Evaluating patients' satisfaction and preferences with a secondary prevention cardiovascular polypill: the Aurora Study

This was an observational, cross-sectional, multi-centric, two-cohort study with the participation of 32 investigator sites in Spain and four investigator sites in Belgium, conducted between July 2016 and September 2018. Each investigator site included consecutive patients in accordance with the selection criteria. Each investigator had to include 12 patients, 6 patients in each study cohort. However, due to the slow recruitment, inclusion became competitive. There were no interventions involving ongoing medications and/or decisions of the physicians. The registry was carried out following the current clinical guidelines and according to current clinical practice and was approved by the Spanish Agency of Medicines and Medical Devices and by the Independent Ethics Committee of the different hospitals involved in both countries.

Patients were enrolled in two different cohorts according to whether they were treated: with the CNIC polypill [5] (cohort A) or with the three monocomponents separately (ASA, a statin and an ACE inhibitor; cohort B). Patients in cohort A had to had been treated with the separate monocomponents before receiving the polypill. Enrollees had to be on their current treatment for at least 3 months prior to enrolment in the study, to prevent from induction to prescription, and to ensure that the patient had been treated for long time enough to have a valid opinion. Patients had to meet the following criteria: age over 18 years, history of cardiovascular disease (CVD) with the last CV event happening >1 year before enrollment, being on preventive treatment for CVD with the CNIC polypill or the three separately monocomponents (ASA, a statin and an ACE inhibitor) for at least 3 months and having signed the informed consent. To ensure comparability across study cohorts and minimize bias, patients from both cohorts were paired based on gender and age (±5 years).

Data collected from the medical record and at the study visit were sociodemographic, anthropometric and clinical characteristics. At the study visit patients were asked to complete two specific and validated questionnaires and one ad-hoc questionnaire. The Treatment Satisfaction Questionnaire for Medication (TSQM-9) [14] evaluated patients' satisfaction with treatment in three domains (effectiveness, convenience and global satisfaction) scored from 0 to 100. The Morisky Medication-Taking Adherence Scale (MMAS-4) [15] (used under permission by Donald E Morisky, ScD, ScM, MSPH) assessed adherence to treatment in a score ranging from 0 to 4, with three levels of medication compliance: highly compliant, moderately compliant and noncompliant with 0, 1–2 and 3–4 points, respectively [16]. The MMAS4 was selected because it is a convenient and easy to use self-reporting questionnaire that has demonstrated reasonable internal consistency and sensitivity and has been validated in different languages, including Spanish and French. In addition, the objective was to measure adherence and not to measure the causes for nonadherence, which case would have justified the need for a more sophisticated questionnaire. An ad-hoc questionnaire was designed to determine patients' preferences with different questions per cohort:

All data were collected in an electronic case report form designed for the study.

The study showed higher satisfaction assessed by the TSQM-9 questionnaire and higher adherence measured using the MMAS-4 in patients receiving the CNIC polypill. In addition, the CNIC polypill was the preferred treatment option for most patients.

The two cohorts of patients were well balanced as shown in Table 1. Sociodemographic and clinical characteristics were similar between both cohorts and with those reported in literature for patients in secondary CV prevention [17]: elderly patients, mainly males, smokers or ex-smokers with an important degree of obesity and with concomitant pathologies. Treatment in cohort B was mainly ASA 100 mg, atorvastatin as the main statin, and ramipril or enalapril as ACEi. Statin strengths were higher in the monocomponents arm compared with the polypill arm. 35.5 and 23.9% were treated with atorvastatin 40 and 80 mg, respectively, compared with 10.3% with 20 mg. At the time of the study, the CNIC polypill was available in the 20 mg atorvastatin strength, although since 2017 the 40 mg atorvastatin strength has also been available.

