Defining Clinically Important Difference in the Atrial Fibrillation Effect on Quality-of-Life Score
The Atrial Fibrillation Effect on Quality-of-Life (AFEQT) questionnaire has recently been validated to measure the impact of atrial fibrillation on quality of life, but a clinically important difference in AFEQT score has not been well defined.
Methods and Results:
To determine the clinically important difference in overall AFEQT (score range= 0 [worst] to 100 [best]) and selected subscales, we analyzed data in the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) registry, a United States-based outpatient atrial fibrillation registry. AFEQT was assessed at baseline and 1 year in a subset of 1347 ORBIT-AF patients from 80 US sites participating in ORBIT-AF from June 2010 to August 2011. The mean change method was used to relate changes in 1-year AFEQT scores to clinically important changes in the physician assessment of European Heart Rhythm Association functional status (1 class improvement and separately 1 class deterioration). Clinically important differences and 95% CI corresponding to either a 1 European Heart Rhythm Association class improvement or deterioration were 5.4 (3.6–7.2) and −4.2 (−6.9 to −1.5) AFEQT points, respectively. Similarly, clinically important difference values were seen for a 1 European Heart Rhythm Association class improvement for the AFEQT subscales Activities of Daily Living and Symptoms: 5.1 (2.5–7.6) and 7.1 (5.3–9.0) AFEQT points, respectively.
Based on the anchor of 1 European Heart Rhythm Association class change, changes in AFEQT score of + or −5 points are clinically important changes in patients’ health.
Clinical Trial Registration:
URL: https://clinicaltrials.gov. Unique identifier: NCT01165710.
WHAT IS KNOWN
Atrial fibrillation frequently impairs patients’ quality of life and many treatment decisions in atrial fibrillation are focused on improving patients’ quality of life.
A clinically important difference in Atrial Fibrillation Effect on Quality-of-Life score has not been well defined.
WHAT THE STUDY ADDS
Changes in Atrial Fibrillation Effect on Quality-of-Life score, of ≈5 points, are clinically relevant.
Study design of therapeutic interventions should take into account the proportion of patients expected to benefit from treatment when applying clinically important difference values.
Atrial fibrillation (AF) increases the risks of stroke, heart failure, cognitive impairment, and death.1 Moreover, AF frequently impairs patients’ quality of life (QoL), often with a magnitude similar to that observed in patients with recent myocardial infarction or heart failure.2 As a result, many treatment decisions in AF (eg, radiofrequency ablation or rhythm control) are focused on improving patients’ QoL.1 Defining clinically meaningful improvements in QoL is, therefore, a priority in assessing AF treatment.
The 20-item Atrial Fibrillation Effect on Quality-of-Life (AFEQT) survey has recently been validated to measure the impact of AF on patients’ QoL.3 The AFEQT’s survey differs from other QoL assessment tools for AF in that it comprehensively assesses patients’ symptom burden, functional impairment, treatment concerns, and treatment satisfaction. Despite its comprehensive assessment, a clinically important difference (CID) in AFEQT score has not been well defined. CIDs are needed to interpret the meaningfulness of observed differences in patient outcomes. By identifying a CID for individual patients, we can also consider mean differences that would be expected in populations or clinical trials, depending on the proportion of patients who experience a CID versus alternative outcomes.
Using the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF), a large, contemporary, prospective, community-based outpatient cohort, we aimed to (1) identify CID values in the overall cohort for the overall AFEQT score and subscales, (2) identify CID values for the overall AFEQT score across predefined subgroups, and (3) characterize the implication to treatment comparisons.
ORBIT-AF is a prospective, nationwide, United States-based outpatient registry enrolling patients from cardiology, electrophysiology, and internal medicine practices. Patients were enrolled from 176 sites from June 2010 to August 2011. The rationale, design, and methods of the ORBIT-AF registry have been reported previously.4 Briefly, eligible patients were ≥18 years old, did not have AF due to a transient, reversible cause, and had a life expectancy >6 months. Patients were followed at 6-month intervals up to a maximum of 3 years. Abstracted data on consecutive patients with AF were entered in the Web-enabled case report form and submitted to the ORBIT-AF registry. The case report form included data on age, sex, race/ethnicity, insurance status, education level, medical history, type of AF, AF treatment strategy (rhythm control versus rate control), medical procedure history, vital signs, laboratory data, and current medication use. Additionally, patient-reported outcome (PRO) questionnaires (including the AFEQT survey) where administered to ≈20% of the ORBIT-AF registry population. Patients in the registry received varying treatments prospectively over time. All subjects provided written informed consent. Institutional review boards of the Duke Clinical Research Institute and the participating enrollment sites approved the study.
