Table 2.
Items included in RoB-PrevMH, suggested rephrasing and guidance
| Item | Question | Elaboration |
| Domain: selection bias | ||
| 1. Representativeness of the sampling frame | Was the sample invited to participate in the study a true or close representation of the target population? | This question is about how well the people invited to participate in the study match the target population in demographic or clinical characteristics that are believed to be associated with the measured condition. |
| 2. Representativeness of the responders | How would you rate the risk of non-response bias? | This question is about the characteristics and assumed presence of the condition in people who were invited but did not respond in a way that enabled the investigators to measure the condition of interest (not necessarily providing complete data). |
| Suggested rephrasing: Was the sample that provided data a true or close representation the sample invited to participate? | ||
| Domain: information bias | ||
| 3. Measurement of the condition | How do you judge the risk of information bias? | This question is about the appropriateness and reliability of the instrument or method used to measure the condition among people who provided the relevant data. Bias might occur when the training of observers of the outcome was not done or the procedures to collect the data are not the same for every participant or every timepoint included in the study. Bias might also occur when questions refer to the past and their answering depends on the condition. |
| Suggested rephrasing: Was the condition measured/detected in an unbiased and reproducible way for all participants? Additional specific questions depending on the context: Was the tool used to measure the condition validated? Were the methods for measuring the condition standardised? Does the measurement of the condition depend on the memory of the participants? | ||
Each of the three items can be given a judgement high, low or unclear.
*This was a requirement for inclusion in the MHCOVID study and hence not included in the current version of the tool.4