Letters
Author Contributions: Dr Cooperberg 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. Study concept and design: All authors. Acquisition, analysis, or interpretation of data: Cooperberg. Drafting of the manuscript: Cooperberg. Critical revision of the manuscript for important intellectual content: All authors. Statistical analysis: Cooperberg. Obtained funding: All authors. Administrative, technical, or material support: Carroll. Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Cooperberg reported receiving grants and personal fees from Myriad Genetics; receiving grants from Genomic Health and GenomeDx; and receiving personal fees from Dendreon, Astellas, and Bayer. No other disclosures were reported. Funding/Support: CaPSURE was funded until 2007 by TAP Pharmaceutical Products Inc. It is currently funded by US Department of Defense grant W81XWH-13-2-0074 and the University of California, San Francisco, Department of Urology resources. Role of the Funder/Sponsor: The sponsor or funding organization 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. 1. Ganz PA, Barry JM, Burke W, et al. NIH State-of-the-Science Conference Statement: role of active surveillance in the management of men with localized prostate cancer. NIH Consens State Sci Statements. 2011;28(1):1-27. 2. Cooperberg MR, Broering JM, Carroll PR. Time trends and local variation in primary treatment of localized prostate cancer. J Clin Oncol. 2010;28(7):1117-1123. 3. Bechis SK, Carroll PR, Cooperberg MR. Impact of age at diagnosis on prostate cancer treatment and survival. J Clin Oncol. 2011;29(2):235-241. 4. Lubeck DP, Litwin MS, Henning JM, et al; CaPSURE Research Panel; Cancer of the Prostate Strategic Urologic Research Endeavor. The CaPSURE database: a methodology for clinical practice and research in prostate cancer. Urology. 1996;48(5):773-777. 5. Cooperberg MR, Broering JM, Carroll PR. Risk assessment for prostate cancer metastasis and mortality at the time of diagnosis. J Natl Cancer Inst. 2009;101 (12):878-887. 6. Womble PR, Montie JE, Ye Z, et al; Michigan Urological Surgery Improvement Collaborative. Contemporary use of initial active surveillance among men in Michigan with low-risk prostate cancer. Eur Urol. 2015;67(1):4450.
William G. Finn, MD E. Blair Holladay, PhD Author Affiliations: American Society for Clinical Pathology, Chicago, Illinois. Corresponding Author: E. Blair Holladay, PhD, American Society for Clinical Pathology, 33 W Monroe St, Chicago, IL 60603 ([email protected]).
COMMENT & RESPONSE
Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Finn reported being the president and Dr Holladay reported being the CEO of the American Society for Clinical Pathology.
Discordant Interpretations of Breast Biopsy Specimens by Pathologists
1. Elmore JG, Longton GM, Carney PA, et al. Diagnostic concordance among pathologists interpreting breast biopsy specimens. JAMA. 2015;313(11):1122-1132.
To the Editor Dr Elmore and colleagues1 reported levels of diagnostic concordance among pathologists in the interpretation of breast biopsy specimens. We believe that the study was flawed in its design, and the conclusions may create unnecessary quality concerns. The study was not designed to evaluate the diagnostic systems currently in place to provide diagnoses by pathologists, but instead to evaluate individual pathologists under unrealistically restrictive conditions. In actual practice, interobserver consultation is commonplace, and an evolving norm is for pathology practices to mandate that certain types of abnormalities (including the types that were the subject of the study) be reviewed by multiple pathologists as part of formal quality assurance programs.2 The level of concordance between the study pathologists and the consensus diagnosis (75%) was no different than the degree of initial concordance reported for the 3 expert breast pathologists who developed the consensus diagnosis. There 82
is no evidence that each individual on the expert panel performed any better in the study than the individual study pathologists as a group. In addition, the interpretation of the study cases was limited to a single slide per case without the opportunity to pursue additional studies, and participants were required to choose a definitive diagnosis on every case, even those with borderline or equivocal features. This approach does not reflect actual pathology practice and clearly contributed to lowering the concordance rate. In actual practice, pathologists have access to, and frequently use, ancillary methods for diagnosis, ranging from simply acquiring additional histological sections from paraffin-embedded tissue blocks to acquiring more tissue from native samples to immunohistochemical analysis. The mix of cases used in the study also paints an unrealistic picture as to the frequency of diagnostic gray zones encountered in practice. Difficult or problematic cases represent only a small minority of breast biopsies encountered in routine clinical practice. This study oversampled cases known to engender diagnostic variability. As a result, the level of concordance reported in the study is not representative of the degree of concordance seen among consecutive breast biopsies encountered in routine clinical practice. We welcome future research that assesses the efficacy of current systems that are in place for the diagnosis of difficult breast lesions, rather than artificial studies that unrealistically isolate incomplete components of those systems.
