Can music reminiscence approaches be used in moderate–severe dementia? A pilot of music mirrors

Project

The project examined the implementation of MM at a specialist care unit in South London. This unit provides specialist assessment, treatment, accommodation and 24-hour nursing care for people living with advanced dementia who have behavioural and psychological symptoms of dementia. The intervention was piloted with three residents with moderate to severe dementia to determine whether it could be developed further. It was hoped that one of the caveats in music-based intervention research could be addressed by taking emotional aspects of well-being into account. See Table 1 for resident demographics.

OPEN IN VIEWER

Table 1.

Resident demographics.

Resident	Age	Diagnosis	sMMSE score (date of sMMSE)	Length of stay in current unit (months)	Other notes
Henry*	80	Amnesiac syndrome Korsakoff’s psychosis	17/30 (10/2016)∼1 year prior to project	77	Tendency to avoid unfamiliar company, rare verbal communication
Jerry*	79	Vascular dementia	16/30 (11/2016)∼1 year prior to project	17	Auditory and visual hallucinations
Dennis*	84	Vascular dementia	Unable to assess	24	Mostly non-verbal

These three residents were selected on the basis that, in comparison to other residents of the unit, each had family members able to assist, were known to enjoy music and were not participating in other interventions.

Procedure

Where residents lacked capacity to consent to participate, families were contacted to obtain consent on their behalf. Residents who were able to communicate their memories and preferences were interviewed to begin to create the MM; interviews with their family members also occurred. Although usually the MM is created online as a digital resource, in this instance, due to inconsistent Wi-Fi access, it was printed out as a hard copy and the music loaded on a CD and as a playlist on a tablet computer (see example in Supplementary Appendix 3).

Residents attended two sessions per week over a 6-week period, facilitated by an Assistant Psychologist and took place in the resident’s bedroom or the communal area. Each memory was read aloud and then the respective track was played. Visual aids in the form of printed out pictures and the video clip of the music were used as material supplementary to that of the original MM to increase communication, promote engagement in the session, as well as stimulate reminiscence. After each song, an attempt was made to engage the resident in conversation about the song and the memories elicited from it.

The Assistant Psychologist rated each resident for mood, activity, engagement and enjoyment before, during and within an hour after each session using an assessment form developed specifically for the project (Supplementary Appendix 2). As a secondary qualitative measure, independent ratings were sought from staff members who were asked to comment on the resident’s mood before and approximately 30 minutes after the session.

Feasibility

Feasibility of the intervention was determined by assessing whether the practical steps to deliver the intervention were possible (i.e. availability of rooms and use of equipment), as well as whether the participants were able to meaningfully engage in a manner consistent with how the intervention was intended to be run. See Table 2.

OPEN IN VIEWER

Table 2.

Feasibility data for music mirror intervention.

Resident	Able to have conversations (% of sessions)	Able to indicate songs remembered (% sessions)	Able to reminisce?
Henry	88.9	100	Yes
Jerry	100	100	Yes
Dennis	0	80	Possibly

As shown in Table 2, the majority of the time, Henry and Jerry were both able to have a conversation (i.e. reciprocal or verbal responses to questions). Furthermore, they were both able to reminisce, meaning that, in at least one session, they both responded to and/or talked about memories evoked by the music/memory, either spontaneously or when prompted with questions.

Due to Dennis’s impaired verbal communication, ability to converse, indicate song recognition and reminisce was judged using non-verbal cues. In eight of 10 sessions the resident gave non-verbal responses to questions, for example nodding and/or giving thumbs up to indicate yes, or taking off the headphones indicating no. No responses occurred in two sessions. In some instances, Dennis indicated that he did remember the song or the memory by nodding his head, when prompted with questions, which could indicate reminiscing.

In terms of practicalities of running the session, although resident’s bedrooms were available, mobility issues and resident preference meant the communal area was the most appropriate space for some sessions. Henry and Jerry utilised both locations, while Dennis only used the communal area. The use of headphones and tablet made this a feasible option.

Acceptability

Acceptability of the intervention was assessed by recording how many sessions each resident attended, how long residents remained in the session, and via Assistant Psychologist rated enjoyment. See Table 3.

OPEN IN VIEWER

Table 3.

Acceptability data for music mirror intervention.

Resident	Number of sessions attended (12 offered)	Mean length of session (minutes)	Enjoyment (mean)	Sustained engagement?
Henry	9	18.75	4.33	N
Jerry	12	26.67	4.5	Y
Dennis	10	27.92	3.3	N

Residents were approached to participate in the sessions up to four times to allow for fluctuations in their presentation and competing environmental demands. Overall, the clients accepted the majority of sessions.

See Table 3 for mean length of sessions. The modal length of engagement per session, measured in 5-minute blocks (e.g. 0–5 minutes, 5–10 minutes, 25–30 minutes etc.), was 15–20 minutes, 25–30 minutes and 15–20 minutes for Henry, Jerry and Dennis, respectively. It is important to note that length of engagement varied significantly between residents and sessions. Jerry was able to sustain engagement for the entirety of the session, whereas Henry and Dennis were unable to do so. Engagement for Henry and Dennis was intermittent throughout the sessions, with length of engagement ranging from 0–5 minutes to 25–30 minutes.

