The Feasibility of a Music Therapy Respiratory Telehealth Protocol on Long COVID Respiratory Symptoms

Abstract

Objective: This study aims to investigate the feasibility of an online music therapy protocol for individuals previously diagnosed with COVID-19, focusing on their perceptions of their respiratory symptoms and the intervention’s impact on psychosocial measures. Methods: A within-subject experimental design was applied to examine an eight-week weekly online music therapy protocol, including singing, wind instrument playing, and music visualizations. All self-report data were collected bi-weekly throughout the 16-weeks study period, including baseline and post-tests. The measures for respiratory symptoms included the Medical Research Council’s Dyspnea Scale (MRC Dyspnea), Chronic Respiratory Questionnaire-Mastery Scores (CRQ Mastery), and Visual Analogue Scale for breathlessness. The measures for the secondary psychosocial outcomes were the Beck Depression Inventory-Short Form, the Generalized Anxiety Disorder 7-item, the Hospital Anxiety and Depression Scale, the Fatigue Severity Scale, the Epworth Sleepiness Scale, the EuroQol 5-Dimension 5-Level, and the Connor-Davidson Resilience Scale. Results: Twenty-four participants were enrolled. The participants perceived a reduction in respiratory symptoms, and shortness of breath (MRC Dyspnea). Planned comparisons showed significant decreases in MRC from baseline to post-treatment (p = 0.008). The mixed-effects model, including pre-baseline and post-treatment, was significant (p < 0.001). Significant changes in Breathing VAS were consistent with improvements in MRC Dyspnea, showing a significant baseline-to-post difference (p = 0.01). The CRQ Mastery showed significant improvements from baseline to Week 12 (p < 0.001). No significant changes were observed in other secondary measures. Conclusions: Our preliminary findings suggest that this protocol is feasible, and as a result, may help individuals previously diagnosed with COVID-19 to cope with lasting respiratory symptoms and improve their perception of shortness of breath. Live music-making, including playing accessible wind instruments and singing, may contribute to an increase sense of control over breathing. As this was a feasibility study, we conducted multiple uncorrected statistical comparisons to explore potential effects. While this approach may increase the risk of Type I error, the findings are intended to inform hypotheses for future confirmatory studies rather than to draw definitive conclusions.

1. Introduction

The SARS-CoV-2 (COVID-19) is a pathogenic virus, which, to date, has taken more than 7 million lives. More than 100 million Americans have had COVID-19 (World Health Organization), and 771 million people across the globe have been affected [1]. According to U.S. Centers for Disease Control and Prevention (CDC), one in five American adults who have had COVID-19 has developed long COVID [2]. Symptoms include fatigue, dyspnea, and other pulmonary dysfunctions [3]. Recent studies have reported the burden that persistent long COVID symptoms have placed on the healthcare system. Pulmonologists have been hard-pressed to develop effective respiratory rehabilitation programs that support recovery for their long COVID patients [4,5].

1.1. Past Studies in Music and Medicine

Several former studies investigated the neurological and physiological benefits of music therapy for patients engaging in breath regulation and with pulmonary dysfunction such as COPD or asthma [6,7,8,9,10,11]. Recent studies show benefits of online singing programs that combine components of vocal exercise, singing, and mindfulness on breathing symptoms and psychological wellness for people who live with long COVID symptoms [12,13]. Interestingly, stories of how playing wind instruments helps breathing in individual cases and community programs have even been popularized in mass media [14,15]. However, online music therapy programs that utilize easy-to-play wind instruments, music imagery, and singing led by music therapists have not been formally investigated. Although both online singing programs and music therapy programs provide music-based intervention, only a music therapy program delivered by established health professionals, such as qualified, credentialed music therapists, can best achieve clinical and evidence-based goals [16].

As COVID-19 affects multiple organs and systems, current evidence shows that the effects of the virus can surpass the acute phase and manifest in residual symptoms (neurological, physical, cognitive, and emotional areas). A guide for primary care physicians categorizes the major persistent antagonists to include respiratory, cognitive, and neurological symptoms [5].