2. Nakhleh RE. Core components of a comprehensive quality assurance program in anatomic pathology. Adv Anat Pathol. 2009;16(6):418-423.
To the Editor I was one of the pathologists who participated in the study by Dr Elmore and colleagues.1 Although the conclusions are appropriately measured based on the inherent constraints imposed by the design, I have significant misgivings about the validity of the study based on my personal experience. My participation was contingent on the promise of receiving relevant feedback to improve my practice. The recruitment material stated the purpose of the study was to “identify ways in which we can improve breast cancer diagnosis.” A single slide is certainly not the way I or most other pathologists practice. The investigators stated that there would be scheduled slides to review on a regular basis. However, more often than not, slides were delayed, sometimes for weeks, sometimes indefinitely. The slides
JAMA July 7, 2015 Volume 314, Number 1 (Reprinted)
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a Carleton University User on 07/08/2015
jama.com
Letters
would then appear when I was busy and I just wanted to get them out of my office. I can honestly say that my effort was uneven. I never received the last set of slides or an explanation of why I did not receive them. I also did not receive the expert opinion feedback, the peer relationship comparison, or the continuing medical education credit that was promised with participation. Cases in the study were disproportionately weighted toward the exact case type that would require additional studies, such as immunohistochemical stains, deeper sections of tissue, and expert or intradepartmental consultation, which were all contingencies that were forbidden by the study. Every breast core biopsy in my practice is reviewed by a second pathologist, which also was not permitted in the study. Although I appreciate the spirit of the study, all pathologists know of the difficulty of diagnosing atypical lesions with differing expert opinions as to what constitutes the criteria for diagnosis. There is no expert consensus. Who is to say that the expert consensus proposed in this study is truly the criterion standard? Where are the outcome data that could either support or refute that assertion? George Leonard, MD, PhD Author Affiliation: Madigan Army Medical Center, Tacoma, Washington. Corresponding Author: George Leonard, MD, PhD, Madigan Army Medical Center, Pathology, 9040 Jackson Ave, Tacoma, WA 98431 (george.t.leonard4 [email protected]). Conflict of Interest Disclosures: The author has completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. Disclaimer: The views expressed in this letter are those of the author and do not reflect the official policy of the US Department of the Army, the US Department of Defense, or the US government. 1. Elmore JG, Longton GM, Carney PA, et al. Diagnostic concordance among pathologists interpreting breast biopsy specimens. JAMA. 2015;313(11):1122-1132.
