Table A3:
Risk of Bias Among Randomized Controlled Trials for the Comparison of Team-Based Model of End-of-Life Care and Usual Care
| Author, Year | Allocation Concealment | Blinding | Complete Accounting of Patients and Outcome Events | Selective Reporting Bias |
|---|---|---|---|---|
| Cheung et al (18) | No limitations | Limitationsc | Limitationsl | No limitations |
| Gade et al (19) | Limitationsa | Limitationsd | Limitationsm | Limitationsf |
| Hanks et al (20) | No limitations | Limitationsg | Limitationsi | No limitations |
| Ahronheim et al (15) | Limitationsa | Limitationsq | Limitationsp | No limitations |
| Brumley et al (27) | Limitationsa | Limitationsc | Limitationsn | No limitations |
| Aiken et al (16) | No limitations | Limitationsk | Limitationsj | No limitations |
| Mitchell et al (22) | Limitationsb | Limitationse | Limitationso | No limitations |
| Jordhoy et al (21) | Limitationsa | Limitationsh | No limitations | No limitations |
| Zimmermann et al (24) | No limitations | Limitationsc | No limitations | No limitations |
| Temel et al (23) | Limitationsa | Limitationsc | Limitationsr | No limitations |
Concealment unclear or not done.
Concealment broken.
Unblinded patients completing satisfaction assessments.
Blinding unclear.
General practitioner blinded to study hypothesis but not to intervention.
Hospital readmissions and emergency department visits not reported.
Satisfaction assessments completed by unblinded patients and staff.
Satisfaction and quality-of-life assessments completed by unblinded participants.
Intention-to-Treat principle was used but 19 patients switched from intervention to control and 2 vice versa; attrition rate (loss to follow up, or died) at 1 week 44% for intervention and 43% for control.
High attrition rate not due to death—intervention group 16%, control group 25%.
The frequency, severity, and distress of symptoms assessed by unblinded patients.
45% return rate of questionnaires—5/10 intervention, 4/10 control.
For ICU-admissions outcome, data reported from 2 of the study's 3 sites.
Reduction in sample size for satisfaction outcomes—baseline 93% of sample used, 30 days 72%, 60 days 56%, 90 days 50%. No method stated for managing missing data.
No protocol for managing missing data; 47% of study population missing (died or withdrew) by 3-week assessment time.
No protocol for managing missing data; attrition rate not documented.
Unblinded health care workers may have been biased in favour of completing advance care planning.
FAMCARE scores not reported due to large proportion of missing data.