The AFEQT 20-item questionnaire is a validated, disease-specific QoL tool that assesses the impact of AF on patients’ QoL (St Jude Medical, St Paul, MN, http://afeqt.org/).3 The questionnaire assesses 4 domains of interest: symptoms, daily activities, treatment concerns, and treatment satisfaction and is completed by the patient or a proxy. The overall AFEQT score is calculated using the first 3 domains and ranges from 0 = worst to 100 = best QoL. Patient satisfaction with treatment is not considered to be a part of a patient’s health status and is not included in the summary score calculation. Additionally, for each domain of interest, there is a corresponding subscale score: symptoms, daily activities, treatment concerns, and treatment satisfaction subscale scores. Each subscale ranges from 0 = worst to 100 = best QoL and is scored separately.
European Heart Rhythm Association Score
The European Heart Rhythm Association (EHRA) score is used to assess symptoms during AF from the physician’s perspective.5 The score was proposed by a panel from the German Atrial Fibrillation Competence Network and the EHRA6 with the need for subsequent validation.7 The EHRA score is similar to the New York Heart Association Functional classification system for heart failure. There are 4 classes: no symptoms (EHRA I); mild symptoms, normal daily activity not affected (EHRA II); severe symptoms, normal daily activity affected (EHRA III); and disabling symptoms, normal daily activity discontinued (EHRA IV) ranging from 1 = no symptoms to 4 = disabling symptoms.
The ORBIT-AF registry enrolled 10 137 patients at baseline from 176 sites. The PRO substudy in which the AFEQT questionnaire was administered consisted of 2008 patients at baseline (20% of the enrolled registry population). For our analysis, we restricted the study sample to patients with baseline and 1-year PRO data for a final study sample of n=1347 patients from 80 sites.
Ms Holmes 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. The data, analytic methods, and study materials will not be made available to other researchers for purposes of reproducing the results or replicating the procedure.
Baseline patient characteristics are presented as frequencies and percentages for categorical variables and continuous variables as mean and SD or median and interquartile range. When comparing groups, such as those included and excluded from the study, standardized differences are presented between the groups and characteristics with a standardized difference <10% are considered to be similar.
Anchor-Based Method—Mean Change
CIDs are defined as differences that are clinically important (as determined by the method of quantification) but not necessarily in any sense minimal.8 Defining CIDs is inherently difficult, as the PRO often does not have a gold standard with which to compare responses. As such, a common approach is first to identify another well-accepted standard as a comparison metric for assessing clinical change, then compare individual patients’ distribution of scores between groups of patients that did and did not experience a clinically important change. This approach is typically referred to as an anchor-based method.9 We chose to use the EHRA score,7,10 an AF-specific clinician-reported outcome, as the anchor with which to compare the change in the AFEQT score.11 This was made on the basis that each category of the EHRA score has substantial face validity in representing clinically important changes from the perspective of the physician caring for that patient. Additionally, the Spearman correlation between the EHRA score and the AFEQT score is −0.39, establishing an association between the 2 scores.8,10 As symptom severity increases, QoL decreases. It’s suggested that the correlation coefficient be at least 0.30 to establish a correlation between the measures.
There is not always agreement between patients and physicians on the clinical importance of change and even physicians often disagree on the health status of their patients.12,13 Moreover, this design can detect clinically important changes but not necessarily the minimal CID in scores. Nevertheless, the mean difference in scores between groups of patients that change by 1 EHRA class would clearly be clinically significant in the eyes of most physicians. In the absence of an additional patient-reported questionnaire/item captured in the registry that could be used as an anchor, the EHRA outcome was the best measure available. We chose to define a CID as opposed to a minimal CID because the EHRA score is a clinician-reported outcome and changes between classes are not necessarily the smallest difference that may be important.
