Using Conjoint Analysis to Investigate Preferences in Horticultural Therapy Programs for Older Adults with Dementia
Article information
Abstract
Background and objective
This study used conjoint analysis to investigate the preferences for various components of a horticultural therapy program for older adults with dementia (OAD). The results of the analysis were used to identify the perceptions and preferences of OAD for horticultural therapy programs to alleviate dementia symptoms, and to provide a basis for the introduction of such programs to alleviate dementia symptoms in OAD.
Methods
In this study, a survey was orthogonally designed through SPSS Statistics with four main attributes (intervention goals, number of sessions per week, total number of sessions, and time per session), and profiles that could be presented according to the level of the attributes. The survey was conducted over a 31-day period from January 29 to February 28, 2024, with an offline survey and an online survey, in which 16 profiles were presented to the participants and they were asked to prioritize their preferences and give a score on a 5-point scale.
Results
First, when analyzing preferences by attribute, the importance of each attribute was highest for number of sessions per week (28.5%), followed by intervention goals (25.7%), time per session (23.0%), and total number of sessions (22.8%; Pearson’s R = .892, p < .001). Second, once a week (.079) was most preferred for the number of sessions per week, psycho-emotional area (.145) for intervention goals, 60 minutes or less (.125) for time per session, and 10 sessions or less (.097) for total sessions.
Conclusion
Dementia facility staff, family caregivers, and horticultural therapists of OAD preferred the following components of horticultural therapy programs for OAD: once a week for number of sessions per week; psychological and emotional domain for intervention goals; 60 minutes or less for time per session; and 10 sessions or less for total number of sessions. It is expected these findings can be used as a basis for developing horticultural therapy programs to alleviate dementia symptoms in OAD.
Introduction
Recent forecasts predict that South Korea’s elderly population will increase from 19.2% of the population in 2024 to 42.5% in 2065, as advanced medical technology and improved living standards extend life expectancy (Statistics Korea, 2024). 89.2% of the elderly live with an average of 2.6 chronic diseases (Seo and Cha, 2016). Therefore, the increase in the elderly population will lead to an increase in the number of patients with geriatric diseases, resulting in high medical expenses and social burden.
In 2023, the population aged 65 and over was 9.46 million, of which an estimated 980,000 had dementia, representing an estimated dementia prevalence rate of 10.4% (Seo et al., 2024). From 610,000 in 2014, the number of people with dementia is increasing rapidly. Dementia is a serious and frightening disease that causes economic damage and psychological distress not only to patients, but also to their families, becoming a social problem (Sun and Oh, 2017). As such, dementia is no longer considered an individual disease but a social problem. This recognition is reflected in the enactment of the Dementia Management Act in 2015 and the introduction of a number of related policy initiatives, including the establishment of centers for dementia care, the provision of dementia prevention programs, and economic support (Joo, 2017).
Various studies and treatments have been conducted on approaches to maintaining memory and orientation in older adults with dementia (OAD). The main treatment method currently is medication, a medical approach, and combinations of two or more medications are suggested depending on the symptoms (Korean Dementia Association, 2012). However, since medications have side effects and do not resolve all symptoms, it is necessary to consider non-pharmacological approaches as a supplementary treatment. Major non-pharmacological approaches include cognitive, behavioral, reminiscence, art, music, occupational, and horticultural therapies (Kim and Choi, 2010).
Horticultural therapy is also known to be an effective non-pharmacological approach for OAD. It has the advantage of being able to provide a holistic approach, including cognitive stimulation and emotional stability, in terms of the characteristic of using living things and approaching the fundamental human longing. Research on horticultural therapy for OAD has been conducted consistently, with findings indicating the effectiveness of such programs. These have included studies showing that horticultural therapy is effective in improving activities of daily living for OAD (Kang, 2014); studies on the effects of horticultural therapy in improving cognitive function, such as attention, in OAD (Yun et al., 2005; Kim and Yun, 2024), as well as its emotional effects, including reducing depression and improving psychological well-being (Kim, 2023; Hwang and Yeo, 2001); studies of its effects on physical function, including hand function and fine motor skills (Yun and Kim, 2009; Chon et al., 2009); and studies of its effects on improving sociality (Yoon et al., 2003; Kim and Han, 2004). However, a comprehensive and objective review of the effectiveness of horticultural therapy is limited by the diversity of research methods, instruments, subjects, and variables, as well as conflicting findings. As the population of OAD continues to grow and the economic and psychological burden on their families increases, there is a need to activate horticultural therapy for OAD. To this end, the design of a horticultural therapy program is important, and a survey of preferences for detailed program attributes for such design is needed.