An observer-rated measure was taken to get some indication of how much the resident appeared to not only accept but also enjoy the sessions. This was judged on a Likert scale with 5 = strongly agree and 1 = strongly disagree with the statement; ‘the resident enjoyed the session’, where possible residents were asked if they enjoyed the session. When responses were given, these were combined with the rater’s observations allowing for more accurate ratings. All three residents had average ratings of agree to strongly agree, suggesting an overall enjoyment of the sessions.

Efficacy

Initial indication of clinical efficacy was assessed by observing any change in participant mood across the session and after the session. Tables 4 and 5 display changes in mood ratings from before the session (T1) to end of the session (T2) and end of the session (T2) to within an hour after the session (T3). For ease of analysis, mood was broadly rated as ‘positive’ (e.g. happy, content, relaxed, talkative and responsive), ‘neutral’ or ‘negative’ (e.g. disinterested, subdued and sad). Mood ratings were based on observed behaviour, facial expressions and any verbal or non-verbal communications. Supplementary Appendix 1 displays the frequency of the observed behaviours during the MM session. This information was used to infer both enjoyment of the session and specific mood. For example a positive mood such as happy would have been inferred from behaviours such as dancing, smiling, singing, tapping etc.

OPEN IN VIEWER

Table 4.

Changes in resident mood from T1 to T2 as a result of music mirror intervention.

Resident	Mood change T1–T2 (frequency)
	Positive	Negative	Remained positive	Remained negative	Remained neutral
Henry	4	0	4	0	1
Jerry	9	0	3	0	0
Dennis	3	4	2	0	3

OPEN IN VIEWER

Table 5.

Changes in resident mood from T2 to T3 as a result of music mirror intervention.

Resident	Mood change T2–T3 (frequency)
	Positive	Negative	Remained positive	Remained negative	Remained neutral
Henry	0	4	3	0	1
Jerry	0	5	7	0	0
Dennis	3	3	0	0	3

On occasions, T3 ratings were not possible due to residents’ schedules.

A positive change was a move from ‘neutral/negative’ to ‘positive’ or ‘negative’ to ‘neutral’. A negative change was a move from ‘neutral/positive’ to ‘negative’ or ‘positive’ to ‘neutral’ (see Figure 1). No change was when the mood of the resident remained positive, negative or neutral.

OPEN IN VIEWER

Figure 1.

Figure illustrating positive and negative changes in mood.

Overall, for Henry and Jerry, the majority of mood changes from T1 to T2 were positive (Table 4). Furthermore, when there was no change, the majority of times this was continued positive mood. This cannot be said for Dennis who had fairly evenly distribution of positive, negative and no changes.

For T2–T3 (Table 5), there was mostly a negative mood change from positive to neutral mood.

Qualitative observations

In addition, significant behavioural observations by the Assistant Psychologist and staff were provided as qualitative outcomes.

Limitations

There are several limitations that warrant comment. Firstly, self-report mood ratings were unable to be effectively ascertained from residents; thus, it was necessary to rely solely on observer-rated measures. There are several difficulties with judging mood purely on observations and behaviours and so the accuracy of the mood reports is attenuated. Furthermore, supplementary reports on mood from staff often came in the form of behavioural observations, such as ‘alert’ or ‘calm’, which then had to be interpreted into moods. This is consistent with research by Fusar-Poli et al., (2017) and Istvandity (2017) who note the difficulty associated with mood/emotion assessment with this population. Some of the data for T3 ratings were missing either due to scheduling, such as unknown doctor appointments or the need for uninterrupted mealtimes meaning that ratings could not be done within the hour. Furthermore, no explicit cognitive or quality of life measures were included.

Secondly, given the advanced dementia of the residents who participated in this project, explaining what it was and the process was not possible, as initial attempts of detailed explanations, or reading the memories in the first person (as they are written) resulted in more confusion or loss of attention.

Thirdly, there were often distractions during the sessions, especially for those that took place in the communal area, for example noise from the TV/someone shouting/singing/talking loudly/alarms. This could impact on both engagement and conversation if the session was interrupted, increasing the likelihood of the resident switching attention. Nonetheless, the use of the tablet and headphones did limit these interruptions significantly.

It is also important to note that having access to a tablet computer and a lot of time within the day to complete the sessions was possible; however, this may not be generalisable to others who attempt to replicate. Therefore, considerations such as scheduling time for sessions (including time to reapproach residents on several occasions), access to headphones and access to Internet and/or portable music players are essential.

Fourthly, the small sample size and lack of variety within the sample (e.g. two out of three residents have vascular dementia and all were male) limits the generalizability of results.

Finally, there is no definitive way of knowing if the positive change seen across sessions is due to the music or due to the regular one-on-one engagement with the Assistant Psychologist.

Future directions

Further controlled studies with more varied participants (e.g. including females and other types of dementia) are needed to determine the efficacy amongst more advanced dementia populations. Nevertheless, in the context of the care unit, findings support the continued use of MM with the residents in this project and others residing at the care unit.