A person-centered framework for designing music-based therapeutic studies in dementia: current barriers and a path forward

Abstract

Objectives:

Music-based interventions have received growing attention to improve quality of life for people diagnosed with dementia. Results of randomized controlled trials and meta-analytic reviews to date, however, reveal a lack of conclusive evidence for or against the effectiveness of such interventions. Herein, we critically review the basic assumptions and methodological issues ingrained in the cultures of research and care as they relate to evaluating music-based treatments for people with dementia, and propose a shift in the methodology by which music interventions are empirically evaluated.

Method:

We begin by reviewing existing barriers to achieving clarity on the effectiveness of music interventions, and we highlight methodological and sociocultural constraints that have limited our ability to reach concrete conclusions in research studies to-date. We then consider several key factors that have demonstrated relevance in matching people to specific music-based interventions. Based on these key factors, we developed a person-centered framework integrating elements from precision-medicine methodology to guide intervention studies.

Results:

Our organizing framework systematically integrates the following factors to inform the design of intervention studies: 1) person-centered goals and desired outcomes; 2) differences among individuals in clinical, cognitive, and historical attributes; and 3) the context of intervention and access to resources.

Conclusion:

Integration of the proposed framework into empirical investigations of music interventions for people living with dementia will inform precise and tailored interventions that will bring clarity to this growing body of research. Another aim of this framework is to foster a more humane, person-centered approach to our culture of care.

Introduction

In light of the lack of effective pharmacological treatments for Alzheimer’s disease and other illnesses that result in the syndrome called dementia (Gaugler et al. 2019), music-based activities have been increasingly investigated as potential interventions to improve cognition, mood, and quality of life in people with dementia (Sihvonen et al., 2017). As mounting evidence for the benefits of music in a wide array of populations continues to emerge - including children and adolescents (Viskontas, 2021), adult professional musicians (Sluming et al., 2002), healthy older adults (Bugos et al., 2007; Rogenmoser et al., 2018), stroke survivors (Särkämö & Soto, 2012), and older adults with dementia (Young et al., 2016) - theoretical models attempting to uncover the underlying mechanisms of these therapeutic effects have grown in sophistication, and have implicated a range of cognitive, motor and neurobiological dimensions (Särkämö et al., 2016). Specifically, researchers have demonstrated that music has an effect on neural activation and neuroplastic reorganization of networks subserving specific cognitive functions such as memory (Cuddy & Duffin, 2005) and language (Merrett et al. 2014); activation of mirror neurons and multimodal integration (Merrett et al., 2014); activation of reward, arousal and emotion networks through limbic/paralimbic brain areas (Blood & Zatorre, 2001; Menon & Levitin, 2005) that modulate the dopamine and mesolimbic system (Salimpoor et al., 2011); reduction of stress responses and of depression (Bradt et al., 2013; Nilsson, 2009b, 2009a; Okada et al., 2009; Sihvonen et al., 2017); and enhanced cognitive reserve from musical engagement (Schneider et al. 2018). More simply put, music engages numerous areas of the brain in a coordinated manner, suggesting the potential for music interventions to strengthen existing or form new memories and/or therapeutic behaviors. Practically speaking, it is widely accepted that music brings pleasure and carries with it little to none of the adverse effects associated with pharmacologic treatment (e.g. gastrointestinal symptoms, dizziness, weight loss; Raina et al., 2008).

Despite the growing interest in music-based interventions and the development of multidimensional models to explain therapeutic mechanisms, systematic reviews continue to arrive at equivocal conclusions regarding the long-term effectiveness of music interventions for people with dementia on specified outcomes. The latest systematic review on music-based therapeutic interventions for people with dementia from the Cochrane Library found only low-quality evidence¹ for improvements in emotional well-being, quality of life, and anxiety, along with low-quality evidence of little to no effect on cognition (van der Steen et al., 2018). More recently, a report led by the Global Council on Brain Health concluded that although music holds great promise to improve brain health and well-being, fundamental questions remain unanswered and much more research is required to close knowledge gaps (Global Council on Brain Health, 2020). Almost all comprehensive reviews have called for more rigorously designed studies with larger sample sizes and a longer duration of treatment (Fusar-Poli et al., 2018; Spiro, 2010; van der Steen et al., 2018). Herein, we highlight several additional important limitations of the current research and propose a new organizing framework, informed by person-centered care and precision medicine methodology, to guide the design of future studies, increase the effectiveness of music interventions within these studies, and improve the clarity of the literature. We believe that the principles outlined in our framework are applicable to a broad range of music-based interventions. Thus, we will focus on studies examining the broader category of music interventions including, but not limited to, interventions delivered by accredited music therapists. This paper is not intended to be a systematic or exhaustive review of the pertinent literature on music-based interventions, but instead highlights areas common to the design of numerous studies that we believe require improvement.

