This study intends to use Play Intervention for Dementia (PID) to promote cognitive functions of people with early to moderate dementia.

This is a cluster randomized controlled trial aims to recruit 38 participants from two study sites. Participants will be randomly allocated into intervention or wait-list controlled groups. Both groups will receive the same content of PID programme at different time. The PID consists of 12 weekly sessions. Each session lasts for 45-75 minutes. The PID will be facilitated by a play specialist, trained elderly volunteers, and centre staffs. Cognitive functions will be evaluated with Montreal Cognitive Assessment, Fuld Object Memory Evaluation and Modified Verbal Fluency Test, conducted by a trained research assistant blinded to group allocation. Centre staffs (not involved in the PID) will be interviewed. Every alternate session will be video-taped for understanding the experience of the participants in the PID programme.

Methodology To investigate the feasibility of using Play Intervention for Dementia for enhancing the cognitive functions of people with dementia.

Research objectives

1. To examine the feasibility of using Play Intervention for Dementia for enhancing cognitive functions.
2. To explore the effects of Play Intervention of Dementia on cognitive functions.
3. To examine the experience of participating in the play intervention by people with dementia.

Research design This study is a feasibility study, designed as a cluster randomized controlled trial. Two study sites (day care centres) have agreed to join this project and will be randomly allocated to either PID group, or wait-list control group.

A mixed methods design will be employed to address the three research objectives. To evaluate the feasibility of Play Intervention, information about the ease of recruitment, attendance, occurrence of adverse events during play intervention will be collected. Staff perceptions upon play intervention will be explored through semi-structured interview. To explore the preliminary effects of Play Intervention, participants' cognitive functions will be measured at baseline (T0), after 2 months (T1), and post intervention (T2). To examine the experience of Play Intervention by people with dementia, every alternate session will be videotaped for analysis.

Participants Participating centres will recruit and screen the eligibility of participants. After obtaining initial consent from the potential participants and their next of kin, his/her eligibility will be verified by the research assistant.

Participants are community dwelling, at the age of 65 or above with a medical confirmed diagnosis of any type of dementia (such as Alzheimer's disease, vascular dementia, etc.). He/she is at the stage of early to moderate dementia, at the stage of 4-6 screened by Global Deterioration Scale, and able to sit independently and participate in table activities; and physically stable.

There was no empirical research conducted on play intervention for dementia. As play intervention is regarded as a general cognitive stimulation activity in this study, sample size will be estimated with reference to related literature. Based on the systematic review of cognitive stimulation on cognitive functions, moderate effect size is found (SMD = 0.41, 95% CI = 0.25, 0.57), with the significance level of 0.05, power 0.8. Assuming of 20% attrition rate, 188 participants / group will be needed. This study is a feasibility study, and therefore only 10% of main study will be recruited, i.e. 19 participants/group, 38 participants in total.

Intervention The Play Intervention for Dementia was developed by Prof A. Ka Tat Tsang of the University of Toronto and his research team in 2013. It was based on the Strategies and Skills Learning and Development (SSLD) model, for promoting the quality of life of individuals with dementia. According to the SSLD model, human actions are driven by unmet needs. Maslow (1943) stated that people are motivated to achieve the needs, including physiological needs, safety needs, belonging needs, esteem needs, and self-actualization needs. People with dementia also have these needs, but they have difficulties to satisfy the needs because of the diminished cognitive abilities. Despite of the neuropathological changes caused by dementia, in play intervention, needs of belonging can be satisfied through social interaction in the group activities; needs of esteem can be contented through playing with toys independently in their own way and being respect throughout the process; and needs of self-actualization are fulfilled by realizing one's ability and creativity.

The PID programme has been piloted in Yee Hong Geriatric Centre, Toronto, Canada and Yan Oi Tong Day Care Centre of Hong Kong. Other than the intended outcomes (i.e. mood and quality of life), general improvement in cognitive functions (including attention, memory, quantitative ability, responsiveness, verbal expression) was noted by the family relatives and the researchers.

This study will replicate the intervention conducted in the previous trials. Each group will consist of 12 - 20 elderly, who will be further divided into three to four subgroups. They will participate in a 12 weekly session for 45 - 75 minutes. In each session, the participants will have warm up activities and play with toys with the other members in the subgroup. The toys have different characteristics that would stimulate gross/fine motor skills, proprioception, neuro-corporeal integration, sensory-motor coordination, cognitive capacity, and emotional expression. Examples of these toys are Jenga, Domino, Board Games, beach ball and plush toy. The session will conclude with playing a percussive music instrument. Play assistant (trained elderly volunteer) will facilitate the play intervention by setting up the play area, encouraging participation, offering assistance with minimal instructions, providing attention, and ensuring safety. They will transfer toys between subgroups every 8 - 10 minutes. It is because Mayer and Griffin (1990) found that people with dementia would show considerable interest on toys and engage in stimulus-seeking and exploring behavior within 10 minutes during unstructured active play. Play assistants will be trained by a Play Specialist. Participants will remain seated at the same table. The intervention fidelity will be monitored through the programme by a Play Specialist.

