Ⅰ. Introduction
Dementia is a condition that affects a large number of the older population worldwide. The elderly population over 65 years of age reached about 850,000 in 2020 and about 16.4% of the total population entering a super-aging society[1]. The prevalence of dementia in the elderly is about 10.3% in 2020 and the mild cognitive impairment elderly is 22.6%, accounting for 1 in 5 elderly people[1].
According to the Diagnostic and Statistical Manual of Mental Disorders fourth edition (DSM-5) for dementia, memory impairment and hypomnesia should be objectively demonstrated. In addition, dementia is deteriorating at least one of high-level cerebral functions such as aphasia, agnosia, and apraxia, and this causes serious obstacles in activities of daily living[2]. Dementia is a chronic and gradually worsening wasting disease. Due to cranial nerve damage, cognitive dysfunction including memory impairment is the main symptom[3]. As the condition worsens, it is accompanied by an emotional problem or a change in personality, causing social problems that indicate inappropriate behavioral behavior. Dementia is not clearly identified due to its characteristics. The progression of the senile dementia is slow so the treatment effect should be expected to maintain the current condition[4].
The therapeutic approach to dementia can be divided into medical treatment and socio-environmental treatment[5]. Medical treatment for dementia has been mainly drug therapy[5]. Recently, the importance of applying non-drug therapy as well as drug therapy has been raised as an approach to improving cognitive function and reducing behavioral psychological symptoms in elderly people with dementia[6]. The advantage of non-drug therapy for dementia patients is that it can reduce the risk of side effects caused by long-term drug use, and its effectiveness has been proven in various studies[6,7]. As the importance and effectiveness of non-drug therapy are raised, interest in various fields of social and environmental therapy is increasing and research for the elderly with dementia is also expanding[8]. An important point in the treatment of cognitive decline is that it is essential to understand and help patients and their families' social problems resulting from the appearance of problematic behaviors as well as deterioration of cognitive abilities[9].
The goal of this psychosocial therapy is to maintain the identity of dementia patients to lead proper social life and maximize cognitive function in the current situation[10]. Among these psychosocial treatments, reminiscence therapy has been reported to have a significant effect on improving cognitive function[11]. Reminiscence therapy improves emotions, behaviors and neurocognitive functions by stimulating the memories and emotions experienced in the life of the elderly[12]. In addition, reminiscence therapy provides opportunities for reintegration through positive experiences among past events, discovers the meaning of life, and ultimately helps social interaction[13].
Therefore, it is necessary to comprehensively verify the effectiveness of reminiscence therapy, which was performed to improve the cognitive function of elderly people with dementia through meta-analysis. In particular, the scope of subjects in this study was limited to institutionalized elderly people with dementia. The reason is that the elderly with dementia are often housed in nursing homes, so research focused on facilities has been conducted. And there is a difference in research environment from elderly people with dementia living in the community[11]. In other words, reminiscence therapy has been traditionally used to preserve cognitive function in elderly people with dementia, but it is reported that there are differences in effectiveness depending on whether it is a community-based or facility-based environment[11]. Therefore, it aims to reduce community living and heterogeneous contexts as much as possible to produce more appropriate research results for the environmental situation of facilities.
The primary objective of this meta-analysis was to assess the effect of reminiscence therapy on cognitive function, which has been conducted on institutionalized elderly people with dementia since 2000. The secondary objective was to identify the effect of reminiscence therapy on cognitive function. The findings should provide bases for establishing guidelines and recommendations for future reminiscence therapy for institutionalized elderly people with dementia.
Ⅱ. Methods
1. Search strategy
The search of the literature for this study followed the recommended guidelines of the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA). This study searched articles published in English or Korean after the year of 2000. The initial search was conducted by two reviewers using eleven electronic databases. The bibliographic databases PubMed, Ovid-MEDLINE, CINAHL, and Cochrane Library and the Korean databases RISS, KISS, and NDSL were used for the systematic search. The keywords used in the search were as follows: (“dementia” OR “Alzheimer’s disease” OR “Alzheimer’s dementia”) AND (“reminiscence” OR “reminiscence therapy” OR “reminiscence intervention” OR “reminiscence treatment” OR “reminiscence program” OR “spiritual reminiscence” OR “memory reminiscence”) AND (“cognitive function” OR “cognition”).