In Reply We emphasized in our article that evaluating the overall diagnostic system was not our objective. We studied diagnostic variation at the level of the individual pathologist reviewing a routinely stained slide because this is the starting point of every microscopic diagnosis. We documented very high variation for breast atypia and ductal carcinoma in situ (DCIS). While our study also explored whether pathologists considered cases to be borderline or would request second opinions or additional information, these results were not reported in the article in detail but will be included in upcoming publications. Among the 6900 interpretations, 1803 (26.1%) were considered borderline and participants indicated they would desire a second opinion on 2451 (35.5%) of the interpretations. As we stated in the article, the 25% diagnostic discordance was calculated based on test cases augmented with more atypia and DCIS diagnoses than typically seen in clinical practice. When our results are applied to the US population, the level of overall disagreement is much less because the majority of breast biopsies are diagnosed as benign without atypia. Nevertheless, the extent of uncertainty
associated with diagnoses of atypia and DCIS is cause for concern. Drs Finn and Holladay misunderstood the comparison between the consensus reference panel and participants. The reference panel’s concordance with the final consensus diagnoses was 90%, whereas participants’ concordance with the same consensus diagnoses was 75%. Although unanimous agreement among the 3 reference pathologists’ independent diagnoses was 75%, we did not report data on the level of unanimous agreement between randomly selected groups of 3 participants for comparison (we subsequently calculated it as 58%). There is no perfect reference standard for a study such as ours, and clinical follow-up cannot be used because removal of tissue by the biopsy might alter outcome. In clinical practice, having 3 experienced breast pathologists independently interpret a case, followed by a consensus meeting if they disagree, would be considered by most to be an ideal practice standard. The commitment of the 115 practicing pathologists who participated in the Breast Pathology study must be commended. Because diagnostic material can vary, even when viewing adjacent, 4- to 6-μm slices obtained from the same tissue block,1 all study pathologists viewed the same glass slide for each case. If a study participant was late in returning slides, we had to reschedule the viewing for the next participant. Dr Leonard’s scheduling challenges were an anomaly that occurred for less than 2% of participants. We are sorry that he missed the approval window granted to us to keep the online continuing medical education activity open; 97% of participants found the personalized feedback useful. We appreciate his participation and will send him his confidential individual results. Ours is not the first study describing diagnostic variability among pathologists, but we wish it could be the last. More than 20 years ago, 76 studies on pathologists’ variability were summarized.2,3 Publications on this topic continue, including a recent report that more than 10% of breast biopsies had a clinically significant change in diagnosis after a second opinion.4 Problems must be acknowledged and quantified before they can be improved. Our data provide a benchmark that hopefully will lead to clinical improvement through a combination of physician education and additional system-level interventions. Joann G. Elmore, MD, MPH Margaret S. Pepe, PhD Donald L. Weaver, MD Author Affiliations: University of Washington School of Medicine, Seattle (Elmore); Fred Hutchinson Cancer Research Center, Seattle (Pepe); University of Vermont School of Medicine, Burlington (Weaver). Corresponding Author: Joann G. Elmore, MD, MPH, University of Washington School of Medicine, 325 Ninth Ave, Seattle, WA 98104 ([email protected]). Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Elmore reported receiving a grant from the National Cancer Institute and serving as a medical editor for the nonprofit Informed Medical Decisions Foundation. No other disclosures were reported.
jama.com
(Reprinted) JAMA July 7, 2015 Volume 314, Number 1
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a Carleton University User on 07/08/2015
83
Letters
1. Knezevich SR, Barnhill RL, Elder DE, et al. Variability in mitotic figures in serial sections of thin melanomas. J Am Acad Dermatol. 2014;71(6):1204-1211. 2. Feinstein AR. A bibliography of publications on observer variability. J Chronic Dis. 1985;38(8):619-632. 3. Elmore JG, Feinstein AR. A bibliography of publications on observer variability (final installment). J Clin Epidemiol. 1992;45(6):567-580. 4. Khazai L, Middleton LP, Goktepe N, Liu BT, Sahin AA. Breast pathology second review identifies clinically significant discrepancies in over 10% of patients. J Surg Oncol. 2015;111(2):192-197.