The mean change method was used to identify CIDs in AFEQT at 1-year follow-up.14,15 The method defines a CID as the mean change in AFEQT score among patients with a 1 EHRA class change.11 This was done for both a 1 EHRA class improvement and 1 EHRA class deterioration on the anchor as studies suggest the CID for improvement may be different in magnitude than for deterioration on the anchor.16 Additionally, CID values were calculated for 2 AFEQT subscales, Activities of Daily Living (ADL) and Symptoms, using for a 1 class improvement in EHRA score because these subscales are closely related to functional status and symptom severity. This method estimates the mean change in AFEQT that is expected in a population where everyone experiences a clinically relevant improvement in the anchor. Therefore, the interpretation of the CID is within the patient in the sense that it reflects differences that would be expected for a patient who has an important improvement. This is quite different from a CID in average AFEQT at the population level or between 2 treatment groups, where some people improve and others do not improve. To illustrate this distinction, we plotted the mean change in AFEQT that would be expected, on average, if a treatment achieved CIDs in some patients but made no difference in others, where the proportion of patients with a CID varied.
Additionally, to demonstrate comparability across different populations of patients with AF, CID values were calculated for the prespecified subgroups of (1) age <75 versus age ≥75, (2) sex, (3) AF type, (4) AF diagnosis ≤1 year versus AF diagnosis >1 year at baseline enrollment, and (5) congestive heart failure. As a sensitivity analysis, we calculated CID values for different gradations of change in EHRA score: (1) 1 EHRA class improvement from 3 to 2 and (2) 1 EHRA class improvement from 2 to 1 and (3) 2 EHRA class improvement to explore whether the CID values are of similar magnitude between different gradations of EHRA classes. A 1 EHRA class improvement from 4 to 3 was omitted because of small number of patients (n=3).
Anchor-Based Method—Receiver Operating Characteristic
As a secondary analysis, the receiver operating characteristics (ROC) method was used to identify CIDs in AFEQT at 1-year follow-up14,15 for a ≥1 class improvement in EHRA score. The ROC method, which plots sensitivity and (1−specificity) across the range of AFEQT scores, identifies a CID as the threshold for change in AFEQT score that best discriminates patients who experienced an important change in the anchor (≥1 EHRA class change) from those who experienced no change. This is the point on the curve that is closest to the upper-left corner (where the sum of the percentages misclassified patients is the lowest).17 At this point, the joint distribution of sensitivity and specificity is maximized. Empirical 95% CIs for the ROC CID were obtained by bootstrapping.18 This method was treated as a secondary analysis rather than primary because the categorization of continuous data results in imprecise estimates of the CID and large CIs. Statistical analyses were performed using SAS software (version 9.4, SAS Institute, Cary, NC).
Study Sample Characteristics
For the study sample (n=1347), the median (interquartile range) AFEQT baseline score was 83.3 (68.5–93.5), and the mean±SD AFEQT baseline score was 79.0±18.5. Additionally, the median (interquartile range) change in AFEQT score from baseline to 1 year is 0.0 (−6.8 to 8.4) and the mean±SD is 1.4±17.0. The mean±SD age was 74±9.8 and the majority of the sample was male 57% and white 91%. Patients in the study sample were more likely to have had new-onset AF and be more symptomatic compared with those excluded from the study sample (Table I in the Data Supplement).
Clinically Important Difference
A dot plot depicting the relationship between change in EHRA score versus change in AFEQT score at 1 year is shown in Figure I in the Data Supplement. Individuals not exhibiting a change in either measure are excluded from the figure. Using the mean change method, a 1 EHRA class improvement corresponded to a 5.4-point increase (95% CI, 3.6–7.2) in overall AFEQT score at 1 year. Similar CID values were seen for the ADL and Symptoms AFEQT subscales (Table 1). With respect to deterioration, a 1 EHRA class deterioration corresponded to a 4.2-point decrease (95% CI, −6.9 to −1.5) in overall AFEQT score at 1 year.