The process of literature selection used the PRISMA flow chart to record the status of each stage according to the criteria of literature selection. After two reviewers independently extracted articles from the literatures, the entire articles were firstly deleted duplicates using the EndNote X7 program. Two reviewers independently reviewed the titles and abstracts of the studies and the final article was selected by applying the selection criteria and exclusion criteria. Two reviewers reviewed all of the articles to determine if the studies fit the inclusion criteria. When the two reviewers disagreed during the study evaluation, disagreement was resolved by discussion with a third reviewer.
2. Selection criteria for review
Prior to performing a systematic literature search, the author developed the following: participants, intervention, comparison, and outcome (PICO) framework[14]. Participants(P) were institutionalized elderly people over 60 years, diagnosis of dementia such as Alzheimer’s disease. Intervention(I) were all kinds of reminiscence therapy applied as an experimental treatment. Comparison(C) were control group who did not receive reminiscence therapy or non-intervention or placebo intervention. Outcome(O) were determined as cognitive function, and cognitive function measurement tools were limited to studies using MMSE. Randomized controlled trials, quasi-experimental studies, and prospective or retrospective cohort studies were included in this review. One-group pretest-posttest design, qualitative research, and experts’ opinions were excluded from this review.
3. Methodological quality assessment
Quality assessment of studies were used the controlled trial method developed by the 2015 Scottish Intercollegiate Guidelines Network (SIGN)[15]. The evaluation was conducted as an individual assessment, with internal validity and total quality assessment. The 10 questions in the internal validity assessment consist of appropriate question, random allocation, adequate concealment, double blind, baseline homogeneity, difference treatment, valid and reliable measurement, adequate analysis, homogeneity of each experimental site, and dropout rate. Items 1 to 9 were evaluated as ‘yes’, ‘no’, and ‘don't know’, and item 10 as a percentage (%). The total quality assessment of the quality of studies was evaluated as ‘++’, ‘+’, ‘-’, and ‘0’ depending on how well each study minimizes errors.
4. Statistical analysis
Statistical values for the posterior mean, standardized deviation, and sample size were collected to calculate the effect size and homogeneity of the intervention study. To calculate the effect size, it was analyzed using R package version 4.0.2[16]. Since there was a difference in the sample size of each study, the weight of each effect size using Standardized Mean Difference(SMD)[17], 95% Confidence Intervals(CI), and inverse of variance was calculated. The average effect size was analyzed by applying a random-effects model.
At this time, according to the standard of Cohen[18], the effect size(ES) of 0.20 or more to less than 0.50 is a small effect, and 0.50 or more to less than 0.80 as a medium effect, and above 0.80 as a large effect. To evaluate the heterogeneity of the effect size, we first looked at the forest plot for visual evaluation. The forest plot was examined to evaluate the heterogeneity of the effect size. To confirm the heterogeneity, the ‘I’ value, which is the ratio of the actual variance to the total variance, was calculated[18]. To explain the heterogeneity of effect size, meta-ANOVA and meta-regression analysis were performed according to gender, frequency of interventions, session of intervention, and duration of intervention as moderator analysis. Finally, the selected study visually identified the presence of publication errors through the funnel plot. If the funnel plot is visually symmetric, the likelihood of publication errors decreases, and if the funnel plot is asymmetric, the likelihood of publication errors increases. Egger's regression test[19], a statistical analysis method, was conducted for objective verification.
Ⅲ. Results
1. Results of the search
As a result of the article search, a total of 4,875 records were searched. The reviewers identified 4,867 records through electronic database(PubMed 4,388 records, CINAHL 304 records, Ovid-MEDLINE 101 records, Cochran Library 14 records, RISS 28 records, DBpia 10 records, NDSL 7 records, KISS 8 records, and NAL 7 records) and 8 records through handwriting search. After removing the duplicates, 1,572 records remained. Within initial screening phase, 3,235 studies were removed which left a total of 68 articles for a full text screening to assess for the eligibility. 12 studies were included in the final qualitative synthesis through the meta-analysis. The selection process for this study is shown in Figure 1.
2. Characteristics of included studies
A total of 12 studies were published in 1 record (8.3%) 2000, 1 record (8.3%) 2005, 1 record (8.3%) 2007, 1 record (8.3%) 2009, 1 record (8.3%) 2013, 1 record (8.3%) 2014, 2 records (16.7%) 2015, 1 record (8.3%) 2016, 2 records (16.7%) 2018, and 1 record (8.3%) 2019. The publication type was all 12 records (100.0%) journal. The study design was 10 records (83.3%) RCT and 2 records (16.7%) NRCT. Among the subjects of the study, 3 records (25.0%) were only targeting female elderly, and 2 records (16.7%) over 60 years old , 1 record (8.3%) over 70 years old, and others over 65 years old. The number of samples was from 5 to 59 in the experimental group, and from 5 to 57 in the control group. The study sites were 2 records (16.7%) day care centers, 8 records (66.7%) nursing homes, 1 record (8.3%) care & rehabilitation center, 1 record (8.3%) nursing home, and 1 record (8.3%) day care center.
