Letters
Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. 1. Ebrahim S, Sohani ZN, Montoya L, et al. Reanalyses of randomized clinical trial data. JAMA. 2014;312(10):1024-1032. 2. National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. Tissue plasminogen activator for acute ischemic stroke. N Engl J Med. 1995;333(24):1581-1587. 3. Wardlaw JM, Murray V, Berge E, del Zoppo GJ. Thrombolysis for acute ischaemic stroke. Cochrane Database Syst Rev. 2014;7:CD000213. 4. Kaydos W. Operational Performance Measurement. Boca Raton, FL: CRC Press; 1999. 5. Krumholz HM, Peterson ED. Open access to clinical trials data. JAMA. 2014;312 (10):1002-1003.
In Reply Drs Platts-Mills and Jones have raised a potential concern with the search strategy used in our study on reanalyses of randomized clinical trial data, specifically with respect to our restriction to studies indexed as a clinical trial publication type in MEDLINE. We agree with them that a strategy with no restriction to clinical trials publication type would be more sensitive in detecting reanalyses. However, performing such a search would yield a total of 112 330 citations (limited to March 9, 2014, as per our original search). Screening this number of citations would be a daunting task and may not be a valuable use of resources if the output of additional studies retrieved would be few in number compared with the studies retrieved in our original search. We have attempted to get an estimate of how many studies we potentially missed by using a simplified search with (reanalysis [ti] OR reanalysis [ti]) NOT Clinical Trial [PT] AND random* (limited to March 9, 2014) that would capture the reanalysis of the NINDS rt-PA Stroke Study. Besides the NINDS study, we also identified eligible reanalyses for 3 more trials. Two of the 4 new eligible trials1-4 had their reanalyses completed by independent authors or committees,2,3 and the other 2 had reanalyses completed by the same research group or included authors from the original study.1,4 Reasons for completing reanalyses included the use of intention-to-treat principles (compared with per-protocol analyses used in the original article),1 correction for subgroup imbalances,3 use of a method for identifying latent trajectories (latent class growth analysis),4 and comparison of the primary end point using a change in the definition of the outcome.2 Of the 4 reanalyses, 1 found that the experimental treatment (bilavirudin) was significantly more effective and had less bleeding complications than control (heparin),1 whereas the original analysis had shown no significant difference in efficacy, so perhaps this would lead to more patients being treated with bilavirudin. Another reanalysis4 found effects for latent classes of participants that were not obvious in the original analysis, so it is possible that different patients may be treated differently based on these results. We have no doubt that a few more trials must have been missed given the imperfect sensitivity of any searches and the lack of a standard tag for reanalyses in PubMed. However, we
believe that it is unlikely that the number of missed published reanalyses is large. Measures to improve awareness of the importance and systematic indexing of reanalyses in PubMed are certainly required. Nonetheless, our main conclusions that published reanalyses are currently rare, almost always involve authors of the original analyses, and about one-third reach different conclusions still hold true. If there were more interest in publishing reanalyses regardless of their results, we suspect that the proportion of those claiming that they had found something different from the original analysis would decrease. Shanil Ebrahim, PhD John P. A. Ioannidis, MD, DSc Author Affiliations: Meta-Research Innovation Center at Stanford, Stanford University, Stanford, California. Corresponding Author: John P. A. Ioannidis, MD, DSc, Meta-Research Innovation Center at Stanford, Stanford University, 1265 Welch Rd, Stanford, CA 94305 ([email protected]). Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. 1. Bittl JA, Chaitman BR, Feit F, Kimball W, Topol EJ. Bivalirudin versus heparin during coronary angioplasty for unstable or postinfarction angina: final report reanalysis of the Bivalirudin Angioplasty Study. Am Heart J. 2001;142(6):952-959. 2. Moses RG. Achieving glycosylated hemoglobin targets using the combination of repaglinide and metformin in type 2 diabetes: a reanalysis of earlier data in terms of current targets. Clin Ther. 2008;30(3):552-554. 3. Ingall TJ, O’Fallon WM, Asplund K, et al. Findings from the reanalysis of the NINDS tissue plasminogen activator for acute ischemic stroke treatment trial. Stroke. 2004;35(10):2418-2424. 4. Lennon MC, McAllister W, Kuang L, Herman DB. Capturing intervention effects over time: reanalysis of a critical time intervention for homeless mentally ill men. Am J Public Health. 2005;95(10):1760-1766.