Accordingly, this study performed a statistical analysis by setting the attributes and levels of a horticultural therapy program for alleviating dementia symptoms in OAD using the conjoint analysis method and extracting profiles from orthogonal files. To comprehensively analyze the preferences for horticultural therapy programs, a survey of the extracted profiles was conducted, targeting subjects who were most involved in horticultural therapy programs for OAD, including horticultural therapists, dementia facility workers, and family caregivers. The survey results were then analyzed by attribute and level. Based on the analyzed results, we aimed to determine the awareness and preferences for horticultural therapy programs to alleviate dementia symptoms in OAD, and provide basic data to support the introduction of such programs.
Research Methods
Survey design
In this study, a meta-analysis of studies on “horticultural therapy for OAD” was conducted in order to develop horticultural therapy programs for the relief of dementia symptoms. The items analyzed were therapeutic intervention goals, size of the participating groups, number of sessions per week, total number of sessions, and time per session. Among these items, the four items that did not have significant results in a meta-ANOVA analysis were selected as attributes for a conjoint analysis: therapeutic intervention goals, number of sessions per week, total number of sessions, and time per session (Table 1).
First, attribute 1 (therapeutic intervention goals) was selected as four levels: psychological and emotional, cognitive, social, and physical domains. Attribute 2 (number of sessions per week) was selected as three levels: once, twice, and three times per week. Attribute 3 (total number of sessions) was selected as three levels: 10 sessions or less, 11–20 sessions, and over 21 sessions. Finally, attribute 4 (time per session) was selected as three levels: 60 minutes or less, 61–90 minutes, and over 90 minutes.
Analysis of the importance of attributes and estimation of part-worth
In the survey, the total number of profile cases that can be presented according to the four main attributes (therapeutic intervention goals, number of sessions per week, total number of sessions, and time per session) and the level of each attribute was 108 (4 × 3 × 3 × 3). However, it is impossible to realistically assess respondents’ preferences for all of these profile cases, as it may be burdensome for them. Therefore, 16 profiles were extracted using the orthogonal design in SPSS Statistics, including four holdout profiles (Table 2). The holdout profiles are included and presented to test the predictive validity by comparing whether they match the actual designed profiles (Heo, 2015).
Survey and analysis procedures
A survey was conducted offline and online for 31 days from January 29 to February 28, 2024. It presented 16 profiles and asked participants to determine their preference priorities. They were asked to rate on a scale of 1–5 points: 5 for most preferred, 4 for preferred, 3 for neutral, 2 for not preferred, and 1 for least preferred. Of the 363 survey responses, 41 responses with unclear or insincere answers were excluded, and 322 responses were analyzed. The analysis was performed after statistical processing according to part-worth function models using SPSS Statistics (IBM SPSS v.19.0 program, USA; Fig. 1). Normality was confirmed using Pearson’s correlation coefficient.
Results and Discussion
Demographic analysis
Looking at respondent demographics, there were 91 males (28.3%) and 231 females (71.7%). By age group, there were 15 respondents (4.7%) in their 20s, 38 (11.8%) in their 30s, 70 (21.7%) in their 40s, 123 (38.2%) in their 50s, and 76 (23.6%) in their 60s or older. Regarding activity or work experience with OAD, the study group included 33 (10.2%) workers in dementia-related facilities, 69 (21.4%) family caregivers of OAD, 104 (32.3%) horticultural therapy program instructors for OAD, 109 (33.9%) workers in elderly care facilities or horticultural therapists with no experience with OAD, and 7 (2.2%) who fell into the “other” category, including forest interpreters, care-farming program instructors, and volunteers for the elderly (Table 3).
Preferences for horticultural therapy programs for OAD by attribute
The maximum utility rule was applied, which means that respondents chose the one with the highest utility, calculated as the sum of the part-worths of each attribute level of each selected profile, including the holdout profiles (Heo, 2015).
Preferences by attribute of horticultural therapy programs for dementia symptom relief, statistically processed according to the part-worth function models, are shown in Table 4. An analysis of preferences by attribute found that the importance of each attribute was highest for number of sessions per week at 28.5%, followed by intervention goals (25.7%), time per session (23.0%), and total number of sessions (22.8%; Pearson’s R = .892, p < . 001).

