Umeå university Department of Psychology Bachelor Thesis Spring 2011

Effect of Musical Intervention in the Treatment of Behavioral, Psychological and Cognitive Problems in Patients with Dementia Ieva Vasionytė

Supervisor: Guy Madison

Effect of Musical Intervention in the Treatment of Behavioral, Psychological and Cognitive Problems in Patients with Dementia Ieva Vasionytė

Abstract As the number of demented patients is growing rapidly and pharmacological intervention is not always effective, more research is needed on non-pharmacological interventions, including music. Previous meta-analyses have not been able to determine if music interventions have a positive effect on behavioral, cognitive, and psychological problems of demented patients. The purpose of this review was to update earlier findings and investigate the extent to which music intervention is effective in the treatment of demented patients. Eighteen studies were included in this meta-analysis, with a total of 533 dementia patients. Both pretest-posttest and controlexperimental group design studies were included, but analyzed separately. The mean effect sizes across these studies were computed and compared. However, the reach of meta-analysis was found to be limited, because a large proportion of music intervention studies have poor methodological quality, and reasonable inclusion criteria leave small samples of studies. We conclude that meta-analyses need to include unpublished studies with larger groups of participants. Music therapy is today a widespread treatment for a wide range of psychological, psychiatric, and physical conditions. It is defined as “a therapeutic medium to address developmental, adaptive, and rehabilitative goals in the areas of psychosocial, cognitive, and sensorimotor behavior of individuals with disabilities” (Hallam, Cross, & Thaut, 2009). The World Federation of Music Therapy describes music therapy as a use of its musical elements (sound, rhythm, melody and harmony) by a qualified music therapist, with a client or group, in a process designed to facilitate and promote therapeutic objectives mentioned above (WFMT, 2010, according to Vink, Bruinsma & Scholten, 2004). Music therapist as a profession was established in 1950. Music therapy and other music interventions such as music listening are widely used both as alternative and supplementary treatment to pharmacological cure. One condition for which music therapy is frequently used is dementia, with the goal of improving everyday functioning of people affected by cognitive deterioration. Because dementia affects a large proportion of elderly population, and music therapy is frequently used for these people, it is important to evaluate the effects of this practice. At the same time a relatively large number of studies addressing these effects are likely to be found that expresses even more a growing interest in this research area. The purpose of the present work is therefore systematically review the effects of music

2

therapy of people with dementia and, if feasible, to subject them to meta-analysis. According to the Encyclopedia of life sciences (2001), “dementia refers to progressive deterioration of thinking abilities severe enough to interfere with social, occupational and intellectual functions. “…Dementia is typically documented by poorer than expected performance on neuropsychological tests which assess memory, general knowledge, language, abstract reasoning and the ability to perform certain tasks of minimal skill, including dressing and simple drawing tasks. …”The prevalence of the disorder rises with age after the age of 65 years” (Nowotny, Kwon, & Goate, 2001). According to the World Factbook, published by the Central Intelligence Agency in the USA (2011) the present population over 65 years of age in the world is 8.1 percent and 17,3 percent in the European Union. Prognoses made by the United Nations (2002) report that the number of people over 60 years of age will reach nearly 2 billion by 2050, as compared with the present 528.8 million. In the interest of improving life quality and everyday efficacy of these individuals there is a great demand for novel interventions and the development of existing ones (World Assembly on Aging, 2002). According to the earliest meta-analysis on music in treatment of dementia, the interest in this area started to grow at the end of 20th century. Brotons and her collegues in 1997 summarized 69 studies since 1985 (Koger, Chapin et al., 1999), including clinical empirical studies of various music interventions, theoretical and philosophical papers, case studies and anecdotal accounts. Brotons made a qualitative review which did not include quantitative data. Although he couldn’t provide effect sizes from his review, it was a background for later researches. He summarized results of the studies about music intervention effectiveness for improvement of social, emotional and cognitive skills, decrease of behavioral problems of demented people (Koger, et al., 1999). A meta-analysis by Koger et al. (1999) updated the aforementioned qualitative review from 1999. Koger (1999) in his meta-analysis selected 21 empirical studies for evaluating whether music is efficient for people with dementia (Koger, Chapin, & Brotons, 1999). It was the first meta-analytical review of quantitative data of music effect on people with dementia. A recent meta-analysis in this area was done in 2004 (Vink et al., 2004). The objectives of that study was “…to assess the effect of music therapy in treatment of behavioral, social, cognitive and emotional problems of older people with dementia” (Vink et al., 2004). The present study will expand Vink et al.’s meta-analysis by including pretest-postest studies and analyzing the effectiveness of music intervention according to parameters such as live vs. recorded music; individualized vs. selected music; music listening (receptive music therapy) vs. active music therapy, and classical/relaxation vs. native/popular music. The need to consider these dimensions was indicated by Vink et al. (Vink et al., 2004). Dementia is an illness that touches a big part of our population, and the number is growing rapidly every year. Therefore, more research on the ways to help those people is required. The patients and their caregivers are faced with a range of problems caused by cognitive decline and brain damages, such as agitation, aggression, mood disorders, eating problems, etc. An umbrella term for these symptoms is ‘behavioral and psychological symptoms of dementia’ (BPSD), according to the International Psychogeriatric Association (Douglas, 2004, according to Finkel et al., 1996). Dementia is currently subject to a wide range of treatments, but the most widespread ones are pharmacological, such as cholinergic neurotransmitter modifying agents; noncholinergic neurotransmitters/neuropeptide modifying agents, and other pharmacological agents (Santaguida et al., 2004). But unfortunately, no curative

3

treatment for dementia is currently available. The effect of drugs is temporary or can only to slow the progression of the disease process. Moreover pharmacological treatment has a range of negative side effects (Caselli et al., 2005). The advantage of non-pharmacological treatments, including music interventions, is of course that they have almost no side effects if applied appropriately. Non-pharmacologic interventions are necessary for professional caretaking to avoid physical illness, such as constipation or infections (Douglas, James, & Ballard, 2004). Many standard treatments of dementia are dependent on patients’ verbal abilities, which unfortunately become very poor in the last stages of dementia. However, the ability to respond to music, such as humming or playing instruments tends to remain even in these late stages (Vink et al., 2004). Thus, summarizing the results in this area will show which topics need more research. Moreover, summarizing the effects will allow to design music therapy sessions for dementia patients in the most effective way. As mentioned above, previous reviews of the field have indicated a need for more precise analysis, taking the specific type or form of varieties of music therapy into account (Vink et al., 2004). The questions posed in this work are therefore how effective music is for the treatment of behavioral, psychological and cognitive problems in patients with dementia. A secondary goal was to assess this efficiency separately for a number of dimensions that commonly vary in music therapy, and that have previously been indicated as important to consider separately (Vink et al., 2004). These dimensions, described in detail in the method section, are music therapy versus music listening, live versus recorded music, pre-selected versus individualized music, and relaxation or classical versus native/popular music.

