Episode 16 - Transforming Clinical Data into a Peer-Reviewed Article (Part 2)

This is the follow-up to our last episode of the BOEI Podcast, Transforming Clinical Data into a Peer-Reviewed Article (Part 1). In this Part 2, we talk with Dr. Angeli Yu, Cornea Fellow at the University of Ferrara, about how to structure collected data and perform data analysis and interpretation for a peer-reviewed article.

As in the last episode, Dr. Yu is joined by Clotilde Jumelle, who is a Regulatory Project Manager and part of our Clinical Writing Team at Medevise, and Kristine Morrill as moderator.

If you didn’t listen to Part 1, listen in for valuable guidance on how to develop a successful clinical manuscript for peer review from Prof. Dr. Stephanie Joachim, who is Head of Experimental Eye Research at Ruhr University in Buchum.

Speakers:

• Kristine Morrill, Founder and President of Medevise Consulting

• Dr. Angeli Yu, Cornea Clinical and Research Fellow, Ferrara University

• Clotilde Jumelle, Regulatory Project Manager at Medevise Consulting

Key Takeaways:

There are three initial questions that are essential to consider before you begin to present clinical data:

1. What are your study goals? The data that you need to collect and the analysis that you need to perform would be based on why you're actually performing the study.

2. What is the appropriate design?

3. How do you want to present your data?

Look Beyond the “P Value”: More than just the p-value, it’s important to consider the data itself to understand whether or not there is a significant difference. Is it clinically meaningful?

What is a P-Value? A p-value is used to determine significance of the results. When the p-value is less than the predetermined level of significance, such as 0.05, this indicates that the results observed in your sample are unlikely to have occurred by chance. It is important is that you report the p-values as the exact p-values

Use Inferential Statistics to Classify Your Data Set. When you have categorical data and unpaired data, you use a chi-square. If there are only two groups to be compared, you can sometimes use a Fisher exact test. When you have paired data on the other hand, then you use a McNemar test or a Cochran's Q. When you have categorical data and you have unpaired data, you use a chi-square. Of course, there's a difference. If there are only two groups to be compared, you can sometimes use a Fisher exact test. When you have paired data on the other hand, then you use a McNemar test or a Cochran's Q.

Continuous Variables: The tests for continuous variables are slightly different. Consider whether your data follows a Gaussian curve or not, and whether the categories are within subject or between subjects, and the number of groups.

---

NOTE: This podcast can also be seen as a webinar recording with additional graphics and slides that illustrate these points. Watch it on our YouTube channel.