (984) of radiology cases with diagnostic allegations, the most common categories were injury/poisoning, circulatory system disorder, and digestive system disorder. Furthermore, 50.1% ($218,477,886/435,800,510) of total paid dollars for radiology cases with diagnostic allegations related to oncology, followed by 13.9% ($60,727,549) for circulatory system disorder, 13.3% ($58,059,198) for injury/poisoning, and 5.7% ($25,013,095) for digestive system disorder. Finally, 37.8% (140/470) of all radiology cases with diagnostic allegations involving CT or MRI related to oncology.
Diagnostic allegations represented 97.4% (772/793) of all oncologic cases with radiology as the primary service vs. 55.0% (984/1,789) of all non-oncologic cases with radiology as the primary service (p<0.001) (Table 1).
Table 1.
Comparison of oncologic and non-oncologic radiology malpractice cases that had diagnostic allegations for 2008–2017
| Oncologic | Non-oncologic | |
|---|---|---|
| Percentage of allegations categorized as diagnostic | 97.4% (772/793) | 55.0% (984/1,789) |
| Severity of harm associated with diagnostic allegations | ||
| High | 79.0% (610) | 42.0% (413) |
| Medium | 19.6% (151) | 52.9% (521) |
| Low | 1.4% (11) | 5.1% (50) |
| Percentage of all paid dollars for diagnostic cases associated with high-severity harm | 90.1% ($196,808,369/$218,477,886) |
80.6% ($175,227,540/$217,332,624) |
| Percentage of diagnostic allegations associated with death | 24.5% (189/772) |
20.5% (202/984) |
Comparison of Oncologic vs. Non-Oncologic Radiology Diagnostic Allegations
Among the 772 oncologic and 984 non-oncologic radiology cases with diagnostic allegations, 79.0% (610) and 42.0% (413), respectively, related to high-severity injuries, 19.6% (151) and 52.9% (521), respectively, related to medium-severity injuries, and 1.4% (11) and 5.1% (50), respectively, related to low-severity injuries (p<0.001; Figure 1). The percentage of all paid dollars represented by high-severity injuries was 90.1% ($196,808,369/$218,477,886) for oncologic radiology cases with diagnostic allegations vs. 80.6% ($175,227,540/$217,332,624) for non-oncologic radiology cases with diagnostic allegations. The percentage associated with death was 24.5% (189/772) for oncologic radiology cases with diagnostic allegations vs. 20.5% (202/984) for non-oncologic radiology cases with diagnostic allegations (p=0.048).
Figure 1.
Comparison of distributions of severity of harm associated with oncologic vs. non-oncologic radiology diagnostic allegations
Misinterpretation of an imaging study was a contributing factor for 80.7% (623/772) of oncologic and 81.3% (800/984) of non-oncologic radiology cases with diagnostic allegations. Among the 623 oncologic radiology cases with diagnostic allegations concerning misinterpretation, the most common final diagnoses for the patients involved were breast cancer (42.9%; 267), lung cancer (16.7%; 104), benign neoplasms (6.3%; 39), and gastrointestinal tract cancers (6.1%; 38). A total of 1,095 radiology cases with diagnostic allegations involving misinterpretation were classified as relating to the following five modalities: CT (28.0%), radiography (26.1%, 286), mammography (19.1%, 209), ultrasound (13.7%, 150), and MRI (13.1%, 143). Table 2 shows the percentages of these cases that were oncologic or non-oncologic and their distribution by modality. Of the 511 that were oncologic, the largest shares related to mammography (40.3%) and CT (21.9%) (Table 2). Of those that were non-oncologic, the largest shares related to radiography (36.8%) and CT (33.4%) (195) (Table 2).
Table 2.
