Effectiveness of a multicentre randomized controlled trial with three different nurse-led intervention management programmes for patients with heart failure: the main results of the MEETinCY study

Abstract

Aims

A single-blinded, multicentred randomized controlled trial (RCT) was employed to examine the effectiveness of a 3-month telephone follow-up, a telephone follow-up with education before discharge, or education only before discharge on the heart failure (HF) knowledge, HF self-care management, and health-related quality of life (HR-QoL) in patients with HF by a nurse specialist.

Methods and results

This is a multicentre RCT with three different intervention groups (IGs) and one control group. Participants in the first IG received education on HF self-management principles before discharge. The second IGs were enrolled to a 3-month telephone follow-up educational programme in addition to the pre-discharge educational session. The third IG received the telephone follow-up intervention only for 3 months. A total of 357 patients with HF were enrolled to the study of whom 262 participants completed the intervention. Outcome measures included HF-related QoL, HF self-care, and HF-related knowledge. Intervention effects were analysed using the Cohen d (rm) coefficient for repeated measurements and ANCOVA. There was a significant improvement in the physical dimension of the HR-QoL (F = 2.7, df = 3, P = 0.046) between the control group and in-person education group. In HF-related self-care, the telephone component alone or in combination with in-person education led to greater improvement than the control group (F = 3, df = 3, P = 0.034). Self-care practices were improved in the education and telephone arms as compared to the control group (P = 0.002).

Conclusion

Education and telephone support on self-care management may improve the physical dimension of HR-QoL.

Registration

ClinicalTrials.gov: NCT01905176

Graphical Abstract

Open in new tabDownload slide

Introduction

Heart failure (HF) is affecting ∼26 million people worldwide.^1,2 The aging of the population due to the improved survival following a HF diagnosis associated with the availability of life-saving evidence-based treatments and the overall longer life expectancy of the general population causes a gradual increase in the effect of HF and the need for various management programmes for these patients. All these are expected to increase the cost burden of HF management by 240% by the year 2030.^3,4 Many of these costs are attributable to hospitalization. Despite the improvement on the survival rate of HF patients⁵ through guideline-based medications and devices, HF patients continue to have a rehospitalization rate approaching 30% within 30 days post discharge, impaired functional status, health-related quality of life (HR-QoL), and social isolation.^4,6–9 Negative effects of HF on HR-QoL make maintenance or improvement of HR-QoL equal or more important than longevity for many patients.¹⁰

The aging population and the prevalence of numerous concomitant chronic diseases, along with the rising incidence of HF, have increased the need for thorough inter-specialty multidisciplinary care. Previous studies have demonstrated that a multidisciplinary disease management programme with structured patient education and continuous post-discharge follow-up can significantly decrease rehospitalization and mortality rates and improve their HR-QoL.^11–13 The focus of such educational programmes should be on HF self-care management strategies including improvement of knowledge towards HF, weight monitoring, dietary and medication compliance, close post-discharge follow-up, and how to address related challenges of self-care management.^14,15 Counselling, home visits, 24-h call centres, reminder systems, and complex interventions have been applied in various configurations to augment HF management, without clarifying which one is actually working and with what kind of combinations.^11,13,16

Even though the challenges of HF disease management are now much better understood, there is still an ongoing need to clarify which interventions exactly are more effective due to the heterogeneity of the existing ones.^11,13 This will help us gain an insight into how outcomes, such as HR-QoL, are influenced by the content of complex interventions and understand the underpinning mechanisms.¹⁷ The aim of the current randomized controlled trial (RCT) was to evaluate the Management of patients with HF using three different interventions: Education, Education and Telephone, Telephone in Cyprus (MEETinCY) compared the standard practice (ClinicalTrials.gov, NCT01905176). Meaning to identify which intervention is effective, education and/or telephone, and minimize the risk of confounding by indication, the interventions were applied individually and in combination as well. The current study is the first RCT in patients with HF in Cyprus.

Aim of the study

The primary aim of the study was to determine the effectiveness of a nurse-led multi-interventional HF management programme (3-month telephone follow-up, telephone follow-up combined with education before discharge, or education only before discharge) compared to standard practice, on the HR-QoL, HF knowledge, HF self-care management in patients with HF. Secondary aims were as follows: (i) describe the sample population at baseline and explore differences among study groups, (b) calculate the difference in each group at 3 months post discharge (overall and subscale scores), and (c) explore the correlation between baseline self-care behaviour, knowledge, and HR-QoL.

