Effects of Recorded Music in Heart Failure Patients

Background. Heart failure is an important public health problem. Many heart failure patients experience symptoms burden and poor quality of life, even after the improvements in pharmacological treatments. Recorded music listening was shown to improve outcomes in cardiovascular patients but it was never tested on heart failure patients. Methods. This study is a multi-centred single-blind randomized controlled trial that will involve 150 patients. Eligible patients will have a diagnosis of heart failure, in New York Hearth Association functional classification I to III, and will be recruited from seven outpatient clinics in Northern Italy. Patients will be randomly allocated in a 1:1 ratio to receive the recorded music listening intervention or the standard care for 3 months. Data will be collected at baseline and first, second and three month during the intervention, and at six month for follow-up. The following variables will be collected from heart failure patients by validated outcome measuring instruments for quality of life (primary endpoint), use of emergency services, re-hospitalization rates, all-cause mortality,self-care, somatic symptoms, quality of sleep ,anxiety and depression, cognitive state. The study has been funded by Italian Heart Failure Association in November 2014 with number 01/04. Ethical approval was gained in December 22, 2015.

Discussion. This study will be a trial to examine the effect of recorded music listening on heart failure patients and will inform clinical practice and will provide empirical data for a new music protocol intervention evidence based. The new framework may be helpful for future research focused on music effect in heart failure patients.

Sample size calculation.The sample size will be based on the primary endpoint of quality of life measured with the Minnesota Living with Heart Failure Questionnaire. Considering two balanced groups (n1 = n2), a medium effect size (d = 0.5), α error of 5% and power of 80% to detect differences between groups, it would be necessary to enrol a total of 128 (n1 = n2 = 64) patients. A medium effect size d = 0.5 implies a sample different expectation of 10 points at the Minnesota Living with Heart Failure Questionnaire, in according to study of Parati and colleges (Parati, Malfatto et al. 2008). Given the ordinal nature of the variable, and assuming a normal distribution of scores, to maintain power to the expected value (80%), we should multiply by π / 3 (asymptotic relative efficiency value), obtaining a total of 134 subjects. Finally, assuming a drop-out of 10% for group (Park, Park et al. 2012), it will needed to enrol a total of 150 patients, 75 subjects per group. Data analysis.

Data from patients will entered into spreadsheet files and checked for data errors independently by another researcher. Analyses will conducted using SPSS 19.0 (IBM Corporation, Armonk,NY). Descriptive statistics, mean, standard deviation, frequencies, median and interquartile ranges will be used to describe scale scores and social-demographic and clinical data. All tests will be two-tailed. A probability value <0.05 will be considered the minimum level of statistical significance. To test the means differences between groups at baseline and at 1st 2nd, 3th month and at 6th month after enrolment, repeated measures ANOVA with Bonferroni post-hoc test will be used. If repeated measures ANOVA will show an effect of treatment, MANCOVA analysis will be used to verify possible association between the outcome and continuous predictor variables. To calculate the correlations between the scores of the different questionnaires the correlation coefficient of Pearson or Spearman will used with. Fisher's exact test or χ2 test will be used to identify differences in use of emergency services, hospitalization and mortality between intervention and control group.