Carson 2016.

Methods	RCT, parallel design, 2‐arm, multicentre study Aim of the study: to determine the effect of informational and emotional support meetings for carer participants led by palliative care specialists on carer‐ and patient‐centred outcomes
Participants	Total number of randomised patient participants: 256 Total number of randomised carer participants: 365 Inclusion criteria: patients ≥ 21 years of age, treated in medical ICUs, requiring ≥ 7 days of mechanical ventilation uninterrupted for ≥ 96 hours, not expected to be weaned or die within 72 hours. Carers who had the responsibility of healthcare decision making for patient (to include multiple family members if they participated together in decision‐making process) Exclusion criteria: patients mechanically ventilated at an outside hospital for > 7 days, or had chronic neuromuscular disease, trauma, or burns. Carer decision‐maker was unavailable or lacked English proficiency, primary clinician refused to grant permission to investigators to approach patient or carer, or investigators were the attending clinician. Patients previously admitted to the study ICU, or who had a palliative care consultation prior to screening Participant baseline characteristics Patient: Intervention group Age, mean (95% CI): 58 (55.2 to 60.8) years Gender, M/F: 64/66 BMI: not reported APACHE II, mean (95% CI): 26.2 (25.2 to 27.3) Mechanically ventilated: Y Primary diagnoses: no details Additional information (e.g. socio‐economic characteristics, ethnicity, primary language): Ethnicity, Hispanic or Latino, N = 17; non‐Hispanic or non‐Latino, N = 112; insurance status, Medicare, N = 60; Medicaid, N = 11; commercial, N = 47; none, N = 9; other, N = 3 Health literacy status (number correct on REALM‐SF): not reported Control group Age, mean (95% CI): 57 (54.0 to 59.7) years Gender, M/F: 61/65 BMI: not reported APACHE II, mean (95% CI): 25.8 (24.6 to 27.0) Mechanically ventilated: yes Primary diagnoses: no details Additional information (e.g. socio‐economic characteristics, ethnicity, primary language): ethnicity, Hispanic or Latino, N = 15; non‐Hispanic or non‐Latino, N = 111. Insurance status, Medicare, N = 57; Medicaid, N = 16; commercial, N = 36; none, N = 11; other, N = 6 Health literacy status (number correct on REALM‐SF): not reported Carer: Intervention group Age, mean (95% CI): 51 (48.8 to 52.8) years Gender, M/F: 56/128 Relationship to patient: spouse (N = 57), parent (N = 18), child (N = 41), sibling (N = 11), other (N = 3) Additional information (e.g. socio‐economic characteristics, nationality, primary language): ethnicity, Hispanic or Latino, N = 28; non‐Hispanic or non‐Latino, N = 155; employment status: employed, N = 103; unemployed, N = 15; homemaker, N = 10; retired, N = 40; disabled, N = 13; student, N = 1 Health literacy status (number correct on REALM‐SF): not reported Control group Age, mean (95% CI): 51 (48.6 to 52.7) years Gender, M/F: 50/131 Relationship to patient: spouse (N = 47), parent (N = 17), child (N = 41), sibling (N = 15), other (N = 6) Additional information (e.g. socio‐economic characteristics, nationality, primary language): ethnicity, Hispanic or Latino, N = 23; non‐Hispanic or non‐Latino, N = 158; employment status: employed, N = 93; unemployed, N = 22; homemaker, N = 16; retired, N = 25; disabled, N = 22; student, N = 3 Health literacy status (number correct on REALM‐SF): not reported Country: USA Setting: multicentre, 4 medical ICUs in 3 tertiary centres
Interventions	Recipient of the information: carer participants Intervention group Patient participants: N = 130; no losses Carer participants: N = 184; losses = 21 (15 refused to participate, data unavailable for 6); analysed = 163 at 3 months Details: carer participants given standardised brochure plus minimum of 2 meetings with palliative care physician and nurse, to include social workers, chaplains, or other disciplines as needed. Meetings structured according to a set of objectives and recommended topics. Intensive care unit clinicians held additional family meetings according to usual practice. Information was for ICU period and ICU transition and included: the nature of the patient participant's illness and treatments; prognosis for outcomes including ventilator independence, function, and life quality; impact of treatment on patient participant experience (including symptom burden); potential complications of treatment; expected care needs after hospitalisation; and alternatives to continuation of treatment. Meetings were held after 7 days of mechanical ventilation, at onset of chronic illness, and when a tracheostomy was considered. First and second meetings separated by 10 days Control group Patient participants: N = 126; no losses Carer participants: N = 181; losses = 32 (15 refused to participate, data unavailable for 17); analysed = 149 at 3 months Details: carer participant given standardised and personalised intervention. Intensive care unit clinicians managed all formal and informal family meetings according to usual practice without input from palliative care specialists. Carers given the same brochure. Intensive care unit clinicians could consult with palliative care team if required. Information was for ICU period and ICU transition. Intervention given at least once after 10 days in the ICU.
Outcomes	Outcomes reported in the review: Anxiety and depression of the carer participant measured at 3 months (using HADS‐A and HADS‐D, 7 items in each questionnaire, total scores range from 0 to 21; lower scores indicate less anxiety and less depression), satisfaction (FS‐ICU 24; scores from 0 to 100; higher score indicates increased satisfaction), ICU length of stay Other outcomes reported in the study: PTSD symptoms of the carer participant (IES‐R) patient‐focused communication regarding the goals of care measured (After‐Death Bereavement Family Interview), Quality of Communications scale score, number of days of mechanical ventilation, ICU length of stay, hospital length of stay, limitations of ICU therapies, hospital mortality, and 90‐day survival Outcomes assessed immediately after second support meeting, then followed up at 90 days.
Notes	Funding/declarations of interest: funded by grant from the NINR. Funders had no role in the design and conduct of the study. One author (Dr Carson) reported a consulting agreement with the Research Triangle Institute related to quality of care in long‐term acute care hospitals. Study dates: October 2010 to November 2014
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Investigators used a computer‐generated, web‐based randomisation method.
Allocation concealment (selection bias)	Low risk	Use of web‐based system ensured allocation concealment.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not feasible to blind personnel and participants to group allocation
Blinding of assessment for carer‐related outcomes (detection bias)	Unclear risk	Assessment of anxiety and depression using HADS, and assessment of satisfaction using FS‐ICU 24. Study authors report that the study primary outcome assessment was blinded, and that interview assessors were also blinded. However, the study authors do not report whether independent investigators were involved in the assessments of carer‐related outcomes. It is not clear whether awareness of group allocation would have influenced carer‐reported outcome data.
Incomplete outcome data (attrition bias) All outcomes	High risk	High attrition of carer participants (overall loss of 14%). Loss is imbalanced between groups.
Selective reporting (reporting bias)	Low risk	Prospective clinical trials registration: NCT01230099. Relevant review outcomes all reported in trial register documents.
Other bias	High risk	The control group had opportunities to meet with the palliative care team, therefore it is feasible that some carer participants in the control group may have received equivalent information to the intervention group.