Risk Profile of Clinical Trial Participants To the Editor In their research letter, Dr Udell and colleagues1 stated that clinical trial participants had a lower risk profile and a more favorable prognosis compared with the broader population included in the National Cardiovascular Data Registry (NCDR) Acute Coronary Treatment and Intervention Outcomes Network Registry–Get With The Guidelines. It is likely that these differences would be more pronounced if clinical trial participants were compared with patients with myocardial infarction in the broader US population because participation in the NCDR is voluntary and includes participation by only 5% of US hospitals.1,2 In addition, the data would be put in better perspective if the age of the patients with myocardial infarction in the NCDR and the rates of diagnostic cardiac catheterization and percutaneous coronary intervention were compared with a more inclusive database such as the Nationwide Inpatient Sample.2 It would also be interesting to compare the characteristics of patients presenting to the NCDR hospitals that did and did not enroll patients in clinical trials because there may be relevant differences in clinical presentation and outcomes. In addition, the better outcomes of participants in clinical trials may be expected from the exclusion criteria of the clinical trial protocols, self-selection of clinical trial participants (who of-
jama.com
JAMA January 6, 2015 Volume 313, Number 1
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of Sussex Library User on 08/10/2015
93
Letters
Table. Risk of In-Hospital Cardiovascular Events Among Trial Participants, Eligible Nonparticipants, and Ineligible Nonparticipantsa Rate of Events, % Eligible Trial Nonparticipants Participants (n = 72 313) (n = 3789) 1.5 2.4
Death Death or myocardial infarction
2.4
3.1
1.37 (1.01-1.86)
8.9
Cardiogenic shock
2.3
3.5
1.50 (1.16-1.96)
6.2
Congestive heart failure
4.4
5.5
1.03 (0.87-1.21)
8.9
Major bleedingc
6.4
6.5
1.11 (0.95-1.30)
11.8
creatinine, and troponin ratio to the upper limit of the local assay normal reference), medical insurance type, within-center correlation and clustering of variable responses using an exchangeable correlation structure covariance matrix, Acute Coronary Treatment and Intervention Outcomes Network bleeding outcome, and medication use (aspirin, clopidogrel, warfarin, β-blocker, angiotensin-converting enzyme inhibitor, angiotensin receptor blocker, aldosterone antagonist, statin, and other lipid-lowering therapy).
Abbreviation: OR, odds ratio. a
Among patients who were not transferred prior to discharge from hospitals that enrolled at least 1 patient in a clinical trial.
b
Adjusted for demographic variables (age, sex, race/ethnicity, and weight), signs and symptoms at presentation (heart failure, heart rate, systolic blood pressure, electrocardiographic findings), medical history (hypertension; diabetes mellitus; current or recent tobacco use; hypercholesterolemia; prior myocardial infarction, coronary revascularization, heart failure, stroke, peripheral artery disease), laboratory results (baseline hemoglobin, serum
ten have better health behaviors), and possibly due to earlier treatment of conditions discovered during protocol-directed examinations.3 William J. Kostis, MD, PhD Sri Ram Pentakota, MD, MPH, PhD Abel E. Morerya, MD Author Affiliations: Cardiology Division, Massachusetts General Hospital, Boston (Kostis); Cardiovascular Institute, Rutgers Robert Wood Johnson Medical School, New Brunswick, New Jersey (Pentakota, Morerya). Corresponding Author: William J. Kostis, MD, PhD, Cardiology Division, Massachusetts General Hospital, 55 Fruit St, Boston, MA 02114 (wkostis@mgh .harvard.edu). Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. 1. Udell JA, Wang TY, Li S, et al. Clinical trial participation after myocardial infarction in a national cardiovascular data registry. JAMA. 2014;312(8):841-843. 2. Healthcare Cost and Utilization Project. Nationwide Inpatient Sample. http://hub .healthdata.gov/dataset/hcup-nationwide-inpatient-sample. Accessed September 10, 2014. 3. Kostis WJ, Cabrera J, Messerli FH, et al. Competing cardiovascular and noncardiovascular risks and longevity in the systolic hypertension in the elderly program. Am J Cardiol. 2014;113(4):676-681.
