I am planning on conducting a systematic review, which will focus on the impact randomised lifestyle interventions (RCTs) on intergenerational health.
Initially, I considered the ROBINS-I tool, which is often used for non-randomised cohort follow-up studies, but the non-randomised element does not align with the randomised nature of RCTs. My next thought was the RoB 2 tool, but it seems more geared towards evaluating the outcome of the trial (e.g. weight loss during trial) rather than the longer-term effects on the cohort.
I have looked online and have not found a specific tool tailored to assessing risk of bias in long-term follow-up cohort studies from RCTs. As it stands, I'm leaning towards the ROBINS-I tool, but any expertise would be appreciated.
Thanks.
This is aside from your questions, but at the recent Public Health Preventative Conference I listened to a researcher who considers you should not use the term lifestyle. The talk is as below. Dr Mark Robinson
Senior Research Fellow
Institute for Social Science Research, University Of Queensland
The lazy language of lifestyles: a call to action.
Thanks Deborah J Hilton . Interesting. Could you elaborate as to why they think we should not use the term "lifestyle" and what we should use instead? I struggle to think of a term that summarises changes in "lifestyle", such as diet and physical activity, in one word...or should we be stating exactly what we mean (e.g. I used "randomised lifestyle interventions", so this would be "randomised diet and physical activity interventions" ?)
Thanks Sam. I'm undecided whether I agree or disagree with this. In some ways some people could choose a more healthy lifestyle, but in some cases the situation they are put in, means they have no choice so I am undecided about usage of the word implying that it is totally the person's own choice. The notes that I took as I listened are below, this is not well written sentences as it is my note taking as I'm sometimes haphazard as I can't keep up with talks anymore as I get older, so it is just some words I've written down as I listened. Messaging and personal responsibility. obesity stragegy; nsw active living strategy
Word permeates our prof. discourse, ubiquitous use of the word lifestyle, framing, value based messaging. How it is characteristic manner in which a person lives or chooses to live his/ her life, zombie hypothesis that one's health is down to one's choices. editorial. Wiley Health promotion journal. challenge discourse of personal responsibility by avoiding using lifestyle language. I tend to think professionals are too quick to label those they are trying to assist putting them in categories whereby the label does imply they are at fault. I heard also Dr Cassandra Wright talk on alcohol in the NT. She mentions priority populations. It is a similar? about use of this word, which some people won't like use of the word priority. Again, I am undecided as to my opinion on the use of priority as a word in the context of populations. I never hear professionals label themselves accordingly. You could say obsessive shopper researchers could you not, about those who have too many suits! Compulsive flying researchers about those who travel inappropriately for conferences. But it never works that way. Hypocrisy I guess like you find in the church also; Ministers praying for the poor, but then they feather their own nests when no one is looking. Take Care.
Hello Sam, my thoughts are that the risk of bias in long-term study arise from drop-out of participants and missing data. The Cochrane ROB 2 tool addresses bias due to missing outcome data, but perhaps researchers need to take a nuanced approach for longer-term outcomes. This might concern how missing data are handled, e.g. missing at randon, missing completely at random etc. MIssing data might be ignored (poor choice) or there might be imputation of previous values or average values, for example. Hope this helps and good luck with the review.
Thank you Paul Baker.
My only issue that I have with the RoB2 tool is about the "missing data". In the studies that I will be reviewing, only sub-samples of the original cohorts were followed-up (e.g. "100 participants from the original 500 were followed-up to achieve a power of etc..."). Therefore, are the hypothetical 400 participants not studied classed as missing, even if they were not recruited for the follow-up?
Hi Sam, I would advise you get advice from a statistician. But, my hunch is that if the sub-group is representative of the overall cohort, not "cherry-picked" and their characteristics and experiences/exposures are relevant to the study question, AND analysis and conclusions refer to the 100 and not the 500, it's not missing data per se. Would welcome comments from others having stuck my neck out!
Thanks again Paul Baker.
I agree with you, your answer was what I was also leaning towards if I go with the RoB 2 tool. Indeed, some of the studies have taken this into account and have done comparisons of the baseline characteristics between those who were and were not followed-up. I guess any studies that have not done this would be scored lower using the RoB 2 tool.
I second Paul's comment for other opinions.
- Badges
- Science topic
- Similar topics
- Higher Education
- Universities