Current barriers to achieving clarity regarding music as an efficacious experience

Heterogeneity in group studies

Randomized controlled trials (RCTs) and other studies on music-based interventions to date have included heterogeneous and incompletely characterized samples, rendering outcomes inappropriate for group-based inferential statistics. In other words, potentially meaningful effects of a given intervention are obscured when individuals are combined into a single treatment group without consideration of personal attributes such as prior music training or clinical and cognitive profiles. Meta-analytic reviews face the added challenge of evaluating studies of different intervention types (e.g. music therapy vs. music-based activities; active vs. passive interventions; individual vs. group) using heterogeneous outcome measures (e.g. behavior vs. mood vs. cognition; global vs. specific cognitive domains; self-report vs. observational) with varied methodological quality (van der Steen et al., 2018). The marked variation in the above factors both within and between studies makes it nearly impossible to evaluate effectiveness, and investigators have begun to call for more focused studies of tailored interventions informed by characteristics of the individuals in question (Chang et al., 2015; Fusar-Poli et al., 2018; van der Steen et al., 2018; Vink et al., 2003).

Narrow approach to defining efficacy

Another barrier is the narrow approach to evaluating and operationalizing efficacious results. Although several authors have recommended longitudinal RCTs (van der Steen et al., 2018; Vink et al., 2003), it is important to consider whether conventional RCTs are the only or best way of demonstrating a meaningful effect (de Medeiros & Basting, 2014). RCTs that evaluate long-term effects after halting the intervention are inappropriate for music-based treatments, as theoretical models of the therapeutic mechanism do not propose that a ‘dosage’ of a music intervention should have a long-term effect similar to that of a vaccine. Like many pharmaceuticals that are used to treat chronic medical conditions, daily or regular dosages of a music intervention are likely required for a lasting effect. Evaluating the effects of an intervention in the context of a degenerative condition, which by its nature is expected to worsen over time, is particularly challenging, as positive results might be reflected by stability or a slower worsening over time (McDermott et al., 2013). Therefore, if RCTs are used to evaluate music interventions in the context of dementia care, they should prioritize the assessment of realistic, clinically-relevant short-term effects over long-term effects, including those observed during the intervention (Dowlen et al., 2021; Weise et al., 2018).

Inappropriate selection of outcome measures

Another important issue contributing to the lack of clarity surrounding the efficacy of music interventions involves the selection of appropriate and meaningful outcome measures. For instance, many studies have explored the effect of music interventions for individuals with dementia on cognition, as measured by traditional neuropsychological tests, and continue to find limited effects (van der Steen et al., 2018). However, we argue that standard tests of cognition may not be the most appropriate outcome measures when exploring the benefits of music in this group. Reasons for this are that (a) performance on cognitive tests may not be of value to the individuals concerned and (b) such tests may inspire anxiety in individuals living with dementia (Keady & Gilliard, 2002). On the other hand, other more immediately-relevant outcomes that prioritize agency, independence, and engagement may be of higher value (Harding et al., 2020; Jennings, 2009). Examples of such non-traditional measures are social cognition, social interaction, conversation, expressions of pleasure or measures of quality of life to name a few (see Harding et al. (2020) for a thorough review of core outcomes identified in consultation with key stakeholders including people living with dementia).