In Reply Dr Kostis and colleagues suggest that patients with myocardial infarction who participate in clinical trials may demonstrate even more pronounced differences in presenting factors, hospital care, and outcomes compared with patients included in the US Healthcare Cost and Utilization Project Nationwide Inpatient Sample. Although such a comparison could potentially bolster our findings, we are not aware of data from the sample documenting clinical trial participation to allow for a valid comparison.1 We agree with Kostis et al that among patients with myocardial infarction, trial participants often receive closer medical supervision by research staff and invested parties than otherwise occurs in routine practice.2 For instance, we observed that clinical trial participants more frequently received care for 94
Adjusted OR (95% CI)b 1.63 (1.11-2.39)
Rate of Events for Ineligible Nonparticipants, % (n = 33 912) 8.1
c
A major bleeding event was defined as bleeding that occurred within the first 24 hours of admission and was a criterion for trial ineligibility.
myocardial infarction meeting quality metrics, noninvasive and invasive diagnostic procedures, discharge counseling, and cardiac rehabilitation referral. Increased medical supervision may be associated with improved outcomes either because high-quality practitioners participate as investigators in clinical trials (eg, physician adherence to guideline recommendations) or as a result of the Hawthorne effect, a phenomenon of alteration and improvement in patient behavior resulting from participant awareness of being under observation (eg, patient adherence to prescribed therapy). The cardiovascular trial literature has documented many examples of clinical research participants faring better than anticipated compared with epidemiological reports.2,3 Although it is possible that trial participants may have been managed in hospitals that practice higher-quality care, our statistical models adjusted for clustering of patients within hospitals. When the multivariable analysis of inpatient outcomes was limited to patients at hospitals that enrolled at least 1 patient in a clinical trial, results were nearly identical (Table). Because the type of clinical trial that patients participated in may have evaluated a drug, strategy, or device, we cannot ascertain how results might differ with participation in a trial of a specific intervention. We nevertheless agree it is likely that a selection bias occurs during trial recruitment at the patient level. There are implicit (sometimes discretionary) characteristics that make a preferable trial participant, such as language comprehension, reliability, and compliance in following trial requirements, as well as general assessments of health and frailty by research personnel that may not be apparent in captured data. Thus, our observations suggest an opportunity for broader efforts to improve trial participation and enroll patients at similar risk to those in whom the therapies will ultimately be used. Specifically, there are large numbers of potentially eligible patients within existing registries to recruit into trials, a strategy increasingly being used successfully in cardiovascular outcomes research.4,5
JAMA January 6, 2015 Volume 313, Number 1
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of Sussex Library User on 08/10/2015
jama.com
Letters
Performing more clinical trial research nested within such registries would allow leveraging of established resources and funding to minimize redundancy in study coordination, data collection and monitoring, clinical follow-up, regulatory issues, and financial barriers to participation from an institutional perspective while simultaneously improving testing of therapies in heterogeneous patient populations seen in routine clinical practice. Jacob A. Udell, MD, MPH Matthew T. Roe, MD, MHS Stephen D. Wiviott, MD Author Affiliations: Women’s College Hospital, University of Toronto, Toronto, Ontario, Canada (Udell); Duke Clinical Research Institute, Duke University, Durham, North Carolina (Roe); Cardiovascular Division, Brigham and Women’s Hospital, Boston, Massachusetts (Wiviott).
CORRECTION Cost Correction: In the article from The Medical Letter on Drugs and Therapeutics entitled “Drugs for MRSA Skin and Soft-Tissue Infections” published in the October 15, 2014, issue of JAMA (2014;312[15]:1583-1584. doi:10.1001/jama.2014 .13385), the cost of tedizolid in Table 1 was reported as $2950 for 10 days’ treatment. Cubist’s tedizolid (Sivextro) received FDA approval for 6 days’ treatment of acute bacterial skin and skin structure infection, which would cost $1770. This article was corrected online. Incorrect Reporting: In the Original Investigation entitled “Effect of Reversible Intermittent Intra-abdominal Vagal Nerve Blockade on Morbid Obesity: The ReCharge Randomized Clinical Trial,” published in the December 3, 2014, issue of JAMA (2014;312[9]:915-922. doi:10.1001/jama.2014.10540), the reason of device removal was incorrectly reported. On page 918, the third sentence of the fifth full paragraph should read, “Three patients in the vagal nerve block did so because 1 had experienced pain at the neuroregulator site; 1 experienced pain with therapy, and 1 experienced heartburn.” The fifth sentence of the same paragraph should read, “Two patients in the vagal nerve block group and 4 in the sham group asked to have the devices removed.” This article was corrected online.