Methods The purpose of meta-analysis is to summarize the empirical results of a collection of studies done in particular research area. This method has been widely used for summarizing the results of empirical researches in health, social and behavioral sciences (Lipsey & Wilson, 2001). The main principles of meta-analysis are to collect all appropriate studies with quantitative data of the topic one is interested in, examine the characteristics of the variables and quantitative findings, calculate the effect sizes of each study, and finally describe the overall results, mostly in terms of a mean effect size across studies. The studies for this meta-analysis were selected from the academic databases and search engines JSTOR, EBSCO, ERIC, SCIRUS, MEDLINE, PsyINFO, Cochrane Library, and ProQuest. We also searched the journal databases SAGE PUB and Cambridge journals. All searches were made during April 2011. The keyword combinations applied were “music AND dementia” and “music AND Alzheimer*”. Only published articles were considered, because they are refereed and usually represent higher quality research than unpublished ones (Lipsey & Wilson, 2001). Studies written in other languages than English were also excluded. No publication time limit was applied, because research on the effect of music is a quite new area, and the oldest articles were from 1990. With regard to study design, no particular inclusion criteria were applied. In order to be sure that the study participants were demented patients, studies were included only if they involved patients having a diagnosis of dementia according to either Mini-Mental State Examination (MMSE, 1975), the GBS rating-

4

scale, the National Institute of Neurological and Communication Disorders-Alzheimer’s Disease and Related Disorders Association criteria (NINCDS-ADRDA), Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) (Fourth Edition, 1994 and later versions), or by any other established diagnostic. All the patients included in the studies were clinically diagnosed as having dementia type disorders.

Interventions Studies were eligible if the interventions reported used active music therapy, receptive music therapy, or music listening. As one of the purposes of the present metaanalysis was to estimate how the effectiveness of the intervention depends on different dimensions, the following dimensions were selected. Division of studies was mainly made according to the Oxford handbook of music psychology (Hallam et al., 2009). Active music therapy versus Music listening: active music therapy is defined as a combination of more than one musical therapy techniques, including active involvement of the participants. The code of active music therapy includes playing musical instruments, singing, song drawing, talking and dancing. Music listening was also included in active music therapy category if it was used together with any other of the mentioned activities. Receptive music therapy is characterized by participants’ lack of activity in therapy process. Music is selected by the therapist or according to participants’ preferences. Participants respond to the music verbally by expressing his or her feelings or memories that are aroused. This intervention was coded under the same “Listening” category together with just music listening without any response of the participant. Both types of interventions are commonly used in music therapy, each having its own area of practice. Active music therapy is mainly more used for arousal of positive emotions and increasing self-confidence (Hallam et al., 2009), whereas music listening is typically used for relaxation and reminiscence music therapy (Grocke et al., 2006). Live versus Recorded music: The included studies were also subdivided according to the dimension live vs. recorded music, in which the music is either reproduced from an audiogram or played or sung by the therapist, by professional musicians, or by the participants themselves. Live music appears to be more effective than recorded, probably because it creates a stronger sense of reality and because the patient can observe the musicians playing and interact with them (e.g., Sherratt et al., 2004). Selected versus Individual music: Selected music is selected by the therapist, without consulting the patient or his relatives or caregivers. Individualized music is selected according to the patient’s preferences, identified by asking herself, her relative, or caregivers. If the music for the intervention was selected using both methods, the study was attributed to the Individualized music category, because in such studies the therapist still chose the music considering the patient’s reaction to the particular music selected, interacting with the individual. Most studies use individualized music purposefully for arousing memories, which is one important ingredient in reminiscence therapy that is commonly used for treatment of dementia patients (Ashida, 2000).

5

Group versus Personal intervention: Personal intervention is based on interaction between the patient and the therapist only, or listening on one’s own. Group intervention refers to the treatment being applied to two or more patients at the same time. Personal intervention is more used for individualized music listening, while group sessions are always a part of active music therapy. Group therapy has been shown to be more effective for improving social and socio-emotional skills of dementia patients (Choi et al., 2009). Classical/Relaxation versus Pop/Native music: This distinction is unclear, but is nevertheless included because it is frequently made in music therapy studies. It is probably inspired by the fact that people tend to choose “easy” listening classical music when attempting to select something that is relaxing. Other music intended to be relaxing is often composed specially for that purpose, and is typically slow, has an unpronounced or absent beat, and uses relatively high-pitch, reverberating sounds. It is also common to use natural sounds, such as from wind, water, or animals. One practical reason for this distinction is that there is a market for relaxing music per se, but less so for the more particular preferences related to native/pop music, and that both classical and “self-composed” music is free from paying charges to an external composer. Native/pop music defines the category in which the music is native for the patients (like folk songs) or popular during a receptive period of the patient’s life. In receptive music therapies, classical or relaxation music is used most often to soothe demented patients. Native or popular music is usually used to arouse and enhance memories. As studies report more and more advantages of preferred music, there is a growing discussion about choice of music style. For example, not all people like classical music.

Outcome measures The outcome measures considered were behavioral, cognitive, psychological, and physiological problems of people with dementia. Although some studies include other measures as well, most do include some of these aspects. They also follow logically from the fact that people with dementia do suffer from a range of problems, and the objective of music therapy for these patients is specifically to improve their everyday functioning and quality of life. Moreover, many behavioral, cognitive, and psychological problems are targeted by particular tests, and physiological problems are often wise readily and relatively objectively assessed. These categories were constructed by dividing outcome measures used in the study into groups according to what they were testing. More detailed information about the outcome measures can be found in Appendix 1.

