Understanding the Impact of ‘Black Box’ Warnings on Patient Behaviors

Sep 29, 2022


A data case study from patients perspective of the impact of FDA safety warnings. 

As the world’s first patient intelligence company aimed at meeting the needs of the healthcare industry, Talking Medicines was established with the mission of giving patients a voice. Our PatientMetRx cloud-based data service combines machine learning with natural language processing to provide science-based insights which cut through the noise from social platforms. By structuring and translating the patient’s voice from these platforms into empowering metrics, PatientMetRx® can give healthcare marketing agencies the edge when it comes winning and retaining business.In September 2021, an FDA approved auto-immune medicine was at the centre of a safety concern, which culminated in a “black box” warning.1 XELJANZ® (tofacitinib) is the first approved pill of its kind (JAK inhibitor) that treats adults with moderate to severe rheumatoid arthritis, active psoriatic arthritis, active ankylosing spondylitis, and moderate to severe ulcerative colitis.2A black box warning is the FDA’s most stringent warning for drugs and medical devices on the market. Black box warnings, or boxed warnings, alert the public and health care providers to serious side effects, such as injury or death.3

We wanted to use PatientMetRx® to see the effect that this regulatory event had on patient experiences, at what lengths it was discussed and whether this impacted patient confidence and opinions expressed during this period of time. Running the data through our platform, we looked at how many opinion posts were talking about the black box warning and which terms and in what quantity these terms were mentioned in opinion posts. Our assumption is that words such as ‘clots’ and ‘FDA’ would be mentioned in relation to this event, which guided our data search. These counts were then visualised as a bar chart to show the portion of posts that contained these terms of interest and a pie chart to show the relative proportion of each term across posts. For instance, how many times the term “FDA” was mentioned across the subset of posts compared to “clot”.

PatientMetRx(R) Data Story

Fig 1:PatientMetRx® Patient Feed view showing opinion posts that contain “clot”, an example of a search term of interest related to the FDA black box warning news.

Having identified the key themes (using tags) we filtered our opinion map by these themes, allowing a more focused overview of patient sentiment and the context in which these patient opinions of XELJANZ® were being mentioned – at a sentence and post level.

Fig 2: PatientMetRx® Opinion Map view – a node style diagram showing only the opinion posts that were found to contain the word “clot” and have been tagged accordingly.

Having identified the key themes (using tags) we filtered our opinion map by these themes, allowing a more focused overview of patient sentiment and the context in which these patient opinions of XELJANZ® were being mentioned – at a sentence and post level.Before we ran the case through PatientMetRx®, we made a number of assumptions, the biggest of which was that the FDA warning would cause a considerable shift in sentiment for XELJANZ®. We also assumed that patients would want to change their medication.

However, what we found through our analysis was quite different. Whilst there were mentions of FDA and safety concerns in our analysis results, such concerns were not common and were only expressed in 20% of the posts analysed. To the contrary, people were more likely to make positive appraisals of XELJANZ®.

Using PatientMetRx®, we were quickly able to establish the impact of the FDA warning on patient perception both contextually and from a sentiment analysis perspective, insights that could easily be used to inform specific communication strategy and language in the face of a regulatory issue. Understanding that the patient response to black box warnings in the context of a chronic condition is not negative is vital information for a brand trying to get to grips with how they should respond and support patients through such an event.

As important as the information itself is the speed with which such information can be obtained. PatientMetRx® does not require users to prepare complicated data queries or extract and analyze information through complex excels. All analysis can be completed in the product with minimal effort due to search, tag and flag functionality that allows users to filter vast volumes of data to relevant patient insight. A data study proved that pitch ready data and insight was available 80% faster than a conventional excel based analysis, a significant difference when it comes to brand’s ability to make an informed decision on how to respond to such an event.

For more information on how PatientMetRx® can help you to gather valuable patient intelligence, quickly and effectively, visit www.patientmetrx.com or contact Mike Strassberg, Chief Customer Officer, mike@talkingmedicines.com

[1] Blood clots

[2] Inflammation & immunology

[3] Black Box warnings

PatientMetRx® may be a super smart AI accelerated machine learning platform, but it’s backed by a team of human beings, passionate about putting the patient voice at the heart of healthcare.

Keep Listening

Looking to do things smarter, faster and more effectively with your marketing? Subscribe to the MetRx Matters newsletter for the latest insights on brands and patient experience.