Failure to consider these non-traditional outcomes that are possibly more meaningful to persons with dementia as well as to their care partners may lead to the false conclusion that some interventions are not efficacious (Fusar-Poli et al., 2018). A case in point of this dilemma is a recent study that evaluated older adult choir singers against a control group of older adults on objective cognitive measures and subjective self-report measures of mood, social engagement, role of music in daily life and quality of life (Pentikäinen et al., 2021). When compared with the control group, the choir singers performed significantly better on only one of ten cognitive tasks (verbal fluency); however, on subjective scales, they nonetheless demonstrated significantly higher social integration, better general health, and higher musical engagement in daily life (Pentikäinen et al., 2021). Had the authors not included subjective measures tapping into constructs like social integration, these results may have been interpreted as largely null. Indeed, few studies have incorporated outcomes on social cognition and interaction (Chang et al., 2015; van der Steen et al., 2018), which is surprising, because selecting, listening to, and playing music can be quintessentially social experiences that support agency and are often done in the presence of others in residential and day centers. In fact, the most recently published meta-analysis on music-based interventions found very low-quality evidence of benefit on social behavior, with only three of the 22 studies including social behavior as an outcome measure (van der Steen et al., 2018). Given the importance of social stimulation for preserving overall brain health and function (Ertel et al., 2008; Mortimer et al., 2012; Norman et al., 2012), and its demonstrated relevance to quality of life for people with dementia (Allan & Killick, 2008; Cohen-Mansfield, 2018; Stansell, 2002; Vernooij-Dassen et al., 2010), this is a gap that is especially significant.

Most studies also fail to evaluate whether benefits are observed on measures of functional abilities (e.g., activities of daily living), yet the US Food and Drug Administration has emphasized that treatments for neurodegenerative disorders must show effects on functional abilities (Kozauer & Katz, 2013). Also missing from the bulk of these studies is the positive effect that may be observed in subjective satisfaction, even if not in a person’s performance on standard tests of cognition. In other words, a person’s enjoyment of a particular symphony or jazz piece may have absolutely no relation to whether or not he or she can improve recall of the date, city, county, state, etc. Indeed, the idea that music would increase joy and pleasure is in line with neuroscience models that emphasize basic reward and emotion systems (Blood & Zatorre, 2001; Menon & Levitin, 2005; Salimpoor et al., 2011). We therefore suggest several non-traditional, potentially more sensitive and relevant outcome measures to consider in future studies of music interventions in Table 1. We believe the incorporation of outcome measures such as these, especially those developed in collaboration with all stakeholders (e.g., persons living with dementia, their care partners), will be an important step towards a paradigm shift regarding outcome measurement in music intervention studies (Dowlen et al., 2018; Harding et al., 2020).

Table 1.

Example Outcome Measures for Music Intervention Studies.

Domains	Example Measures	References
Everyday Functioning	Everyday Cognition (ECog) Scale	Farias et al. (2008)
	Naturalistic Action Test (performance-based test)	Schwartz et al. (2002)
	Virtual Kitchen (computer administered, performance-based test)	Giovannetti et al. (2019)
Quality of Life	Alzheimer’s Disease-Related Quality of Life (ADRQL)	Rabins et al. (1999)
	Dementia Care Mapping Welbeing score	Fossey et al. (2002)
	Zarit Caregiver Burden Scale	Zarit et al. (1980)
Social Cognition/Interpersonal Connectedness	Social Provisions Scale (SPS)	Cutrona and Russell, (1987)
	Engagement and Independence in Dementia Questionnaire (EID-Q)	Stoner et al. (2018)
	Music in Dementia Assessment Scale (MiDAS)	McDermott et al. (2014)
Self-Efficacy	General Self-Efficacy Scale	Schwarzer and Jerusalem (1995)
	Rosenberg Self-Esteem Scale	Rosenberg (1965)
	Family Caregivers’ Self-efficacy	Fortinsky et al. (2002)
Positive Reactions During the Musical Activity	Observational/interview methodology with thematic analysis for ‘in the moment’ experiences	Dowlen et al. (2021)
Integration of Music in Everyday Life	Music Engagement Questionnaire (MusEQ)	Vanstone et al. (2016)
Person-Centered Goal Setting	Goal Attainment Scaling (GAS)	Jennings et al. (2018)

Note: This is not an exhaustive list of suggested outcome measures for music-based intervention studies, but instead summarizes key examples that reflect non-traditional domains that are less commonly used.

