Hello,
I´m in the process of writing a literature review on the topic of cough assessment for specific populations. The goal is to retrieve norm values for those populations before treatment. Therefore a lot of different study types will be included.
Because outcome values after an intervention or the intervention itself are not of interest to my research question, my questions are:
1. How do I assess the quality of the different study types?
2. Do I apply the checklists like the Newcastle Ottawa Scales for the different study types on the intire study although the main objectives (the interventions or outcomes) of most of the studies are not going to be reviewed?
3. Does anyone have experiences or ideas?
Many thanks,
Laura
Hi,
Normally risk of bias is assessed in regard to the study design and which study design(s) you decide to include in your review depends on your specific question. There are several different tools for different study types, you can find more about that in this paper:
https://mmrjournal.biomedcentral.com/articles/10.1186/s40779-020-00238-8
However, your aim does not seem to be to investigate effects or cough assessment as a diagnostic tool or aid, but just specific population values on cough assessments made in various populations. This means that many of the questions and domains in the RoB-tools do not apply to your question. For example, you may assess an RCT as having high RoB based on an instrument, but the cough assessment used in the study was state of the art with high credibility in the numbers reported.
Perhaps you should try to develop your own questions that assesses quality/risk of bias for the cough assessment. You are probably interested in nailing your population so questions on selection and description of group characteristic will be of value as well as if there were attrition/missing values, and how/by whom the assessment was performed and documented etcetera.
If I am understanding correctly you are using the studies to give a description of particular groups/populations. The underlying research questions of the original studies are of little importance to you.
As the 'risk of bias' of any study can never be entirely independent of the question at hand I think you ought to select a single framework that applies the same criteria, irrespective of the study design - that framework should assess the risk of bias in answering the question you have asked.
I think that a framework suited to cross-sectional descriptive studies is what you are after. A simple approach is to use the relevant section of the Mixed Methods Appraisal Tool. It's not perfect but no tool is.
Whatever you use I suggest you take a step back and determine what are the major issues that relate to the ability of this body of research to answer your question (measurement of outcomes, population/samples / sampling strategy/selection bias). Use this to 'populate' the generic items in the framework with your own signalling questions and, at first pass, rather than just rating items also 'extract' the information on which you are basing the judgement.
https://www.nccmt.ca/knowledge-repositories/search/232%20(accessed%20May%202017)
Check this: https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools
- Badges
- Science topic
- Similar topics
- Filing