Methods

Study design

This is a multicentre RCT with three different intervention groups (IGs) and one control group (CG) (usual care explained in more detail below). The first IG included only patients’ education before discharge (E). The second IG included patients’ education before discharge and telephone follow-up after a week, in 2 weeks, and then once a month after discharge for 3 months (ET). The third IG included only telephone follow-up: after a week, in 2 weeks, and then once a month after discharge for 3 months (T). Three IGs were necessary to clarify which intervention was effective. The patients were assigned to the IG and CG with a 1:1 ratio. The random allocation was developed using a software program. To predict selection bias, the researcher was blinded regarding the group allocation of the next patient. To achieve this, closed envelopes were used for the recruitment, which the researcher opened after the patient agreed to participate in the research and signed the informed consent form. ‘Double blinding’ could not be achieved as patients should be informed before signing the consent form to participate in the research. Thus, after opening the envelop, both the patient and the researcher were aware in which group will be the patient.

Participants and setting

The study took place in the four (biggest) public hospitals in the Republic of Cyprus. Participants were approached and consented prior discharge and were screened for eligibility. Afterwards, they were randomly assigned to the IGs or CG. Inclusion criteria included adult patients with HF (evident in their medical record), who were currently hospitalized and ready for discharge of any New York Heart Association (ΝΥΗΑ) category. Exclusion criteria included patients who (i) had dementia or serious mental health problem, (ii) were discharged to hospice or care home, (iii) cannot be contacted by phone, (iv) were scheduled for surgery in the next 3 months, (v) had active cancer, (vi) had a major surgery the last 3 months, (vii) had life expectancy <3 months, (viii) had chronic degenerative diseases (e.g. dementia, Parkinson’s disease), (ix) were on dialysis, and (x) do not speak Greek. Sample size was estimated using G Power software and was set at minimum of 347 patients for all groups. The level of statistical significance (α-value) was pre-set at 0.05 and the statistical power at 80%. Health-related quality of life was considered as the main variable, and the MLHFQ tool was used to determine it. The smallest clinically significant difference of the MLHFQ tool has been found to be on the order of five points, which corresponds to the smallest difference in the values of the specific questionnaire based on previous studies.^18,19

Of the 357 patients approached, 262 (CG = 73, E = 63, ET = 62, T = 64) completed follow-up. Rest of the patients were either lost in follow-up (n = 73), died (n = 19), discontinued due to worsening of HF (n = 1), or withdraw from the study (n = 2). Response rate was at 73.4% (26.6% dropouts). Specifically, the percentage of the CG who did not continue was 17.8%, the E group was 25%, the ET group was 30.3%, and the T group was 32.6% (Figure 1).

Intervention and control group

Three different educational IGs (differing in delivery mode) by a nurse specialist were compared to standard care in in patients with HF at 3 months. All patients included in the study received the educational booklet and CD at the end of the study, including the CG.

E (intervention group)

In the E group, a standardized education was delivered by a specialist nurse of the study team at the baseline before discharge. An individual approach was used with a standardized content in line with ESC guidelines and based on the theoretic models of Orem and Riegel and Dickson.^20–22 The content of the educational session included the importance of monitoring body weight and intake/output, methods for monitoring and detecting the symptoms and signs of HF, drug effects, and the importance of compliance. An educational booklet with all information about HF and HF management was used during the educational session and given to the patient, including a diary, so the patient would be able to record the daily weight, medication, intake/output and note important information about his/her condition. Also, a CD with educational videos that the patients could watch, read subtitles, and listen to was provided as well. Family was involved in the education process whenever this was possible.

T (intervention group)

In the T (Telephone follow-up) group, nurse specialists were explaining via telephone what is HF, the importance of self-care management, and adherence to the therapy, following a structural guide based on the educational booklet referred above. Then, structural follow-up telephone calls were made after a week, in 2 weeks, and then once a month after discharge for 3 months by a specialist nurse. During these contacts, the nurse assessed the patient for any signs of deterioration, enhanced self-care behaviours and education, supported the patient psychologically, and resolved any questions he/she had about self-care management of HF. Even though the guide was used, the conversation was following patients’ needs and queries.

ET (intervention group)

In the ET group, the education was offered as described above for the E group before discharge. Additionally, they received follow-up telephone calls like the T group of the study.

Control group

The CG received the usual care provided by the health care services that did not include structural education and the follow-up included only the visit to the cardiologist.