Suicide Among US Military Personnel To the Editor Dr Friedman’s article highlighted the current research and risk factors associated with suicide among US Army personnel.1 However, I was surprised when the US Special Operations Command (USSOCOM), which includes members from all military branches, was discussed. Friedman suggested that the Command addresses the issue of suicide prevention “through a widespread emphasis on physical fitness” rather than focusing on identification and treatment of soldiers with disorders. His misperceptions of USSOCOM and their physical fitness program should be clarified. The Preservation of the Force and Family initiative was launched by USSOCOM over the past few years as an innovative and holistic means to enable human, psychological, and spiritual performance both on and off the battlefield for the Special Operations tactical athlete (Army Rangers, Army Special Forces, Navy SEALs, Air Force Special Tactics, Marine Corps Critical Skills Operators, etc) and their families.2 This initiative has ushered family- and psychological-embedded support into multiple USSOCOM units. This support system has included social workers, chaplains, and psychologists. The physical fitness aspect is simply one part of a larger holistic program. It has allowed USSOCOM members to receive strength and conditioning training along with physical therapy and rehabilitation on par with the collegiate and even the professional sports medicine model. Having served the past 3 years within one of these units as an embedded physician, I can attest to the benefits of this holistic and preventative approach. This team-based model allows the unit member (over time) to feel comfortable with a practitioner and seek help in a manner that would never be seen in a traditional mental health approach. Some of my most effective mental health sessions with members came about during physical fitness sessions. The average USSOCOM member has gone through extensive and rigorous assessments prior to entry into the Command, thus limiting many of the risk factors discussed in the article. Nevertheless, the threat of suicide remains. The care and treatment of USSOCOM warriors require innovative personal approaches coupled with data-driven guidelines. Philip M. Flatau, MD Author Affiliation: Air Force Special Operations, Hurlburt Field, Florida. Corresponding Author: Philip M. Flatau, MD, Air Force Special Operations, FAWM, 415 Independence Rd, Hurlburt Field, FL 32544 ([email protected]). Conflict of Interest Disclosures: The author has completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. 84
Disclaimer: The views expressed in this letter are the authors’ own and do not necessarily represent the views of the US Air Force, USSOCOM, or the US Department of Defense. 1. Friedman MJ. Risk factors for suicides among army personnel. JAMA. 2015; 313(11):1154-1155. 2. US Special Operations Command. Preservation of the force and family. http: //www.socom.mil/POTFF/default.aspx. Accessed March 22, 2015.
In Reply Suicide is an urgent problem for all US military services, their components, and their commands as well as the United States. Because this is such a high priority, the US Army and the National Institute of Mental Health jointly funded the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS) to identify “modifiable risk and resilience factors for suicidality that could be used to target effective preventative interventions for Army suicides … .”1 In my From The JAMA Network article regarding the first 3 Army STARRS articles, I concluded that these results provide sufficient information to suggest improvements in current suicide prevention and policy. The major finding was that a minority of soldiers, those with mental health problems before enlistment, after enlistment, or both, are responsible for a disproportionate amount of suicidal behavior. More granular analyses identified soldiers having frequent anger attacks (eg, meeting diagnostic criteria for intermittent explosive disorder) as the highestrisk group. I also identified peer and family support as important targets for intervention. These data hopefully will be translated into evidence-based policy and practice and save the lives of those who might otherwise kill themselves. Dr Flatau made 2 interesting statements based on his personal observations: that USSOCOM’s holistic approach promotes more comfortable mental health–seeking behavior than seen with traditional approaches and that his most effective mental health sessions with service members came during physical fitness. Scientific inquiry always begins with sophisticated observations such as these. Therefore, I urge Flatau and his USSOCOM colleagues to test these hypotheses concerning the utility of the holistic approach embodied within USSOCOM’s Preservation of the Force and Family initiative with respect to the prevention or early detection of suicidal behavior. It would be important for the field to have such data to help guide suicide prevention and intervention. Flatau also suggested that Army STARRS findings should not be generalized to USSOCOM because of differences between the military services. The USSOCOM should provide data showing that Army STARRS findings do not apply to their troops before setting suicide prevention policies that assume this to be the case. Until such data are gathered, however, the best scientific evidence available should be used to design preventative strategies. Right now, Army STARRS is the best available. The scientific data point toward a focus on mental health and early detection and treatment of mental illness to address suicide risk. Matthew J. Friedman, MD, PhD