A new person-centered framework

Considering the above limitations endemic to research studies of music interventions (see Figure 1A for a summary), we propose a person-centered framework, integrating elements from precision-medicine methodology, to better match certain people to specific interventions (see Figure 1B). We believe that the application of this framework as a guide for study design will enhance the personalization of interventions for research participants and will increase the chances of informative and efficacious findings. The framework we propose herein systematically considers three elements: (1) person-centered goals and desired outcomes, (2) differences among individuals in clinical attributes and historical preferences/abilities, and (3) the context of intervention and access to resources.

Figure 1.

A. Barriers to achieving clarity on music interventions. B. Proposed person-centered framework to guide music-based intervention studies.

Framework element #1: person-centered goals and desired outcomes

At the outset of treatment for any condition or illness, it is considered best practice to engage in an informed discussion around treatment goals– both short and long term (Stiggelbout et al., 2012). Despite the fact that music-based interventions were originally developed with the aim of accomplishing ‘individualized goals’ (American Music Therapy Association, 2005), this initial consideration seems to be lacking in many published studies of music interventions to date (Ceccato et al., 2012; Chu et al., 2014; Guétin et al., 2009; Raglio et al., 2008). Instead, a single intervention is broadly applied to a relatively heterogeneous group of people, regardless of particular individual goals, and outcome measures determined at the outset are assessed at various subsequent times. This common practice strips individuals of the opportunity to take an active part in goal setting, which can and should be afforded collaboratively with a caregiver, commensurate with the abilities of that particular individual, even when they are diagnosed with moderate to severe dementia. In the majority of studies, predetermined goals of reducing common neuropsychiatric symptoms such as anxiety, agitation, depression, and apathy are identified by clinicians or researchers before interacting with participants (Cheung et al., 2020; Dowson et al., 2019; Garrido et al., 2020; Raglio et al., 2008; Ray & Mittelman, 2017). We raise two critical interim points about this approach.

First, neuropsychiatric symptoms are conceptually distinct and should be targeted differentially. For instance, if the goal is to target psychomotor retardation, apathy, and fatigue by increasing arousal, live or upbeat music may be most effective (O’Connor, Ames, Gardner, & King, 2009; Sherratt et al., 2004). Group settings also increase opportunities for positive interpersonal interactions that can reduce depression symptoms (Chang et al., 2015). Alternatively, if targeting worry, tension, rumination, and/or agitation with the aim of decreasing arousal, soothing music may be selected over other types, and the music should be familiar to that individual to limit confusion and focus attention on a positively experienced and familiar stimulus (Sihvonen et al., 2017). Further, individual versus group listening, potentially using headphones, has been proposed to reduce anxiety and arousal (Sihvonen et al., 2017). Interestingly, in a case study where a brief ‘emergent’ music intervention was used to address a reported high state of agitation and facilitate much needed care for an individual with advanced dementia, the musician performed fast-paced music to instantaneously engage with and match the individual’s current state of high arousal (Garabedian, 2020). Matching the musical mood, rhythm and tempo to the individual’s current physiological state (leveraging psychophysiological responses to music; Khalfa et al., 2008) may be most effective to engage and orient the person to the present moment. In this example, matched-state music provided an opportunity to engage with this individual where other conventional means had failed (Garabedian, 2020). Finally, tailoring music to individual preferences seems to be more effective at reducing ‘agitation’ than applying generic classical or relaxation music (Clark et al., 1998; Spiro, 2010).

