differences in breast cancer survival by race

Copyright 2013 American Medical Association. All rights reserved.

Author Affiliations: University of Pittsburgh Graduate School of PublicHealth, Pittsburgh, Pennsylvania (Christian, King, Belle); National Instituteof Diabetes and Digestive and Kidney Diseases, Bethesda, Maryland(Yanovski); University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania(Courcoulas).

Corresponding Author: Nicholas J. Christian, PhD, University of PittsburghGraduate School of Public Health, 130 DeSoto St, 127 Parran Hall, Pittsburgh, PA15261 ([email protected]).

Published Online: November 4, 2013.doi:10.1001/jama.2013.281043.

Author Contributions: Dr Christian had full access to all of the data in the studyand takes responsibility for the integrity of the data and the accuracy of the dataanalysis.Study concept and design: Christian, Belle.Acquisition of data: Courcoulas.Analysis and interpretation of data: Christian, King, Yanovski, Courcoulas,Belle.Drafting of the manuscript: Christian.Critical revision of the manuscript for important intellectual content: Christian,King, Yanovski, Courcoulas, Belle.Statistical analysis: Christian, Belle.Obtained funding: Belle.Administrative, technical, or material support: Courcoulas, Belle.Study supervision: Belle.

Conflict of Interest Disclosures: The authors have completed and submittedthe ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Courcoulasreported receiving research grants from Allergan, Pfizer, Covidien, EndoGastricSolutions, and Nutrisystem; serving on a scientific advisory board ofEthicon J & J Healthcare system; and serving as a consultant to Ethicon J & JHealthcare system. No other disclosures were reported.

Funding/Support: Longitudinal Assessment of Bariatric Surgery-2 was a coop-erative agreement funded by the following grants from the National Institute ofDiabetes and Digestive and Kidney Diseases: U01 DK066557 (awarded to thedata coordinating center), U01-DK66667 (Columbia University Medical Center;in collaboration with Cornell University Medical Center grant UL1-RR024996),U01-DK66568 (University of Washington; in collaboration with grant M01RR-00037), U01-DK66471 (Neuropsychiatric Research Institute), U01-DK66526(East Carolina University), U01-DK66585 (University of Pittsburgh MedicalCenter; in collaboration with grant UL1-RR024153), and U01-DK66555 (OregonHealth & Science University).

Role of the Sponsor: Scientists from the National Institute of Diabetes andDigestive and Kidney Diseases (NIDDK) contributed to the design and conductof the study. The project scientist from the NIDDK served as a member of thesteering committee, along with the principal investigator from each clinical siteand the data coordinating center. The decision to publish was made by theLongitudinal Assessment of Bariatric Surgery steering committee, with norestrictions imposed by the sponsor. As a coauthor, an NIDDK scientistcontributed to the interpretation of the data and preparation, review, andapproval of the manuscript.

1. Connor Gorber S, Tremblay M, Moher D, Gorber B. A comparison of direct vsself-report measures for assessing height, weight and body mass index: asystematic review. Obes Rev. 2007;8(4):307-326.

2. Villanueva EV. The validity of self-reported weight in US adults: a populationbased cross-sectional study. BMC Public Health. 2001;1:11-20.

3. Merrill RM, Richardson JS. Validity of self-reported height, weight, and bodymass index: findings from the National Health and Nutrition ExaminationSurvey, 2001-2006. Prev Chronic Dis. 2009;6(4):A121.

4. Nyholm M, Gullberg B, Merlo J, Lundqvist-Persson C, Råstam L,Lindblad U. The validity of obesity based on self-reported weightand height: implications for population studies. Obesity (Silver Spring).2007;15(1):197-208.

5. White MA, Masheb RM, Burke-Martindale C, Rothschild B, Grilo CM.Accuracy of self-reported weight among bariatric surgery candidates: theinfluence of race and weight cycling. Obesity (Silver Spring). 2007;15(11):2761-2768.

6. Belle SH, Berk PD, Chapman WH, et al; for the LABS Consortium. Baselinecharacteristics of participants in the Longitudinal Assessment of BariatricSurgery-2 (LABS-2) study [published online March 7, 2013]. Surg Obes Relat Dis.doi:10.1016/j.soard.2013.01.023.

COMMENT & RESPONSE

Differences in Breast Cancer Survival by RaceTo the Editor Dr Silber and colleagues1 reported survival differ-ences among black and white women with breast cancer. How-ever, screening overdiagnosis may weaken the results of thisand prior studies on the topic.

For example, cases detected through screening mammog-raphy are sometimes diagnosed as breast cancer even thoughthey would never be noticed otherwise. These overdiagnosescan lead to unnecessary treatments.2 Recent studies haveprompted disagreement about the extent of overdiagnosis.3 De-spite this attention, the effect of overdiagnosis on survival dis-parities is unclear.

