Participants
The study adopted a quasi-experimental design and was approved from the Research Ethics Review Committee of En Chu Kong Hospital (ECKIRB1090503). The study was conducted in two community long-term care facilities in New Taipei City and were designated one facility as an experimental and the other as comparison group. Participants were recruited via posters and verbal advertisements made by the staff during internal activities. The selection criteria included intact cognition, the ability to understand verbal instruction and the ability to operate a VR joystick independently. The exclusion criteria included a history of hand dysfunction, severe visual and hearing impairment, being allergic to plants or pollen, and a current illness such as epilepsy or stroke.
Sample estimation
G*power 3.1 (HHU; Germany) software [25] was used to estimate sample size using an effect size of 0.35, an α error probability significant level set to 0.05 and power set to 0.8; we found that a sample size of 52 was adequate. After considering 20% attrition rate, we decided to recruit 62 participants, 32 and 30 in the experimental and comparison group, respectively.
Participants’ Enrollment and Assessment
The enrollment and assessment process is shown in Figure 1. After recruiting the two long-term care facilities, a research associate approached the executive director and staff to explain the purpose of the study and research procedure. After obtaining the administration’s consent, we distributed recruiting messages to potential participants and explained the informed consent form face-to-face. Each participant completed the consent form before the collection of baseline data, which was done one-by-one in a quiet and independent room at the facility. During the study period, 1 horticultural therapist, 2 VR operation instructors and 4 staff of station were present to ensure the smooth progress of the study. The participants of the comparison group received scheduled activities, such as physical fitness, paper cutting, etc., without any gardening activities during the intervention and follow-up period.
The Combination of 3D VR and Hands-on Horticultural Activities
The intervention program consisted of 8 two-hour sessions that were conducted once a week for 8 consecutive weeks. Before the intervention implementation, research staff advised the participants on how to wear the VR helmet, operate the VR joystick and familiarized them with the VR scenes. The participants of the experimental group were divided into 4 groups during the weekly activities, and each group was assisted by a facility staff. Before the activity, staff explained the theme of the week and performed the hands-on horticultural activity after the VR operation.
Measurement Instruments
Sociodemographic variables assessed at baseline included age, sex, and educational level.
Rosenberg Self-Esteem Scale
This scale has been used previously in a study targeting older adults [26], and has good reliability and validity [27]. The scale consists of 14 items and each item is scored on a Likert-type scale that ranges from 1 (strongly agree) to 4 (strongly disagree), with higher scores indicating a higher level of self-esteem. A sample item is “Sometimes I don't think I have any merit at all.” The Cronbach's α coefficients were 0.77 and 0. 89 at baseline and post-test in this study.
Geriatric Depression Scale-Short Form (GDS-15)
The Chinese version of the Geriatric Depression Scale [28] was used, which has a total of 15 items. All items were answered as either yes (1) or no (0). The total raw scores ranged from 0 to 15, with a higher score indicating a higher level of depression. Items 1, 5, 7, 11, 13 were reverse scored because they are of a positive nature. The same Chinese version of this scale was used to study depression and its correlates among 195 older adults living in southern rural communities in Taiwan with a Cronbach's α of 0.82 [29]. The Cronbach's α coefficients were 0.82 and 0. 77 at baseline and post-test in this study.
Perceived Isolation
A short-form scale of perceived isolation was used to measure isolation. It consisted of 3 items as follows: 1. In general, how often do you feel that you lack companionship? 2. In general, how often do you feel left out? 3. In general, how often do you feel isolated from others? Each item was scored on a Likert-type scale from 1 (never) to 4 (often), with higher scores indicating a higher level of perceived isolation. The Cronbach's α of the scale during its development was 0.70 [30]. The Cronbach's α coefficients were 0.81 and 0.86 at baseline and post-test in this study.
Perceived Mastery
The Chinese version of the Perceived Mastery Scale [31] with a total of 7 items was used. Each item was scored on a Likert-type scale from 1 (strongly agree) to 4 (strongly disagree). The total raw score ranged from 7 to 28, with a higher score indicating a higher level of perceived mastery. A previous study indicated the Cronbach's α coefficients were 0.82 [32]. The Cronbach's α coefficients were 0.78 and 0.82 at baseline and post-test in this study.
Achievement Motives Scale (AMS)
The AMS consists of two kinds of motives, referred to as Motive to Achieve Success, (MAS) and Motive to Avoid Failure (MAF) [33]. A short-form scale with 10 items was used to measure achievement motives, with 5 items each for MAS and MAF. Each item was scored on a Likert-type scale from 1 (strongly agree) to 4 (strongly disagree). A previous study indicated the Cronbach's α coefficients were 0.88 and 0.86 at MAS and MAF, respectively [34]. The Cronbach's α coefficients were 0.79 and 0.82 at MAS and MAF at baseline and post-test in this study.
Horticultural Knowledge
To examine learned horticultural knowledge, we designed a post-session evaluation with 2 questions according to the learning objectives and contents of each session. The program consisted of 8 sessions, thus, participants completed 16 knowledge questions.
Data Analyses
Descriptive analyses were conducted for demographic and outcome variables. A 2-tailed t test and χ2 test were used to compare differences in age, sex, education level, and chronic disease history between the experimental and comparison groups. A generalized estimating equation (GEE) was used to investigate the effect of time point, group, and their interaction on the outcome variables. GEE analyses enable an understanding on the patterns of change and their effects at both the individual and group levels [35]. Statistical analyses were conducted using SPSS (version 23.0; IBM Corp).