Having a specific research question is among the most critical steps during the systematic review process. Consider the following examples:
- What is the effect of drugs on animal models of neurological disorders?
- What is the effect of drugs on animal models of acute stroke?
- What is the effect of fluoxetine on animal models of acute stroke?
- What is the effect of fluoxetine on functional outcomes in rodent models of acute stroke?
Questions 1 and 2 would likely not only comprise too many eligible publications to manually process but are also too broad to obtain a meaningful answer. Questions 3 and 4 are more focused. Now, the task is to define the right breadth and depth of the question: Answers to broader but shallower questions might be more generalizable to other settings whereas answers to more in-depth but narrow questions might be more specific.
These four questions above are gradually narrowed in terms of the population, intervention, and outcome of interest. Based on this, clinical systematic reviews commonly apply the PICO framework to define their research question:
- Patients, Population or Problem: Which characteristics does the population have, e.g., in terms of species, sex, development stage or other factors? Which problem are you addressing? (In the example above: rodent models of stroke)
- Intervention: Which intervention or exposure is tested in the population? (In the example above: fluoxetine)
- Comparison: What is the comparator to the intervention, e.g., placebo, vector-control, or no treatment. (In the example above: placebo)
- Outcomes: Which outcomes are you interested in, e.g., functional disability, plus maze performance, motor neuron cell count? (In the example above: functional outcomes)
Another example would be: what is the effect of opicinumab (I) on myelin regeneration (O) in multiple sclerosis animal models (P) compared to controls (C)? Since not all research questions fit into this PICO-framework, further frameworks have been developed such as PE(CO), SPIDER, SPICE, or ECLIPSE. Of note, systematic reviews do not necessarily need to address the impact of therapies on health; they can also be applied to observational, diagnostic, prognostic, or, e.g., in the case of preclinical systematic reviews, to methodological or translational questions.
Pitfalls
The most common pitfall in this step is having an ill-defined research question. The danger here is that the realization about this might only happen at late stages in the systematic review process. Thus, it is worthwhile to scrutinize the research question. A solution to facilitate this step is to conduct preliminary searches using relevant keywords, e.g., in PubMed, to probe the available evidence: Will there be likely 10’000 publications or only 10?