Author + information
- Received January 25, 2019
- Revision received March 7, 2019
- Accepted March 14, 2019
- Published online May 20, 2019.
- Benjamin A. Steinberg, MD, MHSa,∗ (, )
- Paul Dorian, MDb,
- Kevin J. Anstrom, PhDc,
- Rachel Hess, MD, MSa,
- Daniel B. Mark, MDc,
- Peter A. Noseworthy, MDd,
- John A. Spertus, MD, MPHe and
- Jonathan P. Piccini, MD, MHSc
- aDepartment of Medicine, University of Utah Health Sciences Center, Salt Lake City, Utah
- bDivision of Cardiology, University of Toronto, St. Michael's Hospital, Toronto, Ontario, Canada
- cDuke Clinical Research Institute, Durham, North Carolina
- dDepartment of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota
- eCardiovascular Research, Department of Biomedical and Health Informatics, Saint Luke's Mid-America Heart Institute, Kansas City, Missouri
- ↵∗Address for correspondence:
Dr. Benjamin A. Steinberg, Division of Cardiovascular Medicine, University of Utah Health Sciences Center, 30 North 1900 East, Room 4A100, Salt Lake City, Utah 84132.
Objectives This study sought to determine how frequently patient-reported outcomes (PROs) are collected in registered clinical studies of atrial fibrillation (AF).
Background Improving symptom burden and quality of life are important goals in the treatment of AF and are best measured with PROs.
Methods Data from Clincaltrials.gov were studied to identify PROs in AF studies. All studies reporting AF as the disease condition were included, and PROs were identified by search terms within the outcome measures field. Generic and AF-specific PROs were identified and assessed by study type and year. Clinicaltrials.gov reporting was compared with published reports of linked studies in PubMed.
Results From 1999 to 2018, 1,709 studies including AF patients were posted; 238 studies (14%) included PROs. Collection of PROs was reported in 22% of trials (n = 83 of 386) studying procedural interventions and 11% of all Phase 3 studies (n = 18 of 168). Among the 238 studies with PROs, most described “quality of life” (n = 194 [82%]), and most (n = 198 [83%]) included only generic (not AF-specific) PROs. Only 17% of studies (n = 40) reporting PROs specified a previously published AF-specific tool, most commonly the AFEQT (Atrial Fibrillation Effect on QualiTy-of-life) form (n = 20 [8.4%]). Among the available PubMed citations of 391 studies, 74 studies (19%) described collecting a specific PRO tool (n = 29 [7.4%]) for an AF-specific PRO.
Conclusions Despite increased emphasis on the importance of PROs in AF, a minority of registered clinical trials reported collecting PROs, with very few using validated, AF-specific PROs. Improving outcomes that are most important to patients will necessitate increased emphasis on these PROs in pivotal clinical studies.
Patient-reported outcomes (PROs) have been defined as “any report of the status of a patient's health condition that comes directly from the patient (i.e., without interpretation of the patient's response by a clinician or anyone else)” (1). Understanding patients’ perspectives is important for atrial fibrillation (AF) because: 1) the subjective experience can vary dramatically among patients (some are asymptomatic whereas others can have debilitating symptoms); 2) AF occurs more commonly in older patients who often prioritize quality of life (QoL) over longevity; and 3) substantial resources are frequently devoted to improving patients’ health status (2,3). The burden of AF symptoms can be comparable to that of ischemic heart disease or congestive heart failure (4), and strategies such as catheter ablation have been shown to improve arrhythmia-free survival and improve health-related quality of life (hrQoL) (5–7). Yet there is limited evidence that aggressive rhythm control (compared to symptomatic treatment with rate control) improves clinical endpoints of stroke and/or mortality, thus the primary indication for pursuing a rhythm control strategy remains improvement of AF symptoms (8,9). Despite this, it is unclear whether pivotal studies of AF are directly assessing hrQoL using PROs (8).
Data from the Clinicaltrials.gov repository were used to assess the reported use of PROs in clinical studies of patients with AF. The objectives of the current analysis were to: 1) measure the proportion of registered clinical studies of AF that included PROs as primary outcome measures; 2) assess trends in the use of PROs over time; and 3) assess the use of AF-specific PROs versus generic instruments in trials of AF. Results of the Clinicaltrials.gov analysis were then validated in PubMed searches for the registered studies.
