How do you justify (or would you justify) a new research study to ensure it doesn’t unnecessarily duplicate previous work or repeat past methodological mistakes?

Before developing a research problem on the topic to be studied, it is essential to conduct a broad literature review on the topic to identify possible gaps, as well as the existence or not of previous studies or even a careful evaluation of the methodological quality of existing articles, which is done in many systematic reviews. Justifications such as a new study design to study the effect of an exposure on the outcome are valid, for example, when we consider that at a given time this relationship was evaluated in a cross-sectional study (which often does not allow establishing temporality, much less causality) and could now be evaluated in a cohort study, we establish a solid justification to bring more robust evidence on the topic. In addition, the literature review allows us to recognize the biases of other studies that can be corrected in a new study based on the sampling process or data collection.

Well prior to starting a new project it is essential doing a very thorough search (literature review) to see how well-studied the project is and to find potential gaps that need to be filled.

To build *meaningfully* on what's known, not just add noise, we must:

1. Listen deeply: Conduct a *thorough*, humble literature review – not just ticking a box, but truly understanding previous findings, limitations, and the voices (especially from LMICs) already in the conversation. 

2. Spot the real gap: Identify a clear, unanswered question that *matters* to communities or science, not just a minor variation. Explain *why* revisiting an area is needed (e.g., new context, past flaws). 

3. Learn from stumbles: Honestly analyze *why* past methods failed (e.g., inadequate local engagement, flawed tools) and design ours to actively address those pitfalls. 

4. Add unique value: Articulate precisely how our approach whether new data, perspective, or method offers fresh insight you can't get by re-analyzing old work. 

It's impossible for all researchers to be aware of all previous studies. Most duplicated research is found published in conferences and weaker journals, generally, whether by the same authors or different ones. The rate of such duplication becomes lower and lower in higher-impact factor journals, as the reviewers there are more robust and more aware of technologies and research.
Looking from another aspect, if an idea is novel, then it's easy to describe. The researcher can easily write about previous weaknesses and the new method's advantages.
For incremental modifications, these can be found within the literature that has been used in the same paper, while some researchers will hide the paper from where they might get some critical knowledge. This is the most difficult situation, where it depends totally on the reviewers' background information.