Regarding the primary aim of the study, patients' satisfaction (as the perception of an individual's experience compared with their expectations), our results showed that secondary prevention patients treated with the CNIC polypill were more satisfied than those patients treated with monocomponents separately, with a significantly higher score in each of the three scales in the TSQM-9 questionnaire: ‘effectiveness’, ‘convenience’ and ‘global satisfaction’. In this regard, a factor that can contribute to a higher satisfaction is the number of prescribed drugs. In our study, patients treated with the polypill took a smaller number of concomitant drugs than patients treated with the monocomponents (6.1 ± 2.9 vs 8.1 ± 3.2, p < 0.0001) which may explain, at least in part, the results on satisfaction. It is also important to notice that the perception of ‘effectiveness’ increased with the polypill, even though patients in cohort A were treated with higher statin strengths. It could be argued that higher satisfaction would be indirectly related with a higher degree of treatment adherence, which may result in a better CV risk factors control. In fact, recent data from two different observational studies in a real-life clinical settings in Mexico [18,19] shed some light into this hypothesis, showing that patients treated with a version of the CNIC polypill containing simvastatin instead of atorvastatin improved their blood pressure and lipid levels, and showed a significant reduction Framingham risk score compared with their previous treatment.

Our results on satisfaction were also consistent with results of another study that compared satisfaction of an antihypertensive fixed dose combination versus separate pills. A high percentage of patients were more satisfied with the antihypertensive polypill therapy, stating than ‘effectiveness’ and ‘convenience’ of the treatment were the most important factors for patients [20].

In terms of patients' preferences, an ad-hoc questionnaire was developed to evaluate this item. Questionnaire results for cohort A revealed that the polypill strategy was relevant for nearly all our studied population (98.9%) who stated to prefer the polypill rather than the monocomponents separately. This could be explained as patients' expectations with the polypill have been satisfied also in improving patient's quality of life, willing to keep the polypill treatment as a long-term option. Most patients also stated that the polypill was easier to take and more trustful than the monocomponents, which could reinforce a better adherence, and ultimately, a better CV risk control.

Conversely, 72.8% of the patients treated in cohort B declared they would switch to the CV polypill while 4.6% answered that they would not consider it. These findings are aligned with published literature which also reflects that the a polypill treatment is an attractive option to patients, who see advantages over their current treatment [21], and highlight the need to ask patients about their preferences. In addition, in our study 63.1% of all the patients reaffirmed their will to switch to the CV polypill even if the out of pocket cost was higher. Previous research supports an increased willingness to pay from patients in order to reduce the pill burden [22].

The adherence to the treatment as measured by the MMAS-4 was higher in the CV polypill group (57.7 vs 41.1% p = 0.0027) when considering highly compliant patients and the adherence difference between groups increased when also considering moderately compliant patients (76.6 vs 50.0% p < 0.0001). It is important to notice that 50.0% of patients treated with the monocomponents were considered non-adherent to treatment, and 49.7% answered ‘no’ to all MMAS-4 items compared with the 23.4% of patients in cohort A. The increased number of drugs could be one of the reasons for the higher non-adherence in the monocomponents cohort of our study. There is also evidence than adherence in chronic diseases is influenced by medication complexity: the simplest the regimen the greater patient adherence is. Our results on adherence are in line with previous studies with several polypills [23,24].

Although the increase in adherence is followed by an increase in CV risk factors control and in a lower incidence of CV events, as recently shown in the PolyIran study [7], a recent study has also shown a synergistic effect within the components of the present CV polypill, in terms of a higher LDL-c reduction, which also may contribute to a better risk factors control [25].

This research, however, is subject to several limitations. One of them is the cross sectional study design, with no longitudinal data obtained. Another limitation is the criteria about time since the event, which was defined once the study was active, and 30 patients who had the last CV event <1 year since study visit were included in the total analyzed population. It must be pointed out that patients in cohort B had a significantly longer median duration of treatment at the time of study compared with those patients in cohort A, potentially having an effect in the comparability of the data between the two cohorts. The impact on comparability can be more significant in the adherence end point as it has been demonstrated that adherence to treatment tends to decrease over time. However, research shows that the decrease in treatment adherence occurs during the first 18 months of treatment, plateauing at a later stage [26]. Finally, the self-reported nature of the three questionnaires employed have intrinsic limitations such as a social desirability bias or the absence of introspective ability. In particular, the use of the MMAS-4 may limit comparability to other adherence studies. MMAS-4 was selected as a self-reported, easy to use and validated instrument for adherence measurement that enabled the comparison between two groups. However, its simplistic nature has been suggested to affect its acceptability and internal consistency, and comparisons with more sophisticated instruments or direct adherence observation must be made with caution.

