A framework for developing PICO research questions with subgroups in clinical practice guidelines

Developing research questions for evidence-based clinical practice guidelines (CPGs) requires careful consideration of PICO (population, intervention, comparator, outcomes) components and subgroups. This can be challenging, especially for test-related questions. For example, CPGs may not specify whether genetic tests are intended to predict differential treatment responses. When this occurs, subgroup interactions may be overlooked, potentially leading to inadequate evidence and less trustworthy recommendations. Currently available approaches limit research question instructions to interventions or diagnostic tests without considering prognostic and predictive tests. We present a comprehensive framework that supports the development of research questions aimed to provide information about the benefits and harms of interventions or tests, including diagnostic, prognostic, and predictive tests. Tests are also known as factors, indicators, models, nomograms, tools, and so on. This framework emphasizes the importance of the PICO format, defining subgroups, and identifying ideal study designs.

This framework, illustrated through a Table and two Figures, was developed by a team of seven guideline methodologists and one clinician during four online brainstorming sessions. It was subsequently reviewed and improved by eight additional experienced methodologists. The framework was used in the development of a Methods Guide for the Program in Evidence-based Care in the Department of Oncology at McMaster University.

Research questions were classified into two categories: (1) interventions and (2) tests that guide interventions, with the tests being further categorized into diagnostic, prognostic, or predictive. For each category, the framework outlines defining characteristics, ideal study designs, and examples to guide question development.

This framework provides a structured approach for guideline developers to generate appropriate research questions, identify key parameters, and anticipate the potential studies that should be included. It aims to streamline research question development and enhance the quality and trustworthiness of CPG recommendations.