To the Editor We write to report a series of pervasive errors in our Research Letter, “Cardiac Rehabilitation Use Among Veterans With Ischemic Heart Disease,” published in the October 2014 issue of JAMA Internal Medicine.1 The errors, once corrected, result in changes to a number of findings. Thus, we have requested that the original article be retracted and replaced.
In the original article,1 we reported the proportion of eligible veterans in the Veterans Health Administration (VA) with ischemic heart disease who participated in cardiac rehabilitation (CR) after hospitalization for myocardial infarction, percutaneous coronary intervention, and/or coronary artery bypass graft; whether the presence of an on-site CR program was associated with greater participation; and the characteristics of the CR participants.
The primary problem with the originally reported study was incorrect cohort identification (ie, we missed many patients who were eligible for cardiac rehabilitation). We did not query all relevant codes used to identify the cohort of patients with ischemic heart disease. For example, some International Classification of Diseases, Ninth Revision (ICD-9), procedure codes were not included in the initial query. Therefore, the cohort originally used had fewer participants than the current cohort when we repeated our query. However, all subsequent coding regarding the analysis of the cohort was unchanged.
The errors in the original article, once corrected, result in a substantial increase in the numbers of eligible patients included in the study, from 47 051 to 88 826. Despite the almost doubling of eligible patients in the corrected analysis, there was only a small change in the proportion of patients participating in CR (from 8.5% to 10.3%), and the overall proportion of patients participating in CR remains low.
However, after correcting for the coding cohort identification errors, there are several changes in the findings reported in the corrected version compared with the originally published version. In the original article,1 we reported no statistically significant difference in CR participation by race, but the corrected version shows a difference: “10.4% of whites (7126 of 68 259) and 9.8% of nonwhites (1610 of 16 397) received any CR (P = .02).” Also, in the original article,1 we reported that participation rates increased from 7.5% to 9.4% (P < .01) from 2007 to 2011, but the corrected version indicates that participation rates “remained stable between 8.5% and 8.7% for all years.”
In addition, there are a number of changes in the results of the multivariate model that assessed participant factors associated with participation in CR, with associations changing statistical significance for participants with income greater than $33 000 and those with heart failure, stroke, peripheral vascular disease, chronic obstructive pulmonary disease, and depression.
As a result of these errors, corrections are needed to the original article’s text, Table, Figure, and eTables. Because of the large increase in the cohort size, many of the associations in eTable 2 and eTable 3 are now statistically different. In eTable 1, the number of participants with corresponding diagnoses has changed with the updated cohort. The statistical differences presented in the Figure depicting the percent participation by diagnosis remain the same, but the individual percentages differ from those reported in the original article.
The overall conclusions of our study remain unchanged. Both the presence of an on-site CR program and patient proximity to a VA facility were associated with greater participation in CR; however, participation was low regardless of the presence or absence of an on-site program.
We sincerely regret these errors as well as the confusion caused to the readers and editors of JAMA Internal Medicine. The text, Table, Figure, and eTables of the original article have been corrected and replaced online.1 An online supplement has been added that includes a version of the original article with the errors highlighted and a version of the replacement article with the corrections highlighted.
Corresponding Author: David W. Schopfer, MD, MAS, San Francisco VA Medical Center, 4150 Clement St, San Francisco, CA 94121 (firstname.lastname@example.org).
Published Online: October 10, 2016. doi:10.1001/jamainternmed.2016.5831
Conflict of Interest Disclosures: Dr Whooley has received research funding from Janssen Healthcare Innovations. No other disclosures are reported.
Schopfer DW, Takemoto S, Allsup K, Helfrich CD, Ho PM, Forman DE, Whooley MA. Notice of Retraction and Replacement. Schopfer DW, et al. Cardiac Rehabilitation Use Among Veterans With Ischemic Heart Disease. JAMA Intern Med. 2014;174(10):1687-1689. JAMA Intern Med. 2016;176(11):1726-1727. doi:10.1001/jamainternmed.2016.5831