In addition to addressing the conceptually distinct nature of neuropsychiatric ‘symptoms’ during the goal-setting phase, our second interim consideration pertains to the overarching framework and biomedicalization of nearly all behaviors demonstrated by people with dementia, often reducing the person to a description of behavioral ‘symptoms’. Such a framework is in direct opposition to a person-centered approach, particularly when it comes to collaborative goal-setting and outcome selection. For example, many conceptualize anxiety and depression as symptoms of dementia in the same way that fever is symptomatic of malaria. The way in which behaviors are viewed is also influenced by the care location (see Framework Element #3 below). The problem herein is that for many people with dementia, anxiety and depression may actually be entirely appropriate reactions to their illness and circumstances and to the meaning of the illness and its consequences (Sabat, 2001). If one could no longer sign one’s name or dress oneself, it would be entirely appropriate to feel sad, anxious, agitated, apathetic, embarrassed, ashamed. Understanding behaviors as reactions to one’s circumstances (e.g. ‘distress’ instead of ‘agitation’), in addition to being correlates of an underlying disease process, could make a great difference in our application of music in this environment, as we could learn to rely on music to offer reassurance, build coping strategies, and promote personhood instead of to ‘manage symptoms’. Further, individuals and their care partners are more inclined to take an active role in the goal-setting process when they are encouraged to express the difficulties they wish to improve upon in practical, personal terms instead of in medicalized terms, and when they are involved from the outset. For instance, there is a notable difference between ‘I wish to feel calmer and more peaceful at night’ or ‘I would like to feel more connected with others’, versus ‘reduced levels of agitation’ or ‘improved verbal fluency’. It is also noteworthy that when speaking of the effects that music has on the mood of people with dementia, phrases such as ‘alleviating mood symptoms’ and ‘reducing disruptive behaviors’ are used, whereas in healthy individuals, it is accepted that music is frequently and intentionally used to regulate our emotions and connect with others (Sloboda, 2010). We argue that while the assessment and identification of medical symptoms is necessary in certain contexts (i.e., chart notes, insurance reimbursement), it can be unhelpful and unmotivating when it comes to goal-setting and outcome selection for personalized music interventions.

Instead of emphasis placed overarchingly on symptom management, we should begin with collaboratively-established, person-centered goals such as what Kitwood (1997) defined as basic human needs, including occupation, inclusion, identity, attachment, and comfort. These needs can be supported by music-based interventions and offer a framework for setting goals and identifying outcomes (Framework Element #1). For example, occupation, defined as active and meaningful involvement in life without boredom and apathy, may be achieved if individuals’ interests, tastes and abilities are considered at the outset of music interventions. Social inclusion can be facilitated by group-based music interventions, which have been shown to increase communication, coordination, cooperation and empathy among group members (Harris & Caporella, 2014; Koelsch, 2014; Vink et al., 2003), and lead to an improved sense of social network and community (Mittelman & Papayannopoulou, 2018; Vink et al., 2003). Secure attachments become especially important as individuals with dementia increasingly find themselves in situations that are ‘subjectively strange’ (Kitwood, 1997). Again, music interventions carried out in group settings or with close friends or family offer a way to foster new or strengthen existing social bonds. Maintaining identity and agency remains important to people with dementia, as autonomy and continuity with the past can rapidly dissolve. Enabling individuals to choose their own musical selections, sing familiar songs and/or share specific autobiographical memories associated with that particular song facilitates reminiscence, reinforces one’s personal narrative, and supports agency (Camerlynck et al., 2021; Cuddy & Duffin, 2005; Garabedian, 2020). The ability to make choices becomes extremely important especially in long-term care settings where people are often unwittingly stripped of agency and directed in most aspects of their life. Identity and agency can also be fostered in cases where musical abilities are preserved where other cognitive abilities are impaired (Brotons et al., 1997; Polk & Kertesz, 1993), thus enabling individuals with even severe cognitive difficulties to experience success and self-efficacy by playing musical instruments. Comfort results from the soothing of pain and sorrow, the calming of anxiety, or the feeling of closeness, which overlaps with both attachment and inclusion (Kitwood, 1997). Music’s ability to instill comfort is well established, and can be achieved with music that is related to one’s past, whose lyrics hold special meaning (in general or in the moment), or whose rhythm and tone promote a state of calm (Spiro, 2010).

Importantly, few studies assess these basic human needs on an individual level at the recruitment or intake phase. Doing so collaboratively has the potential to drive a more meaningful, personalized goal-setting process and improve upon the selection of outcome measures, representing an important gap requiring attention in future studies. One recommended method to collaboratively outline personalized goals includes goal attainment scaling, an approach used to elicit, specify, measure attainment of, and revise personalized goals of care, as described by Jennings et al. (2018; Table 1). Some of the many benefits of goal attainment scaling are that it encourages individuals with dementia and their care partners to identify personally-relevant goals including those focused on quality of life, it enables them to define what is meant by ‘achieving’ these goals using a standardized yet individualized self-report response scale (i.e., less than expected vs. expected vs. better than expected goal attainment levels), and it allows flexible revision of goal scaling or goal selection over time. Additionally, goal attainment scaling has the benefit of enabling comparisons across people within a study, even though they may have very different treatment goals.

