Appendix B.

MMAT Quality Assessment Details.Assessment of Methodological Quality According to the MMAT.

Qualitative	1.1 Are the sources of qualitative data relevant to address the research question?	1.2 Is the process for analyzing data relevant to address the research question?	1.3 Is appropriate consideration given to how findings relate to the context, eg, the setting, in which the data were collected?	1.4 Is appropriate consideration given to how findings relate to researchers’ influence, eg, through their interactions with participants?
Theurer et al, Canada, 2014	Yes. The objective/research question is to evaluate the process, structure and content of the intervention. The sources of data are focus groups, individual interviews, and observations with residents and staff.	Yes. The analysis framework used has 5 stages: (1) listening to recordings, reading the transcripts, and observational notes; (2) identifying thematic framework; (3) Indexing; (4) charting (lifting quotes from original context and rearranging them under themes); (5) mapping and interpretation. The analysis also explored any variations of themes across the 2 facilities.	Insufficient information. A little is written about context on page 405: they state that when comparing the 3 homes, 2 characteristics stood out: prevalence of dementia differed, and level of participation in activities differed.	Yes. A self-reflexive process was used to identify any biases that the principal investigator held during the course of the study and its interpretations. In addition, validity checks of the data interpretations were conducted by the investigative teams. It is also addressed to a limited extent in the discussion p. 410.
Quantitative randomized	2.1 Is there a clear description of the randomization (or an appropriate sequence generation)?	2.2 Is there a clear description of the allocation concealment (or blinding when applicable)?	2.3 Are there complete outcome data (80% or above)?	2.4 Is there low withdrawal/dropout (below 20%)?
Chiang et al, Taiwan, 2009	Yes. Those who consented to participate were randomly assigned to either control or experimental group by permuted block randomization.	No. No sufficient description and/or concealment not done and also no blinding of outcome, participants, or personnel.	No. Attrition/exclusion were reported (20 in the experimental group = 31% dropout rate, 18 in the control group = 28% dropout rate).	No.
Haslam et al, the United Kingdom, 2010	No. They merely write that the study utilized a randomized controlled trial allocating participants to 1 of 3 interventions.	No. Not sufficient description and/or concealment not done. They state that facilitators were not (and could not be) blind to the condition to which participants had been assigned.	Yes. See flowchart in Figure 1, where it is shown that no one was excluded from analysis.	No. Total dropout rate was 37% (29% because of death, 8% because of withdrawal).
Karimi et al, Iran, 2010	No. They merely write that 39 participants were randomly selected. They further describe that to form matched groups in terms of depression severity and gender, participants were systematically divided into 3 groups and were then randomly assigned to the 3 conditions of intervention (p. 882). But the randomization itself is not well described.	No clear description.	No clear description.	No. Total dropout rate was 25.6% (excluded for various reasons including suffering from an illness, not attending at least 60% of the sessions).
Serrani Azcurra, Argentina, 2012	No. They merely write that participants were randomly assigned to 1 of the 3 groups; intervention, active control and passive control. Participants were recruited from 2 privately funded long-term nursing homes sharing structural and functional characteristics.	Yes. To some extent. They state that the design is single-blinded and describe that the psychologists delivering the intervention were blinded to the outcome measures. The outcome measures were administered by independent raters (registered nurses). Further analysis of the data was done by statisticians who were blinded to the subject assignment (p. 426).	Yes. The percentage of missing data for the outcome variable at T0, T1, and T2 was 2.5% and the percentage of missing data for the independent variables was 1.9%.	Yes. Out of the 145 eligible cases, 5 dropped out during the study (death, moved to another facility, refused to participate).
Stinson et al, the United States, 2005	No. They merely write that participants who met all eligibility requirements were randomly assigned with equal probability to either the experimental group or the control group.	No clear description.	No.	No. They state that complete data were obtained for 17 of the STS outcomes (29%) and 18 of the GDS outcomes (25%). Attrition occurred because of difficulty of completing all data across time. During the 6-week period, there were 2 hospitalizations, 1 relocation, and 1 decision from a participant to withdraw from the study.
