Research data with an analyst’s naked eye

Two years ago, I’ve been at a meeting, which after all, changed my life. At this meeting I was basically sitting for 1.5 hours in complete silence, drowning in unfamiliar terms and abbreviations while trying to make some meaningful notes – since I was the business analyst of the team, I had to summarize what the next steps will be, right?

My first notes were like: “Oncology…neurology… CDISC standards (?) Safety reports… EDC?”

So, I landed on a project in the clinical research domain. The company I work at has many clients across multiple industries and so the projects we engage with are very broad and never the same. Prior to that I worked with manufacturer’s data and customer reviews – yes, quite far from clinical trials!

But I did not have the time to feel sorry for myself, we had deadlines. As an analyst, I did what we do for a living – analyzed the situation, and not panicked at all. Not at all…

First, what does a data analyst in clinical research do?

In general, clinical research is an experimentation with developing new treatments, devices or procedures intended for human use – in human volunteers.

Today more and more medical records are stored electronically, researchers can identify patterns and immediately spot anomalies in patients’ information. In medical analytics, business intelligence tools can provide key insights in a second: visuals focusing on patient data can reveal connections of medical history, lab results and adverse reactions related to the treatment. Other visuals aggregate data of all patients and outline the overall effect on the patients’ health and disease. Real-time data allows a continuous access to the results, while BI dashboards point out, where we should look: it has an enormous potential to ensure safety analysis and to improve patient care as well as cost management1. Exciting, isn’t it?

As in many positions, a clinical data analyst can wear many hats from reviewing and validating that a trial follows protocol or coordinating data management effort to leveraging BI in clinical research. But regardless what you are responsible for, you should understand all parts to effectively manage yours. In my case, that was developing the ETL process and dashboards for medical analytics.

I faced many challenges while gaining some knowledge. Looking back now, I tried to get a grip on symptoms instead of realizing the root cause first: what is driving the customer and how the work of my team will effectively serve their needs? Well, to be able to ask that question and read between the lines, I should have been familiar with the industry first: how a study can be designed, what data is collected and related, how statisticians review and report data, what are they looking for – so many pieces which I tried put together in this series to help you to be one step ahead than I was.

If you understand the industry, you will connect faster with your customer. You won’t just execute what they ask for, but you will think together. Take advantage of both the domain knowledge and data analytics best practices! Being involved in that level will not just improve the quality of your dashboards but also engage you with advocating BI and supporting end users how effectively work with it.

It’s hard to start if you don’t have the right questions

What I learned that the internet has the answer for everything. BUT! Only if you know, what you’re supposed to ask, and I didn’t know at the beginning. It took months to interpret the data, the relations and the meaning of visuals by reading hundreds of posts and papers. I wished to have one crash course, where someone already took this path, and gives me one document saying – read this, and you’ll know all you need to know.

That’s why I was eager to start with this post. If you don’t know, what you are looking for, searching takes time and time is the most valuable. In the next articles, I’m going to summarize all the information I gathered over the years in medical review, and point to resources, which can help in understanding the big picture.

I’ll cover on high-level the process of conducting a research, the actors included, and the terms used daily. But as a data analyst I’ll share even more details about the data collected, the data transformations applied and finally the visualizations which serve the primary goal: to secure the patients while trying to develop a ground-breaking treatment for diseases like cancer, the Alzheimer’s or in 2020, the COVID-19.

Finally, if any of my teammates stumbles into this article: I’d like to thank y’all, as you’ve contributed in so many ways in my learning path both personally and professionally!

Resources:

  1. Business Intelligence in Healthcare, Villanova University, 2020.