Corresponding Author: Jacob A. Udell, MD, MPH, Women’s College Hospital, University of Toronto, 76 Grenville St, Toronto, ON M5S 1B1, Canada (jay.udell @utoronto.ca). Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Udell reported receiving honoraria from the American College of Cardiology Foundation. Dr Roe reported receiving grants from Eli Lilly & Company, Daiichi Sankyo, sanofi-aventis, the American College of Cardiology, the American Heart Association, and the Familial Hypercholesterolemia Foundation; and personal fees from Merck, Janssen Pharmaceuticals, Regeneron, AstraZeneca, and Amgen outside the submitted work. Dr Wiviott reported receiving grants from Arena, AstraZeneca, Bristol-Myers Squibb, Daiichi Sankyo, Eisai, Eli Lilly, Merck, and sanofi-aventis; and consulting fees from Aegerion, Angelmed, Arena, AstraZeneca, Boston Clinical Research Institute, Bristol-Myers Squibb, Daiichi Sankyo, Eisai, Eli Lilly, ICON Clinical, Janssen, and Xoma. 1. Healthcare Cost and Utilization Project (HCUP). HCUP National Inpatient Sample Database Documentation. http://www.hcup-us.ahrq.gov/db/nation/nis /nisdbdocumentation.jsp. Accessed October 6, 2014. 2. Kandzari DE, Roe MT, Chen AY, et al. Influence of clinical trial enrollment on the quality of care and outcomes for patients with non-ST-segment elevation acute coronary syndromes. Am Heart J. 2005;149(3):474-481. 3. Steg PG, López-Sendón J, Lopez de Sa E, et al; GRACE Investigators. External validity of clinical trials in acute myocardial infarction. Arch Intern Med. 2007;167 (1):68-73. 4. Fröbert O, Lagerqvist B, Olivecrona GK, et al; TASTE Trial. Thrombus aspiration during ST-segment elevation myocardial infarction. N Engl J Med. 2013;369(17):1587-1597. 5. Rao SV, Hess CN, Barham B, et al. A registry-based randomized trial comparing radial and femoral approaches in women undergoing percutaneous coronary intervention: the SAFE-PCI for Women (Study of Access Site for Enhancement of PCI for Women) trial. JACC Cardiovasc Interv. 2014;7(8):857-867.
Guidelines for Letters Letters discussing a recent JAMA article should be submitted within 4 weeks of the article's publication in print. Letters received after 4 weeks will rarely be considered. Letters should not exceed 400 words of text and 5 references and may have no more than 3 authors. Letters reporting original research should not exceed 600 words of text and 6 references and may have no more than 7 authors. They may include up to 2 tables or figures but online supplementary material is not allowed. All letters should include a word count. Letters must not duplicate other material published or submitted for publication. Letters not meeting these specifications are generally not considered. Letters being considered for publication ordinarily will be sent to the authors of the JAMA article, who will be given the opportunity to reply. Letters will be published at the discretion of the editors and are subject to abridgement and editing. Further instructions can be found at http://jama.com/public /InstructionsForAuthors.aspx. A signed statement for authorship criteria and responsibility, financial disclosure, copyright transfer, and acknowledgment and the ICMJE Form for Disclosure of Potential Conflicts of Interest are required before publication. Letters should be submitted via the JAMA online submission and review system at http: //manuscripts.jama.com. For technical assistance, please contact [email protected]. Section Editor: Jody W. Zylke, MD, Deputy Editor.
