Social Media Analytics in Pharma-Use Case (Part-1)
In today’s era, social media has become an integral part of life and business and it is even making a significant impact on the healthcare and pharma industries. In addition to the newly generated data, various analytics engines are helping the Pharma industry to discover more personalized drugs which more effective and enhance patient care. Through social media analytics, several new dimensions are being explored, which were previously inaccessible due to lack of information. Through proper channels, now Doctors’ and Patients’ data can easily be accessed by many industry competitors.
With the help of social media data, we will explore several different aspects of a medicine in this post.
Key Dimensions and Parameters for a Medicine:
According to our study, there are usually three major dimensions that are important for assessing a medicine’s performance in the market.
- Root of admin
- Familiarity key opinion Leaders
- Ability to meet unmet needs
- Size of the market
- Sales force
- Addressing multiple indications
- Different influencers
- Forum analysis
Top 10 brands of a popular pharma company, named Pfizer:
Benchmark of Portfolio of different products in Clinical, Commercial and Social:
Text Mining for analysis of Social Aspect:
We investigated the social impact of the medicine, ‘Lyrica’. Using some advanced machine learning techniques, we extracted few patient reviews from Drugs.com and analysed them
Here are some important keywords the patients are discussing:
- Side Effects
Then we tried to see what the experiences of those people are, who have given rating between 5 to 8 out of 10, to that medicine.
Patients in this group are mostly discussing:
In order to understand the context of their discussions, we studied word associations in their reviews.
Note: Comparatively BOLDER and THICKER linesbetween two words indicate a stronger association between them compared to other words.
Capturing the appropriate users:
We extracted 3 metrics from a review to better understand the users and medicine effectiveness.
- Sentiment score
We captured the first two directly from the reviews; the third we calculated from the reviews using the Polarity score from Sentiment Analysis.
- From the name field, we can capture customers who dissatisfied with Lyrica and accordingly target them for marketing.
- The ‘usefull’ attribute indicates the issue which commonly experienced patients. For e.g., user ‘Danvic’ has given only ‘1’ rating to Lyrica and his/her review has been found useful by 66 users, so we should focus on what ‘Danvic’ is saying.
In summary, we have learned that social media and data from other websites, can be used to better understand a medicine’s effectiveness and to improve on areas where consumers are not satisfied.
In the next part of the blog series, we will discuss how social media like Facebook and Doctor’s blog can be used to understand how the market scenario is for different well-known drugs. Until then, Happy Learning 🙂
About Atanu Maity
Atanu is a Senior Data Science professional with 8+ years of experience in Applied Statistical and Mathematical Modeling for different Business Use cases. Currently He is working as Senior Consultant in Data Science-AI/ML with Schlumberger (SLB). Also, He is leading the ML-DL vertical in Sogeti India AAA Practice, which consists of 8-9 team members at present. He is experienced in application of Data Mining, AI/ML Techniques in Different Business Domains like Energy, Retail and Financial Sectors. He is Highly skilled in Python Programming, Data Engineering and Application of Machine Learning, Deep Learning, Active Learning in real life business problems. He has exceptional interpersonal and multitasking abilities. Always highly self-motivated towards learning different skill sets for continuous development and exposure growth and YES, Always OPEN FOR CHALLENGES!
More on Atanu Maity.