1.2. Music Therapy

Multidisciplinary long COVID care, including group therapy and support groups, is recommended for treatment [5]. Music therapy has demonstrated the potential to improve function in these aforementioned areas [11]. In the Mount Sinai Health System, the Louis Armstrong Center for Music and Medicine has a history of making provisions and interventions for adults with respiratory diseases, providing a Music for Advances in Respiration (AIR) program to treat patients with lung disease, especially those with COPD [9].

Music for AIR has successfully used wind instrument playing, singing, and music visualization within a music therapy context to ease symptoms of COPD, including dyspnea and fatigue [9]. As dyspnea and fatigue are also common persisting symptoms of patients recovering from COVID-19, Music for AIR interventions and their possible therapeutic benefits for recovering long COVID patients warrant investigation.

The current study aims to examine the feasibility and preliminary efficacy of music therapy, inclusive of wind playing, singing, and music visualizations’ effects on physical function and quality of life for adults experiencing shortness of breath who have been previously diagnosed with COVID-19. The study was the first one that applied wind instruments playing in an online format for long COVID patients. The primary goals were to increase respiratory function, reduce respiratory symptoms, and improve breathing and functional capacity for activities of daily living. We aimed to enhance psychological well-being and quality of life for adults with long COVID symptoms referred by our Center for Post-COVID Care.

2. Methods

2.1. Study Design and Participants

We evaluated the potential efficacy and feasibility of the online music-based protocol, utilizing a full within-subject experimental design, comprising a 16-week study. Participants (N = 24) were recruited primarily through The Mount Sinai Center for Post-COVID Care. Treatment from July 2021 to December 2022. The researchers (JZ and JL) conducted video conferences for physicians and nurse practitioners from the Center for Post-COVID Care. Treatment and provided detailed information and materials for study recruitment. Physicians and nurse practitioners referred their eligible patients. Posters and flyers about the program were also presented at our recruitment sites. The research coordinator (JZ) and team members contacted all referrals and screened them for eligibility.

Patients were included in the program if they: (i) were 18 years or older, (ii) had confirmed COVID-19 diagnosis, (iii) were experiencing residual pulmonary issues (shortness of breath, fatigue), and (iv) received a score above grade 3 from the Dyspnea scale of the Medical Research Council. Participants who had the presence of respiratory issues without a confirmed COVID-19 diagnosis were excluded from the study. The research coordinator (JZ) and team members reviewed the consent form and answered questions about the study from the participants upon their enrollment. None of the participants had prior experience with music therapy respiratory programs. No compensation was offered to the participants.

2.2. Feasibility Measures

Participant retention and adherence monitoring were used as measures of feasibility. Of the 24 participants enrolled, 2 discontinued participating in the music therapy group sessions due to scheduling conflicts. Additionally, five participants ceased responding to the weekly survey post treatment. Attendance was tracked over the 8 weeks of online music therapy groups, with an overall attendance rate of 83.34%. Reported reasons for absence included a new COVID-19 diagnosis and scheduling conflicts.

2.3. Procedure

Data for all measures were collected every two weeks from four weeks before the start of the intervention until four weeks after its conclusion. During the 8-week intervention period, daily responses on the MRC Dyspnea Scale were collected. All participants received surveys via secure REDCap links sent to their email addresses. Automatic email reminders were scheduled to prompt timely survey completion.

Music Therapy Intervention

Participants attended a weekly 45-min online music therapy group utilizing a respiratory protocol for eight weeks [8,11]. Prior to beginning the group, participants received a Yamaha recorder and a slide whistle to utilize within the group. The weekly music-based group was delivered over Zoom, led by music therapists from the Louis Armstrong Center for Music and Medicine who hold Board Certification in music therapy and their interns who were trained in and provided the music therapy respiratory telehealth protocol (MTRP) protocol.