JAMA Medical Student Elective in Medical Editing JAMA offers a 4-week elective in medical editing to third- and fourthyear US medical students. The elective is full-time and must be taken for academic credit. For information and an application, please email [email protected].
jama.com
JAMA January 6, 2015 Volume 313, Number 1
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of Sussex Library User on 08/10/2015
95
Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. 1. Ebrahim S, Sohani ZN, Montoya L, et al. Reanalyses of randomized clinical trial data. JAMA. 2014;312(10):1024-1032. 2. National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. Tissue plasminogen activator for acute ischemic stroke. N Engl J Med. 1995;333(24):1581-1587. 3. Wardlaw JM, Murray V, Berge E, del Zoppo GJ. Thrombolysis for acute ischaemic stroke. Cochrane Database Syst Rev. 2014;7:CD000213. 4. Kaydos W. Operational Performance Measurement. Boca Raton, FL: CRC Press; 1999. 5. Krumholz HM, Peterson ED. Open access to clinical trials data. JAMA. 2014;312 (10):1002-1003.
In Reply Drs Platts-Mills and Jones have raised a potential concern with the search strategy used in our study on reanalyses of randomized clinical trial data, specifically with respect to our restriction to studies indexed as a clinical trial publication type in MEDLINE. We agree with them that a strategy with no restriction to clinical trials publication type would be more sensitive in detecting reanalyses. However, performing such a search would yield a total of 112 330 citations (limited to March 9, 2014, as per our original search). Screening this number of citations would be a daunting task and may not be a valuable use of resources if the output of additional studies retrieved would be few in number compared with the studies retrieved in our original search. We have attempted to get an estimate of how many studies we potentially missed by using a simplified search with (reanalysis [ti] OR reanalysis [ti]) NOT Clinical Trial [PT] AND random* (limited to March 9, 2014) that would capture the reanalysis of the NINDS rt-PA Stroke Study. Besides the NINDS study, we also identified eligible reanalyses for 3 more trials. Two of the 4 new eligible trials1-4 had their reanalyses completed by independent authors or committees,2,3 and the other 2 had reanalyses completed by the same research group or included authors from the original study.1,4 Reasons for completing reanalyses included the use of intention-to-treat principles (compared with per-protocol analyses used in the original article),1 correction for subgroup imbalances,3 use of a method for identifying latent trajectories (latent class growth analysis),4 and comparison of the primary end point using a change in the definition of the outcome.2 Of the 4 reanalyses, 1 found that the experimental treatment (bilavirudin) was significantly more effective and had less bleeding complications than control (heparin),1 whereas the original analysis had shown no significant difference in efficacy, so perhaps this would lead to more patients being treated with bilavirudin. Another reanalysis4 found effects for latent classes of participants that were not obvious in the original analysis, so it is possible that different patients may be treated differently based on these results. We have no doubt that a few more trials must have been missed given the imperfect sensitivity of any searches and the lack of a standard tag for reanalyses in PubMed. However, we
believe that it is unlikely that the number of missed published reanalyses is large. Measures to improve awareness of the importance and systematic indexing of reanalyses in PubMed are certainly required. Nonetheless, our main conclusions that published reanalyses are currently rare, almost always involve authors of the original analyses, and about one-third reach different conclusions still hold true. If there were more interest in publishing reanalyses regardless of their results, we suspect that the proportion of those claiming that they had found something different from the original analysis would decrease. Shanil Ebrahim, PhD John P. A. Ioannidis, MD, DSc Author Affiliations: Meta-Research Innovation Center at Stanford, Stanford University, Stanford, California. Corresponding Author: John P. A. Ioannidis, MD, DSc, Meta-Research Innovation Center at Stanford, Stanford University, 1265 Welch Rd, Stanford, CA 94305 ([email protected]). Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. 1. Bittl JA, Chaitman BR, Feit F, Kimball W, Topol EJ. Bivalirudin versus heparin during coronary angioplasty for unstable or postinfarction angina: final report reanalysis of the Bivalirudin Angioplasty Study. Am Heart J. 2001;142(6):952-959. 2. Moses RG. Achieving glycosylated hemoglobin targets using the combination of repaglinide and metformin in type 2 diabetes: a reanalysis of earlier data in terms of current targets. Clin Ther. 2008;30(3):552-554. 3. Ingall TJ, O’Fallon WM, Asplund K, et al. Findings from the reanalysis of the NINDS tissue plasminogen activator for acute ischemic stroke treatment trial. Stroke. 2004;35(10):2418-2424. 4. Lennon MC, McAllister W, Kuang L, Herman DB. Capturing intervention effects over time: reanalysis of a critical time intervention for homeless mentally ill men. Am J Public Health. 2005;95(10):1760-1766.