Methodological quality of the studies The main conclusion of Vink et al. (2011) was that many studies reported insufficient statistical information for computing the meta-analysis statistics, and that their methodological quality was too poor to draw reliable conclusions. Nine of the studies included in that report were included in the present meta-analysis too. Although the additional 13 studies reported enough statistical data to compute the effect sizes, not all of them used randomized selection of the participants and half of the studies used samples less than 20. Some studies had to be excluded because they lacked sufficient statistical information for computing effect sizes. For example Raglio et al. (Raglio et

6

al., 2008) study reported means but not standard deviations. Validity and reliability information of the instruments used in the studies to estimate outcomes of dementia was available in the articles themselves or in other publications, such as test manuals or test standardizations. Only one study used an inventory created by the authors themselves. Nair et al. used a so-called Behavior chart in their study of baroque music and behavioral disturbances in dementia, but reported no psychometrical data for it. A major limitation of most of the studies was that they used small samples, resulting in weak statistical power and poor generalizability of the results. The sample size across studies varied from 10, which was our limit for inclusion, to 87 participants. We decided against applying a higher limit, as 10 studies already had samples of 20 or less participants.

Data Analysis and Synthesis Coding scheme After excluding inappropriate studies according to the criteria mentioned above, the remaining studies were coded according to the following categories. It included Sample; Selected/Individualized, Recorded/Live music; Type of music (Classical, relaxation, native, pop, mixed); Group/individual intervention; Category of outcomes (behavioral, psychological, cognitive functioning, physiological); Outcomes (what exactly was measured, like depression, anxiety etc.); Measurement inventory (inventories, used to measure the outcomes); Statistical tests used; Means (M); Standard deviations (SD); p-value; t-value. The sample was coded as the number of participants in each group in the case of control-experimental groups design, and as the total number in one group pretest-posttest design. Design of meta-analysis As a methodological guide in the present study the book “Practical meta-analysis” by Lipsey & Wilson (2001) was used. According to them, meta-analysis is applied only to empirical research studies. It cannot include theoretical papers, conventional research reviews and the like. What is more, it applies only to research studies producing quantitative data (Lipsey & Wilson, 2001). Exclusion criteria are available in Appendix 2. The studies were divided into two groups according to their study design. Group contrast (GC) studies included studies with experimental and control groups when the dependent variable was measured on two groups, control and experimental and then compared across them. Mostly experimental and quasi-experimental studies provide results in this form. Pretest-posttest contrast (PPC) studies included those using a one group pretest – posttest design. These studies compare the central tendency of the variable that is measured at two or more points in time, typically before and after the intervention. GC studies that reported significant differences between control and experimental groups before the intervention or reported significant changes in the control group were attributed to the PPC group. They were consequently assigned to the same formula as for PPC studies, using pre-intervention and post-intervention scores of the experimental group. The sample sizes vary considerably across these studies, which makes it difficult to compare them directly. We therefore used the inverse variance weight (w) to compensate for this (Lipsey & Wilson, 2001). Each effect size is weighted by this statistic in order to control for sample size in computing the mean effect size.

7

Independence and homogeneity of effect sizes In order to keep effect sizes (ES) statistically independent, only one ES for each construct (behavioral, psychological, cognitive functioning outcomes) was considered for each study. In fact, more than one ES was available from several studies. For example, Raglio et al. (2008) study measured both psychological and behavioral outcomes for the effect of live music, so two effect sizes were available from the same study. In this case, we computed the mean effect size. To test the significance of mean effect sizes, 95 percent confidence intervals (α=.05) were calculated (Lipsey & Wilson, 2001). The homogeneity of the effect size distribution was also considered, in terms of the Q statistic, as a proxy for whether different ES belong to the same population of effect sizes (Lipsey & Wilson, 2001). Effect sizes smaller than 0.2 were considered small, larger than 0.2 and up to 0.6 medium, and larger than 0.6 were large, following Lipsey and Wilson (2001).

Results The literature searches yielded 90 articles, out of which 62 were excluded because they did not meet the inclusion criteria. Twenty-eight articles were selected as suitable. Ten out of the remaining 28 articles were not included in the meta-analysis, because they could not be delivered from the publishers or other sources in time for the completion of the thesis. Nair et al., (2010) in his study used two samples from different units under the same intervention, so in meta-analysis both of these two samples were included, counting two mean effect sizes for the study. Exclusion criteria and references of excluded studies can be found in Appendix 2.

Dependent variables: Music intervention effect on different outcomes Meta-analysis of music intervention effectiveness on different outcomes gave the results shown in Table 1, which is divided in sections according to both the dependent variable and the type of design. Statisticians recommend separate analyses of studies with different designs, such as pretest-posttest and control-experimental group designs. Mean effect sizes are therefore calculated separately for group contrast (GC) and pretest-posttest contrast (PPC) studies. Each effect size larger than 2 was decreased to 2 in order to avoid distortion of the results. In such cases the original value is shown within parentheses. The mean ES of the section is given below that section, together with its 95% confidence interval. Table 1. Effect sizes for different outcomes. First author

Sample size

Outcome measure

Type of the effect size

Effect size (ES)

Standard error (SE)

Inverse variance weight (w)

Behavioral Suzuki, M.

23

MOSES

GC

0.68

1.88

0.28

Chang, F.Y.

41

CMAI

GC

0.04

1.37

0.53

Raglio,

59

NPI-Q

GC

0.17

3.92

0.07

8

A., et al. Svansdo 38 ttir, H.B. et. Al. Choi, 20 A.A.

BEHAVEAD

GC

0.34

2.71

0.41

NPI-Q

GC

0.69

1.56

0.41

Total Sample: 181 Nair, 38 Behaviour B.K. chart Nair, 37 Behaviour B.K. chart Svansdo 38 BEHAVEttir, H.B. AD Ledger, 46 CMAI A. J

Mean ES: 0.39 PPC 0.21

CI (95%): (-1.11 ; 1.89) 0.215 21.62

PPC

0.12

0.24

30.64

GC

0.34

2.71

0.136

PPC

0.13

1.05

0.91

HicksMoore, S.L.

PPC

2 (2.39)

0.35

8.24

30

CMAI

Total Sample: 151

Mean ES: 0.46

CI (95%): (0.21 ; 0.71)

Psychological Sung, H.C.

52

RAID

GC

0.09

1.29

0.6

Choi, A.A.

20

GDS

GC

0.42

1.93

0.27

Hamilton’s scale

GC

2 (2.43)

1.86

0.29

Guetin, 15 S. Total Sample: 87 Cooke, 24 M. Sherratt, 24 K. Ashida, S.

