Background
The high costs of oncology care have refocused efforts to identify and reduce the processes of care that are considered of low value. In particular, ASCO recently participated in the American Board of Internal Medicine (ABIM) Foundation's Choosing Wisely initiative to identify a list of 5 health care processes that should prompt discussions between clinicians and patients regarding the value of that service.1
One of the measures in this initiative addressed low-value radiologic services in early-stage breast cancer (ESBC), stating, “Don't perform PET, CT, and radionuclide bone scans in the staging of early breast cancer at low risk for metastasis.”1 For these recommendations, the Choosing Wisely initiative defines early-stage disease as stage IIb or lower.
The NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) for Breast Cancer ask clinicians to “consider” advanced imaging for worrisome signs or symptoms of early-stage disease, similarly defined.2,3 These 2 recommendations differ in their wording, and may be confusing to clinicians who are weighing “consider” from the NCCN Guidelines with “do not” from ASCO's Choosing Wisely recommendations.
Caution regarding regular radiologic scanning of ESBC originates from an evidence base and subsequent consensus-based recommendations that highlight the limited efficacy of advanced imaging in this population. For example, little evidence shows that advanced imaging can improve long-term survival or detection of metastatic disease in asymptomatic patients.4 This was further demonstrated by a recent meta-analysis showing that preoperative MRI did not reduce long-term breast cancer recurrence.5 Furthermore, increased imaging can have inadequate specificity, leading to false-positives and unnecessary invasive procedures.6–8 For example, PET/CT has been shown to be less sensitive for the detection of axillary lymph node metastases than axillary lymph node dissection or even physical examination.9 Finally, advanced imaging is costly and increases radiation exposure.
However, no study has examined the intersection between guideline recommendations that aim to reduce low value services and oncologists' intuition that performing these services will provide valuable information that may alter clinical management. Although some investigators have begun to describe the incidence from registry data,10 registry data cannot adequately differentiate clinical intention and thus are insensitive to differentiating necessary from unnecessary tests. Few have closely evaluated the clinical charts to understand the clinical reasoning for why imaging was ordered. Why do oncologists order imaging in early-stage disease, and how often is this related to routine staging versus other reasons?
To answer these questions, we studied the frequency and clinical rationale for advanced imaging in patients with ESBC who had undergone diagnostic workup at a large tertiary academic cancer center. We aimed to understand the clinical utility of advanced imaging studies. Further, we sought to propose benchmarks for adherence to these guideline recommendations to form a foundation for comparison across our discipline.
Patients and Methods
Patient Selection
All patients seen at the Duke Cancer Institute with an ICD-9-CM code for a reportable breast tumor found within the tumor registry between January 1, 2010, and June 30, 2012, were included. This date range is mostly before the 2012 announcement of the ASCO Choosing Wisely measures. Patients with early-stage invasive breast cancer (≤ stage IIb) were included in the primary analysis and compared with an age-matched cohort of patients without cancer. The list of medical records was uploaded to the data warehouse tool DEDUCE for the purpose of exporting them with the selected Current Procedural Terminology (CPT) codes for CT of the brain, chest, abdomen, or pelvis; PET with or without CT; MRI of the brain; or nuclear bone scans that occurred within 60 days after diagnosis. When multiple qualifying CPT codes were found, the first code closest to the diagnosis date was selected for the analysis.
Chart Abstraction
To provide a robust sample for abstraction, we randomly selected 40% of charts for review. Two abstractors reviewed the medical charts of a sample of patients who were initially identified. Abstractions were performed by a medical oncology fellow (T.Z.) and a quality administrator with tumor registry experience (S.P.). Results of their abstractions were presented to an attending medical oncologist (A.K.), who was blinded to the identity and response of the reviewer, who also determined intent. Majority findings are reported.
Primarily, chart abstractions focused on the medical reason for ordering the imaging study in question, which providers ordered the imaging studies, and whether the results of imaging were positive for any abnormal findings. Results noted as positive were reviewed by a breast oncologist to determine whether the findings were clinically relevant and resulted in a change in clinical management. Examples of management change include significant additional follow-up tests or further biopsy procedures that significantly delayed or altered the treatment plan, or modification of the treatment regimen (eg, chemotherapy, hormonal therapy) altogether.