Risk Profile of Clinical Trial Participants To the Editor In their research letter, Dr Udell and colleagues1 stated that clinical trial participants had a lower risk profile and a more favorable prognosis compared with the broader population included in the National Cardiovascular Data Registry (NCDR) Acute Coronary Treatment and Intervention Outcomes Network Registry–Get With The Guidelines. It is likely that these differences would be more pronounced if clinical trial participants were compared with patients with myocardial infarction in the broader US population because participation in the NCDR is voluntary and includes participation by only 5% of US hospitals.1,2 In addition, the data would be put in better perspective if the age of the patients with myocardial infarction in the NCDR and the rates of diagnostic cardiac catheterization and percutaneous coronary intervention were compared with a more inclusive database such as the Nationwide Inpatient Sample.2 It would also be interesting to compare the characteristics of patients presenting to the NCDR hospitals that did and did not enroll patients in clinical trials because there may be relevant differences in clinical presentation and outcomes. In addition, the better outcomes of participants in clinical trials may be expected from the exclusion criteria of the clinical trial protocols, self-selection of clinical trial participants (who of-
jama.com
JAMA January 6, 2015 Volume 313, Number 1
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of Sussex Library User on 08/10/2015
93
Letters
Table. Risk of In-Hospital Cardiovascular Events Among Trial Participants, Eligible Nonparticipants, and Ineligible Nonparticipantsa Rate of Events, % Eligible Trial Nonparticipants Participants (n = 72 313) (n = 3789) 1.5 2.4
Death Death or myocardial infarction
2.4
3.1
1.37 (1.01-1.86)
8.9
Cardiogenic shock
2.3
3.5
1.50 (1.16-1.96)
6.2
Congestive heart failure
4.4
5.5
1.03 (0.87-1.21)
8.9
Major bleedingc
6.4
6.5
1.11 (0.95-1.30)
11.8
creatinine, and troponin ratio to the upper limit of the local assay normal reference), medical insurance type, within-center correlation and clustering of variable responses using an exchangeable correlation structure covariance matrix, Acute Coronary Treatment and Intervention Outcomes Network bleeding outcome, and medication use (aspirin, clopidogrel, warfarin, β-blocker, angiotensin-converting enzyme inhibitor, angiotensin receptor blocker, aldosterone antagonist, statin, and other lipid-lowering therapy).
Abbreviation: OR, odds ratio. a
Among patients who were not transferred prior to discharge from hospitals that enrolled at least 1 patient in a clinical trial.
b
Adjusted for demographic variables (age, sex, race/ethnicity, and weight), signs and symptoms at presentation (heart failure, heart rate, systolic blood pressure, electrocardiographic findings), medical history (hypertension; diabetes mellitus; current or recent tobacco use; hypercholesterolemia; prior myocardial infarction, coronary revascularization, heart failure, stroke, peripheral artery disease), laboratory results (baseline hemoglobin, serum
ten have better health behaviors), and possibly due to earlier treatment of conditions discovered during protocol-directed examinations.3 William J. Kostis, MD, PhD Sri Ram Pentakota, MD, MPH, PhD Abel E. Morerya, MD Author Affiliations: Cardiology Division, Massachusetts General Hospital, Boston (Kostis); Cardiovascular Institute, Rutgers Robert Wood Johnson Medical School, New Brunswick, New Jersey (Pentakota, Morerya). Corresponding Author: William J. Kostis, MD, PhD, Cardiology Division, Massachusetts General Hospital, 55 Fruit St, Boston, MA 02114 (wkostis@mgh .harvard.edu). Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. 1. Udell JA, Wang TY, Li S, et al. Clinical trial participation after myocardial infarction in a national cardiovascular data registry. JAMA. 2014;312(8):841-843. 2. Healthcare Cost and Utilization Project. Nationwide Inpatient Sample. http://hub .healthdata.gov/dataset/hcup-nationwide-inpatient-sample. Accessed September 10, 2014. 3. Kostis WJ, Cabrera J, Messerli FH, et al. Competing cardiovascular and noncardiovascular risks and longevity in the systolic hypertension in the elderly program. Am J Cardiol. 2014;113(4):676-681.