Outcome measures and tools

The primary aim of the study was to determine the effectiveness of a 3-month telephone follow-up, a telephone follow-up with education before discharge, or education only before discharge, on the knowledge, self-care management, and HR-QoL in patients with HF. The secondary outcomes were to describe the sample population at baseline and explore differences among study groups and to explore the correlation between baseline self-care behaviour and knowledge. Socio-demographic and clinical variables of the participants were collected using a form developed for the study. The following data were collected by the research team at two points: before discharge (baseline) and 3 months after discharge. At the second time, the questionnaires were filled via telephone by all study groups.

The Greek version of the 9-item European Heart Failure Self-care Behaviour Scale

It was used to measure the level of self-care management regarding HF. The 9-item European Heart Failure Self-care Behaviour Scale (EHFScBS-9) is a 5-point Likert scale, from 1 (‘I completely agree’) to 5 (‘I do not agree at all’), that measures HF-related self-care behaviours. The total score is calculated by summing the ratings for each item. The total score ranges from 9 to 45 with higher scores indicating poorer self-care management behaviours. Gr9-EHFScB is better supported by a one-factor model, and it is thus preferable to be used as a whole.²³ Cronbach’s alpha coefficient of the total scale was found to be 0.66, and factor score determinacy coefficients were 0.57 (adhering to recommendations), 0.75 (fluid and sodium management), and 0.62 (physical activity and recognition of deteriorating symptoms).²⁴

The Greek version of the Minnesota Living with Heart Failure Questionnaire

It was used to measure HF HR-QoL. Minnesota Living with Heart Failure Questionnaire (MLHFQ) consists of 21 items. It is a 6-point Likert scale, from 0 to 5. The total score ranges from 0 to 105, with higher scores reflecting poorer HR-QoL. The Greek validation of MLHFQ provided a 3-factor solution explaining 64.15% of the variance (physical, emotional, and social subscales). The internal consistency of the total scale was 0.95 and 0.8–0.94 for the subscales.²⁵

Dutch Heart Failure Knowledge Scale

It was developed by van der Wal et al.²⁶ to measure the patients’ level of knowledge regarding HF. The Dutch Heart Failure Knowledge Scale (DHFKS) is a 15-item self-administered questionnaire that includes items concerning general HF knowledge, knowledge regarding HF treatment, HF symptoms, and symptom recognition. The score of the scale ranges between 0 and 15. Higher scores indicate higher levels of HF disease knowledge. Cronbach’s alpha of the original DHFKS was 0.62. The scale was translated and culturally adapted by the research team using Brislin’s²⁷ methodology.

Statistical analysis

Descriptive statistics were used to describe the demographic and clinical characteristics of the study’s population. Missing values were dealt in accordance with the instructions of the questionnaires, where available, whereas advanced methods such as multiple imputations were used for the rest of the data set in order to address the issue of missing data.²⁸ Participants with missing follow-up data were excluded from the analysis. The score difference in each group at 3 months post discharge for HF-related QoL, HF self-care management, and HF-related knowledge (overall and subscale scores) was studied by Cohen’s d (rm) coefficient for repeated measurements (Cohen’s d = 0.2 represents a small effect size, 0.5 represents a medium effect size, 0.8 represents a large effect size, and 1.3 represents a very large effect size).²⁹ ANCOVA was employed to detect differences the among the three different types of educational interventions for each outcome as compared to standard care while adjusting for baseline scores and several factors identified in the literature that are potentially associated with each of the outcomes (e.g. NYHA score, educational background, and marital status). Homogeneity of variance was tested and confirmed for all the F-tests. Bonferroni correction test was carried out for pairwise comparisons. Pearson’s correlation coefficient r was calculated to explore the correlation between baseline HF self-care behaviour (Gr9-EHFScBS) and HF knowledge (DHFKS) scores. All statistical analyses were performed using SPSS (version 25) software, and the level of statistical significance was set at 0.05. Analyses were performed on an intention-to-treat basis.

