Article
ROB-OPS: a first draft for a new risk-of-bias tool to assess the applicability and risk of bias in overall prognosis studies
Search Medline for
Authors
Published: | August 30, 2022 |
---|
Outline
Text
Background/research question: Overall prognosis research examines the course and future outcomes of a health-related condition in the current context of care, which is the basis for prioritization, clinical decision making and future research. Systematic reviews of overall prognosis studies on one health condition provide an overview across various settings, health care systems, and populations. An integral part of systematic reviews is the assessment of the applicability and risk of bias (RoB) of included studies, which is the potential for systematic errors from the truth. As for overall prognosis studies, in contrast to prognostic factor and model studies, no standardized tool for the assessment of overall prognosis studies is yet available, we aim to develop such a tool. Here, we present the first milestone in developing the new tool, a draft which will be presented to a wider audience.
Methods: First, we set up a steering group consisting of seven experts on prognosis and systematic reviews. We decided on relevant domains for applicability and RoB, then collected initial items from known tools in the field and additional tools identified by a scoping search, removed duplicate items, rephrased items where necessary, and reflected on answering options through several rounds of discussion.
Preliminary/expected results, outlook: The current draft is structured into separate sections, starting with the three applicability domains “participants”, which includes one signaling question, “context”, which includes one signaling question, and “outcome” with three signaling questions. The section on RoB contains four domains. The domain “participants”, which represents selection bias, includes two signaling questions, the “outcome” and “analysis” domains consist of four signaling questions respectively, and the “selective reporting” domain is represented by one item. We added an explanatory note where necessary and decided on answer options in line with other RoB-tools (yes, probably yes, probably no, no, not sufficient information), and a short elaboration on when to use which option.
Conclusion
The current efforts resulted in a first draft for a tool to evaluate overall prognosis studies under agreement of the steering group. Next development step will be a stakeholder survey to involve a wider group of experts to gather a larger range of opinions regarding the domains, signaling questions, potentially missing topics, and answer options. Based on this feedback, the tool will be further revised.
Competing interests: No financial conflicts of interest to declare.