When multiple images were ordered on the same date, the abstracters reviewed the overall intent of the portfolio of images ordered to determine appropriateness.
Comparison Versus Controls
To obtain a baseline rate of the performance of advanced imaging in noncancer populations, we compared our findings with 2 age-matched control populations, which were patients without a primary diagnosis of any malignancy evaluated at Duke University. These time periods were randomly selected from the time period of the study. The first population was from June 1 to July 30, 2011; the second from December 1, 2011, to January 31, 2012. This comparison was performed to account for baseline rates of imaging performed in general populations for noncancer indications, such as infections or trauma.
Statistical Analysis
Basic descriptive statistics were calculated for this study. For the comparison of frequency of image ordering between patients with cancer and the noncancer control population, a chi-square test was performed. Univariate logistic regression was used to determine patient and disease characteristics associated with the ordering of imaging. For determination of the attribute agreement between the 2 chart abstractors, a κ statistic was calculated. All analyses were performed using JPM Statistical Software (SAS Institute, Cary, NC).
Results
A total of 1,143 patients with ESBC were evaluated. Most patients were white, older than 50 years, and had private insurance as the primary payer. A total of 250 patients (21.8%) had at least one advanced imaging study (including chest CT, abdomen/pelvis CT, brain CT, brain MRI, PET/CT, and nuclear bone scan) performed in the first 60-day period after diagnosis. Of those who received an advanced imaging study, 58 (23%) had 1, 88 (35%) had 2, and 104 (42%) had 3 or more studies performed. Patients who underwent advanced imaging were more likely to be hormone receptor–negative, triple-negative, and younger (age <50 years), and have higher stage disease (stage IIb vs ≤ stage IIa; all P<.001) (Table 1).
Two age-matched control cohorts of 1,142 patients without cancer, all seen at Duke University Hospital, were selected. Of the patients without cancer in the June to July cohort, 47 advanced imaging studies were performed on 42 patients (3.7%); 5 patients had 2 imaging studies each. Of the patients without cancer selected in the December to January cohort, 51 procedures were performed on 36 patients (3.2%); 11 of these patients had multiple imaging studies, including one patient who had 4 imaging studies ordered during this 60-day period.
Characteristics of Patients With Early-Stage Breast Cancer
Comparing patients in our analysis to noncancer controls, a statistically significant higher frequency of advanced imaging was seen between the cancer and noncancer cohorts (21.8% vs 3.7%; P<.0001). There was no seasonal difference in the frequency of advanced imaging when comparing patients with and without cancer in either of the 2 noncancer cohorts.
Manual chart abstraction was performed on 95 of the 250 patients (38%) who had advanced imaging studies. Clinical reasoning for obtaining the scans was abstracted (Figure 1). Routine staging without documentation of any worrisome findings was the reason for radiologic images being ordering for 59 of these patients (62%). A total of 20 patients (22%) had symptoms or concerning physical examination findings prompting the imaging study, and 16 other patients (17%) had a concurrent disease (eg, history of ovarian cancer, sarcoidosis with chest involvement) warranting the imaging study. Of the 59 patients who had scans obtained for routine staging purposes, 9 (15%) had a significant abnormal finding related to the underlying cancer, which resulted in a change in clinical management. Combined, the patients who had a clinical reason for their imaging study and those who had abnormal results found on staging scans accounted for 45% of the abstracted cases.
The manual chart abstraction also captured the type of clinical provider who ordered the advanced imaging studies (Figure 2). Of the 95 cases, advanced imaging was ordered by the medical oncologist in 55 (58%), a surgical oncologist in 27 (28%), an emergency physician in 8 (8%), and a radiation oncologist in 1 (1%), and 4 scans (4%) were ordered by other types of physicians.
Of the 95 reviewed studies, there were discrepancies in the abstractor-determined reasons for ordering the scan in 35 cases (36.8%), for a κ statistic of 0.26. On review by a breast oncologist, 24 of the cases included clinical nuances between typical and atypical findings in ESBC not easily adjudicated through administrative data review. Sample discrepancies
Clinical reasoning for ordering advanced imaging.
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 14, 8; 10.6004/jnccn.2016.0106
Clinical reasoning for ordering advanced imaging.
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 14, 8; 10.6004/jnccn.2016.0106
Clinical reasoning for ordering advanced imaging.