JAMA July 7, 2015 Volume 314, Number 1 (Reprinted)
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a Carleton University User on 07/08/2015
jama.com
Author Contributions: Dr Cooperberg 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. Study concept and design: All authors. Acquisition, analysis, or interpretation of data: Cooperberg. Drafting of the manuscript: Cooperberg. Critical revision of the manuscript for important intellectual content: All authors. Statistical analysis: Cooperberg. Obtained funding: All authors. Administrative, technical, or material support: Carroll. Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Cooperberg reported receiving grants and personal fees from Myriad Genetics; receiving grants from Genomic Health and GenomeDx; and receiving personal fees from Dendreon, Astellas, and Bayer. No other disclosures were reported. Funding/Support: CaPSURE was funded until 2007 by TAP Pharmaceutical Products Inc. It is currently funded by US Department of Defense grant W81XWH-13-2-0074 and the University of California, San Francisco, Department of Urology resources. Role of the Funder/Sponsor: The sponsor or funding organization 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. 1. Ganz PA, Barry JM, Burke W, et al. NIH State-of-the-Science Conference Statement: role of active surveillance in the management of men with localized prostate cancer. NIH Consens State Sci Statements. 2011;28(1):1-27. 2. Cooperberg MR, Broering JM, Carroll PR. Time trends and local variation in primary treatment of localized prostate cancer. J Clin Oncol. 2010;28(7):1117-1123. 3. Bechis SK, Carroll PR, Cooperberg MR. Impact of age at diagnosis on prostate cancer treatment and survival. J Clin Oncol. 2011;29(2):235-241. 4. Lubeck DP, Litwin MS, Henning JM, et al; CaPSURE Research Panel; Cancer of the Prostate Strategic Urologic Research Endeavor. The CaPSURE database: a methodology for clinical practice and research in prostate cancer. Urology. 1996;48(5):773-777. 5. Cooperberg MR, Broering JM, Carroll PR. Risk assessment for prostate cancer metastasis and mortality at the time of diagnosis. J Natl Cancer Inst. 2009;101 (12):878-887. 6. Womble PR, Montie JE, Ye Z, et al; Michigan Urological Surgery Improvement Collaborative. Contemporary use of initial active surveillance among men in Michigan with low-risk prostate cancer. Eur Urol. 2015;67(1):4450.
William G. Finn, MD E. Blair Holladay, PhD Author Affiliations: American Society for Clinical Pathology, Chicago, Illinois. Corresponding Author: E. Blair Holladay, PhD, American Society for Clinical Pathology, 33 W Monroe St, Chicago, IL 60603 ([email protected]).
COMMENT & RESPONSE
Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Finn reported being the president and Dr Holladay reported being the CEO of the American Society for Clinical Pathology.
Discordant Interpretations of Breast Biopsy Specimens by Pathologists
1. Elmore JG, Longton GM, Carney PA, et al. Diagnostic concordance among pathologists interpreting breast biopsy specimens. JAMA. 2015;313(11):1122-1132.
To the Editor Dr Elmore and colleagues1 reported levels of diagnostic concordance among pathologists in the interpretation of breast biopsy specimens. We believe that the study was flawed in its design, and the conclusions may create unnecessary quality concerns. The study was not designed to evaluate the diagnostic systems currently in place to provide diagnoses by pathologists, but instead to evaluate individual pathologists under unrealistically restrictive conditions. In actual practice, interobserver consultation is commonplace, and an evolving norm is for pathology practices to mandate that certain types of abnormalities (including the types that were the subject of the study) be reviewed by multiple pathologists as part of formal quality assurance programs.2 The level of concordance between the study pathologists and the consensus diagnosis (75%) was no different than the degree of initial concordance reported for the 3 expert breast pathologists who developed the consensus diagnosis. There 82
is no evidence that each individual on the expert panel performed any better in the study than the individual study pathologists as a group. In addition, the interpretation of the study cases was limited to a single slide per case without the opportunity to pursue additional studies, and participants were required to choose a definitive diagnosis on every case, even those with borderline or equivocal features. This approach does not reflect actual pathology practice and clearly contributed to lowering the concordance rate. In actual practice, pathologists have access to, and frequently use, ancillary methods for diagnosis, ranging from simply acquiring additional histological sections from paraffin-embedded tissue blocks to acquiring more tissue from native samples to immunohistochemical analysis. The mix of cases used in the study also paints an unrealistic picture as to the frequency of diagnostic gray zones encountered in practice. Difficult or problematic cases represent only a small minority of breast biopsies encountered in routine clinical practice. This study oversampled cases known to engender diagnostic variability. As a result, the level of concordance reported in the study is not representative of the degree of concordance seen among consecutive breast biopsies encountered in routine clinical practice. We welcome future research that assesses the efficacy of current systems that are in place for the diagnosis of difficult breast lesions, rather than artificial studies that unrealistically isolate incomplete components of those systems.