According to results, we can conclude that secondary prevention patients treated with a CV polypill showed a significantly higher degree of satisfaction and significantly higher medication adherence compared with patients that were treated with the monocomponents separately. Patients treated with the polypill strategy stated their perception of the polypill strategy as a convenient treatment enabling a high degree of confidence, preferring it over being treated with the separate monocomponents. A majority of patients treated with separate monocomponents expressed preference toward the polypill strategy when explained the concept.

By asking the patient about its preferences the patient becomes judgmental of the posed treatment options and may become more involved, committed and satisfied with the received treatment positively impacting persistence and adherence to treatment which in turn could lead to improved treatment outcomes. Therefore, patient preferences should be considered, especially when establishing chronic treatments.

Our work settles the ground for further research in this area, aiming to clarify the long-term relationship between patient satisfaction to the treatment with increased adherence and treatment outcomes. Nonetheless, considering the relevant burden of CV disease and the treatment gaps in cardiovascular prevention, the implications of our findings become relevant providing further evidence of the potential role polypill strategies can play in cardiovascular prevention.

All authors have contributed to the conception, design and acquisition of this article. Additionally, they have drafted and critically revised this manuscript. They all agree to be accountable for all aspects of work ensuring integrity and accuracy, as well as its final approval.

The authors wish to acknowledge all the cardiologist from the Aurora investigator group for their effective contribution in this study: A Peset (H Catstellon); VJ Girbes (HL Alcanyis); V Ortiz (H de Sagunto); M Quintanilla (H Gerenal Elche); A Tamayo (H Elda); N Leal (H General de Catalunya); M Gimenez (H Clinic de Barcelona); M Ribells (H General de Granollers); FJ Muñoz (H de Mollet); M Borras (H Arnau de Vilanova); MA Araujo (H Severo Ochoa); A Curcio (H de Fuenlabrada); PP Casado (H la Princesa); A Villanueva (H Universitario Rey Juan Carlos); N Lorenzo (H Infanta Cristina); N Muñoz (H Infanta Leonor); D Gudiño (H Nuestra Señora de la Candelaria); JA Alarcon (H Universitario Donostia); EM Pereira (H Lucus Augusti); C Peña (H Clinico Santiago); M Gutierrez (Complejo H Universitario Ourense); A Rodriguez (H Arquitecto Macide); J Idoate (H de Leon); M Amado (H General Segovia); E Martin (H Universitario Guadalajara); JL Bardaji (H Virgen de la Luz); J Lopez (H Reina Sofia); M Padilla (H de Jaen); P Santiafo (H la Loja); F Sabatel (H Santa Ana); B Cosyns (UZ Brussel) and R Casado (Hôspital Erasme).

Thanks are also extended to A Molina, C Ruiz, F Rodriguez, J Ferrer and M Garcia for their great motivation and commitment with this project. Acknowledgment to Ferrer Internacional for the support on this work.

Financial support for the conduction of the AURORA study and for the creation of this paper was provided by Ferrer Internacional. JCS has received research grants and speaker honorarium from Ferrer Internacional. JMMZ, HOPR, FGG and KH have received research grants from Ferrer Internacional. GR is full-time employees in Ferrer Internacional. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Writing assistance was provided by Anna del Prado, IQVIA, Madrid (Spain) and assistance with submission has been provided by Content Ed Net, Madrid (Spain).

The data underlying this article will be shared on reasonable request to the corresponding author.

This work is licensed under the Attribution-NonCommercial-NoDerivatives 4.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-nd/4.0/