The respiratory protocol was developed from the clinical practice of music therapy in pulmonary rehabilitation used in former studies with COPD and asthma patients treated by our team [8,9]. The protocol involved Music-assisted Relaxation through Music Visualization (MrMuV), wind instrument playing, and singing. In MrMuV, the music therapists play repeated indicated chord progressions of various harmonic constructions, based on an assessment of presentation and mood in arpeggiated style on guitar or piano, entrained to the participants’ breathing while verbally guiding the participants to focus on breathing and to visualize a place of meaning (such as pleasant/desired/favorite place to be). This was followed by wind instrument play, whereby the therapists facilitated the participants’ playing of a recorder or slide whistle freely, implementing clinical improvisatory idioms that led to successful, creative exploration. This seemed to prompt pulmonary incentive and was followed by progressively prolonged breath exhale duration in a relaxation-inducing mode, providing musical structure on guitar or piano, with harmonically supportive cadences to instill supportive exhalation.

The interventionists led vocal play on ascending and descending pitch patterns. The range and duration of the phrasing were adapted to the observed participants’ singing capacity. The group participants were encouraged to select and sing their preferred songs, which were reflective of personally meaningful associations, known as ‘songs of kin’ [17], once their voices were warmed-up. The cultural and emotional meanings and personal relevance of the songs were shared, as verbal processing post-singing provided a forum whereby therapeutic community-building led support to the participants through the songs. The building of common themes provided the participants with meaningful commiseration on the many frustrations of enduring long COVID symptoms with the facilitation of a music therapist.

2.4. Outcome Measures

2.4.1. Primary Measures

The Dyspnea Scale by the Medical Research Council (MRC Dyspnea Scale) [18] was the primary outcome measure to assess the efficacy of weekly online group music therapy for individuals experiencing residual respiratory complications following COVID-19 (or presumed) diagnosis in improving the perception of respiratory symptoms. The self-reported score on the MRC Dyspnea Scale was collected bi-weekly throughout the 16-week engagement of the study and daily during the 8-week music therapy group.

2.4.2. Secondary Measures

The secondary measures were implemented to assess the efficacy of weekly online group music therapy for individuals experiencing residual respiratory complications following COVID-19 (or presumed) diagnosis for depression, anxiety, quality of life, fatigue, sleep, and resilience.

The secondary measures included the Chronic Respiratory Disease Questionnaire-Self Reported (CRQ-SR, CRQ Dyspnea, CRQ Fatigue, CRQ Emotion, CRQ Mastery) [19], the Visual Analogue Scale for breathlessness (VAS) [20], the BDI-SF (Beck Depression Inventory-Short Form) [21], the Generalized Anxiety Disorder 7-item (GAD-7) [22], the Hospital Anxiety and Depression Scale (HADS Anxiety, HADS Depression) [23], the Fatigue Severity Scale (FSS) [24], the Epworth Sleepiness Scale (ESS) [25], the EuroQol 5 Dimension 5 Level (EQ-5D-5L) [26], and the Connor-Davidson Resilience Scale (CD-RISC) [27]. The self-reported scores on all the secondary measures were collected bi-weekly throughout the 16-week engagement of the study.

2.5. Statistical Analysis

Descriptive statistics (mean, sd or n, %) were calculated for all participant characteristics and outcome measures. Distributions of outcome measures were examined for normality; if necessary, transformations (log, square root) would have been applied to normalize the measure; none required transformation. To measure the preliminary efficacy of the music therapy intervention, we evaluated a series of exploratory within-subject mixed effects models with random subject effects examining change in MRC scores over time. These types of models are particularly appropriate for Intent-To-Treat (ITT) designs, because all available data are used to estimate model fit and complete cases are not required. Thus, no estimation of missing data (e.g., imputation) was necessary beyond establishing that the data were missing at random.