20

Irish, M. 10

GDS

Mean ES: 0.64 PPC 0.04

Real-time PPC multiple event recorder Cornels scale PPC for depression in Dementia STAI PPC

Total Sample: 78

CI (95%): (-1.18 ; 2.46) 0.71 1.99

0.26

7.14

0.02

0.82

0.94

1.13

0.46

2.01

0.25

Mean ES: 0.86

CI (95%): (-0.14 ; 1.86)

Cognitive Suzuki, M. Choi, A.A.

23

MMSE

GC

0.31

2.68

0.14

20

MMSE

GC

0.54

1.23

0.66

Mean ES: 0.5 Koger, S.M.

20

CI (95%): (-1.7 ; 2.7)

WAB

PPC

0.33

0.38

7.14

Irish, M. 10

AMI

PPC

2 (4.56)

0.44

5.10

Thomps on, R.G.

Category fluency output Visual spatial task

PPC

0.24

0.9

1.22

PPC

0.394

0.41

6.06

16

Johnson, 17 J.K. Total Sample: 63

Mean ES: 0.78

CI (95%): (0.33; 1.23)

* Type of Effect size: GC - Group contrast; PPC - Pretest – Posttest.

9

Mean effect sizes behavioral outcome values from 11 studies were computed. Even if studies reported subscale results of the test inventory, only the total test results were taken into account. The mean ES of PPC design studies were significant and of medium size on behavioral outcomes (ES= 0.46; CI (95%) = 0.21, 0.71). GC design studies showed a non-significant medium effect of musical intervention. Results of 8 studies showed high but non-significant effects of musical intervention on psychological problems (ES=0.64 for GC; ES=0.86 for PPC design studies). Again, the most likely reason for the lack of statistically significant mean effects seems to be the small samples, since most of the mean effect sizes themselves tend to be quite large. Despite the small number of studies on the effect of music on cognitive problems, they indicate strong and significant effects of music interventions for PPC design studies (ES=0.78; CI (95%) = 0.33; 1.23). Only two GC studies were included, showing a non-significant, medium (ES = 0.5) effect of the intervention. Only one study examining the effect upon physiological outcomes, with a sample of 87 patients was included, indicating a large (ES = 0.67), but non-significant mean ES (CI (95%) = -0.23; 1.58).

Independent variables: Types of intervention All 20 studies included in the meta-analysis were homogeneous in terms of Q values that did not exceed the critical value for a chi-square with k-1 (k – the number of effect sizes) degrees of freedom (Lipsey & Wilson, 2001), and that did not reject the null hypothesis of homogeneity. But as Lipsey and Wilson write, the Q-test might not estimate correctly if the samples are very small, such are most of the studies reported in this meta-analysis. Table 4 reports the meta-analysis results for various interventions on people with dementia, based on one ES from each study. If a study reported more than one ES, we used the one that was measured by a more established or commonly used instrument. Significant ES are indicated by bold confidence intervals. Table 2. Effectiveness of different interventions. Characteristic of intervention

Type of ES*

Listening Active Music Therapy Recorded music Live music Group intervention

Mean ES

CI (95%)

Q (α=.05)

GC

Sample (number of studies) 67 (2)

0.7

-0.75 ; 2.15

0.09

PPC

213 (9)

0.57

-0.02 ; 1.16

0.76

GC

160 (5)

0.71

-1.37 ; 2.79

0.71

PPC

90 (3)

0.67

0.4 ; 0.94

25.24**

GC

39 (2)

1.9

-1.24 ; 5.04

0.04

PPC

189 (7)

0.66

0.39 ; 0.93

34.23**

GC

192 (5)

0.52

-0.87 ; 1.91

0.25

PPC

114 (4)

0.39

-0.04 ; 0.82

0.44

GC

140 (5)

0.7

-0.75 ;2.15

0.1

PPC

260 (8)

0.51

0.24 ; 0.78

22.57

GC

67 (2)

0.71

-1.37 ; 2.79

0.7

Individual intervention

PPC

43 (3)

1.04

0.49 ; 1.59

8.04**

Selected music

GC

101 (4)

0.71

-0.78 ; 2.2

0.06

PPC

209 (8)

0.6

0.35 ; 0.85

34.07**

GC

67 (2)

0.71

-1.37 ; 2.79

0.71

Individualized music

10

PPC Classical/Relaxation music Popular/Native music

24 (1)

1.5

-1.24 ; 4.24 -

-

0.39 ; 0.8

26.66**

GC

-

-

PPC

189 (7)

0.66

81 (3) 0.73 -0.98 ; 2.58 GC 90 (3) 0.4 -0.19 ; 0.99 PPC * Type of ES: GC – group contrast; PPC – pretest-posttest contrast. **Q significant, ES heterogeneous.

0.02

0.34 0.45

The analysis of different interventions failed to show any significant results. Even though the effect of active music therapy, recorded music, individual sessions, selected music, and classical relaxation music appeared to be significant according to the confidence intervals, the Q statistic indicated that these sets of ES were heterogeneous. That means that not all of the averaged ES estimate the same population ES. Remaining results for dimensions of intervention were homogenous but insignificant and showed medium to high effect on people with dementia.

Discussion The main aim of this meta-analysis was to estimate the mean ES of music interventions on behavioral, psychological, and cognitive problems, and on physiological outcomes. Meta-analysis results showed significant medium effect of PPC design studies on behavioral outcomes, confirming results of previous meta-analyses (Vink et al., 2011) and therapy practice. Music therapy for dementia patients is usually used to reduce agitation or other behavioral disturbances (Ragneskog, 2001). Preferred (individualized) music listening has a sedative effect. It might be effective because of the memories it arouses, focusing the patient energy on remembering pleasant moments from the past, rather than on repetitive movements or disruptive vocalizations. Another significant, high music effect rate found was on cognitive problems. The studies behind the significant mean effect size of cognitive outcomes examined short-term effects on spatial task (attention), category fluency and autobiographical memory. Effects were immediate, so long duration of intervention was unnecessary although long-term effects could have been considered. That studies using the Mini mental state examination (MMSE) did not report significant results is trivial: The MMSE is a measure of cognitive functioning, used to evaluate severity of dementia patients. It is in other words a measure of the dementia itself, which is irreversible. However, the language subscale of the MMSE did reflect an effect (ES= 0.88) of music in one study (Suzuki et al., 2004). Psychological symptoms were generally not sensitive to music interventions. High effectiveness rates were insignificant. However, such results are contradictory to various literature recourses that evidence music effectiveness on psychological symptoms. Music therapy was at the very beginning of its history used with psychiatric populations mainly for treating mood, personality, and anxiety disorders. Only later its effectiveness for behavioral problems was discovered (Hallam et al., 2009). But as mentioned in earlier sections of the paper, small sample sizes considerably decrease the chance of obtaining significant results although means ES is large. At the beginning of meta-analysis, it was planned to measure music effectiveness on physiological outcomes of demented patients. However, only one study met all

