Table 1 Risk of bias assessment table for the systematic review of smoking and alcohol interventions in head and neck cancer
Domain | Low risk | High risk |
|---|---|---|
Sequence generation | Describes a random component, e.g. random number generator, coin toss and shuffling envelopes. | Includes a systematic, non-random component, e.g. hospital record number, DOB, participant preference and based on lab results. |
Allocation concealment | Participants/investigators could not foresee allocation, e.g. central allocation system (web or phone based), sequentially numbered supplement/placebo containers and sequentially numbered sealed envelopes. | Participants/investigators could foresee allocation, e.g. an open random allocation schedule, DOB, case record numbers or assignment envelopes without appropriate safeguards. |
Blinding of participants, personnel and outcome assessors | No blinding conducted but this is deemed not likely to outcome or outcome measurement. No blinding but this is deemed either not appropriate or possible, e.g. not blinding participants in PA or non-supplement-based dietary intervention. Blinding of participants and study personnel where appropriate | No/incomplete blinding where the outcome measurement is likely to have been affected. Blinding conducted but likely that the blinding could have been broken. |
Incomplete outcome data | Reason for missing outcome data unlikely to be related to true outcome. If participants are reported and analysed in the groups they were allocated to irrespective of non-compliance/drop out. Up to 10% loss through drop out or loss-to follow-up. | Reason for missing outcome data likely to be related to true outcome. Participants are not analysed based on the group they were allocated. |
Selective outcome reporting | If all pre-specified outcomes have been reported on. | Reporting on outcomes that were not pre-specified. Outcomes of interest are reported incompletely Fails to include results for an expected key outcome. |
Other sources of bias | The study appears to be free of any other sources of bias. | Other possible cause of bias, e.g. extreme baseline imbalance, claim of fraudulence and conflict of interest for investigators not addressed. |