The BrackenData team is excited to announce an update of TrialTrends, an interactive tool we’re providing to the public for free. This tool visualizes the number of clinical trials marked completed on ClinicalTrials.gov each year, with the ability to change the dataset in real-time by phase, and trial type such as drug or device. TrialTrends users also have the ability to drill-down to a specific quarter or month on the timeline, or group data by all months or quarters on the timeline.
This tool provides users with a quick and visual analysis about macro clinical trial activity over the past decade. While anyone can access it, were particularly excited about the insight it may provide to:
- Pharmaceutical executives
- Industry analysts
- Clinical development planners and medical affairs teams
- Academic centers
In this post we’ll quickly address frequently asked questions about TrialTrends, provide details about its forecast, and walk through its features.
Open the tool by clicking here or on the button below, and read along with this how-to guide.
Skip to a section of this guide:
- How to Read the Chart
- About the Forecast
- Why Does the Forecast Start In 2016?
- Why the Drop in Clinical Trial Volume in 2017?
- Why Is There Only a Forecast Number for 2017?
- Why Use ClinicalTrials.gov As a Data Source?
- How to Use the Phase Selector
- How to Use the Trial Type Filter
- How to Drill-down or Group by Time Period
How to Read the Chart
The main function of TrialTrends is to visualize past and future clinical trial activity on ClinicalTrials.gov. By default, we have Year on the X-axis, and Count of NCT ID on the Y-axis. An NCT ID is the unique number given to each trial registered on ClinicalTrials.gov. We've filtered the dataset to show only trials marked completed. Therefore, Count of NCT ID by Year shows the number of trials completed in a year on ClinicalTrials.gov.
The solid green line represents the number of clinical trials year after year while the red dotted line is the linear trend of this clinical trial activity.
The black solid line is the forecasted final count of clinical trials completed in years 2016 through 2020 using a probability calculation.
The gray shaded area outlines the area of probability of the forecast, within 95% statistical confidence. For example, the tool predicts there will be 14989 drug trials completed on ClinicalTrials.gov in 2018. We calculate with 95% statistical confidence that there will be between 12757 and 17220 trials completed through the year.
About the Forecast
The forecast use the last three years of data available to calculate a forecast. The upper and lower bounds of the forecast outline what is probable to happen with 95% statistical confidence, and is represented by the grey shaded area.
Why Does the Forecast Start in 2016?
This forecasted number is more complicated than we can summarize in a few sentences, due to the ClinicalTrials.gov submission process.
Getting a study approved on ClinicalTrials.gov can be a quick process, but in some cases it takes as long as a few months. Additionally, some trials are submitted until after they’ve started. We’ve proven this by comparing submission dates of trials to the dates trials show up on ClinicalTrials.gov (read more about that analysis in our post Why ClinicalTrials.gov Isn't a Rubber Stamp to Get an NCT Number). In conclusion, while data around trial completion rates is available in real-time (such as it is in our TrialFinder product), analyzing these rates on a macro level isn’t reliable until over a year out.
Why the Drop in Clinical Trial Volume in 2017?
One reason there’s a drop in the number of clinical trials recorded on ClinicalTrials.gov in 2017 because not all trials have been marked completed yet (read above).
We predict that there are 4851 trials from 2017 that have yet to show up. It is common for some trials to retroactively be added to the online registry. This is reflected in TrialTrends in the delta between the Count of NCT ID in 2017 and the Forecast for 2017.
With or without these 4851 trials however, 2017 wasn’t a growth year for clinical trial activity.
Why, if the current year is 2018, is there only a forecast number for 2017?
For the reasons listed above, though 2017 is finished there are still a number of clinical trials from 2017 that have yet to show up on ClinicalTrials.gov registry. These trials represent teams who are submitting their clinical trials retroactively as is the case sometimes, or who are waiting for the NIH to officially approve their submission.
Including 2017 data would skew the chart, and misrepresent actual activity.
Why Use ClinicalTrials.gov As a Data Source?
We believe that for reasons described in our blog post Why Isn't There a Global Source for Clinical Trial Data that ClinicalTrials.gov contains over 80% of the worlds clinical trials. Any team producing a study that is to be appear in international medical journals, or commercialize a product in the US, is required to register with ClinicalTrials.gov. As an example of it's exhaustiveness, ClinicalTrials.gov has more European clinical trials than the official online registry for European Trials, EudraCT.
ClinicalTrials.gov is the best data source to use when analyzing trends of global clinical trial activity, and these analyses draw more accurate conclusions when not combining multiple data source.
How to Use the Phase Selector
By default, TrialTrends looks at all phase types. The phase selector tiles shows users a pie chart that breaks down all trials by phase. Interact with this chart to change what data is reflected in the timeline.
For example, click on “Phase 2” to analyze volumes of phase 2 trials over time. The forecast will also only reflect our Phase 2 prediction.
This works in conjunction with the other filters.
How to Use the Trial Type Filter
When uploading a trial protocol, clinical trial teams can mark a category for their clinical trial. Use our “Type” filter to change the chart by trial type.
By default, TrialTrends looks at drug trials. Try changing the view to device trials to see how the data and forecast changes in real-time.
To see all trial types, leave all boxes unchecked.
How to Drill-down or Group by Time Period
By default TrialTrends looks at all data available in a single timeline, but you are able to get more granular with this. To group by time period, or drill-down into a time period, refer to the buttons in the top corners of the tool.
The buttons in the top left of the chart allow you to group the data analysis by time period. When you mouse over these buttons, the dialog box will refer to periods of time as "hierarchies". Going down one hierarchy allows us to look at how clinical trials have been completed by quarter, for all trials in the system. Going down a hierarchy further allows us to look at completions by month. This reveals that December is the most common time to register a trial. There are 6,573 trials in the database registered in December. February is the least common month.
The arrow button in the top right of the chart allows you to drill down into a specific time period. Toggle this button on by clicking it, then click the year you'd like to drill-down into. For example if you want to explore 2013 trials in more detail, click the arrow in the top right, then click on 2013 in the chart.