Available data from the Clinicaltrials.gov database were analyzed to identify registered AF studies that reported use of PROs as outcomes. Data for all studies registered under the condition “atrial fibrillation” were downloaded and were up to date as of July 10, 2018. No other search filters were used. The studies were further restricted to only those that indicated “atrial fibrillation” in the “conditions” field. Studies were further stratified by inclusion of PRO terms in the “outcome measures” field (see Online Search Terms in the Online Appendix). The presence of any PRO term indicated reported use of PROs in the study. These included generic PRO measurement tools, generic terms for PROs (e.g., “quality of life”), as well as named, AF-specific PRO tools (4,10–12). Clinician-reported outcomes such as the European Heart Rhythm Association or Canadian Cardiovascular Society classifications of AF were not included (see Online Search Terms in the Online Appendix for more detailed methods).
Characteristics of studies reporting PRO collections were compared to those that did not report PROs as outcome measurements. Reporting of PRO collection was assessed over the duration of the reporting and stratified by study type (interventional vs. observational studies; expanded access was classified as observational) and intervention type (device or procedure only vs. drug therapy only; studies with both types of interventions [or other interventions] were excluded from that stratification). Among studies that did report PROs, the use of generic PRO tools versus AF-specific PROs is described.
In order to assess under- or over-reporting of PROs in the Clinicaltrials.gov database, PubMed was searched for any related citation for each of the identified AF studies. Citations were matched based on Clinicaltrials.gov identifier, and full-text manuscripts were reviewed (when accessible). Citations were examined for PRO collection as part of the study methods, as well as the reporting of PROs results in any citation linked to the respective identifier. Because the objective was to assess study use and collection of PROs, all registered studies were included in the search (both completed and incomplete), and “Rationale and Design” manuscripts were reviewed for outcomes measurements. Because the description of possible PROs is variable, citations were examined for any report of measuring QoL outcomes, for the use of a specific tool to collect PROs, and for the use of a PRO tool specific to AF. Simply, the collection of “symptomatic AF” as an outcome (as interpreted by the clinician) was not considered a collection of PROs or QoL.
Categorical variables were summarized as number (percentage), and univariate comparisons were performed using the chi-square test. A 2-sided p value of <0.05 was considered significant. All analyses were performed using R version 3.5.2 software (R Foundation, Vienna, Austria) and RStudio version 1.1.463 software (R Foundation, Vienna, Austria) (13) with applicable, additional packages (14).
A total of 1,709 entries were identified in Clinicaltrials.gov that included patients with AF and were first posted from October 28, 1999, to June 29, 2018. Overall, 43% (n = 735) were reported completed; 35.4% (n = 260) of completed studies had a linked citation in PubMed; and 90% (n = 1,534) did not have results available in Clinicaltrials.gov. Among the 1,709 registered AF studies, 14% (n = 238) reported PRO terms in their outcome measurements (Table 1). Studies reporting PROs as outcomes versus those that did not collect PROs were more likely to be classified as interventional (82% vs. 64%, respectively; p < 0.001), to be of smaller size (11% targeting >1,000 participants vs. 19%, respectively; p < 0.001), and more likely to involve a procedural intervention (35% vs. 21%, respectively; p < 0.001). However, among the 386 studies with specifically procedural interventions, only 22% (n = 83) reported collecting PROs.
Proportions of studies reporting PRO collection by year grouping are shown in Figure 1 (stratified by study type) and in the Central Illustration (stratified by intervention type). During the period where Clinicaltrials.gov data were available, the number of posted studies including patients with AF increased from 382 for the first 10 years of reporting (1999 to 2009) up to 705 in the most recent 3 years (2015 to 2018).
Reporting of generic PRO tools and terms versus AF-specific instruments are shown in Table 2, stratified by study type. Of the 238 entries that reported PROs, 40 (17%) referred to an AF-specific instrument, and this was not significantly different between interventional and observational studies (p = 0.90). Among AF-specific PRO measurements, the AFEQT (Atrial Fibrillation Effect on QualiTy-of-life) tool was most commonly cited (n = 20 [8.4% of studies with PRO terms]). Thirty-five studies reported collecting both an AF-specific and a generic PRO measurement, none of which were Phase 3 studies.
Among the 1,709 studies of AF in Clinicaltrials.gov, 23% (n = 391) had at least 1 citation in PubMed associated with the Clinicaltrials.gov registration number (Table 3), including either “Rationale and Design” papers or publication of some results; the majority of these had >1 full-text article that was available and accessible (n = 371 [95% of studies with citations]).