The investigation of the association between baseline clinical and demographic characteristics and the HR-QoL of the patients was conducted using multiple linear regressions. The following variables were used as independent factors: baseline score, NYHA class, gender, age, diabetes, hypertension, previous HF hospitalizations, implantable cardioverter defibrillator/cardiac resynchronization therapy defibrillator (ICD/CRT-D), ejection fraction (EF%), first-line drugs (ACE-inhibitors/ARBs and β-blockers), baseline social support score, education level, baseline emotional subscale score of the Gr-MLHFQ, baseline social subscale score of the Gr-MLHFQ, and baseline DHFKS knowledge score. These variables were selected after the review of relevant articles, as they are considered to be associated with the HR-QoL.^30,31

Ethical considerations

This research was established on a voluntary basis for each patient; thus, an informed consent form has been signed from each participant. A brief description of the research study and the answers to possible questions were given to the patient. Every patient was informed that he/she could step out of the research at any time he/she wished. All procedures were in line with the instructions given by the Data Protection Commissioner for maintaining confidentiality. The study involved no risk or harm to the participants, and the investigation conforms to the principles outlined in the Declaration of Helsinki.

Approvals by the Cyprus Bioethics Committee (EEBK/EΠ/2009/05) and the Data Protection Office were sought and granted. Licence by the Scientific Committee for the Promotion of Research and by the Ministry of Health of Cyprus (Y.Y. 5.14.02.4) were also sought and granted. Each hospital’s management office and cardiology departments were informed and given the opportunity to review the study protocol and make suggestions before approvals.

Results

Demographic and baseline characteristics

A total of 357 individuals were enrolled to the study between 2008 and 2017. Among those who were randomized, 19 died, 1 discontinued due to worsening HF, 2 withdrew consent, and 73 were lost to follow-up. There were no significant differences in the number of incomplete cases among groups (χ² = 5.9, P = 0.116). Participants who completed the follow-up successfully (73.4% n = 262) had a mean age of 69 (SD 11) years. The majority were men (65%), had primary school education (58%), were married (74%), retired (72%), and were living with their family (69%). A percentage of 63% had moderately impaired functionality (NYHA functional classes II and III). More than half of the patients had a previous history of myocardial infraction (MI) (57%) (Table 1).

Heart failure health-related quality of life

As for HR-QoL, higher scores on MLHFQ indicate poorer HQ-QoL. Over the 3-month period, QoL improved in the patients who were in the IGs as compared to CG. The calculated effect size was found to be medium to large for IGs (education—Cohen’s d = 0.68; telephone—Cohen’s d = 0.58; education and telephone—Cohen’s d = 0.62) (Table 2). However, the difference among groups in the adjusted analysis for the overall score was not significant (F = 2.6, df = 3, P = 0.054), but only marginally. Subscale analysis showed significant improvement only in the physical dimension of HR-QoL (F = 2.7, df = 3, P = 0.046); this was detected between the CG and in-person education group following the Bonferroni correction (P = 0.04). The differences among the groups in the emotional and social dimensions of HR-QoL were not significant [(F = 0.8, df = 3, P = 0.479) and (F = 2.2, df = 3, P = 0.095), respectively] (Table 3).

Heart failure self-care management

In HF-related self-care, higher scores indicate poorer self-care management. Greater improvement in the overall mean HF self-care management scores at 3 months was observed in the patient groups that received the telephone component, either alone [mean difference 5.4 (SD 8.6)] or in combination with the in-person education [mean difference 4.6 (SD 7.6)], which was also reflected in the corresponding Cohen’s d (0.67 and 0.62, respectively, that represents a medium to large effect size) (Table 2). The adjusted analysis showed significant differences among the groups in the overall HF self-care management scores (F = 3, df = 3, P = 0.034). Pairwise comparisons demonstrated that the declared self-care practices were improved in the education and telephone arms as compared to the standard care arm (P = 0.002). Similar results were obtained for the sodium and fluid management (F = 6.1, df = 3, P < 0.001) and physical subscale (F = 1.16, df = 3, P = 0.034), but not for the adherence subscale (Table 3).

Heart failure-related knowledge

Higher HF-related knowledge score indicates better knowledge. As shown in Table 2, the mean difference in knowledge scores at 3 months ranged between 1.1 and 1.5 in the intervention arms, with telephone intervention alone exhibiting the greatest improvement 1.5 (SD 2.6) and CG the lowest 1.0 (SD 3.0). The calculated effect size in the telephone intervention was 5.7, which is considered a medium to large according to Cohen. Adjusted analysis however did not yield significant differences in the 3-month knowledge score among the groups (F = 2.5, df = 3, P = 0.060) (Table 3).

Self-care management and knowledge correlation

According to the Pearson correlation coefficient r and as shown on the scatter plot, there was a weak correlation between HF-related self-care management and HF-related knowledge (r = 0.29, P < 0.001) (Figure 2).