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 14, 8; 10.6004/jnccn.2016.0106
Discussion
We found that approximately one-fifth of patients with ESBC received an advanced radiologic image within 60 days of diagnosis. Imaging was more often performed in patients who were younger, had triple-negative cancer subtype, or had higher-stage disease (eg, stage IIb). Most imaging studies were ordered by medical oncologists and were not part of an extended preoperative evaluation by surgical oncologists.
These findings further add to a growing body of evidence regarding incidence of advanced imaging in ESBC that is inconsistent. For example, Simos et al11 recently reported a rate as high as 85% for advanced imaging based on a single-institution experience of 200 patients.11 The same authors performed a larger, province-wide registry analysis of patients in Ontario, finding a rate of 89%.12 Notably, these studies included patients with stage III breast cancer, demonstrating a similar odds ratio for more advanced disease stage, for which imaging is recommended, compared with stage II. Hahn et al13 recently
Medical providers who ordered advanced imaging.
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 14, 8; 10.6004/jnccn.2016.0106
Medical providers who ordered advanced imaging.
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 14, 8; 10.6004/jnccn.2016.0106
Medical providers who ordered advanced imaging.
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 14, 8; 10.6004/jnccn.2016.0106
Both the ASCO Choosing Wisely initiative and NCCN Guidelines aim to maximize high-value care while reducing (but not necessarily eliminating) low value services that do not improve outcomes important to patients, caregivers, or oncology professionals. Practice guidelines in general naturally spark conversations among clinical decision-makers; Choosing Wisely is one of the first large-scale initiatives to engage patients within these discussions. Others have previously demonstrated that oncology clinicians are familiar with and generally agree with the recommendation against advanced imaging in all patients with ESBC, yet 95% believe that exceptions should be made for suspicious history or clinical examination findings.14 Further, most patients with breast cancer prefer to have radiologic imaging performed, with 70% indicating in a survey that they would feel “uncomfortable” if their physician did not order imaging.15 When assessing the use of imaging studies, these factors should be weighed when determining value.
Through a careful analysis of associated clinical documentation, our data support that performing imaging in patients with ESBC occasionally provides clinical utility, and highlight the important limitations of evaluations of quality of oncology care that incorporate administrative data only. For example, without the manual chart abstraction, and relying on the tumor registry and electronic health record information alone, we would have overestimated the frequency of unnecessary imaging. Importantly, review by oncology clinicians reduced the categorization of advanced imaging as “for staging alone” from 21.8% to 12.1%, signifying that nuances regarding “intention” are often not accurately reflected in structured data sets. Similarly, Ryoo et al16 recently reported on the Veterans Health Administration experience, drawing conclusions regarding the quality of processes of care, noting significant limitations in electronic health data in reflecting patient preferences and choices.
There are several limitations to this study. First, we report the experience of a single tertiary care academic cancer center. Second, our analysis searched only the administrative clinical databases of our main medical center, and not those of community-based hospitals in our region; patients who had imaging performed at other regional hospitals or clinics were not captured in our search. Third, our analysis relied on reviewing the documentation of busy oncology clinicians. We suspect that justifications for imaging, such as localized pain with concerns for metastatic disease, may not have been routinely documented. Thus, we may overestimate the frequency of unwarranted imaging. Although incomplete documentation is a reality of busy clinical practice, it is important to remember that payers, regulators, and accreditors often rely solely on the written word of clinical notes to make decisions regarding appropriateness of care. Fourth, having reviewers determine whether results of advanced imaging led to clinically meaningful changes is inherently subjective, and reflective of the interpretation of clinical documentation that is reviewed.
Published clinical practice guidelines and recommendations serve to heighten clinician awareness about value-based care, limit unneeded studies or interventions, and serve as a platform to communicate the rationale of such decisions with patients. However, clinical medicine is a complex endeavor. Our data examine how clinician judgement can influence obtaining imaging studies for ESBC, and support allowing clinician flexibility for ordering advanced imaging in ESBC. As we move into an era of more efficient and quality-focused care delivery models, we must be careful to develop benchmarks that do not excessively proscribe advanced imaging in this setting.