Of the 1,471 studies without PROs described as outcomes in Clinicaltrials.gov, 3% (n = 46) had PubMed entries that described PRO collection. Among these 46 studies, 36 described a specific PRO tool (e.g., EQ-5D), and 16 described an AF-specific PRO tool. Overall, among all of the 1,709 registered studies, 3.8% (n = 65) describe the use of an AF-specific PRO in either the registered outcomes or a linked PubMed citation.
Over the past few decades, there has been increasing recognition of the importance of PROs in the assessment of outcomes of clinical research studies, especially in chronic, nonfatal conditions. Our understanding of the best methods to create and validate PRO tools has evolved, and new disease-specific measurement tools for AF have been introduced. However, in contrast to our expectations, this analysis of AF studies in the Clinicaltrials.gov registry found a low proportion of studies reporting PRO collection, and this proportion has not improved over time. Thus, although there has been substantial growth in the number of studies of AF patients, only 1 in 7 studies includes PRO terms as outcomes measurements. Pivotal Phase 3 and intervention-based studies do not show any substantially higher use of PROs. Finally, among studies that appeared to report PROs, only 17% included AF-specific PRO tools (vs. more generic tools or terms).
Although controlling symptoms in patients with AF can be challenging, it is often the most important outcome to patients. Interventions to improve symptoms, particularly those geared toward rhythm control (e.g., catheter ablation) can be costly and are not without risk of harm. Furthermore, although emerging data suggest improved thromboembolic and reduced mortality outcomes in subgroups of AF patients undergoing ablation (15), uncertainty remains as to whether these findings are applicable to the broader AF population.
Therefore, the current objective of rhythm control strategies in most patients remains symptom control; yet measurements of these vital outcomes remain challenging and are not being routinely collected or reported in the largest repository of clinical trials. Furthermore, “treatment success,” as defined by subsequent AF burden and freedom from recurrence, is poorly correlated to QoL. Thus, “arrhythmia burden” is not an appropriate surrogate for PROs, and this has limited the interpretation of many AF trials to date. It is not clear that the endpoints most important to patients are measured; therefore, it is difficult to know which interventions are of highest value to them. The use of arrhythmia burden in existing trials is a valuable endpoint for clinicians and researchers but of little value in understanding hrQoL implications of our interventions, which requires PROs.
These data may also reflect the state of AF clinical research generally. A recent analysis of only interventional AF studies in the Clinicaltrials.gov dataset demonstrated some variability in the quality of trials, many of which are of modest size, and the use of randomization and blinding is inconsistent (16). The present analysis focused on an even broader cohort of registered AF studies (all phases, all designs, any status), and these shortcomings appear to extend to a lack of PROs, particularly a lack of AF-specific PROs, despite the development and validation of several AF-specific PRO measurement tools (17) that have been studied extensively. Furthermore, there are robust data supporting the use of AF-specific PRO tools over general hrQoL assessments (18). Last, the latest consensus on AF ablation “recommends that all clinical trials incorporate some measure of patient reported outcomes and preferably measure them using both a general and an AF specific measurement scale” (9). More detailed guidance on PRO incorporation in clinical trials has recently been provided by the SPIRIT-PRO (Standard Protocol Items: Recommendations for Interventional Trials - Patient Reported Outcomes) group (19). Nevertheless, the present data demonstrate suboptimal implementation of PROs in AF research, according to the Clinicaltrials.gov registry.
Although the Clinicaltrials.gov dataset may be limited and exclude smaller or observational studies that are not required to be registered, it remains the de facto registration repository for landmark and regulatory pathway studies, and registration is often required by journals before considering publication of study results. Therefore, the absence of PROs among studies included in this dataset represents an informative if potentially incomplete assessment of their use in AF clinical research. Because there has been increasing scrutiny of the use of brief-duration AF as an endpoint in clinical trials (20) and because such events are often clinically inconsequential to patients, PROs offer a much better opportunity to demonstrate value of an intervention to patients. Furthermore, optimal and efficient implementation of PROs in studies, both generic and disease-specific, could be greatly enhanced with their routine incorporation into clinical care and the electronic health record (21,22). This would also facilitate additional, much needed research, such as: 1) identifying minimum clinically important changes in PROs for AF; 2) understanding the correlation between PROs and more “traditional” outcomes of arrhythmia recurrence, arrhythmia burden, and clinical events (e.g., thromboembolism, heart failure); and 3) using PROs to guide treatment decisions.