Comparison of oncologic and non-oncologic radiology malpractice cases with diagnostic allegations involving misinterpretation of imaging, among those that could be classified as relating to five primary radiology modalities
| All cases (n=1,095) | Oncologic cases (n=511) | Non-oncologic cases (n=584) | |
|---|---|---|---|
| CT | 28.0% (307) | 21.9% (112) | 33.4% (195) |
| X-Ray | 26.1% (286) | 13.9% (71) | 36.8% (215) |
| Mammography | 19.1% (209) | 40.3% (206) | 0.5% (3) |
| Ultrasound | 13.7% (150) | 12.5% (64) | 14.7% (86) |
| MRI | 13.1% (143) | 11.4% (58) | 14.6% (85) |
The percentage of radiology cases with diagnostic allegations that were deemed oncologic in 2008 (47.8% [43/90]) was nearly identical to that in 2017 (47.7% [103/216], Table 3). Variation in this percentage across the intervening years was not statistically significant (p=0.218).
Discussion
Using a large medical malpractice claims database in the United States, we showed that over a ten-year period, oncology (i.e., neoplasms) accounted for nearly half of radiology cases with diagnostic allegations, making it the most common category of such cases. Furthermore, oncology claims accounted for half of all paid dollars for radiology cases with diagnostic allegations, with no other categories accounting for more than 20%. Injuries associated with radiology cases that had diagnostic allegations were approximately twice as likely to be of high severity when they concerned oncologic as opposed to non-oncologic conditions (of the oncologic subset, approximately 80% were of high severity and roughly a quarter led to death). While radiology has previously been recognized as contributing to errors in cancer diagnosis (7), our findings indicate that oncology is a key contributor to radiology diagnostic errors overall, in terms of both frequency and impact.
The fact that 44.0% of radiology cases with diagnostic allegations were oncologic is particularly worrisome given that only 3.9% of all radiologic imaging nationally is oncologic, according to a recent estimate (8). Thus, the high proportion of radiology errors that are oncologic is not attributable to a high proportion of radiologic imaging being oncologic; rather, oncologic imaging appears to have a genuinely higher risk profile than non-oncologic imaging. Furthermore, the discrepancy cannot be attributed solely to the high-risk nature of breast imaging, given that 37.8% of all radiology cases with CT/MRI-based diagnostic allegations were oncologic—a percentage highly disproportionate to the estimated 9.5% of all advanced imaging (i.e., CT, MRI and PET/CT) that is oncologic nationally (8).
It is also striking that diagnostic failure accounted for nearly all oncologic radiology cases but only about half of non-oncologic radiology cases. Past work showed that CT was the most common modality associated with radiology claims (5), potentially relating to the very high volumes of CT imaging overall. Our study confirms that finding. However, our current analysis also demonstrates that among oncologic radiology cases with diagnostic allegations involving misinterpretation, specifically, the modality most commonly involved was mammography; this aligns with the observation that breast cancer was the most common final diagnosis associated with these cases as well. Indeed, considerable attention has been given to strategies for addressing error in the realm of breast imaging (9–11). Nonetheless, the majority of the oncologic radiology cases with diagnostic allegations involving misinterpretation related to a wide spectrum of other malignancies, indicating the risk of error inherent in any oncologic imaging.
Based on our findings, actions to address diagnostic error as it pertains to oncologic imaging are warranted. The large majority (80.7%) of oncologic radiology cases with diagnostic allegations related to a misinterpretation, and the fraction of radiology cases with diagnostic allegations that were due to oncology did not vary significantly over the 10-year study interval. Strategies to improve interpretation of oncologic imaging examinations could include increasing the availability of dedicated oncologic imaging fellowships, creating subspecialty certification pathways in oncologic imaging, centralizing interpretation through academic medical centers and major cancer centers in ways that leverage teleradiology solutions, forming second-opinion networks of regional expert oncologic imagers to support clinicians within community settings, establishing peer learning networks throughout the clinical community to provide feedback and lifelong learning opportunities, and integrating artificial intelligence-based applications into the routine clinical evaluation of oncologic imaging examinations (12–14). We believe these initiatives would promote consistently high-quality imaging services and safer, less error-prone care for oncology patients across all practice settings, including community settings, where a large volume of oncologic imaging is performed (8).