11

inclusion criteria. Okada and colleagues (2009) found music to be effective on physiological outcomes. Effect size counted was highly significant (ES = 0.67), with a sample of 87 patients. Literature resources, such as literature review by Watkins (1997) report music effect on such physiological parameters as heart rate, breathing rate and blood pressure. Suzuki et al. (2004) in his study reported music effectiveness on endocrinological outcomes of demented patients. Regardless these and many other research findings, there are not so many research made about music efficiency on physiological outcomes of demented patients. Considering that, more research should be done in this area. Besides the main aim of present study, effectiveness of music intervention dependently on various characteristics was also analyzed. Meta-analysis did not succeed to get any significant results. Not significant homogeneity test could proclaim sampling bias of the studies as Lipsey and Wilson (2001) say, but it also could be caused by other reasons. In included studies, independent variable was entire music intervention and not dimensions of music separately. That means, that ES of one dimension, e.g. live music in different studies was dependent on other dimensions of which music intervention consisted (e.g. by individual intervention, selected music in one study and by group intervention, individualized music in another study). In order to make this kind of analysis valid, studies, manipulating separate dimensions must be meta-analyzed. To summarize, music as an intervention has a medium effect for behavioral problems. Besides this, it has a high immediate effect on separate cognitive tasks like autobiographical memory testing, visual-spatial and category fluency tasks. Language capacities also appeared to be sensitive to music intervention.

Limitations and implications for future work The judgment process might differ from coder to coder and it is important to replicate systematic reviews and meta-analyses after a period of time. One should also preferably use two or more different coders, but this was not possible within the scope of this thesis work. One must also consider the so-called publication bias, as only published articles were included. This would constitute a risk for over-estimating effects, since published studies more often than unpublished ones report positive effects. The quality of the included studies is also important to consider. As long as not every study report all statistics required, the value of meta-analysis results gets poorer, since the effect sizes of those articles cannot be computed and included in the meta-analytic review. Another obstacle is the very small sample sizes that are common in music therapy studies. Larger samples are necessary for generalizing to the population and increasing statistical power of meta-analysis. The present results have some implications for future studies of music therapy. We found only six studies of cognitive outcomes, so they need more study. Additionally, more research may lead to the new findings and extend application of music interventions for cognitive problems. We also mentioned in the introduction that metaanalyses of music on patients with dementia could help to improve music interventions. The present study could not provide definitive results for comparing the effects of various dimensions of music intervention, because of the limitations mentioned above. When more research in this area will be done, more precise meta-analysis with bigger samples could realize this intention.

12

References Aldridge, D. (1993). Music therapy research I. A review of the medical research literature within a general context of music therapy research. The Arts in Psychotherapy. Arts in Psychotherapy, 20, 1135. Aldridge, D. (1994). Alzheimer's Disease: rhythm, timing and music as therapy. Biomedicine & Pharmacotherapy, 48, 275-281. Ashida, S. (2000). The effect of reminiscence music therapy sessions on changes in depressive symptoms in elderly persons with dementia. Journal of Music Therapy: 37, 170-182. Berger, G., Bernhardt, T., Schramm, U., Müller, R., Landsiedel-Anders, S., Peters, J., Kratzsch, T., & Frolich, L. (2004). No effects of a combination of caregivers support group and memory training/music therapy in dementia patients from a memory clinic population. International Journal of Geriatric Psychiatry, 19, 223-231. Brod, M., Stewart, A.L., Sands, L., Walton, P. (1999). Conceptualization and measurement of quality of life in dementia: the dementia quality of life instrument (DQoL). Gerontologist, 39, 25-35. Brotons, M. & Marti, P. (2003). Music therapy with Alzheimer's patients and their family caregivers: a pilot project. Journal Of Music Therapy, 40, 138-150. Casby, J. A. & Holm, M. B. (1994). The effect of music on repetitive disruptive vocalizations of persons with dementia. American Journal of Occupational Therapy, 48, 883-889. Caselli, R.J. (2005). Dementia medication overview (cont.). eMedicineHealth Site. Available: http://www.emedicinehealth.com/dementia_medication_overview/article_em.htm Central Intelligence Agency. (2011). World Fact Book. Washington, DC. Available: https://www.cia.gov/library/publications/the-world-factbook/index.html Chang, F.Y., Huang, H.C., Lin, K.C., & Lin, L.C. (2010). The effect of a music programme during lunchtime on the problem behaviour of the older residents with dementia at an institution in Taiwan. Journal of Clinical Nursing, 19, 939-948. Choi, A. N., Lee, M. S., Cheong, K. J., & Lee, J. S. (2009). Effects of group music intervention on behavioral and psychological symptoms in patients with dementia: A pilot-controlled trial. International Journal of Neuroscience, 119, 471-481. Clair, A. A. & Bernstein, B. (1990). A preliminary study of music therapy programming for severely regressed persons with alzheimer's-type dementia. Journal of Applied Gerontology, 9, 299-311. Clair, A. A. & Bernstein, B. (1995). The effect of no music, stimulative background music and sedative background music on agitated behaviors in persons with severe dementia. Activities, Adaptation & Aging, 19, 61-70. Clair, A. A. (2002). The effects of music therapy on engagement in family caregiver and care receiver couples with dementia. American Journal of Alzheimer's Disease and other Dementias, 17, 286-290. Clair, A. A., Tebb, S., & Bernstein, B. (1993). The effects of a socialization and music therapy intervention on self-esteem and loneliness in spouse caregivers of those diagnosed with dementia of the Alzheimer type: A pilot study. American Journal of Alzheimer's Disease and other Dementias, 8, 24-32. Clark, M. E., Lipe, A. W., & Bilbrey, M. (1998). Use of music to decrease aggressive behaviors in people with dementia. Journal of Gerontological Nursing, 24, 10-17. Cooke, M., Moyle, W., Shum, D., Harrison, S., & Murfield, J. (2010). A randomized controlled trial exploring the effect of eusic on quality of life and depression in older people with dementia. Journal of Health Psychology, 15, 765-776. Douglas, S., James, I., & Ballard, C. (2004). Non-pharmacological interventions in dementia. Advances in Psychiatric Treatment, 10.

