Introduction to Clinical Research

A quick start guide for students, residents, and fellows

A guide to getting started on a clinical research project
Author
Affiliations

Brian Locke, MD, MSc

Department of Pulmonary & Critical Care, Intermountain Medical Center

Overview

This workbook provides a practical, step-by-step guide for medical students, residents, and fellows embarking on a clinical research project.

The focus is specifically on clinical epidemiology and health services research, reflecting both my expertise and the suitability of these areas for clinician-researchers (discussed further in What Type of Project Should You Do?).

While numerous excellent resources exist, this guide uniquely emphasizes:

  • Accessibility for clinical trainees (particularly in Internal Medicine and related specialties) who may have little prior experience with statistics, programming, or research but who wish to develop foundational skills.
  • The often-overlooked “tacit knowledge” required to efficiently conduct research. Unlike PhD or MPH students, clinical trainees typically receive less structured mentorship or formal research training, potentially leading to challenges in conducting rigorous and impactful studies. Additionally, most trainees have less time to devote to in-depth study, so this resource is meant to only focus on the highest-value topics.
  • the key lessons and recurrent issues I’ve encountered. This guide is meant to be an opinionated “lessons learned” that offers practical advice as I see it, rather than sticking to universally accepted advice.

Ideally, this guide serves as a complement to active mentorship, where a dedicated mentor can address specific questions and contextual challenges. However, as detailed in What Kind of Mentor(s) Do You Need?, mentorship can vary significantly in quality and depth. Some mentors might lack content-specific expertise (methodologists are unfortunately rare and highly sought-after) or sufficient involvement, making self-directed resources particularly valuable for trainees eager to independently advance their research skills.

Must you learn to do your own stats?

Strictly speaking, no. Not every successful clinician-researcher masters statistics or study design. Research increasingly involves collaborative teams, where methodologists and/or statisticians can handle technical details. In fact, some advocate for involving a statistician in every research project, a model that is more feasible in well-resourced settings.

However, relying exclusively on others for statistical expertise has important drawbacks:

  • Speed: Conducting at least basic analyses yourself substantially accelerates research.
  • Effectiveness: Understanding key statistical concepts enables more productive interactions with the rest of the research team, as you’re more likely to understand key issues and understand the proposed solutions.
  • Independence: Dependence on external expertise restricts the range of research questions you can feasibly explore.
  • Educational Value: A stated reason (see note) that many training programs retain research requirements is that familiarity with study design and statistics directly enhances clinical practice by improving your interpretation of research literature. Outsourcing related tasks will not help you critically evaluate evidence or apply findings effectively to patient care.

Note: There are less complimentary reasons that explain why such requirements actually persist.

For these reasons and others, I strongly recommend acquiring statistical and study design skills, even though it’s not strictly required in all cases. I still recommend it even if you have access to statisticians and methodologists because it will help you interact with them, and it’ll help you appraise the research.

Learning Objectives:

By the end of this module you will be able to:

  1. Select a feasible, worthwhile research topic that balances rigor, resources, and personal motivation.
  2. Draft a reproducible study plan. This entails formulating a testable question, choosing an appropriate design, outlining data acquisition, and creating (and possibly preregistering) an analytic protocol.
  3. Perform and interpret (frequentist) statistical analyses, including interpreting p-values, classifying data-types, and and avoiding common pitfalls.
  4. Build and explain basic regression models (linear, logistic, survival), including understanding when adjustment is required/desired, interpreting the regression coeficients, and communicating results clearly.
  5. Prepare a manuscript suitable for peer review, applying reporting guidelines (e.g., STROBE) and constructing a coherent argument from introduction through discussion.

About the author

I am an Assistant Professor of Research in the Department of Critical Care at Intermountain Medical Center and core faculty in the University of Utah Pulmonary and Critical Care Medicine Fellowship Program and closely involved with the University of Utah Internal Medicine Residency. I earned a Master of Science in Clinical Investigation in 2024. My academic research primarily addresses the epidemiology of hypercapnic respiratory failure, clinical reasoning, and research informatics. For more information about my current projects, visit reblocke.github.io

TODO List:

# Trouble shoot regression and chi2 equivalence... may be easier with linear regression.
# Create an example of using chatgpt to troubleshoot an error
# Find statisticians in all projects citation - and set up references/bibliography

If you find errors or have suggestions (content, resources, corrections, anything), email me and {first}dot{last}at imail.org