Instrument review
Overall, 25 unique PRO instruments were identified through searches of the various sources including OVID publications, clinical trials, PROQOLID, clinical organizations, and HTA documents. The concepts the PROs assessed included mainly HRQoL, treatment satisfaction, and mental health.
The OVID literature review resulted in a total of 16 measures identified from 35 publications (Fig.
1). The measures most used were the EuroQol (EQ)-5D [
16] and 36-Item Short Form Survey (SF-36) [
17]. The Anti-Clot Treatment Scale (ACTS) [
18], which was used in ACS and SPAF, was the measure most identified to capture symptomatic treatment effects associated with anticoagulant and antiplatelet use. The Nuisance Bleeding Scale (NBS) [
19], which captures minor bleeding in anticoagulated patients, was also identified. Several other PRO instruments used in similar conditions were also identified.
A total of 16 clinical trials were identified; 13 trials were from ClinicalTrials.gov, and 3 trials were from ClinicalTrialsRegister.eu. Twelve different PRO instruments were identified which were being used as an outcome measure across these trials; 8 of which were unique and not identified in the OVID instrument search. The majority of trials with PROs were those of CAD (n = 10). No measures of symptomatic treatment effects associated with anticoagulant and antiplatelet use such as the ACTS or NBS were identified in any of the trials although, treatment satisfaction [e.g., Treatment Satisfaction Questionnaire for Medication (TSQM)] [
20] was measured in one of the trials.
No PROs were included in any of the 9 FDA/EMA medical product labels that were reviewed which included apixaban, clopidogrel, dabigatran, edoxaban, prasugrel, rivaroxaban, ticagrelor, vorapaxar, and warfarin.
No regulatory guidance documents, including FDA and EMA guidance documents, provided any recommendations on PRO instruments to capture symptomatic treatment effects associated with anticoagulant and antiplatelet use or physical function in the thrombotic indications.
Two organizations, International Consortium for Health Outcomes Measurement (ICHOM) and American College of Cardiology (ACC) PRO Forum, provided recommendations on use of PRO instruments in several related conditions and were reviewed. ICHOM has several PRO instrument recommendations for use in CAD, AF, and stroke in general. For CAD, ICHOM recommends use of the Seattle Angina Questionnaire (SAQ-7) [
21] to measure angina, Patient Health Questionnaire-2 (PHQ-2) [
22] for depression, and Rose Dyspnea Scale for Dyspnea [
23‐
25]. For AF, 12-Item Short Form Survey (SF-12) [
26] and Patient-Reported Outcomes Measurement Information System (PROMIS) global health [
27] to measure HRQoL and physical and emotional functioning [
28,
29]; Atrial Fibrillation Effect on Quality of Life (AFEQT) [
30] or Atrial Fibrillation Severity Scale [
31] for exercise tolerance and symptom severity; and Work Productivity and Activity Impairment Questionnaire (WPAI General Health v2.0) [
32] for ability to work are recommended. For stroke, ICHOM also recommends the PROMIS-10 [
33] to measure overall mental wellbeing including cognitive, psychiatric, and social functioning, as well as overall physical wellbeing including pain, fatigue, and general health status [
34]. The ACC PRO Forum [
35] identified the SAQ (19-item version) [
36] as a well-validated PRO measure used in CAD clinical trials [
37,
38] and the SAQ-7 (7-item version) [
20] for use in clinical practice. Additionally, the AFEQT was identified in AF [
39] although further validation to improve the measure is needed, and the Cardiac Self-Efficacy Questionnaire (CSEQ) [
40], HeartQoL [
41], MacNew Heart Disease Health-Related Quality of Life Questionnaire (MacNew) [
42,
43], and SAQ [
36] were identified as being developed in patients with coronary heart disease.
The search in PROQOLID resulted in the SAQ for CAD. The indications ACS, SSP, and SPAF provided no results. Following this, 2 separate searches were conducted using the keywords
atrial fibrillation and
stroke, which resulted in 5 stroke and 5 AF PRO measures being identified. The stroke measures included Stroke Impact Scale (SIS v2.0) [
44], SIS v3.0 [
45], SIS-16 [
46], Newcastle Stroke-Specific Quality of Life Measure (NEWSQOL) [
47], Stroke-Specific Quality of Life Measure (SS-QOL) [
48]. The AF measures included Atrial Fibrillation Impact Questionnaire (AF Impact) [
49], Atrial Fibrillation Severity Scale (AFSS) [
50], Quality of Life Questionnaire for Patients with Atrial Fibrillation (AF-QOL18) [
51], AFEQT, and Perception of Anticoagulant Treatment Questionnaire (PACT-Q) [
52].
HTA evidence use analysis
The targeted HTA website searches identified 27 evaluation documents. A total of 5 different PROs were identified in the search, and all HTA agencies (NICE, SMC, HAS, G-BA, INESSS, CADTH) discussed EQ-5D consistently in their documents. NICE and SMC discussed ACTS and TSQM in the review of rivaroxaban (Xarelto); NICE mentioned the SAQ Angina Frequency and Physical Limitations scale scores and the London School of Hygiene Dyspnoea Questionnaire score in its review of prasugrel (Efient). There was no further discussion of the above-mentioned instruments by NICE and SMC in their recommendations.
Reducing the risk of bleeding was consistently discussed in France (HAS), Germany (G-BA), Canada (CADTH, INESSS), and UK (SMC, NICE) reviews. The UK assessment identified the inability to demonstrate HRQoL loss associated with stroke and major bleed as one of several limitations of the health economic evidence cited for not adequately demonstrating an economic case for the product. The INESSS review discussed concern about some patients with AF not currently receiving anticoagulant therapy due to fear of bleeding (e.g., patient at high risk of falls or having a history of gastrointestinal ulcer as well as alcoholism). The G-BA mentioned that the collection of HRQoL could have been used to provide information as to whether bleeding events impact HRQoL significantly, which might have altered their recommendation for listing the product. Finally, PRO data did not contribute to the overall decision if the efficacy data do not show a significant and relevant improvement.