|Anchor Method||N||Overall AFEQT Score||Activities of Daily Living Subscale||Symptoms Subscale|
|Mean change||344||5.4 (3.6–7.2)||5.1 (2.5–7.6)||7.1 (5.3–9.0)|
|ROC||1153||1.9 (0.9–9.3)||6.3 (2.1–27.0)||4.2 (4.1–8.3)|
CID at the Population Level
At the population level, as the percentage of patients experiencing a CID varies from 0% to 100%, and all other patients experience no change, the average change in AFEQT ranges from 0 to the CID. If the CID in AFEQT is taken to be 5.4, a treatment that induced a CID in 20% of the population but no difference in the other 80% would result in a difference of mean AFEQT equal to 1 point (Figure 1).
CIDs in Subgroups
CIDs using the mean change method across various subgroups are presented for improvement and deterioration on the anchor in Figures 2 and 3, respectively. Across various subgroups, CIDs were similar to the CID estimates for the overall study sample except for newly diagnosed AF patients for improvement on the anchor. Patients diagnosed with AF within 1 year of enrollment in the registry had a CID value of 10.1 (95% CI, 5.7–14.5) for improvement using 1 EHRA class as the anchor. Furthermore, this CID value was significantly different from the CID for patients diagnosed >1 year before enrollment in the registry 3.7 (95% CI, 0.9–5.5), P=0.008. For deterioration on the anchor, the CID value across various subgroups was similar to the CID value for the overall study sample.
Sensitivity and Secondary Analyses
In additional sensitivity analyses using the mean change method, CID values for a 1 EHRA class improvement from 3 to 2 and from 2 to 1 were similar to the CID value for an overall 1 EHRA class improvement (Table 2). The CID value for a 2 EHRA class improvement was approximately double that of a 1 EHRA class improvement 13.4 (95% CI, 8.2–18.6).
|EHRA Class Change*||EHRA Class Change Type||N||Mean Change in AFEQT Score|
|No change||…||731||0.2 (−0.9 to 1.2)|
|1 class improvement||3 to 2||78||7.2 (3.2 to 11.1)|
|2 to 1||263||5.0 (3.0 to 7.0)|
|2 class improvement||(4 to 2) or (3 to 1)||73||13.4 (8.2 to 18.6)|
As a secondary analysis using the ROC method, a 1 EHRA class improvement corresponded to a 1.9-point increase (95% CI, 0.9–9.3) in overall AFEQT score at 1 year. The C index (95% CI) was 0.59 (0.57–0.62). A 1 EHRA class deterioration corresponded to a 7.2-point decrease (95% CI, −13.9 to 3.7) in overall AFEQT score using the ROC method. ROC curves and sensitivity and specificity values for a subset of change in AFEQT score at 1-year values are presented in Figures II and III and Tables II and III in the Data Supplement. CID values corresponding to a 1 EHRA class improvement for the AFEQT subscales ADL and Symptoms (Table 1) were more than double the CID value for the overall AFEQT score.
The use of PROs in clinical trials, clinical care, and quality assessment is growing rapidly. However, most PROs are unfamiliar to many practitioners, creating a need for a deeper understanding of their interpretation. To facilitate the use of a new, recently validated, disease-specific measure for AF, AFEQT, we explicitly sought to define what magnitude of change in AFEQT scores is clinically important. Using the mean change method, based on the clinician-reported EHRA score in over 1300 outpatients with AF, we found that changes in the overall AFEQT and subscale scores of ≈5 points are clinically important.
Our findings significantly extend the original efforts of Dorian et al19 to define a CID for AFEQT. They conducted an analysis to assess the minimal important difference in the AFEQT score in 210 patients. The investigators used the patient-reported global change in QoL and the AF symptom scale as their anchors and found that a 17- to 19-point change in AFEQT was the minimal important difference. In contrast, in our study, we found that a change of ≈5 points in the AFEQT score corresponded to a change in functional status as indicated by a 1 class change in the EHRA score (similar to the New York Heart Association functional classification system for heart failure). Whether this difference reflects different populations of patients, different anchors, greater power in our much larger study, or random error will require further investigation. Nevertheless, our use of the EHRA score supports the clinical importance of a 5-point change in score from physicians’ perspectives.
