Blinding is often a challenge, particularly when working with people with cognitive impairment. I'd be really interested to hear any comments on the effectiveness of blinding in this study, and any suggestions to keep researchers blinded.
Many Thanks
Clare
In general blinding (single/double/triple) wherever feasible strengthens the scientific validity of interventional studies, particularly to control for placebo and other effects. In some cases bliniding is not methodologically feasible, and if it cannot be done some alternate strategies should be put in place to limit the implications of non-blinding, and the weakening of the scientific validity should be acknowledged in the results. Hard to be more specific without better knowledge of the study in question.
Hi,
This is a single-blind study, in which the subjects but not the researchers know which subjects are receiving the cognitive rehabilitation vs treatment as usual. The researchers taking the measures will be blind to allocation, as will the data analysts.
Despite careful consideration of methods to blind subjects in trials, situations will invariably arise when some or all groups of subjects simply cannot ethically be blinded.
If blinding is not possible, researchers can choose to incorporate other methodologic safeguards (e.g. use objective, reliable outcomes if possible, consider duplicate assessment etc.) but should understand and acknowledge the limitations of these strategies.
Best regards,
Qin Xiang
Hi Qin and Douglas
Thanks for your responses.
I was interested about the experiences of blinding in this trial in particular, from a practical perspective. In my experience, I have found that whilst researchers can be blinded and ask participants not to reveal what group they were allocated to, participants do sometimes inadvertently reveal this information anyway when they see the researcher at follow-up.
Best
Clare
It is sometimes possible to set a social science intervention study up so that participants, research assistants and even the PI don't know exactly what group a specific participant is in. (Prospective participants must of course be asked to consent to being randomised, and consent to not knowing what group they are in until they are debriefed after the study). (There are also ethical obligations that a PI owes to untreated control groups if an intervention is shown to be effective, but that is a different question). As you say, and as mentioned in both responses above, sometimes it is not possible to blind (for ethical or practical reasons), or to prevent inadvertent disclosure or unblinding. In biomedical clinical trials unblinding should be declared, and sometimes data from unblinded participants is excluded from final analysis. In general, social scientists are not as rigid about blinding as biomedical researchers, paradoxically where it is often more important scientifically because outcome measures are less categorical and more subjective than in much biomedical research. In any event, if, as mentioned above, double or triple blinding is not practically possible, it shoud be mentioned as a study limitation in the discussion of results.
In a situation where the researcher may be able to guess which group a patient belongs to, one strategy is to assess the extent to which the study was blind by getting those doing the assessments to record their guess about which group each patient belonged to. These data should be mentioned when the data are published. There are many published studies that are allegedly blind, but no assessment of blinding was done, so they may not have been blind.
- Badges
- Science topic