Gudex et al, Denmark, 2010	Yes. They write that the study was undertaken as a randomized, matched intervention study. Ten nursing homes were matched by the project team into 2 groups on the basis of location, type, and size. The 2 lists of 5 nursing homes were then placed in 2 blank sealed envelopes; a colleague external to the project group was asked to arbitrarily choose one envelope; the 5 nursing homes named in this envelope became the intervention group, who implemented reminiscence. The remaining 5 nursing homes became the control group who continued with usual nursing care.	Yes. Nursing homes were not told of their group until after the second baseline data collection was completed. Blinding of nursing staff with respect to intervention was not possible, although the project interviewers were not formally told which group the nursing homes were in.	No.	No. The dropout rate over the project period was 32% for residents. Of the 111 residents who failed to complete the study, 98 had died, 11 had moved out of the nursing home, 1 withdrew study participation, and 1 had become too ill. There were no significant differences in sociodemographic characteristics between the 348 who started the study and the 237 who completed the study.
Quantitative nonrandomized	3.1 Are participants (organizations) recruited in a way that minimizes selection bias?	3.2 Are measurements appropriate (clear origin, or validity known, or standard instrument; and absence of contamination between groups when appropriate) regarding the exposure/intervention and outcomes?	3.3 In the groups being compared (exposed vs nonexposed; with intervention vs without; cases vs controls), are the participants comparable, or do researchers take into account (control for) the difference between these groups?	3.4 Are there complete outcome data (80% or above), and, when applicable, an acceptable follow-up rate for cohort studies (depending on the duration of follow-up)?
Chao et al, Taiwan, 2006	Yes. Purposive sampling was done to recruit participants; participants in experimental and control group were matched for demographic characteristics, depression, self-esteem, and life satisfaction.	Yes. Validated measures are applied.	Yes. Groups have been matched on relevant measures.	Yes. Attrition/exclusion were reported (2 in the experimental group and 4 in the control group) at follow-up (after the 9 weekly group sessions). There is a relatively short follow-up period. However, the follow-up time might be suitable for the specific outcome measures in the study (depression, self-esteem, life satisfaction).
Tsai et al, Taiwan, 2010	Yes. Nursing homes were selected based on size and accessibility to the researcher. To compare participants in experimental and control groups, at least 30 participants in each group were needed. A list of 20 medium-large nursing homes were first obtained; these were all over Taiwan and were accessible to the researchers. Each of these nursing homes was assigned a number. Nursing homes for the control group were randomly selected by number and were approached to recruit participants until the goal of 30 participants was reached. The same procedure was followed for the experimental group. Six nursing homes rejected participation and few residents wanted to participate, so further nursing homes were approached to reach the 30 participants.	Yes. Validated measures are applied.	Yes. However, I am a bit uncertain here. They have a demographic table comparing the 2 groups on age, gender, marital status, education, residency, activities of daily living (ADL), and MMSE where for most of the variables the distribution is very similar (no tests). Furthermore they write that the same inclusion/exclusion criteria were used for both control and experimental group. They do not seem to control statistically for potential differences.	Yes. During the 3 months of this study, the control group lost 5 participants (28 remained, a 15% loss to follow-up), and the experimental group lost 3 participants (21 remained, 12.5% loss to follow-up). They state that participants who withdrew from the 2 groups did not differ significantly from those who remained in any demographic characteristic except age.
Winningham et al, the United States, 2008	Yes. They write that all participants in a given facility were assigned to either the CEP or the control group. Participants within a given facility were all assigned to the same conditions, because they had previously observed that nonparticipating residents in the same facility might have been exposed to some aspects by hearing participants discuss parts of the intervention. Facilities were assigned to be part of the CEP or control simply based on location of the researchers and availability of a large room.	Yes. Validated measures are applied.	Yes. 73 participants began the study but 16 dropped out for various reasons. A series of analyses was carried out to assess whether participants who dropped out differed from those who completed the study. There were indications that those who completed the study were younger and had higher SS-A scores than those who dropped out, but there were no differences in SS-B scores and loneliness.	No. With 16 out of 73 dropping out, the dropout rate is 21.9%.

Note. MMAT = McGill Mixed Methods Appraisal Tool; STS = Self-Transcendence Scale; GDS = Geriatric Depression Scale; MMSE = Mini-Mental State Examination; CEP = cognitive enhancement program; SS-A = Social Support Appraisal; SS-B = Social Support Behaviors.