Framework element #2: characterizing the study sample based on clinical variables, historical preferences, and abilities

Once key person-centered goals are collaboratively established, it is crucial to tailor the intervention based on clinical and cognitive attributes, as well as personal experiences with music, to increase chances of success. Regarding cognitive abilities, matching intervention procedures to a person’s level and type of cognitive impairment will likely increase participation and efficacy. For instance, active approaches (i.e., when participants are actively involved in music-making by playing instruments, singing, or dancing) may be more beneficial for individuals with mild-moderate impairment, whereas receptive approaches (i.e., when participants listen to music that is played by someone else) may be more appropriate at later stages (Fusar-Poli et al., 2018; van der Steen et al., 2018). If the music intervention’s goal is to improve cognition, care should be taken to consider the specific systems that will be targeted. For instance, instruments may be incorporated to engage and train visuomotor systems or to facilitate motoric compensatory strategies, if those abilities remain sufficiently intact. Others have examined interventions designed to target specific cognitive domains (e.g., memory and attention; Ceccato et al., 2012). Before beginning a targeted intervention, cognitive deficit profiles should be considered to reduce frustration and promote engagement. For example, music programs that require learning new song lyrics may be frustrating to individuals with episodic recall impairment or language difficulties. On the other hand, leveraging implicit procedural memory systems (Squire, 1984) by asking these individuals to learn and reproduce other elements of the song, such as the associated manual movements, rhythms and tones, may be more rewarding, pleasurable, and ultimately efficacious. An example would be encouraging individuals to whistle, clap, or hum to songs instead, or to play songs using an instrument should they have prior experience performing on that instrument. Matching cognitive interventions to people’s specific deficit profiles has been supported in a range of complex neuropsychological syndromes, including aphasia, hemi-spatial neglect and difficulties performing goal-directed activities (Bowie et al., 2014; Giovannetti et al., 2015; Riddoch & Humphreys, 1994; Sohlberg & Mateer, 2001; Wilson, 1999). Future large-scale studies on music-based interventions for people living with dementia should consider this approach that is commonly implemented in real-world, clinical and rehabilitative settings.

A person’s history of music exposure and preferences also should be especially considered to make the intervention engaging and optimize its potential as a purposeful and meaningful activity (Kitwood, 1997; Weise et al., 2018). Previous training in music is particularly relevant for interventions that involve music-making. There is sufficient evidence demonstrating spared implicit/procedural musical memory in people with dementia, even during later stages (Baird & Samson, 2015). Thus, active interventions that involve playing an instrument should be selected for individuals with any degree of past music training, both to strengthen existing memory systems and to provide an opportunity for occupation, identity and reminiscence (Kitwood, 1997). Type and level of past training are important attributes to match people to targeted interventions and also should be examined as potential moderators of treatment efficacy to inform future intervention studies. The role of music engagement in daily life should be assessed using standardized measures such as the Music Engagement Questionnaire (MusEQ; Vanstone et al., 2016). Considering personal preferences and understanding what sorts of music were enjoyable and meaningful to a person in the decades of life preceding the diagnosis of dementia are also key to adopting a person-centered approach. The fact that a person has been diagnosed with dementia does not mean that he or she is no longer able to recognize and enjoy music that was previously enjoyed, or dislike music that was previously disliked. To use a receptive intervention consisting of only classical music with someone who has never before enjoyed listening to classical music might limit positive change and may cause frustration. Further, to ignore the person-centered factor of preference could easily obscure effects in group-based studies, again arguing against a one-size-fits-all approach. Personal preferences are routinely considered in clinical practice and although some studies have considered ‘relevant personal information’ when developing a music-based intervention (Chu et al., 2014; Guétin et al., 2009), others have explicitly stated their aim to broadly exercise cognitive function ‘without considering sociodemographic characteristics’ (Ceccato et al., 2012). We recommend against research protocols that allow little room for personalization and urge inclusion of personal preferences and music experience as key elements to the intervention design.