In addition to identifying CIDs associated with a 1 EHRA class improvement, we identified CIDs associated with 1 EHRA class deterioration in AFEQT score to be ≈−4 which is similar in magnitude to the CID of 5 identified for a 1 class improvement. Clinicians are often interested in identifying targets for improvement in QoL. However, treatments can also have a negative impact on QoL, highlighting the importance of additionally defining CIDs associated with a 1 EHRA class deterioration in health status. By identifying CIDs associated with deteriorating QoL, clinicians can identify negative changes in QoL that should result in discontinuation or modification of treatment.16,20 For overall AFEQT and ADL and symptoms subscale scores, an improvement or deterioration of ≈5 may be clinically important.
Across various subgroups, CIDs were similar to the CID for the overall study sample except for AF diagnosed more than or less than a 1 year from enrollment in ORBIT. For patients with AF diagnosed within 1 year of enrollment, CIDs were approximately double than the CID for the overall study sample. The mean and median baseline AFEQT score in this subgroup were 74 and 78 points, respectively. This is likely because their mean baseline AFEQT score is lower than the baseline score for the overall study sample. Thus, there was more opportunity for the AFEQT score to increase at 1 year resulting in a larger CID value than the overall study sample.
In the sensitivity analysis determining CID values for a 1 EHRA class improvement from 3 to 2 and from 2 to 1 using the mean change method, the CID values were similar to the CID value for any 1 EHRA class improvement. Therefore, the gradations between adjacent categories were of similar magnitude. Similarly, an improvement of 2 EHRA classes resulted in a CID value of ≈2.5 times what was seen for 1 EHRA class improvement.
In the secondary analysis using the ROC method, the points estimate for the CID was ≈2 points, less than half the value identified using the mean change method. However, the CIs were wide and included the 5-point estimate from the mean change method. The CID values for the ADL and Symptom subscales were more similar to the CID value identified using the mean change method. These differing values may be because of the fact that different calculation methods can lead to different CID values as seen with Terwee et al21 analysis. In their analysis, they calculated minimal important change values across 5 different methods and obtained a different minimal important change value for each method.
The CID value observed in ORBIT-AF of ≈5 points, corresponds to the mean change in AFEQT in a subgroup where everyone experienced a clinically relevant improvement in EHRA score. This threshold may define an important target for success in improving QoL and provides a starting point for study design of therapies that seek to improve QoL of patients with AF. However, it is important to note that a successful therapy will not necessarily impact all patients. As illustrated in Figure 1, a therapy which results in a CID for some patients, but not all, will often result in a population level mean change in AFEQT that is substantially smaller than the CID. Clinical trials and other population-average comparisons need to consider both the CID and the distribution of changes across the population. For instance, a treatment for AF can markedly improve health status compared with placebo (as measured by those who achieve the CID of 5 points at follow-up) but may appear to have only a modest increase in the population-level AFEQT when fewer than 50% of patients had a CID change of 5 or more points. Study design of therapeutic interventions should, therefore, take this into account. More work into defining the patient-level changes in scores associated with small, moderate, and large clinically important improvements is an important area for future investigation.
Our findings should be interpreted in the context of several potential limitations. The anchor used in this study, the EHRA score, is a widely accepted measure of functional status, but assessed from the physicians’, rather than patients’, perspectives. Typically, anchors used to identify clinical important changes in QoL instruments are ideally assessed from the patient’s perspective.22 Furthermore, the time between the initial and follow-up measurements of 1 year is relatively long, although the use of cross-sectional assessments (AFEQT from the patients and EHRA from the clinicians) avoid recall bias and may mitigate this concern. Having a shorter window of time between measurements would enable the use of global assessments of change from both the patients’ and providers’ perspectives, an alternative technique for establishing CIDs. Additionally, because patients in the ORBIT registry have a well preserved QoL at baseline, these patients may not necessarily be representative of all patients with AF, and that there may well be a ceiling effect with respect to the potential for improvement in QoL. Also, patients excluded because of missing 1-year AFEQT follow-up data could cause potential bias. Finally, the CIDs identified represent changes on an absolute scale. CIDs may be different based on a relative scale.