The primary outcome model included baseline, mid-treatment (week 4), and post-treatment (week 8) scores, with planned comparisons among each pair of assessments. We then extended this model to include the two pre-baseline assessments (at −4 and −2 weeks) and the two post-treatment assessments (at 10 and 12 weeks), again with planned comparisons among assessments, with pre-baseline assessments compared to baseline assessments to examine stability, and subsequent assessments compared to each other. Since this is a preliminary study, we were interested whether we could identify points at which improvement was likely to occur, as well as which assessments were most likely to show notable change. LSD corrections were applied to all pairwise comparisons.

We applied similar models to the secondary outcome measures. We examined the association of demographic characteristics and MRC to identify potential covariates. We evaluated change in the context of MCID and examined potential associations with baseline characteristics. Alpha was set at p < 0.05 throughout. All analyses were conducted using IBM SPSS v. 29.

3. Results

3.1. Demographics

In total, 24 participants consented and were enrolled. Please see

Table 1. Demographic characteristics of the participants.

Variable	Number	Percentage (%)
Gender
Female	17	70.8
Male	5	20.8
Not reported	2	8.4
Age
Mean (SD), Range	45.9 (13.6), 22–70
Race
Asian	1	4.2
Black or African American	5	20.8
White	13	54.2
More Than One Race	1	4.2
Not Reported	4	16.6

for the demographic information.

Mean MRC scores with 95% confidence intervals are presented in Figure 1. The primary outcome analysis focused on the three assessments during the treatment period, weeks 0, 4, and 8. The overall Time effect was significant (F(2,41.36) = 4.47, p = 0.018), with an average drop in MRC of 0.6 points. Planned comparisons showed significant decreases in MRC from baseline to post-treatment (∆ = 0.58, p = 0.008) and from midpoint to post-treatment assessments (∆ = 0.52, p = 0.021). The comparison from baseline to midpoint was not significant (∆ = 0.07, p = 0.738), suggesting that, on average, the treatment effects were most potent after the midpoint.

We expanded the mixed effects model to include the pre-baseline and follow-up assessments. Again, the overall Time effect was significant (F(6,110.79) = 4.25, p < 0.001), confirming the overall trend of decreasing MRC scores throughout the study period. Planned comparisons confirmed the baseline-post effect (∆ = 0.59, p = 0.002); this effect remained significant at both follow-up assessments (baseline-10 weeks ∆ = 0.47, p = 0.046; baseline-12 weeks ∆ = 0.44, p = 0.027). The two pre-baseline visits did not differ significantly from baseline (∆ = 0.22, p = 0.239; ∆ = 0.08, p = 0.679, respectively), confirming the stability of the pre-treatment MRC level. Although the pattern of means suggests a slight increase in MRC scores at follow up, neither assessment showed a significant return of symptoms compared to the level achieved at post-treatment (post-10 week ∆ = 0.12, p = 0.598; post-12 week ∆ = 0.15, p = 0.447).

In addition to examining statistically significant change during the intervention, we were interested in the proportion of participants who achieved and/or maintained clinically meaningful change, defined as a ≥1 point reduction of MRC score. Overall, 11 of the 24 participants (45.8%) achieved threshold MCID, with 6 (25.0%) meeting the threshold by the midpoint assessment. Once achieved, all 11 participants remained above the threshold through the remaining follow-up assessments.

We wondered if we could identify any baseline characteristics that would distinguish between these two groups: those who achieved MCID and those who did not. We found two measures that significantly distinguished between the groups: MRC Dyspnea (t(1,22) = 3.22, p = 0.004) and ESS Total (t(1,15.97) = 2.37, p = 0.031), and one indicating a strong trend, FSS Total (t(1,19.39) = 2.07, p = 0.052). Participants with a clinically meaningful decrease in MRC scores (≥1 point) scored significantly higher on each measure. As these measures represent Dyspnea, Sleepiness, and Fatigue Severity, it seems that greater initial symptomatology may be associated with achieving clinically significant improvement, perhaps indicating a ceiling effect for participants who began the study with less impairment (and therefore less room for improvement, at least at the MCID level). This finding may point to important considerations for inclusion criteria in subsequent clinical trials.