Framework element #3: context of intervention and access to resources

The context and environment in which a music intervention occurs is directly tied to questions of feasibility and sustainability. Group-based interventions that occur in long-term care homes may face competing logistical and medical priorities, where staff’s concerns frequently must take precedence over person-centered needs due to the professional guidelines and constraints to which staff are subject (Garrido et al., 2020). A qualitative study involving three care homes found that staff and residents expressed differing views about what makes activities meaningful: whereas staff prioritized activities that maintained physical abilities, residents prioritized the quality of an experience and its ability to address psychological and social needs (Harmer & Orrell, 2008). The use of medicalized terms and emphasis on physical abilities and safety, whose originations stem from the language of insurance reimbursement policies, often places the focus on symptoms rather than on persons, and also creates negative expectations and beliefs among the individuals concerned that can have negative effects on their cognition and behavior (Schwarz et al., 2016; Suhr & Gunstad, 2002).

The culture of care among staff frequently focuses intervention outcomes away from the person-centered factors outlined above. The discrepancy in staff’s versus residents’ priorities is also troubling because care home residents often feel frightened, lonely, or annoyed by being in a place where they did not choose to live, and it is this type of setting where activities designed to foster agency, allow choice and improve subjective quality of life are most needed and can be quite beneficial in terms of physical as well as mental health (Langer & Rodin, 1976). Ideally, efforts to merge these two sets of priorities and work together towards a shared understanding of a multidimensional approach to health would be most effective. Researchers could benefit from an enhanced understanding of the culture of care in such settings and organize their approach accordingly, whereas professional staff could, by learning about the researchers’ approaches and foci, offer insightful suggestions to organize the work seamlessly within the confines of the care setting. There have been recent calls for improvement in care homes to provide sufficient creative arts-based activities (Spiro, 2010), as well as calls to counteract assumptions that people with dementia have limited capacity for meaningful and intentional experiences (Kontos & Grigorovich, 2018). We hope that this collaborative effort continues to gain momentum so that interventions like music-based activities can be provided in a manner that is safe while also aligning with residents’ priorities.

Considering financial resources, group-based music interventions may be more readily available to individuals who have the resources to attend day centers or community settings where visiting musicians, instruments or sound systems are readily available. This highlights the need to ensure music interventions are accessible to everyone, regardless of socioeconomic status. It is also critical to consider the context of an individual’s day to day life and whether their environment would feasibly support long-term benefits through continued engagement with the intervention. For instance, when testing an intervention involving playing of musical instruments, investigators should consider the likelihood of adoption outside of the study context – that is, does this person have the resources to (a) purchase the instrument, and (b) continue playing the instrument if they require assistance from a paid or family caregiver? Fortunately, advancements in music delivery mechanisms through increasingly accessible digital devices (e.g. smartphones, online personalized libraries) are enabling greater access to music than ever before, and digital platforms could serve as a cost-effective approach to disseminate music-based interventions. Further, remote availability of community and healthcare-related resources (such as group music-making and music-listening sessions via virtual videoconference platforms) has expanded as a result of the COVID-19 pandemic, and it is possible that these mechanisms will continue to enable a wider group of individuals to access these services who would otherwise not be able to attend in-person. Still, digital technology is not affordable to all and may require training (e.g., use of new software, charging headphones, comfort with videoconferencing, etc.), and these factors must be considered when designing and selecting interventions.

Conclusion and future directions

In this paper, we began by highlighting three main barriers to evaluating music-based interventions in dementia: (a) heterogeneity in extant studies with respect to intervention type, study sample, and outcome measures; (b) a narrow approach to defining efficacy in these studies; and (c) inappropriate selection of outcome measures based on a culture of care that prioritizes a medical/symptom-centered versus person-centered approach. To redress these barriers, we proposed a person-centered framework to focus future studies with more uniformly characterized subgroups and matching individuals to interventions that have a greater likelihood of effecting a positive, ongoing influence on their quality of life. This framework requires consideration of (1) person-centered goals and desired outcomes, (2) differences in clinical, cognitive, and historical attributes, and (3) the context of intervention and access to resources.