It is worth noting that a recent study found a positive association between paid medical malpractice claims and subsequent advanced imaging utilization for Medicare beneficiaries on a state-by-state level in the United States; specifically, the authors showed that each additional paid malpractice claim over the course of 5 years was associated with an average of 1,389 additional advanced imaging examinations among Medicare beneficiaries in the following year (15). They hypothesized that practicing in an environment with a rising level of medical malpractice litigation led physicians to order additional (and likely unnecessary) imaging studies as a defensive strategy. Thus, it is possible that reducing the relatively high level of oncology-related radiology malpractice claims could also help reduce the unnecessary use of imaging (15).
Limitations
This study has a number of limitations. First, the data were obtained from a consolidated national claims database. While clinical patient records were used at the time of initial data entry, the medical records were not directly reviewed at the time of performing this investigation. It is thus possible that, in spite of rigorous governance and audits, the information was incorrectly or incompletely coded. Also, it is unknown from our analysis whether the oncologic allegations related to an initial cancer diagnosis or a previously diagnosed cancer; such information could be useful to guide efforts to increase oncologic imaging expertise in the community. In addition, the utilized database does not represent all medical malpractice claims in the U.S., and medical malpractice claims do not represent all radiology errors that occur.
It should also be noted that our analysis did not include claims involving nuclear medicine or interventional radiology. A recent study based on federal and state legal cases from a 35-year period found that instances of complex litigation (including malpractice and other types of litigation) involving IR physicians were rare but increasing (16).
Figure 2.
Percentages of radiology malpractice cases with diagnostic allegations that were oncologic by year
Take-Home Points.
-Oncology accounts for the largest share of errors leading to allegations of diagnostic failure in radiology—a share greatly disproportionate to the share of imaging that is oncologic.
-Diagnostic errors in radiology are significantly more likely to be high-impact when they concern oncology as opposed to other conditions.
-Allegations of malpractice in radiology relate to diagnostic errors nearly always when they concern oncologic indications but only about half the time when they concern non-oncologic indications.
-Among oncologic radiology diagnostic errors, breast cancer continues to represent the most common missed or improperly diagnosed entity and mammography the most commonly involved modality; however, in relation to how often they are used, CT and MRI are disproportionately involved in radiology cases that concern oncology, which relate to a broad range of malignancies.
-Actions to increase the availability of oncologic imaging expertise are needed to reduce misinterpretations of oncologic diagnostic radiology examinations, thereby promoting safer, higher-quality patient care.
Funding Sources:
This work was supported in part by the Cancer Center Support Grant P30 CA008748 from the National Institutes of Health/National Cancer Institute to Memorial Sloan Kettering Cancer Center.
Disclosures:
Dr. Hricak reports service on the Board of Directors of Ion Beam Applications (IBA), for which she receives annual compensation, as well as uncompensated service on the External Advisory Board, Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins, on the International Advisory Board, University of Vienna, on the Board of Trustees and Scientific Committee, DKFZ (German Cancer Research Center), and on the Scientific Advisory Board, Euro-BioImaging, outside of the submitted work. Dr. Hricak and Ada Muellner both report funding from the National Cancer Institute to their institution, Memorial Sloan Kettering Cancer Center (also noted under “Funding Sources” above). Dr. Nass reports grant funding from Bristol-Myers Squibb, Flatiron Health, Merck, Novartis, Pfizer, and Sanofi, outside the submitted work. The remaining authors declare no conflicts of interest.