2. Nakhleh RE. Core components of a comprehensive quality assurance program in anatomic pathology. Adv Anat Pathol. 2009;16(6):418-423.
To the Editor I was one of the pathologists who participated in the study by Dr Elmore and colleagues.1 Although the conclusions are appropriately measured based on the inherent constraints imposed by the design, I have significant misgivings about the validity of the study based on my personal experience. My participation was contingent on the promise of receiving relevant feedback to improve my practice. The recruitment material stated the purpose of the study was to “identify ways in which we can improve breast cancer diagnosis.” A single slide is certainly not the way I or most other pathologists practice. The investigators stated that there would be scheduled slides to review on a regular basis. However, more often than not, slides were delayed, sometimes for weeks, sometimes indefinitely. The slides
JAMA July 7, 2015 Volume 314, Number 1 (Reprinted)
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a Carleton University User on 07/08/2015
jama.com
Letters
would then appear when I was busy and I just wanted to get them out of my office. I can honestly say that my effort was uneven. I never received the last set of slides or an explanation of why I did not receive them. I also did not receive the expert opinion feedback, the peer relationship comparison, or the continuing medical education credit that was promised with participation. Cases in the study were disproportionately weighted toward the exact case type that would require additional studies, such as immunohistochemical stains, deeper sections of tissue, and expert or intradepartmental consultation, which were all contingencies that were forbidden by the study. Every breast core biopsy in my practice is reviewed by a second pathologist, which also was not permitted in the study. Although I appreciate the spirit of the study, all pathologists know of the difficulty of diagnosing atypical lesions with differing expert opinions as to what constitutes the criteria for diagnosis. There is no expert consensus. Who is to say that the expert consensus proposed in this study is truly the criterion standard? Where are the outcome data that could either support or refute that assertion? George Leonard, MD, PhD Author Affiliation: Madigan Army Medical Center, Tacoma, Washington. Corresponding Author: George Leonard, MD, PhD, Madigan Army Medical Center, Pathology, 9040 Jackson Ave, Tacoma, WA 98431 (george.t.leonard4 [email protected]). Conflict of Interest Disclosures: The author has completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. Disclaimer: The views expressed in this letter are those of the author and do not reflect the official policy of the US Department of the Army, the US Department of Defense, or the US government. 1. Elmore JG, Longton GM, Carney PA, et al. Diagnostic concordance among pathologists interpreting breast biopsy specimens. JAMA. 2015;313(11):1122-1132.