As the global incidence of dementia continues to rise, clinicians, caregivers and researchers in the fields of aging and dementia are in dire need of effective, safe, and cost-efficient interventions that meaningfully improve the lives of individuals with dementia. Music-based interventions offer a promising option if targeted and evaluated effectively. There are, however, systemic barriers in U.S. culture, at least. For example, at present, ‘music therapy’ (in its official accredited form) has not gained blanket coverage. In the private healthcare sector, most insurance companies review music therapy claims on a case-by-case basis. With regard to government funded programs (i.e., Medicare and Medicaid in the U.S.), music therapy is considered a covered service only in select situations. These include in a Partial Hospitalization Program as part of its psychiatric service, where Medicare funding funnels through the facility to the music therapist, and the Medicare Prospective Payment System which allocates a set amount of funds to a facility for a given diagnosis in a given year. Neither of these scenarios allow music therapists to bill insurance directly. Further, Medicaid waivers provide reimbursement for music therapy delivered in a home, community care or rehabilitation setting; however, these waivers are not consistently available across states. In fact, many states do not recognize music therapy credentials. Herein, however, we are addressing all music-based interventions in addition to accredited forms of traditional music therapy.

To secure more consistent funding and increase support and dissemination of broader music-based therapeutic interventions (i.e., beyond traditional ‘music therapy’ approaches) to a wider range of individuals, we must do more than simply show that music interventions are benign. We must demonstrate sufficiently strong scientific evidence of music’s ability to make significant, meaningful changes in quality of life for people living with dementia. To do this, we must recognize the meaningful heterogeneity among individuals with dementia, as opposed to applying a one-size-fits-all intervention and outcome measurement approach. One possible solution would be for grant-funding agencies to impose requirements for applications on music interventions to incorporate person-centered elements into their research design. As previously discussed, person-centered approaches are prevalent and successful in the field of rehabilitation psychology for a number of conditions including stroke and aphasia (Fu et al., 2020; Riddoch & Humphreys, 1994; Sohlberg & Mateer, 2001; Wilson, 1999), including the use of music-based approaches such as melodic intonation therapy (Norton et al., 2009). However, person-centered approaches have not yet gained traction in the empirical study of music interventions for people with dementia (to our knowledge). Person-centered statistical approaches, such as latent profile analysis, may also be useful in identifying more homogeneous subgroups of individuals based on dynamic patterns of an array of personal attributes (Zhang et al., 2018). Additionally, there is reason to believe that our suggested framework would apply to a number of other non-pharmacologic interventions for dementia, including art, dance, poetry, gardening, and other forms of creative expression (Deshmukh et al., 2018; Newton et al., 2021; Robertson & McCall, 2020; Swinnen & de Medeiros, 2018; Young et al., 2016). We encourage the integration of our framework and pursuit of a person-centered approach in the context of any intervention, wherever possible. It is important to note the possibility that application of our framework may yield more specific, but less generalizable results (i.e., increasingly targeted research studies designed for specific samples). Another possible concern is that interacting collaboratively with research participants, prior to beginning the intervention, could lead to expectation biases among participants and/or researchers. Although the advantages to a collaboratively-established, individualized intervention appear to far outweigh the possible disadvantages, these questions represent important areas for future research.

Finally, we must question what it means for an intervention to be clinically meaningful while using conventional outcome measures from medical research (e.g., reduced ‘agitation’, reduced reliance on pain medication, or standard cognitive test scores). This may warrant adoption of new psychosocial outcome measures including basic human needs, emotional well-being, social cognition and interaction, increased self-expression and everyday functioning/independence to achieve a more naturalistic assessment of how music can enhance the quality of the everyday lives and social wellbeing of individuals diagnosed with dementia (Sabat, 2018). As Oliver Sacks noted, ‘music is part of being human’ and ‘has a unique power to express inner states or feelings. Music can pierce the heart directly; it needs no mediation’ (Sacks, 2007). Indeed, the intervention ought to be evaluated in a way that captures its essence, and in terms of it being personally meaningful and enriching to the individuals for whom it is being provided. Through this approach, we can bring the person-centered focus into the foreground. We are, after all, seeking to increase the number of enlivening moments for all concerned.