13

Finkel, S. I., Costa e Silva, J., Cohen, G., Miller, S. and Sartorius, N. (1996). Behavioral and psychological signs and symptoms of dementia: a consensus statement on current knowledge and implications for research and treatment. International Journal of Geriatric Psychiatry, 12, 1060-1061. Garland, K., Beer, E., Eppingstall, B., & O'Connor, D. W. (2007). A comparison of two treatments of agitated behavior in nursing home residents with dementia: simulated family presence and preferred music. The American Journal Of Geriatric Psychiatry: Official Journal Of The American Association For Geriatric Psychiatry, 15, 514-521. Geer, E., Vink, A. C., Schols, J. M., & Slaets, J. P. (2009). Music in the nursing home: hitting the right note! The provision of music to dementia patients with verbal and vocal agitation in Dutch nursing homes. International Psychogeriatrics, 21, 86-93. Gerdner, L. A. (2000). Effects of individualized versus classical and relaxational music on the frequency of agitation in elderly persons with alzheimer's disease and related disorders. International Psychogeriatrics, 12, 49-65. Gerdner, L. A. (2010). Individualized music for elders with dementia. Journal of Gerontological Nursing, 36, 7-15. Goodall, D. & Etters, L. (2005). The therapeutic use of music on agitated behavior in those with dementia. Holistic Nursing Practice, 19, 258-262. Götell, E., Brown, S., & Ekman, S. L. (2009). The influence of caregiver singing and background music on vocally expressed emotions and moods in dementia care. International Journal of Nursing Studies, 46, 422-430. Grocke, D., Wigram, T., Dileo, C. (2006). Receptive methods in music therapy: techniques and clinical applications for music therapy clinicians, educators and students. London: Jessica Kingsley Publishers. Groene, R. W., II (1992). Effectiveness of music therapy intervention with individuals having senile dementia of the Alzheimer's type. Doctoral dissertation, University of Minnesota. Guetin, S., Portet, F., Picot, M. C., Pommie, C., Messaoudi, M., Djabelkir, L., Olsen, A.L., Cano, M.M., Lecourt, E., and Touchon, J. (2009). Effect of music therapy on anxiety and depression in patients with Alzheimer's type dementia: randomised, controlled study. Dementia and Geriatric Cognitive Disorders, 28, 36-46. Hallam, S., Cross, I., & Thaut, M. (2009). The Oxford handbook of music psychology. New York: Oxford university press. Han, P., Kwan, M., Chen, D., Yusoff, S. Z., Chionh, H. L., Goh, J., & Yap, P. (2010). A controlled naturalistic study on a weekly music therapy and activity program on disruptive and depressive behaviors in dementia. Dementia And Geriatric Cognitive Disorders, 30, 540-546. Harrison, S., Cooke, M., Moyle, W., Shum, D., & Murfield, J. (2010). Delivering a music mntervention in a mandomized controlled trial involving older people with dementia. Music and Medicine, 2, 214-218. Hicks, S. (2002). Research corner: relaxing music: what effect does it have on agitation at mealtime among nursing home patients with dementia? Info Nursing, 33, 17. Hicks-Moore, S. L. & Robinson, B. A. (2008). Favorite music and hand massage: two interventions to decrease agitation in residents with dementia. Dementia, 7, 95-108. Hicks-Moore, S. L. (2005). Relaxing music at mealtime in nursing homes: effect on agitated patients with dementia. Journal of Gerontological Nursing, 31, 26-32. Holmes, C., Knights, A., Dean, C., Hodkinson, S., & Hopkins, V. (2006). Keep music live: music and the alleviation of apathy in dementia subjects. International Psychogeriatrics, 18, 623-630. Irish, M., Cunningham, C.J., Walsh, J.B., Coakley, D., Lawor, B.A., Robertson, I.H., and Coen, R.F. (2006). Investigating the enhancing effect of music on autobiographical memory in mild Alzheimer's disease. Dementia and Geriatric Cognitive Disorders, 22, 108-120. Jennings, B. and Vance, D. (2002). The short-term effects of music therapy on different types of agitation in adults with Alzheimer's. Activities, Adaptation & Aging, 26, 6, 27-33.