Discussion

The main findings of the study were as follows: HF-related self-care management in general was improved significantly at 3 months in patients who received the telephone component either alone or in combination with the in-person education as compared to CG. Additionally, HF-related self-care management related to fluid and sodium management and physical activity was improved significantly at 3 months in all the IGs as compared to CG. Finally, HR-QoL physical dimension was improved significantly at 3 months of patients who were in the IGs as compared to CG, but not in the emotional or the social dimension of the HR-QoL.

The current research project has set the basis for HF nursing research in Cyprus (the first RCT in the country). The profile of the Greek–Cypriot population with HF has been described for the first time and characteristics that are potential barriers during the education process, such as low literacy similar to other populations, have been identified.^32,33 The mean age (x̄ = 64) of the current study is not that usual. Usually, patients are presented in an older age. Also, the NYHA class reported in most of the studies (NYHA II and III) is the same as most of the similar studies.³⁴ This may be explained by the fact that most people of that age and older in the country are not educated and they do not feel safe participating in studies. They are also taken care of by their family (children) of which the majority live (69%). Also, women (35%) seem to be under-represented in this study as in the most studies internationally.³⁵

The development of disease management programmes for HF care is an absolute requirement for effective care of this condition.¹⁵ Regardless of the model of intervention, an important focus of chronic HF disease management is optimizing evidence-based therapies, providing support to patients and their families, promoting HR-QoL, and avoiding adverse events. A HF management programme needs to be based on patient’s needs, financial resources, administrative policies, and adapted to local priorities.¹ However, the most important challenge is to find an optimal model that is cost-effective in taking care of the growing group of HF patients and addressing their specific needs. Heart failure clinics still have a place including new components such as tele-monitoring or tele-education.¹ The study of Liu et al.³⁶ proposed patient transitions from hospital to community with HF clinical and home monitoring by a HF nurse specialist. The COACH-2 study concluded that HF patients can be discharged to primary care with close collaboration with HF physicians and the emphasis should be on maintaining guideline-directed therapy.³⁷ On the other hand, the NorthStar study conducted in Denmark showed no difference in mortality rate and rehospitalizations between primary care physicians and disease management programmes.

As shown from this RCT, a structured nurse-led HF disease management programme significantly improved patients’ self-care management, HF symptom management, and HR-QoL. Similarly, Liu et al.³⁶ stated that HF nurses can improve patients’ self-awareness, knowledge, self-care, and symptom management as they continuously communicate with patients and their family. A relevant meta-analysis showed that HF nurses can help patients to learn about disease-related self-care technologies and meet unexpected patient needs with structured post-discharge follow-up.³⁸ Heart failure nurses can collaborate with multidisciplinary teams in order to provide patients with education regarding diet, disease pathophysiology, self-care management, physical activity, medication, and responses to signs of deterioration.³⁹

According to Healy et al.,¹⁵ there is an important gap in evidence, which includes important information on the main points that has to be included in a HF management programmes such as who, where, and to whom such a programme should be offered. In the literature, some of the advice given on HF management programmes are based on the experiences and are not always supported by RCTs. The current RCT is giving a lot of helpful information on the aspect of how these programmes have to delivered. It has been shown that HF self-care practices in general and HR-QoL improved significantly at 3 months in patients who received the telephone component either alone or in combination with the in-person education as compared to CG. All these are directly highlighting the importance of telephone contact during follow-up in the current population. The study provides evidence that regular communication and support (even via telephone) are important for persons with HF. Koehler et al.⁴⁰ showed a benefit of more evolved telehealth strategies, while there is a need to further clarify to whom and when this form of surveillance should be applied. Telephone regular follow-up gave the opportunity to patients to have a regular communication with a nurse specialist, resolve queries, and get support for HF self-care management. Continuing support has been raised important for patients with HF through the HF trajectory.⁴¹ Other studies have also highlighted the opportunity of personalization of education and motivation for self-care management through telephone follow-up or mobile applications.^42,43 Telephone calls seem to give the opportunity to provide support in different populations and health care systems with different economic possibilities and restrictions.⁴⁴