Each length of intervention ranged from a minimum of 30 minutes to a maximum of 120 minutes. 6 records (50.0%) for 30 minutes to 50 minutes, 4 records (33.3%) for 60 minutes, 1 record (8.3%) for 90 minutes, and 1 record (8.3%) for 120 minutes. The session of intervention was the most with 4 records for 6 times (33.3%), with a minimum of 6 times and a maximum of 20 times. The duration of intervention was at least 4 to 15 weeks, with 8 weeks being the most common with 4 records (33.3%). A tool for measuring cognitive function used MMSE (Mini Mental State Examination), which is the most widely used as a dementia screening tool<Table 1>.
3. Effect of reminiscence therapy
The effect size of the selected 12 studies was calculated and presented as a forest plot(Figure 2). The effect size of the overall study is SMD=0.64 (95% CI 0.30~0.98, p<.001), showing a statistically significant moderate effect size. The heterogeneity of the overall effect size was I2=73% (Q=24.34, df=11, p<.01), showing moderate heterogeneity.
4. Heterogeneity test of effect size
The heterogeneity of the overall effect size showed moderate heterogeneity (I2=73%) using meta-ANOVA was performed to explore why the effect sizes are different. First, the effect size (SMD=1.18) in the case of including both female and male elderly was greater than the effect size (SMD=2.66) for only female elderly, but the difference was not statistically significant (Qb=1.41, df=1, p=.235). Second, the effect size (SMD=1.72) when the frequency of reminiscence therapy exceeded 10 times was greater than the effect size (SMD=0.63) less than 10 times, but the difference was not statistically significant (Qb=2.11, df=1, p=.146). Third, the effect size (SMD=1.96) when the session of reminiscence therapy exceeded 50 times was greater than the effect size (SMD=1.25) less than 50 times, but the difference was not statistically significant (Qb=2.73, df=1, p=.099). Fourth, the effect size (SMD=1.72) when the duration of reminiscence therapy lasted 11 months was greater than the effect size (SMD=1.25) less than 5 months, but the difference was not statistically significant (Qb=0.93, df=2 p=.628)<Table 2>.
5. Publication bias analysis
Publication bias analysis was performed through funnel plot analysis to verify the validity of the selected research results. Looking at Figure 3, it can be seen that the left and right sides are symmetrical around the center line. For objective publication bias analysis, Egger's regression test was conducted on the relationship between the effect size and standard error of each study. As a result, there was no publication bias statistically with bias=4.55 (t=2.17, df=11, p=.056, 95% CI 0.429~8.661).
Ⅳ. Discussion
This study calculated the effect size of reminiscence therapy on institutionalized elderly people with dementia through meta-analysis, and identified specific characteristics of the intervention study. It was attempted to provide practical data for the development of the reminiscence therapy for the institutionalized elderly people with dementia. This was attempted to provide practical data for the development of reminiscence therapy for institutionalized elderly people with dementia.
A total of 4,875 papers were first searched, and duplicate papers were removed and the final 12 papers were selected according to the selection criteria. The characteristics of the selected papers were almost insufficient as one published in the early 2000s, and since 2005, the papers published have accounted for the majority of the 11 selected papers. As for the characteristics of the selected papers, only one paper published in the early 2000s was insufficient, and the papers published after 2005 accounted for the majority of the papers selected as 11 papers. The selected papers were one published in the early 2000s, and most of them were eleven published since 2005. This is presumed to be due to a relatively recent increase in awareness of the reminiscence therapy of dementia. The research design was relatively more than other meta-analysis study[19] where RCT studies identified the effectiveness of the reminiscence therapy, and the research biases were minimized by adequate concealment and double blind. This reflected that the effect of reminiscence therapy was more accurately verified. Nursing home was the most frequent research place, and in addition, day care center, care & rehabilitation center. This may be due to the use of an accessible place centering on the admission facilities of the study subjects. Among the various cognitive function measurement tools selected in this study, MMSE was the most used tool. MMSE (Mini-Mental State Examination) is the most widely used dementia screening test in the world and is recommended by various clinical practice guidelines. The advantage of MMSE is that it can be tested even if it is simply educated, the test time is short, and it can excellently screen for cognitive impairment. The quality assessment of the selected studies in this study carried out a research method to increase the internal validity, resulting in a low risk of biases. And there was no problem in synthesizing the research results and deriving the results. In individual quality, 8 RCT studies have applied the adequate concealment and double blind about treatment allocation. Baseline homogeneity was performed in all 12 selected papers, and result variables were measured using a measurement tool with reliability and validity secured to measure the outcome variables.