3.2. Secondary Measures

We used the same mixed-effects model as described above for each of the secondary outcomes, examining significant changes over time for the treatment-related assessments (baseline, midpoint, and post-treatment). We also examined pairwise comparisons for each study visit, including pre-baseline and follow-up assessments, with LSD correction for multiple comparisons. Because these are exploratory analyses, we did not correct for experiment-wide multiplicity error.

Two secondary measures (See

Table 2. Results of the secondary measures.

	Mean Difference	p-Value	Cohen’s d	Confidence Intervals
CRQ Dyspnea	0.02	0.95	0.042	[−0.57, 0.61]
CRQ Fatigue	0.26	0.16	−0.449	[−0.11, 0.64]
CRQ Emotion	0.03	0.88	−0.023	[−0.37, 0.42]
CRQ Mastery	0.60 *	0.01	−0.602	[0.17, 1.02]
Breathing VAS	−1.15 *	0.01	0.625	[−1.97, −0.33]
BDI-SF	−0.50	0.17	0.290	[−1.22, 0.22]
GAD-7	−0.09	0.90	0.087	[−1.57, 1.38]
HADS Anxiety	−0.32	0.61	0.146	[−1.60, 0.95]
HADS Depression	−0.29	0.61	0.095	[−1.43, 0.86]
FSS Total	−3.55	0.29	0.490	[−10.30, 3.20]
ESS	−1.24	0.11	0.419	[−2.79, 0.30]
Eq_5d_5l	4.90	0.19	−0.282	[−2.48, 12.27]
CDRISC Total	1.67	0.35	−0.243	[−1.91, 5.25]

Note: the results indicate changes from 0–8 weeks. * indicates statistically significant changes.

) showed significant treatment effects during the intervention period: CRQ Mastery and Breathing VAS are presented in Figure 2 and Figure 3. CRQ Mastery scores showed steady gains, with a significant overall Time effect (F(2,36.98) = 4.08, p = 0.030). Planned comparisons showed significant differences from baseline to post-treatment (∆ = 0.60, p = 0.01); the midpoint did not differ from either assessment. The inclusion of pre-baseline and follow-up assessments shows that treatment gains remain, and indeed continue to improve, with significant differences from baseline (baseline-week 10 ∆ = 0.77, p < 0.001; baseline-week 12 ∆ = 0.78, p < 0.001). Neither of the pre-baseline assessments differed significantly from baseline (∆ = 0.34, p = 0.08; ∆ = 0.01, p = 0.82), although the −4-week assessment trended in that direction.

The significant changes in Breathing VAS (See Figure 4) are consistent with the improvement in the primary measure, the MRC Dyspnea score. The overall mixed effects model revealed a significant effect for Time (F(2,37.61) = 4.11, p = 0.020). Breathing VAS scores showed consistent decreases during the intervention period, with a significant baseline-post difference (∆ = 1.15, p = 0.01). Additional comparisons showed that the treatment effect was still evident at week 12, with a significant baseline-week 12 difference (∆ = 0.82, p = 0.04). Of note is the high average score at follow up week 10: the increase from baseline shows a significant loss of treatment effect (∆ = 0.99, p = 0.02), but the effects are regained at week 12. Regardless of these variations, the average remains lower than the pre-treatment assessments, which did not differ from baseline.

The analyses of the remaining secondary measures (CRQ Dyspnea, CRQ Fatigue, CRQ emotional function, ESS Total, HADS Anxiety, HADS Depression, GAD 7 Total, CD-RISC Total, BDI7 Total, FSS Total, and EQ-5D-5L) did not show significant changes during the study period.