Changes in AFEQT score of ≈5 points, regardless of direction, are likely to be clinically relevant. This can help inform the interpretation of mean differences between groups and lay the foundation for developing responder analyses to interpret the mean changes in scores across patients. Moreover, the study design of therapeutic interventions should take into account the proportion of patients expected to benefit from treatment when applying CID values. Last, particular attention should be paid to unique patient population characteristics which may lead to higher or lower CIDs than in the overall population. More studies are needed to help identify a range of values to assess meaningful improvement in QoL in patients with AF.
We thank the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) Registry staff and participants for their important contributions to this work.
Sources of Funding
The Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) registry is sponsored by Janssen Scientific Affairs, LLC, Raritan, NJ.
Dr Piccini reports research support from Boston Scientific, ResMed, ARCA Biopharma, St Jude Medical Center, Gilead Sciences, Johnson & Johnson, Spectranetics, and Janssen (all significant) and consultancies for Janssen Scientific Affairs (significant), Spectranetics (significant), Medtronic (significant), Forest Laboratories (modest), Pfizer (modest), Glaxo SmithKline (modest), and Amgen and Allergan. Dr Allen reports research grants from PCORI, NIH/NHLBI and American Heart Association (all significant) and consultancies for ACI clinical (significant), Boston Scientific (modest), Cytokinetics/Amgen (modest), Duke Clinical Research Institute (modest) and Janssen (modest). Dr Fonarow reports consulting for Janssen (significant). Dr Gersh reports consultancies with Janssen Scientific Affairs (significant), Cipla Limited and Armetheon Inc. Data Safety Monitoring Board for Mount Sinai St Lukes, Boston Scientific Corporation (modest), Teva Pharmaceutical Industries, St Jude Medical, Janssen Research & Development, Baxter Healthcare Corporation, Thrombosis Research Institute, Duke Clinical Research Institute, Duke University, Kowa Research Institute and Cardiovascular Research Foundation (all modest) and advisory board member for Medtronic (modest). Dr Kowey reports consulting for Johnson & Johnson (significant), Daiichi-Sankyo (modest), Bristol-Myers Squibb, and Boehringer Ingelheim. Dr O’Brien reports research grants from Janssen Scientific Affairs, BMS, Novartis (significant) and Sanofi (modest). Dr Reiffel reports research grants from Medtronic and Janssen and is a consultant for Medtronic, Portola, InCardia Therapeutics, and Acesion. Dr Naccarelli reports a research grant from Janssen Scientific Affairs (significant) and is a consultant to Janssen (significant), Glaxo Smith Kline (significant), Sanofi (modest), Novartis (modest), Portola (modest), Acesion (modest), and Omeicos (modest). Dr Ezekowitz reports research grants from Bristol-Myers Squibb, Boehringer Ingelheim and Pfizer (all significant) and is a consultant for Medtronic, Boehringer Ingelheim, Pfizer, Sanofi, Portota, Daiichi-Sankyo, Johnson and Johnson, Janssen Scientific Affairs, Merck, Pfizer, and Bristol-Myers Squibb (all modest). Dr Chan is supported by the National Heart Blood and Lung Institute (1R01HL123980). Dr Singer reports contract research with Bristol-Myers Squibb and Boehringer Ingelheim (both significant) and is a consultant/advisory board member for Boehringer Ingelheim (significant), Bristol-Myers Squibb (significant), Medtronic (modest), Johnson and Johnson (modest), Merck (modest), Pfizer (modest), and CVS Health (modest). Dr Spertus reports research support from Bayer (significant) and consultant for Janssen Scientific Affairs, Bayer, Novartis, United Healthcare, AstraZeneca (all modest). Dr Peterson reports significant research support from Eli Lilly & Company, Daiichi Sankyo, and Janssen. Dr Thomas reports research support with Novartis, Boston Scientific, Gilead Sciences, Inc, and Janssen Scientific Affairs (all modest). The other author reports no conflicts.