Preferences by attributes of horticultural therapy programs for older adults with dementia (N = 322)
The highest preference for number of sessions per week was once per week, followed by twice and three times per week. The greatest preference for the level of intervention goals was the psychological and emotional domain, followed by the cognitive, social, and physical domains. The highest preference for level of time per session was 60 minutes or less, followed by 61–90 minutes, and over 90 minutes. The greatest preference for the level of total number of sessions was 10 sessions or less, followed by 11–20 sessions, and over 21 sessions. The finding that 60 minutes or less is the most preferred time per session is consistent with Cho’s (2022) finding in a survey of horticultural therapy preferences that older adults preferred short intervention times. In addition, a meta-analysis of the effects of horticultural therapy programs for OAD conducted by Park (2023) showed that once a week was the most common. This appears to be in line with the highest preference for once per week in this study.
Preferences for horticultural therapy program attributes by OAD-related activity type
Table 5 shows the results of an analysis of preferences for horticultural therapy program attributes by type of activity related to OAD. First, for dementia facility workers, the preference for each attribute was highest for intervention goals at 27.4%, followed by the number of sessions per week (26.8%), total number of sessions (26.1%), and time per session (19. 7%). Second, for family caregivers, the importance of each attribute was greatest for the number of sessions per week (29.5%), followed by intervention goals (24.4%), time per session (23.3%), and total number of sessions (22.8%). Third, for horticultural therapists, it was highest for number of sessions per week (29.5%), followed by intervention goals (25.5%), total number of sessions (22.9%), and time per session (22.1%). Fourth, among those with no experience in OAD-related activities, it was highest for the number of sessions per week at 27.2%, followed by intervention goals (26.6%), time per session (24.6%), and total number of sessions (21.6%). Fifth, for others engaged in volunteer work related to the elderly, or care farming, it was greatest for number of sessions per week at 33.7%, followed by total number of sessions (23.5%), time per session (22.5%), and intervention goals (20.3%).

Preferences for horticultural therapy program attributes by activity type among older adults with dementia
For dementia facility workers, the importance of intervention goals was the highest, while for all types except for dementia facility workers, the preference of the number of sessions per week was the greatest.
In terms of preference by level of intervention goals, all types had a preference for the psychological and emotional domain as an intervention goal. Regarding the number of sessions per week, dementia facility workers, family caregivers, and horticultural therapists preferred once a week, while those with no experience in OAD-related activities preferred twice a week, and others preferred three times a week. In terms of total number of sessions and time per session, all types preferred 10 sessions or less and 60 minutes or less, respectively.
Lee’s (2024) meta-analysis of studies from 2003 to 2022 reported that the most common frequency for horticultural therapy programs was once a week, with 55 cases. This is consistent with the results of this study, which found that horticultural therapists preferred once-a-week horticultural therapy programs for OAD. In addition, the findings of Cho (2022), who reported that the elderly preferred horticultural activities of 60 minutes or less to those of 60 minutes or more, appear to be consistent with those of this study.
Conclusion
Conjoint analysis is a representative method for analyzing consumer utility in the field of marketing. It is an analytical technique that predicts the products that will be preferred by customers by estimating the utility that they attach to each attribute of products (Ha and Lee, 2012). Recently, however, conjoint analysis has been used in various fields other than marketing: estimation of the economic value of using weather observation equipment and improving the accuracy of forecasting models (Kim et al., 2020); and components of horticultural education programs for kindergarten students using conjoint analysis (Jung, 2021).
To develop a horticultural therapy program to alleviate dementia symptoms in older adults with dementia (OAD), we conducted a meta-analysis of studies on “horticultural therapy for OAD.” Among the items analyzed, four items that did not have significant results in the meta-ANOVA analysis were selected as attributes for conjoint analysis, and the levels for each attribute were also selected. After extracting 16 profiles from orthogonal files, 322 valid responses were analyzed.
Through analyzing preferences by attribute, it was found that the importance of each attribute was highest for number of sessions per week (28.5%), followed by intervention goals (25.7%), time per session (23.0%), and total number of sessions (22.8%; Pearson’s R = .892, p < . 001). Level preference for each attribute showed the highest importance for each level as follows: once a week for number of sessions per week; psychological and emotional domain for intervention goals; 60 minutes or less for time per session; and 10 sessions or less for total number of sessions. The following components of horticultural therapy programs for OAD were found to be preferred by dementia facility workers and family caregivers and horticultural therapists: once a week for number of sessions per week; psychological and emotional domain for intervention goals; 60 minutes or less for time per session; and 10 sessions or less for total number of sessions.
The significance of this study is as follows. First, the preferences for components of horticultural therapy programs to alleviate dementia symptoms in OAD were investigated and analyzed using the conjoint analysis method. Second, it is expected that these findings can be used as basic data for the development of horticultural therapy programs to alleviate dementia symptoms in OAD.