Martindale‐Adams 2013.

Methods	Randomised controlled trial (February 2005‐June 2007 (Final data collection date for primary outcome))
Participants	Caregivers of people with dementia (n = 154), Memphis, USA. Caregiver age in years ranged from 37.9‐86.5 (mean age 65.6, SD 12.4). The sample consisted of 83.3% (n = 129) females and 16.2% (n = 25) males of which n = 38 (24.7%) were employed and n = 132 (85.7%) were married. The majority of the caregivers were white (n = 108, 70.1%), black (n = 45, 29.2%), other (n = 1, 0.6%). Most were married (n = 132, 85.7%); n = 38 (24.7%) were employed and mean years of education were 12.8 (SD = 2.0) (i.e. almost a year beyond high school).
Interventions	Intervention: CONNECT (individual and group delivery, 5‐6 caregivers/group; 15 groups) (n = 77) Aim: To determine if telephone support groups for dementia caregivers have an effect on bother with patient behaviours, burden, depression, and general well‐being Interventionist(s): 3 group leaders each with a case load (one with an MSc in divinity, one an MSc in psychology, one an MA in Sociology) Mode of delivery: Telephone Duration: One year (biweekly for 2 months and monthly thereafter for a year, for a total of 14 hour‐long sessions) Content: Content and structure of the intervention were based on the 6‐month REACH II intervention of 12 individual in‐home and by‐telephone sessions and five telephone support group sessions. Session materials consisted of a Caregiver Notebook and commercially available pamphlets. The Notebook, initially developed for a primary care intervention comprised behaviour management chapters and 17 caregiver stress/coping chapters. Each participant received a one‐on‐one introductory telephone call. Like REACH, the multicomponent intervention targeted caregiving risks, including risks associated with emotional and physical well‐being, safety, burden, social support, and patient behaviour management. Standardisation: Training and certification helped to ensure consistency across group leaders. During the final certifying role play, each prospective Group Leader provided the entire first session and two additional educational presentations. Study investigators evaluated behaviourally anchored ratings of specific procedural techniques (e.g. correct use of forms) and clinical skills (e.g. active listening). Comparison group: Caregivers received pamphlets on dementia and safety as well as telephone numbers for local resources. At the end of the study, they received the Caregiver Notebook and a workshop focusing on knowledge, safety, health, well‐being, behaviour management, and stress (n = 77).
Outcomes	1. Burden: The 12‐item Zarit Burden Interview (ZBI) assessed caregiver burden. Scoring was 0 (never) to 4 (nearly always); a higher score indicated greater burden. 2. Psychological health (depression): The 10‐item Center for Epidemiological Studies Depression Scale (CES‐D) assessed depressive symptoms within the past week. Scoring was 0 (rarely, none of the time) to 3 (most, almost all the time), for a score of 0 to 30; higher scores indicated greater symptoms. 3. Satisfaction: Perceived satisfaction with practical and other supports: Nineteen social support items measured received support and negative interactions, satisfaction, and social networks.The first three social support domains used a scale of 0 (never, not at all) to 3 (very often, very). Social network items used a scale of 0 (none) to 5 (9 or more). Social support items summed to 0 through 69; higher scores indicated more support. Satisfaction with the intervention: After final data collection, participants were asked by telephone about their satisfaction with the groups and components (e.g. format, length, information), any difficulties (e.g. talking to unseen members, distractions), and benefits (e.g. confidence, ability to provide care). Responses were scored from 1 (not at all) to 5 (extremely). Higher scores indicated greater satisfaction. Outcome data were collected at the end of intervention which was the 12‐month post‐discharge time point.
Notes	For the outcome 'satisfaction', the satisfaction scores from the 19‐item social support items were used in the analysis for this review. The author provided additional information on the interventionist training and outcome data for satisfaction with supports. Funding source: This work was supported by the Veterans Health Administration, Health Services Research and Development Service, US Department of Veterans Affairs with additional support from the Memphis Veterans Affairs Medical Center (p.47).
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	Overall attrition low, n = 15 (9.7%), Reasons for attrition similar for both groups: Intervention group, N = 8 (refused contact, n = 3; not interested n = 2; illness, n = 1; other reasons n = 2); control group, N = 7 (refused contact, n = 2; illness, n = 3; other reasons n = 2) (p.38 figure 1 and p.39).
Selective reporting (reporting bias)	High risk	Physical health outcomes not analysed
Other bias	Low risk	No significant group differences reported at baseline and baseline data on outcome measures was similar for both groups (p.40 table 1 and p.41 table 2).