14

Kelleher, A. Y. (2001). The beat of a different drummer: music therapy's role in dementia respite care. Activities, Adaptation & Aging: 25, 75-84. Klotter, J. (2001). Music and Alzheimer’s. Townsend letter for doctors and patients: April: The Townsend letter group. Available: http://findarticles.com/p/articles/mi_m0ISW/is_2001_April/ai_72297149/ Koger, S. M. & Brotons, M. (2000). Music therapy for dementia symptoms. Cochrane Database Of Systematic Reviews (Online), (2), CD001121. Koger, S. M., Chapin, K., & Brotons, M. (1999). Is music therapy an effective intervention for dementia? A meta-analytic review of literature. Journal of Music Therapy, 36, 2-15. Kydd, P. (2001). Using music therapy to help a client with Alzheimer's disease adapt to long-term care. American Journal of Alzheimer's Disease and other Dementias, 16, 103-108. Kyle, W. (2000). Music therapy: achieving positive clinical outcomes in dementia care. Vision, 6, 8-11. Ledger, A.J., Baker, F.A. (2007). An investigation of long-term effects of group music therapy on agitation levelsof people with Alzheimer’s disease. Aging & Mental Health, 11, 330-338. Lindenmuth, G. F., Patel, M., & Chang, P. K. (1992). Effects of music on sleep in healthy elderly and subjects with senile dementia of the Alzheimer type. American Journal of Alzheimer's Disease and other Dementias, 7, 13-20. Lipe, A. W., York, E., & Jensen, E. (2007). Construct validation of two music-based assessments for people with dementia. Journal of Music Therapy, 44, 369-387. Lipsey, M. W. & Wilson, D. B. (2001). Practical Meta-analysis. New Delhi: Sage publications. Lloyd, S. (1992). Finding the key... how music helped a man with dementia. Nursing Times, 88 (32), 48. Munk-Madsen, N. M. (2001). Assessment in music therapy with clients suffering from dementia. Nordic Journal of Music Therapy, 10, 205-208. Nair, B., Heim, C., Krishnan, C., D'Este, C., Marley, J., & Attia, J. (2010). The effect of baroque music on behavioural disturbances in patients with dementia. Australasian Journal on Ageing, 30, 11-15. Norberg, A., ,elin, E., & Asplund, K. (2003). Reactions to music, touch and object presentation in the final stage of dementia: an exploratory study. International Journal of Nursing Studies: 40, 473-479. Nowotny, P., Kwon, J.M., Goate, A.M. (2001). Alzheimer’s disease. Encyclopedia of life sciences, 20. Nature publishing group. Okada, K., Kurita, A., Takase, B., Otsuka, T., Kadani, E., Kusama, Y., Atarashi, H., and Mizuno, K. (2009). Effects of music therapy on autonomic nervous system activity, incidence of heart failure events, and plasma cytokine and catecholamine levels in elderly patients with cerebrovascular disease and dementia. International Heart Journal, 50 (1), 95-110 Park, H. & Pringle Specht, J. K. (2009). Effect of individualized music on agitation in individuals with dementia who live at home. Journal of Gerontological Nursing, 35, 47-55. Park, H. (2010). Effect of music on pain for home-dwelling persons with dementia. Pain Management Nursing, 11, 141-147. Pinkney, L. (1997). A comparison of Snoezelen environment and a music relaxation group on the mood and behaviour of peoples with senile dementia. British journal of occupational therapy, 60 (5), 209212. Raglio, A., Bellelli, G., Traficante, D., Gianotti, M., Ubezio, M. C., Villani, D., & Trabucchi, M. (2008). Efficacy of music therapy in the treatment of behavioral and psychiatric symptoms of dementia. Alzheimer Disease And Associated Disorders, 22, 158-162. Ragneskog, H., Asplund, K., Kihlgren, M., & Norberg, A. (2001). Individualized music played for agitated patients with dementia: Analysis of video-recorded sessions. International Journal of Nursing Practice, 7, 146-155.

15

Ragneskog, H., Brane, G., Karlsson, I., & Kihlgren, M. (1996). Influence of dinner music on food intake and symptoms common in dementia. Scandinavian Journal of Caring Sciences, 10, 11-17. Richeson, N. E. & Neill, D. J. (2004). Therapeutic recreation music intervention to decrease mealtime agitation and increase food intake in older adults with dementia. American Journal of Recreation Therapy, 3, 37-41. Santaguida, P.S., Raina, P., Booker, L., Patterson, C., Baldassarre, F., Cowan, D., Gauld, M., Levine, M., & Unsal, A. (2004). Pharmacological treatment of dementia (Summary), (Rep. No. 97). Rockwille, United States: AHRQ Publication. Sherratt, K., Thornton, A., & Hatton, C. (2004b). Music interventions for people with dementia: a review of the literature. Aging & Mental Health, 8, 3-12. Smith, S. (1990). The Unique power of music therapy benefits Alzheimer's patients. Activities, Adaptation & Aging, 14, 59-64. Smith-Marchese, K. (1994). The effects of participatory music on the reality orientation and sociability of Alzheimer's residents in a long-term-care setting. Activities, Adaptation & Aging, 18, 41-55. Spiro, N. (2010). Music and dementia: Observing effects and searching for underlying theories. Aging & Mental Health, 14, 891-899. Sung, H. C. & Chang, A. M. (2005). Use of preferred music to decrease agitated behaviours in older people with dementia: a review of the literature. Journal of Clinical Nursing, 14, 1133-1140. Sung, H. C., Chang, A. M., & Abbey, J. (2006). P4-195: Preferred music intervention decreases agitated behaviors of institutionalized elders with dementia. Alzheimer's and Dementia, 2, S573-S574. Suzuki, M., Kanamori, M., Nagasawa, S., & Takayuki, S. (2007). Music therapy-induced changes in behavioral evaluations, and saliva chromogranin A and immunoglobulin A concentrations in elderly patients with senile dementia. Geriatrics and Gerontology International, 7, 61-71. Suzuki, M., Kanamori, M., Watanabe, M., Nagasawa, S., Kojima, E., Ooshiro, H., & Nakahara, D. (2004). Behavioral and endocrinological evaluation of music therapy for elderly patients with dementia. Nursing & Health Sciences, 6, 11-18. Thompson, R. G., Moulin, C. J. A., Hayre, S., & Jones, R. W. (2005). Music enhances category fluency in healthy older adults and Alzheimer's disease patients. Experimental Aging Research, 31, 91-99. Tomaino, C. M. (1998). Music on their minds: a qualitative study of the effects of using familiar music to stimulate preserved memory function in persons with dementia. (No other information is available. EBSCO suggests to cite in this way according to APA). Topo, P., Maki, O., Saarikalle, K., Clarke, N., Begley, E., Cahill, S., Arenlind, J., Holthe, T., Morbey, H., Hayes, K., & Gilliard, J. (2004). Assessment of a music-based multimedia program for people with dementia. Dementia, 3, 331-350. Tow, D. (2006). Music is magic for residents with Alzheimer's. Nursing homes: Long Term Care Management, 55, 40-41. Van de Winckel, A., Feys, H., De Weerdt, W., & Dom, R. (2004). Cognitive and behavioural effects of music-based exercises in patients with dementia. Clinical Rehabilitation, 18, 253-260. Vink, A.C., Bruinsma, M.S., & Scholten, R.J.P.M. (2004). Music therapy for people with dementia. Cochrane Database of Systematic Reviews, Issue 3. Art. No.: CD003477. DOI: 10.1002/14651858.CD003477.pub2. Wall, M. & Duffy, A. (2010). The effects of music therapy for older people with dementia. British Journal of Nursing, 19, 108-113. Watkins, G., (1997). Music Therapy: proposed physiological mechanisms and clinical implications. Clinical nurse specialist, 11(2), 43-50. Chicago, Illinois: Williams & Wilkins. Witzke, J., Rhone, R. A., Backhaus, D., & Shaver, N. A. (2008). How sweet the sound - Research evidence for the use of music in Alzheimer's dementia. Journal of Gerontological Nursing, 34, 45-52.

16

World Assembly on Aging. (2002).World population aging: 1950-2050. United Nations. Available: http://www.un.org/esa/population/publications/worldageing19502050/ World Fact book (2011). Central intelligence agency (CIA) [On-line]. Available: https://www.cia.gov/library/publications/the-world-factbook/index.html

17

Appendix 1. Main characteristics of the included studies. First Author

Year Design

Recorded Type of vs. live music

Suzuki, M.