Here is a simplified example to explain how overdiagno-sis could alter survival. Silber and colleagues1 reported 5-yearsurvival of 55.9% for black patients and 59.5% for treatment-matched white patients, a 3.6% difference (Table 2 in article).For illustrative purposes, suppose 7.5% of black patients and15% of matched white patients were overdiagnosed withharmless tumors and did not die over the next 5 years. Sur-vival for true breast cancer is estimated by subtracting thepercentage of overdiagnosis from survival numerators anddenominators. Specifically, estimates would be (55.9−7.5)/(100−7.5) = 52.3% and (59.5−15)/(100−15) = 52.3% for blackand matched white patients respectively, a 0% disparity. Con-versely, supposing 15% of black patients and 7.5% of matchedwhite patients were overdiagnosed, the corrected survivaldifference would be 8.1%, which is more than twice the esti-mate by Silber and colleagues.

For clarity, we made several simplifications in this ex-ample. It ignores lead-time bias and assumes that overdiag-nosed patients have little risk of death. Yet with Bleyer andWelch4 reporting that 31% of breast cancers were overdiag-nosed as of 2008, we believe our example illustrates how es-timates of survival disparities are sensitive to the precise ex-tent of overdiagnosis. In the study by Silber and colleagues,1

it appears that unrecognized overdiagnosis could signifi-cantly alter the results.

Overdiagnosis bias is equally relevant to past studies ofbreast cancer survival disparities, including those based on ran-domized controlled trials. Indeed, the meticulous approachused by the authors appears more robust than previous stud-ies because matching tumor characteristics lessens differ-ences in screening, which could reduce bias from overdiag-nosis. Perhaps the authors could recreate Table 2 including onlypatients known not to have been screened recently before theirdiagnoses.

For 25 years, it has been reported that much of the differ-ence in breast cancer survival between black and white womenis unexplained, but a variety of studies have concluded thatthe disparity is largely unrelated to differences in treatments.1,5

Yet overdiagnosis may have obscured the real burden of in-equalities in care.

Charles Ford Harding, ABFrancesco Pompei, PhDRichard Wilson, DPhil

Letters

2456 JAMA December 11, 2013 Volume 310, Number 22 jama.com


Downloaded From: http://jama.jamanetwork.com/ by a UB-Universitaets Landesbibliothek Dusseldorf User on 12/18/2013


Author Affiliations: Independent data scientist (self-employed), Seattle,Washington (Harding); Department of Physics, Harvard University, Cambridge,Massachusetts (Pompei, Wilson).

Corresponding Author: Richard Wilson, DPhil, Harvard University Departmentof Physics, 17 Oxford St, Cambridge, MA 02138 ([email protected]).

Conflict of Interest Disclosures: The authors have completed and submittedthe ICMJE Form for Disclosure of Potential Conflicts of Interest. Mr Hardingreported receiving funding from Exergen Corp for work on this letter; previousconsulting for Janssen Pharmaceuticals, Johnson & Johnson Pharmaceuticals,Lifeline Screening, Novartis Pharmaceuticals, and Innovative Science Solutions;previous employment by Innovative Science Solutions; and receiving paymentfor manuscript editing from Cactus Global. Neither the consulting, employment,or editing pertained to the topics of this letter. Dr Pompei reported being CEOof Exergen Corp, a manufacturer of noninvasive thermometry devices. No otherdisclosures were reported.

1. Silber JH, Rosenbaum PR, Clark AS, et al. Characteristics associated withdifferences in survival among black and white women with breast cancer. JAMA.2013;310(4):389-397.

2. Esserman LJ, Thompson IM Jr, Reid B. Overdiagnosis and overtreatment incancer: an opportunity for improvement. JAMA. 2013;310(8):797-798.

3. Marmot MG, Altman DG, Cameron DA, Dewar JA, Thompson SG, Wilcox M.The benefits and harms of breast cancer screening: an independent review. Br JCancer. 2013;108(11):2205-2240.

4. Bleyer A, Welch HG. Effect of three decades of screening mammography onbreast-cancer incidence. N Engl J Med. 2012;367(21):1998-2005.

5. Albain KS, Unger JM, Crowley JJ, Coltman CA Jr, Hershman DL. Racialdisparities in cancer survival among randomized clinical trials patients of theSouthwest Oncology Group. J Natl Cancer Inst. 2009;101(14):984-992.

In Reply Our study examined all 7375 black women olderthan 65 years diagnosed with breast cancer between 1991and 2005 in the Surveillance, Epidemiology, and EndResults database. Although Mr Harding and colleaguesquote various numbers to illustrate the potential effect ofoverdiagnosis on racial disparities in breast cancer survival,these are hypothetical.