January CT, Wann LS, Alpert JS, Calkins H, Cigarroa JE, Cleveland JC, Conti JB, Ellinor PT, Ezekowitz MD, Field ME, Murray KT, Sacco RL, Stevenson WG, Tchou PJ, Tracy CM, Yancy CW; ACC/AHA Task Force Members. 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines and the Heart Rhythm Society.Circulation. 2014; 130:2071–2104. doi: 10.1161/CIR.0000000000000040LinkGoogle Scholar
Dorian P, Jung W, Newman D, Paquette M, Wood K, Ayers GM, Camm J, Akhtar M, Luderitz B. The impairment of health-related quality of life in patients with intermittent atrial fibrillation: implications for the assessment of investigational therapy.J Am Coll Cardiol. 2000; 36:1303–1309.CrossrefMedlineGoogle Scholar
Spertus J, Dorian P, Bubien R, Lewis S, Godejohn D, Reynolds MR, Lakkireddy DR, Wimmer AP, Bhandari A, Burk C. Development and validation of the Atrial Fibrillation Effect on QualiTy-of-Life (AFEQT) questionnaire in patients with atrial fibrillation.Circ Arrhythm Electrophysiol. 2011; 4:15–25. doi: 10.1161/CIRCEP.110.958033LinkGoogle Scholar
Piccini JP, Fraulo ES, Ansell JE, Fonarow GC, Gersh BJ, Go AS, Hylek EM, Kowey PR, Mahaffey KW, Thomas LE, Kong MH, Lopes RD, Mills RM, Peterson ED. Outcomes registry for better informed treatment of atrial fibrillation: rationale and design of ORBIT-AF.Am Heart J. 2011; 162:606–612 e1. doi: 10.1016/j.ahj.2011.07.001CrossrefMedlineGoogle Scholar
- 5. European Heart Rhythm Association, European Association for Cardio-Thoracic Surgery,
Camm AJ, Kirchhof P, Lip GY, Schotten U, Savelieva I, Ernst S, Van Gelder IC, Al-Attar N, Hindricks G, Prendergast B, Heidbuchel H, Alfieri O, Angelini A, Atar D, Colonna P, De Caterina R, De Sutter J, Goette A, Gorenek B, Heldal M, Hohloser SH, Kolh P, Le Heuzey JY, Ponikowski P, Rutten FH. Guidelines for the management of atrial fibrillation: the Task Force for the Management of Atrial Fibrillation of the European Society of Cardiology (ESC).Eur Heart J. 2010; 31:2369–2429. doi: 10.1093/eurheartj/ehq278CrossrefMedlineGoogle Scholar
Kirchhof P, Auricchio A, Bax J, Crijns H, Camm J, Diener HC, Goette A, Hindricks G, Hohnloser S, Kappenberger L, Kuck KH, Lip GY, Olsson B, Meinertz T, Priori S, Ravens U, Steinbeck G, Svernhage E, Tijssen J, Vincent A, Breithardt G. Outcome parameters for trials in atrial fibrillation: recommendations from a consensus conference organized by the German Atrial Fibrillation Competence NETwork and the European Heart Rhythm Association.Europace. 2007; 9:1006–1023.CrossrefMedlineGoogle Scholar
Wynn GJ, Todd DM, Webber M, Bonnett L, McShane J, Kirchhof P, Gupta D. The European Heart Rhythm Association symptom classification for atrial fibrillation: validation and improvement through a simple modification.Europace. 2014; 16:965–972. doi: 10.1093/europace/eut395CrossrefMedlineGoogle Scholar
King MT. A point of minimal important difference (MID): a critique of terminology and methods.Expert Rev Pharmacoecon Outcomes Res. 2011; 11:171–184. doi: 10.1586/erp.11.9CrossrefMedlineGoogle Scholar
McGlothlin AE, Lewis RJ. Minimal clinically important difference: defining what really matters to patients.JAMA. 2014; 312:1342–1343. doi: 10.1001/jama.2014.13128CrossrefMedlineGoogle Scholar