In Reply We emphasized in our article that evaluating the overall diagnostic system was not our objective. We studied diagnostic variation at the level of the individual pathologist reviewing a routinely stained slide because this is the starting point of every microscopic diagnosis. We documented very high variation for breast atypia and ductal carcinoma in situ (DCIS). While our study also explored whether pathologists considered cases to be borderline or would request second opinions or additional information, these results were not reported in the article in detail but will be included in upcoming publications. Among the 6900 interpretations, 1803 (26.1%) were considered borderline and participants indicated they would desire a second opinion on 2451 (35.5%) of the interpretations. As we stated in the article, the 25% diagnostic discordance was calculated based on test cases augmented with more atypia and DCIS diagnoses than typically seen in clinical practice. When our results are applied to the US population, the level of overall disagreement is much less because the majority of breast biopsies are diagnosed as benign without atypia. Nevertheless, the extent of uncertainty
associated with diagnoses of atypia and DCIS is cause for concern. Drs Finn and Holladay misunderstood the comparison between the consensus reference panel and participants. The reference panel’s concordance with the final consensus diagnoses was 90%, whereas participants’ concordance with the same consensus diagnoses was 75%. Although unanimous agreement among the 3 reference pathologists’ independent diagnoses was 75%, we did not report data on the level of unanimous agreement between randomly selected groups of 3 participants for comparison (we subsequently calculated it as 58%). There is no perfect reference standard for a study such as ours, and clinical follow-up cannot be used because removal of tissue by the biopsy might alter outcome. In clinical practice, having 3 experienced breast pathologists independently interpret a case, followed by a consensus meeting if they disagree, would be considered by most to be an ideal practice standard. The commitment of the 115 practicing pathologists who participated in the Breast Pathology study must be commended. Because diagnostic material can vary, even when viewing adjacent, 4- to 6-μm slices obtained from the same tissue block,1 all study pathologists viewed the same glass slide for each case. If a study participant was late in returning slides, we had to reschedule the viewing for the next participant. Dr Leonard’s scheduling challenges were an anomaly that occurred for less than 2% of participants. We are sorry that he missed the approval window granted to us to keep the online continuing medical education activity open; 97% of participants found the personalized feedback useful. We appreciate his participation and will send him his confidential individual results. Ours is not the first study describing diagnostic variability among pathologists, but we wish it could be the last. More than 20 years ago, 76 studies on pathologists’ variability were summarized.2,3 Publications on this topic continue, including a recent report that more than 10% of breast biopsies had a clinically significant change in diagnosis after a second opinion.4 Problems must be acknowledged and quantified before they can be improved. Our data provide a benchmark that hopefully will lead to clinical improvement through a combination of physician education and additional system-level interventions. Joann G. Elmore, MD, MPH Margaret S. Pepe, PhD Donald L. Weaver, MD Author Affiliations: University of Washington School of Medicine, Seattle (Elmore); Fred Hutchinson Cancer Research Center, Seattle (Pepe); University of Vermont School of Medicine, Burlington (Weaver). Corresponding Author: Joann G. Elmore, MD, MPH, University of Washington School of Medicine, 325 Ninth Ave, Seattle, WA 98104 ([email protected]). Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Elmore reported receiving a grant from the National Cancer Institute and serving as a medical editor for the nonprofit Informed Medical Decisions Foundation. No other disclosures were reported.
jama.com
(Reprinted) JAMA July 7, 2015 Volume 314, Number 1
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a Carleton University User on 07/08/2015
83
Letters
1. Knezevich SR, Barnhill RL, Elder DE, et al. Variability in mitotic figures in serial sections of thin melanomas. J Am Acad Dermatol. 2014;71(6):1204-1211. 2. Feinstein AR. A bibliography of publications on observer variability. J Chronic Dis. 1985;38(8):619-632. 3. Elmore JG, Feinstein AR. A bibliography of publications on observer variability (final installment). J Clin Epidemiol. 1992;45(6):567-580. 4. Khazai L, Middleton LP, Goktepe N, Liu BT, Sahin AA. Breast pathology second review identifies clinically significant discrepancies in over 10% of patients. J Surg Oncol. 2015;111(2):192-197.