Acknowledgments
The authors would like to acknowledge the Duke Cancer Institute for providing support for this study. Technical editing was provided by Dr. Donald Kirkendall, LLS, through the Center for Learning Health Care at the Duke Clinical Research Institute.
Dr. Kamal has disclosed that he receives consulting fees from Pfizer, Inc. and INSYS Therapeutics, and grant support from AHRQ, Michigan Oncology Quality Consortium, Retirement Research Foundation, California HealthCare Foundation, and Cambia Healthcare Foundation. Dr. Zhang has disclosed that he receives research support from Janssen Research & Development. The remaining authors have disclosed that they have no financial interests, arrangements, affiliations, or commercial interests with the manufacturers of any products discussed in this article or their competitors.
References
- 1.↑
Schnipper LE, Smith TJ, Raghavan D et al.. American Society of Clinical Oncology identifies five key opportunities to improve care and reduce costs: the top five list for oncology. J Clin Oncol 2012;30:1715–2724.
- 2.↑
Theriault RL, Carlson RW, Allred C et al.. Breast cancer, version 3.2013: featured updates to the NCCN guidelines. J Natl Compr Canc Netw 2013;11:753–760; quiz 761.
- 3.↑
Zagouri F, Liakou P, Bartsch R et al.. Discrepancies between ESMO and NCCN breast cancer guidelines: an appraisal. Breast 2015;24:513–523.
- 4.↑
Barry MC, Thornton F, Murphy M et al.. The value of metastatic screening in early primary breast cancer. Ir J Med Sci 1999;168:248–250.
- 5.↑
Houssami N, Turner R, Macaskill P et al.. An individual person data meta-analysis of preoperative magnetic resonance imaging and breast cancer recurrence. J Clin Oncol 2014;32:392–401.
- 6.↑
Puglisi F, Follador A, Minisini AM et al.. Baseline staging tests after a new diagnosis of breast cancer: further evidence of their limited indications. Ann Oncol 2005;16:263–266.
- 7.
Norum J, Andreassen T. Screening for metastatic disease in newly diagnosed breast cancer patients. What is cost-effective? Anticancer Res 2000;20:2193–2196.
- 8.↑
Fuster D, Duch J, Paredes P et al.. Preoperative staging of large primary breast cancer with [18F]fluorodeoxyglucose positron emission tomography/computed tomography compared with conventional imaging procedures. J Clin Oncol 2008;26:4746–4751.
- 9.↑
Wahl RL, Siegel BA, Coleman RE et al.. Prospective multicenter study of axillary nodal staging by positron emission tomography in breast cancer: a report of the staging breast cancer with PET Study Group. J Clin Oncol 2004;22:277–285.
- 10.↑
Crivello ML, Ruth K, Sigurdson ER et al.. Advanced imaging modalities in early stage breast cancer: preoperative use in the United States Medicare population. Ann Surg Oncol 2013;20:102–110.
- 11.↑
Simos D, Hutton B, Clemons M. Are physicians choosing wisely when imaging for distant metastases in women with operable breast cancer? [published online ahead of print November 12, 2014]. J Oncol Pract, pii: JOP.2014.000125.
- 12.↑
Simos D, Catley C, vanWalraven C et al.. Are physicians “choosing wisely” when imaging for distant metastases in women with early stage (1 or 2) breast cancer? A population study [abstract]. Presented at the 37th Annual CTRC-AACR San Antonio Breast Cancer Symposium; December 9–13, 2014; San Antonio, TX. Abstract P1-10-01.
- 13.↑
Hahn EH, Tang T, Lee JS et al.. Use of imaging for staging of early breast cancer in two integrated health care systems [abstract]. J Clin Oncol 2014;32(Suppl):Abstract 184.
- 14.↑
Simos D, Hutton B, Graham ID et al.. Imaging for metastatic disease in patients with newly diagnosed breast cancer: are doctor's perceptions in keeping with the guidelines? J Eval Clin Pract 2015;21:67–73.
- 15.↑
Simos D, Hutton B, Graham ID et al.. Patient perceptions and expectations regarding imaging for metastatic disease in early stage breast cancer. Springerplus 2014;3:176.
- 16.↑
Ryoo JJ, Ordin DL, Antonio AL et al.. Patient preference and contraindications in measuring quality of care: what do administrative data miss? J Clin Oncol 2013;31:2716–2723.
