The risk of a clinical trial failing is around a whopping 80% because it is hard to find enough trial participants in a timely manner. Clinical research organizations (CROs) have to contact the right patients but without breaching the privacy of medical records. The best way to do this is to work with physicians as influencers and brand ambassadors, as they are legally entitled to offer various treatment options to their patients, including clinical trial research. The natural language processing (NLP) social graph technique is one possible solution for this situation.
The social graph technique in data science refers to a method of data analysis derived from using social networks to find influencers; those people engaging with the largest and most relevant audience. It’s most often represented as a map with nodes (influencers and followers) connected with lines (various kinds of subscriptions on social media).
There are multiple data sources that can serve as reference points for creating social graphs. For example, social graphs can be developed from or enriched by the data obtained from public datasets, such as PubMed, ClinicalTrials.gov, or H-CUP, and from web sources such as Google Scholar, vitals.com, ratemds.com, etc. With the help of NLP, the data from these datasets is structured, semantically parsed, and pre-processed with extracted keywords and relationships between nodes, and then ranked by its impact. Such analysis can be visualized in the form of heatmaps that show the geographical locations of doctors/influencers and the most prospective sites.
All kinds of brands actively use this approach when marketing their products. They pay influencers to showcase particular items, like a clutch/purse, a dress, a car, or any other commodity, to their audience and recommend them for purchase. The issue here is that people rarely follow a single influencer and there are overlaps in audiences where multiple influencers address the same gender/age/social strata/etc groups. Thus, a potential customer could receive advertising posts about the same brand from multiple influencers, which might result in customer backlash and an unwillingness to buy the product.
→ Read our article on Pharma Manufacturing – Improving the risk-reward calculus for clinical trials: How natural language processing and machine learning can boost success in drug development by Michael DePalma and Igor Kryglyak
The social graph example below explains the weight of every node in terms of degree, or the number of connections and rank, or the value. As you can see, node #6 is the most valuable as it has 4 connections, and node #2 is the runner-up with 3 connections. Node #7 is an overlap and can be omitted.
NLP can be applied to social graphs to programmatically analyze the audiences of various influencers to find matching names and highlight them on the graph. This helps determine the influencers whose audiences overlap as little as possible with those of other influencers and then promote products or services through them. As a result, the brands optimize their marketing expenses by partnering with the smallest number of influencers needed to cover the biggest possible target audience.
→ Read an overview of digital tools for drug commercialization to accelerate the drug development lifecycle and speed up the overall time to market.
Coming back to the issue at hand, social graph NLP can be of great use to CROs, sponsors in clinical trials, and medical institutions. While medical details of every patient’s condition and disorder are protected by HIPAA, physicians should keep public the records of their overall patient flow. In addition, many doctors like cosmetologists, nutritionists and plastic surgeons have an active following on social media.
By analyzing publicly available data, CROs can find the doctors with the most relevant audiences, based on their area of expertise and geographical location, and then advertise to them directly. These doctors can then inform their patients about an opportunity to take part in a clinical trial that could help solve their health issues. All this is done without breaching the patient’s privacy, as the CROs do not know all their personal medical details and cannot address the patients directly. It is up to the patient to decide whether they want to progress with patient recruitment through their doctor and take part in a trial.
Using the NLP social graph for clinical research helps to meet clinical trial patient guidelines and quickly gather large pools of eligible patients. This is a win-win situation for all parties, as clinical trial sponsors can approach the doctors with the most relevant patient pools. The doctors can promote participation in clinical trials to their most relevant patients and the patients get an amazing chance to overcome their health problems using the latest medical innovations.
→ Explore Avenga’s expertise in clinical trial software development.
The issue here is the lack of readymade tools for social graph NLP analysis. Well, this is one of the main reasons why 80% of clinical trials fail as the sponsors cannot enroll enough patients and thus the trials fall through, leading to investment losses worth millions of dollars.
Therefore, every CRO or clinical trial sponsor needs to build a social graph NLP platform from scratch, which requires in-depth expertise in software development and data science. As this expertise is not freely available, the best way to obtain it is to hire a trustworthy technology provider with a proven record of successful projects in the field.
Avenga, with its 20+ years of expertise in delivering transformational digital solutions to clients in the pharmaceutical industry, might be your best bet. Avenga has provided Machine Learning (ML) and Artificial Intelligence (AI) platforms, Big Data analytics, data science and business intelligence solutions, cloud transitions, and DevOps services for different companies across the globe and . . . this is just the tip of the iceberg of what Avenga does.
Discover 7 natural language processing and understanding techniques that Avenga uses to solve challenges which promote quick identification of influential Principal Investigators and accelerated Subject recruitment.