The Play Specialist possesses a professional qualification in one of the human service professions by licensing or regulatory authorities (e.g. registered nurse, registered social workers), and has successfully completed at least 30 hours of SSLD and Play Intervention training in a recognized or accredited programme.

Wait-list control condition The control group will receive standard care only until the last data collection was conducted in the experimental group. Standard care consists of medical and personal care, provision of basic needs, and activities carried out as usual.

Measurements Demographic data including age and gender, socioeconomic and educational background, medical history, living arrangement and marital status will be collected.

The global cognitive functions of the participants will be assessed with the Hong Kong version of Montreal Cognitive Assessment which has excellent inter-rater reliability and good discriminatory ability. It consists of 11 questions and takes around 15 minutes to complete.

Fuld Object Memory Evaluation (FOME) will be used for evaluating memory functions of encoding, storage, and recall across five recall trials and a delayed recall trial. This instrument has been validated in the Hong Kong Chinese population with good test-retest reliability, convergent validity, and discriminative power, and independent to educational background. Verbal fluency will be assessed by the Modified Fuld Verbal Fluency (MVFT). This instrument is modified from the original Fuld Verbal Fluency Test with cultural adaptation, which is a part of the FOME. The FOME and MVFT will take only 15 minutes to complete.

Semi-structured interview

At post-intervention, semi-structured interviews with staff will be conducted to explore their perception of play intervention and its effects on participants with the intervention group. Responsible programme workers or social workers from two study sites will be invited. The questions will be but not limited to:

1. Do you think play intervention can have a positive or negative effect on the participants?
2. Have you seen a change in participants' symptoms or well-being since they start receiving play intervention?

Video-taped The engagement of participants in the play intervention will be analyzed by watching the videotapes and filling in a log-sheet by the play assistants. In the log sheet, (a) total time spent, (b) names of all the activities participants engaged in and time spent on each toy, (c) level of engagement, (d) assistance participants needed, (e) participants' verbal and nonverbal reactions, and (f) barriers to engagement in activities will be recorded.

Data collection procedure A trained research assistant will be responsible for data collection and the semi-structured interview. He/she will be blinded to group allocation.

Ethical considerations Ethical approval will be sought from the University and participating organization. The participation will be entirely voluntary. Verbal explanation and information sheet will be given to the participants and their next of kin. Written consent will be sought from next of kin. Staff to be interviewed will also be explained about the study, and their informed consent will be sought too. Participants and their next-of-kin, and staff are free to withdraw anytime of the study without any penalty. Before each session, participates will be invited to join the play intervention. If he/she declined twice, his/her decision will be respected, but the play assistant and centre staff will invite again in the later session.

The play intervention is regarded safe and no side effect has been reported. However, if there is emotional upset or any other adverse reaction, the participant will be brought away from the venue and referred to the centre social worker for follow-up. The participant may choose to continue or withdraw from the next intervention session. The questions being asked by the research assistant during data collection are neutral and should not arouse any negative feeling. However, if emotional upset is observed, the research assistant will try to comfort the participant and refer to the social worker immediately. All the data collected will be kept confidential, while only the research team can retrieve. Individual identity would not be revealed in any published materials.

Data analysis strategies

The statistical software SPSS 23.0 will be used for data entry and analysis. As this study will adopt cluster randomization, characteristics of participants from the same centre may be correlated. The Generalized Estimating Equation (GEE) model will be used to conduct the repeated measurement analysis and correct the within cluster correlation. In all analysis, 5% significance level will be used. 95% Confident interval will be calculated.

The interviews will be audio-taped, transcribed and then subjected to thematic analysis by two researchers to elicit concisely described patterns and themes.

Inclusion Criteria:

• Community dwelling, at the age of 65 or above
• With a medical confirmed diagnosis of any type of dementia (such as Alzheimer's disease, vascular dementia, etc.).
• At the stage of early to moderate dementia, at the stage of 4-6 screened by Global Deterioration Scale

Exclusion Criteria:

• Unable to sit independently and participate in table activities
• Physically unstable
• Dislikes group activity

8 weeks

We do not yet know whether additional sessions or more frequently held sessions of the CoS-Play would lead to greater efficacy.