Powell 2014.

Methods	A randomised 2‐group design (June 2008‐April 2013)
Participants	Caregivers of adult patients with traumatic brain injury (TBI) from an inpatient rehabilitation unit of a level I trauma centre, Washington, USA. Caregiver age ranged from 19 to 89 years (mean 49.7; SD 13.5). The sample comprised of 82% female, 18% male, of which 79% were of a white non‐Hispanic race and 69% were married. The majority of caregivers had post‐high school education (75%), with employment ranging from working full‐time at time of injury (49%); working part‐time (18%); student (not working) (1%); unemployed (5%) and not in workforce or other (27%). Most were spouses or partners (54%) of the care‐recipient and 35% were parents.
Interventions	Intervention: An individualised education and mentored problem‐solving intervention plus usual care (n = 77). Aim: To investigate the effect of a solely telephone‐based, individualised, combined education and problem‐solving intervention on the quality of life (QoL) and emotional well‐being of caregivers of persons with moderate to severe TBI. Interventionist(s): A master’s level social worker with experience in prior studies of TBI and problem‐solving treatment approaches. Mode of delivery: Telephone Duration: Planned maximum 10 calls (20 weeks), with a target of 8 calls and 2 additional calls at the caregiver’s discretion. Content: The experimental intervention combined education and mentored problem‐solving for topics relating to caregiving and TBI recovery and management (in addition to usual care). The focus of the intervention was on self‐management of issues by the caregivers through applied problem‐solving rather than the provision of solutions or direction to resources, or both, by study personnel. The study interventionist began each call by asking open‐ended questions to ascertain what, if any, issues had arisen or had been resolved since the last call. The caregiver was then asked to identify the concern that he or she wished to address on the call. The final choice of the concern(s) to be targeted on each call was left up to the caregiver, with no requirement to address a new issue or a previously targeted one with action plans in progress. The interventionist then mentored the caregiver in a problem‐solving approach aimed at addressing the concern. Standardisation: No details provided, only one interventionist Comparison group: Usual care (n = 76)
Outcomes	1. Quality of life: Adapted Bakas Caregiving Outcomes Scale (BCOS), a 15‐item, 7‐point scale that measures change in social functioning, emotional well‐being, and physical health related to caregiving. Higher scores reflected better quality of life. 2. Psychological health (depression): Brief Symptom Inventory (BSI‐18), an 18‐item instrument designed to quantify symptoms of somatisation, depression, and anxiety. Respondents use a 5‐point Likert scale to indicate the extent to which each symptom bothered them over the preceding 2 weeks. Higher scores indicated greater symptoms of depression. 3. Psychological health (coping): Modified Caregiver Appraisal (MCA); higher scores indicated better coping. 4. Health Status and Well‐Being (social activity): The PART‐O (Participation Assessment with Recombined Tools–Objective) as a measure of community participation; higher scores indicated greater social activity. 5. Knowledge and understanding (knowledge): No instrument specified, stated structured interview and Likert ratings; no detail provided on the scoring system Data were collected at the end of intervention which was 6 months after discharge of the survivor to the community.
Notes	Funding source: Funded by the Department of Education, National Institute on Disability and Rehabilitation Research, TBI Model Systems: University of Washington Traumatic Brain Injury Model System (H133A070032) (p.180). Author confirmed that the abstract was linked to the study. Additonal information requested from the author but not provided at the time of submission of this review.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated (p.182)
Allocation concealment (selection bias)	Low risk	Password‐protected database, "The study coordinator entered identifying information into the database and was given the group assignment (p.182).
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	Low risk	"An examiner blinded to the group allocation conducted the follow‐up assessments..." (p.182).
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Sample size estimates allowed for a 10% loss to follow‐up. Loss to follow‐up, however, was greater, and there was an imbalance in loss to follow‐up across the groups (23% in the intervention group and 13% in the control group). Withdrawn from the intervention group (n = 4); withdrawn from the control group (n = 0) (Figure 1, p.185).
Selective reporting (reporting bias)	High risk	Unclear how or if the prespecified secondary outcome of knowledge was reported; satisfaction not assessed due to insufficient responses at follow‐up (Table 4, p.187)
Other bias	High risk	Fewer caregivers in the intervention group providing direct financial support which could potentially influence outcome measures of QoL and emotional well‐being (p.185)