Gallagher‐Thompson 2007.

Methods	Randomised clinical trial (Study dates not reported)
Participants	Chinese family caregivers of people with dementia in the San Francisco Bay area, United States of America (USA). All caregivers were female (n = 55, 100%) and were included in the study if they were at least 21 years of age, caring for a family member with significant memory loss or deterioration in cognitive abilities, spending at least 8 hours/week caregiving for at least 6 months, owned a phone, planning to remain in the area for 6 months, and agreed to random assignment to both conditions. Care‐recipients were required to have a score of 23 or less on the Mini‐Mental State Examination (MMSE) and be unable to perform one or more activities of daily living (ADLs) and two or more Instrumental ADL (IADLs), or have a documented dementia diagnosis. The mean age of the caregivers who completed the study was 59.3 (SD 12.23), were in the USA for 31.13 years (SD ‐20.93), and had a mean number of years in education of 13.42 (SD 4.10). Most were non‐spousal family relationships (mean spouse caregiver in telephone support condition (TSC) (7, SD 30.4), in‐home behavioural management program (IHBMP) (7, SD 31.8); mean non‐spouse caregiver TSC (16, SD 69.6) IHBMP (15, SD 68.2). Average duration of caregiving was roughly four years (TSC (41.26 months, SD 29.77), IHBMP group (48.32 months, SD 42.86)). More than 75% of caregivers were married: TSC (single (n = 2, 8.7%) , married (n = 18, 78.3%), widowed (n = 1 4.3%), divorced (n = 2, 8.7%)); IHBMP group (single (n = 4, 18.2%), married (n = 17, 77.3%), widowed (n = 1, 4.5), divorced (0)). About 80% of them had children: TSC: (n = 18, 78.3%), IHBMP (n = 18, 81.8%)). At least 75% reported that they had some help with caregiving (n = 20, 87% of the TSC and n = 6, 27.3% of the IHBMP group). About 30% said they were having financial difficulties and over 30% said that they assumed the primary caregiver role because no one else was available.
Interventions	Intervention: Telephone support groups (n = 28) Aim: To evaluate the efficacy of an in‐home intervention, based on cognitive behaviour therapy principles, to relieve stress and depression in female Chinese‐American caregivers Interventionists: Advanced doctoral students in psychology from a local university program Mode of delivery: Telephone Duration: 12 weeks (six phone calls at two week intervals) Content: Calls began with a general inquiry as to caregiver and care‐recipient well‐being, then one or more problems were identified for discussion. Common themes were incontinence, incessant questioning, temper outbursts, and nocturnal awakenings. The interventionist remained empathic and supportive, and at a comfortable moment, indicated that written information was available to help. Consumer‐friendly materials (in Chinese or English) were mailed if requested. The next phone call was scheduled and the session ended with expressions of concern for the welfare of caregiver and care‐recipient. Standardisation: No detail provided Comparison group: IHBMP is comprised of six modules that focus on learning new skills to help the caregivers cope with caregiving stress. Each module required one or more 90‐minute sessions; one additional session was used for any module requiring extra time (n = 27).
Outcomes	1. Psychological health (depression): 20‐item Center for Epidemiological Studies Depression scale (CES‐D). Higher scores indicated higher level of depression. 2. Psychological health (stress): The 10‐item Perceived Stress Scale and the Conditional Bother Subscale (CBS) is derived from the Revised Memory and Behavior Problems Checklist (RMBPC). Higher scores indicated higher level of stress. 3. Health status and well‐being (self‐efficacy): The revised self‐efficacy scale (SE). Higher scores indicated better perceived self‐efficacy. Time point for data collection not stated
Notes	For the outcome 'stress', the results from the 10‐item Perceived Stress Scale were used in the analysis for this review. Funding source: Research grant from the National office of the Alzheimer’s Association, Chicago – grant IIRG‐01‐3157 to DGT (p.433). Unpublished data sought via email but not received; author did provide information to enable categorisation of the study during the data screening process.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Insufficient information to assess
Allocation concealment (selection bias)	Unclear risk	Insufficient information to assess
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Unable to blind due to nature of the intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Insufficient information to assess
Incomplete outcome data (attrition bias) All outcomes	Low risk	Ten of the 55 (18%) participants dropped out either before or in the early stages of treatment — five from each group (p.427).
Selective reporting (reporting bias)	Low risk	Reports on all prespecified outcomes (p.431).
Other bias	Low risk	“The two groups were statistically equivalent at baseline, although the IHBMP appears to be higher than the TSC on CES‐D" (p.431).