The response rate to the surveys declined from 24 participants at the start to 19 by the end of the study, resulting in a 79% adherence rate to survey completion. The eleven secondary measures offered an opportunity to gain a comprehensive understanding of the psychological impact of the MTRP protocol. However, the volume of surveys may have contributed to the treatment burden, which could have prevented five participants from adhering to survey completion over the 16-week period.

4. Discussion

This study is among the first to present a music therapy respiratory telehealth protocol’s (MTRP) effect on dyspnea symptoms in patients previously diagnosed with COVID-19.

Interestingly, in comparing this study with recently published studies that used singing-based interventions [12,13], the MTRP is the first to demonstrate the preliminary effect on perceived dyspnea. Former research on a singing instructional program enhanced participants’ self-ratings on a Visual Analogue Scale (VAS) for breathlessness. Our study protocol, MTRP, included singing but also implemented music therapy techniques that connected participants’ musical themes with their personal experiences. Their breathing capacity was further expanded and enhanced through clinical improvisational wind instrument play.

The impact of music therapy treatment on dyspnea showed improvement from week four to week eight. It is reasonable to deduce that these positive effects were in response to the MTRP’s ardent adherence to gentle respiratory strengthening, under which the effect was productively motivated through the participants’ therapeutic wind-play, and singing. These two activities were implemented as consistent creative incentives throughout our weekly sessions. Although in-person wind instrument playing was considered to be unsafe due to COVID-19-airborne risk [28], our online sessions prevented the airborne transmission of COVID-19 and made MTRP accessible to participants who were socially isolated.

Since the participants had never played recorder and slide whistle before, they learned and played those instruments for the first time in the study, which may have contributed to the significant improvements in CRQ Mastery. The sense of mastery is one’s perception that one has control of one’s life [29]. Many participants in our study described a loss of sense of control due to the radical changes that COVID-19 brought to their lives.

A sense of mastery seemed to be a determinant of positive mental health and a protective factor that built resilience during the COVID-19 pandemic [30]. Public health agencies recommended increasing public health measures that could enhance individuals’ capacity to cope and alleviate the mental health disparities resulting from the COVID-19 pandemic.

This feasibility study explored the effects of music therapy, specifically the Music Therapy Respiratory Protocol (MTRP), on perceived dyspnea and psychological factors among individuals previously diagnosed with COVID-19. While the preliminary findings suggest potential benefits of the MTRP, the lack of a control group and the small sample size limit the generalizability of the results. To accommodate individuals recovering from COVID-19, the study utilized self-report measures, which were convenient during the pandemic. Biophysiological measures were not employed due to the likelihood of participant self-isolation. As a result, the reliance on self-report data limited insights into potential biophysiological changes in pulmonary function associated with the intervention. Future research should incorporate a control group, a larger sample size, and biophysiological measures to more robustly assess the intervention’s effects.

We want to acknowledge the limitation of multiple statistical comparisons without adjustment for multiplicity. While this approach allowed us to explore a range of outcome measures relevant to long COVID recovery and the potential impact of the music therapy sessions, it increases the likelihood of a Type I error, i.e., the chance of falsely identifying a statistically significant result. We chose not to apply formal corrections because this was a feasibility study with a small sample size and exploratory aims, where overly conservative adjustments could obscure potentially meaningful signals. Nevertheless, this analytic choice limits the interpretability of p-values and raises the risk of false-positive findings. Accordingly, these results should be interpreted with caution and viewed as preliminary, pending replication in larger, adequately powered confirmatory studies.

5. Conclusions

The MTRP demonstrated promising potential as an effective intervention that enhanced participants’ sense of coping and perceived reduction in dyspnea, even though other secondary measures did not show significant changes. It is possible that one session of music therapy per week was insufficient to produce lasting improvements in the psychosocial outcomes measured. Future studies may consider modifying the dosage of music therapy sessions and investigating how engaging in music-making—such as instrument play and singing at home—can offer a creative and cost-effective way to strengthen overall pulmonary capacity for individuals experiencing symptoms of long COVID.