However, HF knowledge was not improved at 3 months in the group that received telephone and in-person education as compared to CG, and there was a weak correlation between baseline HF self-care management and knowledge. In previous studies, it seemed that having sufficient knowledge of HF is linked to good self-care practices in previous studies.^45–47 On the other hand, a question is raised: If the noticeable improvement in self-care management is not linked with improved knowledge in the telephone group, what is it? Is it the attention or the social aspect of the telephone intervention? Most of our patients had only completed their primary education. Based on this, they may have found it challenging to grasp the intricacies of their disease despite their efforts in self-care management. Sometimes, patients may be able to effectively apply self-care practices without having in-depth knowledge about their disease. Until recently (previous health care system in Cyprus), patients with HF had the opportunity to see their cardiologist every 6 months for their follow-up. No structured education was provided. For the first time, they had the opportunity to be educated and contact specialists in HF. The current study, especially the arms with the telephone follow-up, gave them the opportunity to be educated and resolve queries on a regular basis for 3 months, enhancing them for self-care management and maintaining their HR-QoL. The fact that only the physical dimension of the HR-QoL has been improved raise awareness for complementary support for patients’ social and psychological support.

Considering recent events during the management of the COVID-19 pandemic, the effective utilization of telephone follow-up has brought attention to the wider possibilities of remote monitoring technologies in the management of HF and other chronic ailments. We have learnt more than ever how crucial it is to stay in touch with patients, even if it is done remotely, because of the difficulties the pandemic created. As a result of the pandemic, tele-medicine and other modes of remote communication (even merely using the telephone) have increased in popularity among healthcare practitioners.^48,49 Hence, the last European guidelines for the management of HF include tele-monitoring for the first time, emphasizing the need for innovative approaches, and the utilization of tele-monitoring options in managing of HF. It is feasible to collect daily information remotely, such as body weight, heart rate and rhythm, blood pressure, symptoms, and medication compliance, and it serves the purpose of guiding patients, assessing their condition, and providing immediate intervention.⁸ This shows that tele-monitoring’s potential advantages and chances for enhancing HF management are becoming more widely acknowledged.

Limitations

This study has several limitations. Baseline responses to the study questionnaires were self-reported whereas the 3-month follow-up responses were collected via telephone interview; this could have introduced bias to the measurement as it was not always blinded. Loss to follow-up (73/357) could have introduced selection bias. Private sector hospitals were excluded from the study; thus, the generation of the results should be done with caution. Improvements in the IGs could be a result of their awareness of being observed (‘Hawthorne effect’) as the study was not double blinded. Finally, the majority of the participants had low health literacy level, which could have influence the accuracy of the outcome measurements during the completion of the questionnaires.

Conclusions

The current research project has set the basis for HF nursing research in Cyprus (the first RCTs in the country). The most important finding of the study was that regular communication and support (even via telephone) are important for persons with HF. Moreover, education and telephone support on self-care management may improve overall the physical dimension of HR-QoL, but not the emotional and social dimensions. That means that focusing only on knowledge and self-care management does not mean that all dimensions of HR-QoL and outcomes are established. And if the aim is to provide person-centred care, we need to design more effective and tailored management programmes focusing in socio-cultural manners as well.

Author contributions

Ekaterini Lambrinou (Conceptualization [equal], Data curation [equal], Formal analysis [equal], Methodology [equal], Supervision [equal], Visualization [equal], Project administration [equal], Writing—review & editing [equal]), Andreas Protopapas (Data curation [supporting], Formal analysis [supporting], Methodology [supporting], Project administration [supporting], Software [supporting], Supervision [equal], Visualization [equal], Writing—review & editing [equal]), Lefkios Paikousis (Data curation [equal], Formal analysis [supporting], Methodology [supporting], Software [equal]), Nicos Middleton (Data curation [equal], Formal analysis [supporting], Methodology [equal], Software [equal], Supervision [supporting]), Elizabeth D. E. Papathanassoglou (Conceptualization [equal], Data curation [supporting], Supervision [supporting]), Panayota Sourtzi (Conceptualization [equal], Data curation [supporting], Supervision [supporting]), and Fotini Kaloyirou (Conceptualization [equal], Data curation [equal], Formal analysis [equal], Methodology [equal], Project administration [equal], Software [equal], Supervision [equal], Validation [equal], Writing—review & editing [equal])

Funding

The study was funded by the Cyprus University of Technology [Startup Fund EX2007 (04)].

Data availability

The data that support the findings of this study are available on request from the corresponding author (E.L.). The data are not publicly available due to restrictions, e.g. their containing information that could compromise the privacy of research participants.

References

Author notes