As a result, it is unlikely that the results of the study will change as internal validity was obtained in the individual quality assessment of this study. In other words, it means that most of the criteria are satisfied, there is little risk of biases, and there is little possibility that the results will change in future studies. The overall average effect size of the intervention studies selected in this study showed statistically significantly large effect sizes. This is consistent in that the effect size of reminiscence therapy showed a statistically significant effect size in Kim & Lee's meta-analysis study[20] on reminiscence therapy for elderly peoples with dementia (SMD=-0.62, 95% Cl –0.92~-0.31).
Reminiscence therapy had a small-size effect on cognitive functions (g=0.18, 95% CI 0.05– 0.30) in elderly people with dementia[11]. Reminiscence is of potential to become a nonpharmacological therapy for the elderly who suffer from cognitive dysfunction[21]. Reminiscence has been described as “the volitional or non-volitional act or process of recollecting memories of oneself in the past,”[22] and is a process of recalling and re-experiencing one's life events[21]. Moreover, this process itself will help the elderly to extract and repeat and thereby strengthening personal memories[21]. Therefore, it is very suitable for the elderly people with dementia[23]. Furthermore, compared to conventional drug treatment, reminiscence treatment has no severe side effects[24]. Accordingly, reminiscence therapy was shown to be more effective than everyday conversations in the treatment of elderly people with dementia[20, 21, 25].
All 12 articles selected from Huang et al. study[14] were RCT studies. This study showed a slightly larger effect than Huang et al. study[11], which showed the results of the study, reminiscence therapy had a small-size effect on cognitive functions. Kim & Lee's study[20] also reported that reminiscence in old age showed a larger effect size than the result of this study, which is a result of including studies with high degree of bias in the study methodology. The reason that the effect size of this study result was higher than that of Huang et al.'s study[11] may be because these studies were conducted only on RCT studies, unlike this study. Therefore, it will be necessary to analyze RCT studies and NRCT studies separately in future studies. As a result of meta-ANOVA analysis, there were no statistically significant differences in the subject's gender, time of each intervention, number of interventions, and duration of intervention.
These results are thought to have provided an important basis for selecting the type of intervention and constructing a program when developing reminiscence therapy for elderly people with dementia. This study presented the effect size of reminiscence therapy targeting elderly people with dementia through meta-analysis. Particularly, as a result of verification in reminiscence therapy intervention, a large effect size was revealed, and as a result of analysis of moderation effects, there was no difference between groups according to the intervention method.
This presented an important basis for selecting an intervention method when applying reminiscence therapy, which improves cognitive function of elderly people with dementia. However, it has limitations in collecting data through a limited period and limited database search. As a result of efforts to collect relevant papers as much as possible, no publication error occurred. In addition, the effect of reminiscence therapy for elderly people with dementia conducted in various fields on cognitive function was systematically confirmed. This was first attempted to provide practical evidence for the application of reminiscence therapy in the facility. Therefore, the study is meaningful in that the basis for the development of an intervention program for improving cognitive function was prepared based on the effect size of the intervention program.
Ⅴ. Conclusions
This study conducted a meta-analysis to verify the effectiveness of reminiscence therapy for 12 studies for institutionalized elderly people with dementia since 2000. According to the results of this study, it was shown that reminiscence therapy significantly improved MMSE in institutionalized elderly people with dementia. This is the result of demonstrating that during non-drug therapy to improve cognitive function, reminiscence therapy is effective in improving cognitive function of institutionalized people with dementia. As a result of the publication bias analysis, it was confirmed that there was no publication bias because it was not statistically significant. The results of this study presented a practical basis for the development of a cognitive function improvement program targeting institutionalized elderly people with dementia.
It was also suggested that reminiscence therapy as a non-drug therapy can improve cognitive function in institutionalized elderly people with dementia. However, all the articles included in this study are not complete RCT studies in allocation concealment or double-blindness, and attention should be taken to generalize the effect of reminiscence therapy due to the limitations of the cognitive function measurement tool used in the analysis. In future studies, a meta-analysis of the effect size of each intervention method will be needed to prepare standard guidelines for reminiscence therapy to improve evidence-based practice.