In Reply Dr Kostis and colleagues suggest that patients with myocardial infarction who participate in clinical trials may demonstrate even more pronounced differences in presenting factors, hospital care, and outcomes compared with patients included in the US Healthcare Cost and Utilization Project Nationwide Inpatient Sample. Although such a comparison could potentially bolster our findings, we are not aware of data from the sample documenting clinical trial participation to allow for a valid comparison.1 We agree with Kostis et al that among patients with myocardial infarction, trial participants often receive closer medical supervision by research staff and invested parties than otherwise occurs in routine practice.2 For instance, we observed that clinical trial participants more frequently received care for 94
Adjusted OR (95% CI)b 1.63 (1.11-2.39)
Rate of Events for Ineligible Nonparticipants, % (n = 33 912) 8.1
c
A major bleeding event was defined as bleeding that occurred within the first 24 hours of admission and was a criterion for trial ineligibility.
myocardial infarction meeting quality metrics, noninvasive and invasive diagnostic procedures, discharge counseling, and cardiac rehabilitation referral. Increased medical supervision may be associated with improved outcomes either because high-quality practitioners participate as investigators in clinical trials (eg, physician adherence to guideline recommendations) or as a result of the Hawthorne effect, a phenomenon of alteration and improvement in patient behavior resulting from participant awareness of being under observation (eg, patient adherence to prescribed therapy). The cardiovascular trial literature has documented many examples of clinical research participants faring better than anticipated compared with epidemiological reports.2,3 Although it is possible that trial participants may have been managed in hospitals that practice higher-quality care, our statistical models adjusted for clustering of patients within hospitals. When the multivariable analysis of inpatient outcomes was limited to patients at hospitals that enrolled at least 1 patient in a clinical trial, results were nearly identical (Table). Because the type of clinical trial that patients participated in may have evaluated a drug, strategy, or device, we cannot ascertain how results might differ with participation in a trial of a specific intervention. We nevertheless agree it is likely that a selection bias occurs during trial recruitment at the patient level. There are implicit (sometimes discretionary) characteristics that make a preferable trial participant, such as language comprehension, reliability, and compliance in following trial requirements, as well as general assessments of health and frailty by research personnel that may not be apparent in captured data. Thus, our observations suggest an opportunity for broader efforts to improve trial participation and enroll patients at similar risk to those in whom the therapies will ultimately be used. Specifically, there are large numbers of potentially eligible patients within existing registries to recruit into trials, a strategy increasingly being used successfully in cardiovascular outcomes research.4,5
JAMA January 6, 2015 Volume 313, Number 1
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of Sussex Library User on 08/10/2015
jama.com
Letters
Performing more clinical trial research nested within such registries would allow leveraging of established resources and funding to minimize redundancy in study coordination, data collection and monitoring, clinical follow-up, regulatory issues, and financial barriers to participation from an institutional perspective while simultaneously improving testing of therapies in heterogeneous patient populations seen in routine clinical practice. Jacob A. Udell, MD, MPH Matthew T. Roe, MD, MHS Stephen D. Wiviott, MD Author Affiliations: Women’s College Hospital, University of Toronto, Toronto, Ontario, Canada (Udell); Duke Clinical Research Institute, Duke University, Durham, North Carolina (Roe); Cardiovascular Division, Brigham and Women’s Hospital, Boston, Massachusetts (Wiviott).
CORRECTION Cost Correction: In the article from The Medical Letter on Drugs and Therapeutics entitled “Drugs for MRSA Skin and Soft-Tissue Infections” published in the October 15, 2014, issue of JAMA (2014;312[15]:1583-1584. doi:10.1001/jama.2014 .13385), the cost of tedizolid in Table 1 was reported as $2950 for 10 days’ treatment. Cubist’s tedizolid (Sivextro) received FDA approval for 6 days’ treatment of acute bacterial skin and skin structure infection, which would cost $1770. This article was corrected online. Incorrect Reporting: In the Original Investigation entitled “Effect of Reversible Intermittent Intra-abdominal Vagal Nerve Blockade on Morbid Obesity: The ReCharge Randomized Clinical Trial,” published in the December 3, 2014, issue of JAMA (2014;312[9]:915-922. doi:10.1001/jama.2014.10540), the reason of device removal was incorrectly reported. On page 918, the third sentence of the fifth full paragraph should read, “Three patients in the vagal nerve block did so because 1 had experienced pain at the neuroregulator site; 1 experienced pain with therapy, and 1 experienced heartburn.” The fifth sentence of the same paragraph should read, “Two patients in the vagal nerve block group and 4 in the sham group asked to have the devices removed.” This article was corrected online.