Suicide Among US Military Personnel To the Editor Dr Friedman’s article highlighted the current research and risk factors associated with suicide among US Army personnel.1 However, I was surprised when the US Special Operations Command (USSOCOM), which includes members from all military branches, was discussed. Friedman suggested that the Command addresses the issue of suicide prevention “through a widespread emphasis on physical fitness” rather than focusing on identification and treatment of soldiers with disorders. His misperceptions of USSOCOM and their physical fitness program should be clarified. The Preservation of the Force and Family initiative was launched by USSOCOM over the past few years as an innovative and holistic means to enable human, psychological, and spiritual performance both on and off the battlefield for the Special Operations tactical athlete (Army Rangers, Army Special Forces, Navy SEALs, Air Force Special Tactics, Marine Corps Critical Skills Operators, etc) and their families.2 This initiative has ushered family- and psychological-embedded support into multiple USSOCOM units. This support system has included social workers, chaplains, and psychologists. The physical fitness aspect is simply one part of a larger holistic program. It has allowed USSOCOM members to receive strength and conditioning training along with physical therapy and rehabilitation on par with the collegiate and even the professional sports medicine model. Having served the past 3 years within one of these units as an embedded physician, I can attest to the benefits of this holistic and preventative approach. This team-based model allows the unit member (over time) to feel comfortable with a practitioner and seek help in a manner that would never be seen in a traditional mental health approach. Some of my most effective mental health sessions with members came about during physical fitness sessions. The average USSOCOM member has gone through extensive and rigorous assessments prior to entry into the Command, thus limiting many of the risk factors discussed in the article. Nevertheless, the threat of suicide remains. The care and treatment of USSOCOM warriors require innovative personal approaches coupled with data-driven guidelines. Philip M. Flatau, MD Author Affiliation: Air Force Special Operations, Hurlburt Field, Florida. Corresponding Author: Philip M. Flatau, MD, Air Force Special Operations, FAWM, 415 Independence Rd, Hurlburt Field, FL 32544 ([email protected]). Conflict of Interest Disclosures: The author has completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. 84
Disclaimer: The views expressed in this letter are the authors’ own and do not necessarily represent the views of the US Air Force, USSOCOM, or the US Department of Defense. 1. Friedman MJ. Risk factors for suicides among army personnel. JAMA. 2015; 313(11):1154-1155. 2. US Special Operations Command. Preservation of the force and family. http: //www.socom.mil/POTFF/default.aspx. Accessed March 22, 2015.
In Reply Suicide is an urgent problem for all US military services, their components, and their commands as well as the United States. Because this is such a high priority, the US Army and the National Institute of Mental Health jointly funded the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS) to identify “modifiable risk and resilience factors for suicidality that could be used to target effective preventative interventions for Army suicides … .”1 In my From The JAMA Network article regarding the first 3 Army STARRS articles, I concluded that these results provide sufficient information to suggest improvements in current suicide prevention and policy. The major finding was that a minority of soldiers, those with mental health problems before enlistment, after enlistment, or both, are responsible for a disproportionate amount of suicidal behavior. More granular analyses identified soldiers having frequent anger attacks (eg, meeting diagnostic criteria for intermittent explosive disorder) as the highestrisk group. I also identified peer and family support as important targets for intervention. These data hopefully will be translated into evidence-based policy and practice and save the lives of those who might otherwise kill themselves. Dr Flatau made 2 interesting statements based on his personal observations: that USSOCOM’s holistic approach promotes more comfortable mental health–seeking behavior than seen with traditional approaches and that his most effective mental health sessions with service members came during physical fitness. Scientific inquiry always begins with sophisticated observations such as these. Therefore, I urge Flatau and his USSOCOM colleagues to test these hypotheses concerning the utility of the holistic approach embodied within USSOCOM’s Preservation of the Force and Family initiative with respect to the prevention or early detection of suicidal behavior. It would be important for the field to have such data to help guide suicide prevention and intervention. Flatau also suggested that Army STARRS findings should not be generalized to USSOCOM because of differences between the military services. The USSOCOM should provide data showing that Army STARRS findings do not apply to their troops before setting suicide prevention policies that assume this to be the case. Until such data are gathered, however, the best scientific evidence available should be used to design preventative strategies. Right now, Army STARRS is the best available. The scientific data point toward a focus on mental health and early detection and treatment of mental illness to address suicide risk. Matthew J. Friedman, MD, PhD
JAMA July 7, 2015 Volume 314, Number 1 (Reprinted)
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a Carleton University User on 07/08/2015
jama.com