2004

Live

Native old Group songs

Behaviora GC l and Cognitive

Cooke, M.

2010

Mixed

Not Group mentioned

Psycholog PPC ical

Nair, B.K. 2010 Svansdotti 2006 r, H.B. Okada, K. 2009

Diagnosi Interven Intervent Selected s tion ion vs. duration Individua lized Experimental Alzheime 8 weeks Singing, Individual and control r's, playing ized groups vascular type dementia Randomized Early to 8 weeks Singing, Mixed control trial; exp mid-stage playing, & contr. Dementia listening Groups; crossover Randomized, MMSE 12 weeks Listening Selected cross over trial 15/30 Control and Moderate 6 weeks Singing, Selected experimental to severe listening, groups, dementia playing, randomized moving Control and Advanced 6 weeks Singing, Selected experimental, dementia listening, non-randomized playing groups QuasiSevere 6 weeks Listening Individuaexperimental, 13, lized Control and Moderate experimental -severe 8, groups Moderate 8 experimental, Moderate 12 weeks Listening Individual withinto severe ized participants, dementia Pre-Post test QuasiVascular 8 weeks Listening Selected experimental, dementia, pre-test, postAlzheime test. r's, other types. Control and Alzheime 5 weeks Singing, Selected experimental r's, listening, groups vascular, playing other type

Sung, H.C.

2010

Sherratt, K.

2004

Chang, F.Y.

2010

Choi, A.A.

2009

Koger, S.M.

2000 Pretest and posttest.

HicksMoore, S.L.

2005 Quasiexperiment, pretest and postest, week intervals Raglio, A. 2008 Control and experimental groups, randomized Ashida, S. 2000 Pretest and posttest.

Alzheime r's and related diorders, mild to severe dementia Alzheime r's and Irreversib le/severe dementia Alzheime r's type, vascular dementia Various types of dementia

12 weeks Singing, talking

Selected

2 weeks Listening Selected

16 weeks Music playing, talking.

Recorded Classical music Live Popular native music

Group

Live

Group

Popular native music

Recorded Popular native music

Live

Group

Behaviora PPC l Behaviora GC l Physiolog GC ical

Individual Psycholog GC ical

Not Group mentioned

Psycholog PPC ical

Recorded Relaxatio Group nal

Behaviora PPC l

Mixed

Not Group mentioned

Live

Mixed

Group

Cognitive, GC Psycholog ical, Behaviora l Cognitive PPC

Recorded Relaxatio Group n

Behaviora PPC l

Not Live mentioned

3 weeks Music Selected listening, playing, talking

Group vs. Category Type of Individua of ES l outcomes

Mixed

Not Group mentioned 19801990

Group

Behaviora GC l, Psycholog ical Psycholog PPC ical

18

Guetin, S. 2009 Randomized, control study; exp. and control groups Ledger, 2007 Nonrandomized A. J Repeated measures design. Control and experimental groups Irish, M. 2006 Repeated measures, control and exp groups Thompso 2005 control and exp n, R.G. groups, nonrandomized Jennings, 2002 Pretest and B., posttest. Vance, D. Johnson, 2002 Pretest and J.K., et al. posttest.

Mild to 16 weeks Listening Individual Recorded Various Individual Psycholog GC moderate ized ical Alzheime r's disease Alzheime 53 weeks Singing, Mixed Live Not Group Behaviora PPC r's type, playing, mentioned l Mild 4, listening Moderate 7, Severe 16 Alzheime 2 weeks Listening Selected r's type, Mild

Recorded Classical

Individual Cognitive, PPC Psycholog ical

Alzheime r's type, MMSE M= 29,5 Alzheime r's type

Not Listening Selected mentione d

Recorded Classical

Individual Cognitive PPC

4 weeks Singing

Recorded Mixed

Group

Recorded Classical

Individual Cognitive PPC

Selected

Alzheime Not Listening Selected r's type, mentione Mild d

Behaviora PPC l

19

Appendix 2. Exclusion criteria and references of the excluded studies. Exclusion criteria

References

Literature reviews

Gerdner, 2000; Sung & Chang, 2005; Wall & Duffy, 2010

Meta-analysis

Koger, 1999; Vink et al., 2011

Discussion papers

Smith, 1990; Aldridge, 1993; Aldridge, 1994; Clark et al., 1998; Brotons & Koger, 2000; James et al., 2000; Kyle, 2000; Klotter, 2001; Hicks-Moore, 2005; Goodall & Etters, 2005; Tow, 2006; Witzke et al., 2008; Spiro, 2010

Case studies

Lindenmuth & Patel, 1992; Lloyd, 1992; SmithMarchese, 1994; Casby & Holm, 1994; Kelleher, 2001; Hicks, 2002; Han et al., 2010

Sample less than 10

Ragneskog et al., 1996; Pinkney, 1997; Kydd, 2001; Munk-Madsen, 2001; Norberg et al., 2003; Suzuki et al., 2006

No quantitative data reported

Groene, 1992; Ragneskog, 2001; Tomaino, 1998; Van de Winckel et al., 2004; Topo et al., 2004; Sung et al., 2006; Götell et al., 2009; Lipe, 2007; Chang et al., 2008; Geer et al., 2009; Harrison, 2010; Gerdner, 2010

Do not provide separate results of music intervention

Clair & Bernstein, 1995; Richeson & Neill, 2004; Holmes et al., 2006; Garland, 2007; Hicks-Moore & Robinson, 2008

Lack of sufficient statistical information for computing effect sizes

Jennings et al., 2002; Raglio et al., 2008; Park et al., 2010

Examined the effect of music therapy on caregivers Clair & Bernstein, 1990; Brotons et al., 2003; Clair rather than on the demented people themselves et al., 1993; Berger, 2004 Did not meet the purpose of meta-analysis

Clair, 2002 (estimates music effect on relations between caregiver and demented patients)

The same sample was used in another study (Park, 2010)

Park, 2009

Assessment of music program itself, rather than the Dupuis & Pedlar, 1993; Götell et al., 2002 effect on demented patients. Food consumption*

Chang et al., 2008; Choi et al., 2009; Thomas & Smith, 2009

* There were too few studies estimating effect on those outcomes to make a reliable meta-analysis

20