The data analyzed here are derived from a publicly available registry, where data are entered manually and often voluntarily by study personnel. Not all data elements are required (requirements change over time), and many studies are completed or not without results reported; only primary (and not secondary) outcomes are downloaded (23). Additionally, changes over time in the patterns of trial registration (e.g., study types and so forth) and availability of PROs could influence the assessment of chronological trends. Finally, responsible persons may not have included PROs in the Clinicaltrials.gov registration, even if they were collected. However, this implies that such outcomes were neither the primary outcome, nor were they believed to be of sufficient import to include in this registry (although they may have been identified in the present PubMed review). This in itself reflects an under-emphasis of outcomes that are of primary concern to patients with AF. As technologies rapidly develop to more effectively detect and treat AF, it is incumbent on the field to ensure we are improving outcomes that matter to patients.
Despite increased emphasis on research for AF interventions, a minority of registered AF clinical trials reported the use of PROs as outcome measurements, and very few described collecting validated, AF-specific PROs. These data reflect an under-emphasis of outcome measurements most relevant to patients and among the most important studies in AF. Improving the care of patients with AF will necessitate increased emphasis on these PROs in the pivotal clinical trials of contemporary interventions.
COMPETENCY IN MEDICAL KNOWLEDGE: Despite recommendations from international governing bodies that PROs be included in all clinical studies of AF, a minority of trials cited them as primary outcomes in Clinicaltrials.gov registrations.
TRANSLATIONAL OUTLOOK: Although there is contemporary evidence suggesting improved clinical events and reduced mortality among patients undergoing rhythm control for AF, the primary objective of this approach remains symptom control. However, validated AF PROs are cited as primary outcomes in a minority of these registered trials. Understanding the impact of interventions on these outcomes that frequently matter most to patients will require greater emphasis and commitment to PROs in clinical studies of AF.
Supported by U.S. National Institutes of Health/National Heart, Lung, and Blood Institute award K23HL143156 to Dr. Steinberg. The content is the sole responsibility of the authors and does not necessarily represent the official views of the NIH. Dr. Piccini has received research funding from Abbott, Boston Scientific, and Gilead; and is a consultant for Abbott, Allergan, ARCA Biopharma, Biotronik, Johnson & Johnson, LivaNova, Medtronic, Sanofi, and Phillips. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- atrial fibrillation
- health-related quality of life
- patient-reported outcomes
- Received January 25, 2019.
- Revision received March 7, 2019.
- Accepted March 14, 2019.
- 2019 American College of Cardiology Foundation
- Kim M.H.,
- Johnston S.S.,
- Chu B.C.,
- Dalal M.R.,
- Schulman K.L.
- Dorian P.,
- Jung W.,
- Newman D.,
- et al.
- Piccini J.P.,
- Lopes R.D.,
- Kong M.H.,
- Hasselblad V.,
- Jackson K.,
- Al-Khatib S.M.
- Al Halabi S.,
- Qintar M.,
- Hussein A.,
- et al.
- January C.T.,
- Wann L.S.,
- Alpert J.S.,
- et al.
- Spertus J.,
- Dorian P.,
- Bubien R.,
- et al.
- Wokhlu A.,
- Monahan K.H.,
- Hodge D.O.,
- et al.
- ↵(2017) R: A Language and Environment for Statistical Computing. Vienna, Austria. R Foundation for Statistical Computing, Available at:. http://www.R-project.org/. 2032019.
- ↵Yoshida K, Bohn J. Tableone: Create “Table 1” to Describe Baseline Characteristics (R package). 0.9.2 ed, 2018. R Foundation for Statistical Computing, 2017. Available at: http://www.R-project.org/. Accessed March 20, 2019.
- Marrouche N.F.,
- Brachmann J.,
- Andresen D.,
- et al.
- Patel R.B.,
- Venkateswaran R.V.,
- Singh A.,
- et al.
- Mark D.B.
- Steinberg J.S.,
- O'Connell H.,
- Li S.,
- Ziegler P.D.
- Stehlik J.,
- Rodriguez-Correa C.,
- Spertus J.A.,
- et al.
- Biber J.,
- Ose D.,
- Reese J.,
- et al.
- Tse T.,
- Fain K.M.,
- Zarin D.A.