Corresponding Author: Jacob A. Udell, MD, MPH, Women’s College Hospital, University of Toronto, 76 Grenville St, Toronto, ON M5S 1B1, Canada (jay.udell @utoronto.ca). Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Udell reported receiving honoraria from the American College of Cardiology Foundation. Dr Roe reported receiving grants from Eli Lilly & Company, Daiichi Sankyo, sanofi-aventis, the American College of Cardiology, the American Heart Association, and the Familial Hypercholesterolemia Foundation; and personal fees from Merck, Janssen Pharmaceuticals, Regeneron, AstraZeneca, and Amgen outside the submitted work. Dr Wiviott reported receiving grants from Arena, AstraZeneca, Bristol-Myers Squibb, Daiichi Sankyo, Eisai, Eli Lilly, Merck, and sanofi-aventis; and consulting fees from Aegerion, Angelmed, Arena, AstraZeneca, Boston Clinical Research Institute, Bristol-Myers Squibb, Daiichi Sankyo, Eisai, Eli Lilly, ICON Clinical, Janssen, and Xoma. 1. Healthcare Cost and Utilization Project (HCUP). HCUP National Inpatient Sample Database Documentation. http://www.hcup-us.ahrq.gov/db/nation/nis /nisdbdocumentation.jsp. Accessed October 6, 2014. 2. Kandzari DE, Roe MT, Chen AY, et al. Influence of clinical trial enrollment on the quality of care and outcomes for patients with non-ST-segment elevation acute coronary syndromes. Am Heart J. 2005;149(3):474-481. 3. Steg PG, López-Sendón J, Lopez de Sa E, et al; GRACE Investigators. External validity of clinical trials in acute myocardial infarction. Arch Intern Med. 2007;167 (1):68-73. 4. Fröbert O, Lagerqvist B, Olivecrona GK, et al; TASTE Trial. Thrombus aspiration during ST-segment elevation myocardial infarction. N Engl J Med. 2013;369(17):1587-1597. 5. Rao SV, Hess CN, Barham B, et al. A registry-based randomized trial comparing radial and femoral approaches in women undergoing percutaneous coronary intervention: the SAFE-PCI for Women (Study of Access Site for Enhancement of PCI for Women) trial. JACC Cardiovasc Interv. 2014;7(8):857-867.
Guidelines for Letters Letters discussing a recent JAMA article should be submitted within 4 weeks of the article's publication in print. Letters received after 4 weeks will rarely be considered. Letters should not exceed 400 words of text and 5 references and may have no more than 3 authors. Letters reporting original research should not exceed 600 words of text and 6 references and may have no more than 7 authors. They may include up to 2 tables or figures but online supplementary material is not allowed. All letters should include a word count. Letters must not duplicate other material published or submitted for publication. Letters not meeting these specifications are generally not considered. Letters being considered for publication ordinarily will be sent to the authors of the JAMA article, who will be given the opportunity to reply. Letters will be published at the discretion of the editors and are subject to abridgement and editing. Further instructions can be found at http://jama.com/public /InstructionsForAuthors.aspx. A signed statement for authorship criteria and responsibility, financial disclosure, copyright transfer, and acknowledgment and the ICMJE Form for Disclosure of Potential Conflicts of Interest are required before publication. Letters should be submitted via the JAMA online submission and review system at http: //manuscripts.jama.com. For technical assistance, please contact [email protected]. Section Editor: Jody W. Zylke, MD, Deputy Editor.
JAMA Medical Student Elective in Medical Editing JAMA offers a 4-week elective in medical editing to third- and fourthyear US medical students. The elective is full-time and must be taken for academic credit. For information and an application, please email [email protected].
jama.com
JAMA January 6, 2015 Volume 313, Number 1
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of Sussex Library User on 08/10/2015
95