We matched patients for cancer stage and many other char-acteristics, in which stage was determined from the databasebased on chart review. That chart review included informa-tion from biopsy and postsurgical pathology. On this basis, pa-tients with ductal carcinoma in situ (DCIS) were excluded fromour data set. Although overdiagnosis may lead to unneces-sary biopsies or unneeded treatment of patients with DCIS, nei-ther of these possibilities are relevant to our study. Harding andcolleagues cite an Editorial by Esserman et al1 that referred tothe problematic diagnosis of DCIS when estimating cancer sur-vival and suggest that premalignant conditions “(eg, ductal car-cinoma in situ or high-grade prostatic intraepithelial neopla-sia) should not be labeled as cancers or neoplasia, nor shouldthe word ‘cancer’ be in the name.” Harding and colleagues sug-gest an additional analysis that takes account of breast cancerscreening.

As suggested, in each of our 3 matched black and white pa-tient comparisons, we adjusted for the indicator of mammo-graphic screening and obtained results qualitatively similar tothose reported in Table 2 of our article. Initially, black and whitepatients had different survival prospects, but the majority ofthis difference was removed by comparing black and whitewomen with similar cancers (eg, stage, size, grade, estrogen-receptor status) and similar comorbidities (eg, congestive heartfailure, diabetes). Black women received somewhat inferior

cancer treatment, but this explained only a small portion ofthe disparity in survival. These results are not changed by ad-justment for screening.

The adjustment used the Cox proportional hazardsmodel for paired survival data as used several times forother analyses in our study (eg, the adjustment for income).After adjustment for screening, the black-white hazard ratiowas 1.41 (95% CI, 1.32-1.50) in the demographic match,1.11 (95% CI, 1.04-1.18) in the presentation match, and 1.05(95% CI, 0.98-1.11) in the treatment match. Consistent withTable 2, there is a large initial disparity that is mostlyexplained by differences in presentation, not differences intreatment.

Discussions of our study by Harding and colleagues andothers2 have confused 2 different questions. First, does treat-ment matter for survival? Second, do disparities in treatmentexplain most of the disparity in survival? To explain the black-white disparity in survival following a diagnosis of breast can-cer, treatment would have to matter for survival and also besubstantially different for black and white patients with simi-lar disease.

Jeffrey H. Silber, MD, PhDPaul R. Rosenbaum, PhDKevin R. Fox, MD

Author Affiliations: Center for Outcomes Research, Children’s Hospital ofPhiladelphia, Philadelphia, Pennsylvania (Silber); Department of Statistics,University of Pennsylvania, Philadelphia (Rosenbaum); Leonard and MadlynAbramson Cancer Center, University of Pennsylvania, Philadelphia (Fox).

Corresponding Author: Jeffrey H. Silber, MD, PhD, Children’s Hospitalof Philadelphia, 3535 Market St, Ste 1029, Philadelphia, PA 19104([email protected]).

Conflict of Interest Disclosures: The authors have completed and submittedthe ICMJE Form for Disclosure of Potential Conflicts of Interest and none werereported.

1. Esserman LJ, Thompson IM Jr, Reid B. Overdiagnosis and overtreatment incancer: an opportunity for improvement. JAMA. 2013;310(8):797-798.

2. Mandelblatt JS, Sheppard VB, Neugut AI. Black-white differences in breastcancer outcomes among older Medicare beneficiaries: does systemic treatmentmatter? JAMA. 2013;310(4):376-377.

Therapy for Posttraumatic Stressand Alcohol DependenceTo the Editor In a randomized clinical trial, Dr Foa andcolleagues1 examined prolonged exposure psychotherapy forco-occurring posttraumatic stress disorder (PTSD) and alco-hol dependence. Prolonged exposure therapy was studied inrelation to supportive counseling, use of naltrexone, and a pillplacebo. The article provides an example of how the same find-ings can be interpreted quite differently when considered froma public health perspective.

Results of the trial were null for prolonged exposuretherapy; it did not show a main effect on substance use disor-der or on PTSD compared with supportive counseling. Yet theauthors concluded simply that prolonged exposure therapy didnot exacerbate substance use disorder. Thirty-five studies ofPTSD with substance use disorder, ranging from pilot studiesto multisite trials, have shown that treating PTSD in the con-text of substance use disorder does not worsen either disorder.2

Letters

jama.com JAMA December 11, 2013 Volume 310, Number 22 2457


Downloaded From: http://jama.jamanetwork.com/ by a UB-Universitaets Landesbibliothek Dusseldorf User on 12/18/2013

differences in breast cancer survival by race

Documents