Freeman JV, Simon DN, Go AS, Spertus J, Fonarow GC, Gersh BJ, Hylek EM, Kowey PR, Mahaffey KW, Thomas LE, Chang P, Peterson ED, Piccini JP, Outcomes Registry for Better Informed Treatment of Atrial Fibrillation I and Patients. Association between atrial fibrillation symptoms, quality of life, and patient outcomes: results from the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF).Circ Cardiovasc Qual Outcomes. 2015; 8:393–402. doi: 10.1161/CIRCOUTCOMES.114.001303LinkGoogle Scholar
Copay AG, Subach BR, Glassman SD, Polly DW, Schuler TC. Understanding the minimum clinically important difference: a review of concepts and methods.Spine J. 2007; 7:541–546. doi: 10.1016/j.spinee.2007.01.008CrossrefMedlineGoogle Scholar
Shafiq A, Arnold SV, Gosch K, Kureshi F, Breeding T, Jones PG, Beltrame J, Spertus JA. Patient and physician discordance in reporting symptoms of angina among stable coronary artery disease patients: insights from the Angina Prevalence and Provider Evaluation of Angina Relief (APPEAR) study.Am Heart J. 2016; 175:94–100. doi: 10.1016/j.ahj.2016.02.015CrossrefMedlineGoogle Scholar
Raphael C, Briscoe C, Davies J, Ian Whinnett Z, Manisty C, Sutton R, Mayet J, Francis DP. Limitations of the New York Heart Association functional classification system and self-reported walking distances in chronic heart failure.Heart. 2007; 93:476–482. doi: 10.1136/hrt.2006.089656CrossrefMedlineGoogle Scholar
Jaeschke R, Singer J, Guyatt GH. Measurement of health status. Ascertaining the minimal clinically important difference.Control Clin Trials. 1989; 10:407–415.CrossrefMedlineGoogle Scholar
Deyo RA, Centor RM. Assessing the responsiveness of functional scales to clinical change: an analogy to diagnostic test performance.J Chronic Dis. 1986; 39:897–906.CrossrefMedlineGoogle Scholar
Wright A, Hannon J, Hegedus EJ, Kavchak AE. Clinimetrics corner: a closer look at the minimal clinically important difference (MCID).J Man Manip Ther. 2012; 20:160–166. doi: 10.1179/2042618612Y.0000000001CrossrefMedlineGoogle Scholar
de Vet HC, Ostelo RW, Terwee CB, van der Roer N, Knol DL, Beckerman H, Boers M, Bouter LM. Minimally important change determined by a visual method integrating an anchor-based and a distribution-based approach.Qual Life Res. 2007; 16:131–142. doi: 10.1007/s11136-006-9109-9CrossrefMedlineGoogle Scholar
Carpenter J, Bithell J. Bootstrap confidence intervals: when, which, what? A practical guide for medical statisticians.Stat Med. 2000; 19:1141–1164.CrossrefMedlineGoogle Scholar
Dorian P, Burk C, Mullin CM, Bubien R, Godejohn D, Reynolds MR, Lakkireddy DR, Wimmer AP, Bhandari A, Spertus J. Interpreting changes in quality of life in atrial fibrillation: how much change is meaningful?Am Heart J. 2013; 166:381–387.e8. doi: 10.1016/j.ahj.2013.04.015CrossrefMedlineGoogle Scholar
Guyatt GH, Osoba D, Wu AW, Wyrwich KW, Norman GR; Clinical Significance Consensus Meeting Group. Methods to explain the clinical significance of health status measures.Mayo Clin Proc. 2002; 77:371–383. doi: 10.1016/S0025-6196(11)61793-XCrossrefMedlineGoogle Scholar
Terwee CB, Roorda LD, Dekker J, Bierma-Zeinstra SM, Peat G, Jordan KP, Croft P, de Vet HC. Mind the MIC: large variation among populations and methods.J Clin Epidemiol. 2010; 63:524–534. doi: 10.1016/j.jclinepi.2009.08.010CrossrefMedlineGoogle Scholar
Cook CE. Clinimetrics corner: The Minimal Clinically Important Change Score (MCID): a necessary pretense.J Man Manip Ther. 2008; 16:E82–E83. doi: 10.1179/jmt.2008.16.4.82ECrossrefMedlineGoogle Scholar