Footnotes
Statement of data access and integrity: The authors declare that they had full access to all of the data in this study and take complete responsibility for the integrity of the data and the accuracy of the data analysis.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Institute of Medicine. 2000. To Err Is Human: Building a Safer Health System. Washington, DC: The National Academies Press. 10.17226/9728. [DOI] [PubMed] [Google Scholar]
- 2.Newman-Toker DE, Schaffer AC, Yu-Moe CW, Nassery N, Saber Tehrani AS, Clemens GD, Wang Z, Zhu Y, Fanai M, Siegal D. Serious misdiagnosis-related harms in malpractice claims: The “Big Three” - vascular events, infections, and cancers. Diagnosis (Berl) 2019;6(3):227–240. doi: 10.1515/dx-2019-0019 [DOI] [PubMed] [Google Scholar]
- 3.National Academies of Sciences, Engineering, and Medicine. 2015. Improving Diagnosis in Health Care. Washington, DC: The National Academies Press. 10.17226/21794. [DOI] [Google Scholar]
- 4.Department of Health US and Human Services, Health Resources and Services Administration. NPDB Guidebook. Rockville, Maryland: U.S. Department of Health and Human Services; 2018. October. [Google Scholar]
- 5.Siegal D, Stratchko LM, DeRoo C. The role of radiology in diagnostic errors: a medical malpractice claims review. Diagnosis (Berl.) 2017;4:125–131. [DOI] [PubMed] [Google Scholar]
- 6.Guideline for Implementation of Medical Professional Liability Closed Claim Reporting (GDL-1077). National Association of Insurance Commissioners; 2010:p. 38. Available at: http://www.naic.org/store/free/GDL-1077.pdf. Accessed: 22 November 2020. [Google Scholar]
- 7.Singh H, Sethi S, Raber M, Petersen LA. Errors in cancer diagnosis: current understanding and future directions. J Clin Oncol 2007;25(31):5009–5018. doi: 10.1200/JCO.2007.13.2142 [DOI] [PubMed] [Google Scholar]
- 8.Rosenkrantz AB, Chaves Cerdas L, Hughes DR, Recht MP, Nass SJ, Hricak H. National Trends in Oncologic Diagnostic Imaging. J Am Coll Radiol 2020;17(9):1116–1122. doi: 10.1016/j.jacr.2020.06.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Regev GS, Ser AM. Breast cancer medical malpractice litigation in New York: The past 10 years. Breast 2019;46:1–3. doi: 10.1016/j.breast.2019.03.009 [DOI] [PubMed] [Google Scholar]
- 10.Arleo EK, Saleh M, Rosenblatt R. Lessons Learned From Reviewing Breast Imaging Malpractice Cases. J Am Coll Radiol 2016;13(11S):R58–R60. doi: 10.1016/j.jacr.2016.09.028 [DOI] [PubMed] [Google Scholar]
- 11.Fileni A, Magnavita N, Pescarini L. Analysis of malpractice claims in mammography: a complex issue. Radiol Med 2009;114(4):636–644. doi: 10.1007/s11547-009-0394-6 [DOI] [PubMed] [Google Scholar]
- 12.National Academies of Sciences, Engineering, and Medicine. 2018. Improving cancer diagnosis and care: Patient access to oncologic imaging and pathology expertise and technologies: Proceedings of a workshop. Washington, DC: The National Academies Press. doi: 10.17226/25163. [DOI] [PubMed] [Google Scholar]
- 13.Nass SJ, Cogle CR, Brink JA, Langlotz CP, Balogh EP, Muellner A, Siegal D, Schilsky RL, Hricak H. Improving Cancer Diagnosis and Care: Patient Access to Oncologic Imaging Expertise. J Clin Oncol 2019;37(20):1690–1694. doi: 10.1200/JCO.18.01970 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Nass SJ, Cohen MB, Nayar R, Zutter MM, Balogh EP, Schilsky RL, Hricak H, Elenitoba-Johnson KSJ. Improving Cancer Diagnosis and Care: Patient Access to High-Quality Oncologic Pathology. Oncologist 2019. doi: 10.1634/theoncologist.2019-0261 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Villalobos A, Horný M, Hughes DR, Duszak R Jr. Associations Over Time Between Paid Medical Malpractice Claims and Imaging Utilization in the United States. J Am Coll Radiol 2020;S1546–1440(20)30788–2. doi: 10.1016/j.jacr.2020.04.035. Online ahead of print. [DOI] [PubMed] [Google Scholar]
- 16.Branach C, Tharp K, Duszak R. Distribution and Characteristics of Malpractice and Nonmalpractice Litigation Involving Interventional Radiologists in the United States from 1983–2018. Curr Probl Diagn Radiol. 2020. Nov 15;S0363–0188(20)30207–3. doi: 10.1067/j.cpradiol.2020.11.003. Online ahead of print. [DOI] [PubMed] [Google Scholar]