Understanding the Impact of New Technologies in Pharma CI Data Journeys: Reversing Traditional Challenges

Drug companies are constantly driven by the need to innovate, discover drugs, and market them seamlessly globally. With average drug life cycles spanning two decades, right from the drug development, approval and commercialization, commercial competitive intelligence (CI) poses unique challenges. Traditional CI research includes practices such as wargaming, ratio analysis and Porter’s five forces. But these data management and research techniques in pharma CI teams operate with multitude of users in a highly regulated market environment. The increased technical nature of drugs and health products also complicates the data quality.

Traditionally, sources of data were disparate and data feeds separately entered the database as opposed to data collection in real-time. With multiple databases and sources, there was a lack of well-tested system of data management in pharma. Data journeys were also fraught with many discrepancies. From the start at the data collection stage, to the data cleaning and processing and the data visualization stage, a lot of time was spent on completing complex tasks that were labor intensive. The ROI was very slow, and the final journey ended at analysis and recommendation with limited time spent on insight generation.

Data sources include news alerts, updates on company websites, clinical trial repositories, earning reports and other unstructured data sources. All these sources had to be manually reviewed, tagged for relevance and the insights derived from them. A lot of times, key information could be missed out since it is highly probable that a key source of information has been overlooked by someone who is manually reviewing the information. these let to delays in information being presented to key stakeholders.

CI teams found the traditional data management processes all too complex with limited time spent on insights generation that supported decision making in companies. The traditional approach led to the bulk of the time spent on data aggregating and management. About 10 per cent of the time would then be spent on analysis and recommendation, insight generation and interpretation. This was hardly adequate given the competitive drug marketing and commercialization environment.

Data integrity is also a chief concern with traditional CI research. Data retrieval another key area, that needs to be available on request, but was not the case. As a result, the industry overall decided to implement innovative analytical solutions to not just increase returns on competitive research but also for swifter data retrieval.

Since data management in CI needs to be holistic and agile, pharma companies are increasingly automating and modernizing the process. They aim to pursue smart data-driven outcomes that are a result of scalable data management systems yielding intelligence and insight faster and more efficiently. With the use of emergent technology, the cleaning and organization of these processes have become accelerated and advanced. CI teams now spent about 45 per cent of their time on data collection and cleaning, while more than 55 per cent of their time was now spent on analysis and insight generation.

With the induction of cloud, Artificial Intelligence (AI), Natural Language Processing (NLP) and Text mining, data management in CI teams has become more robust and successful. These technologies are helping pharma companies create automated data mining pipelines, structuring of data sources that allow for better management and search retrieval, and generate better insights. Data sources such as news websites can be mined for information with text mining algorithms. Some other data sources such as clinical trial repositories provide API connectors or XML database updates which can be automated for periodic refresh. Structuring these data sources using good data management practices can help in retaining historical data with powerful search and retrieval. A good data management system also helps the CI teams build good predictive algorithms using historical data and provide better insights to key stakeholders. Good data management practices also help companies reduce any chance of data privacy violation or misuse, which is a key concern in this industry.

Setting up a centralized and automated data management practice enables CI teams to collaborate among themselves and use syndicated data and other data sources in conjunction to secondary and primary CI information to provide more holistic insights.

Companies use each of these technologies to inform their competitor analyses, social media marketing, product development or even advertising strategy. Leveraging each of these technologies, CI teams can now bring about best practices in data management. Advanced visualization practices can help filter and drill down data for business intelligence teams to integrate and collaborate efficiently. Data analysis becomes a seamless process that is augmented and aided by intelligent technology platforms and CI tools.

Text mining algorithms are now capable of attending to three important facets of CI: the event alerts, filtering, and search. Together with NLP, Information Retrieval, and Data Mining, CI tasks become easy and useful for pharma companies. Each of these standard data management technologies can be “assembled” within the whole competitor analysis work stream. Cloud-based CI platforms on the other hand, allow an immediate, customizable, and interactive way of storing CI data. CI data can be accessed from multiple devices through secure and seamless authentication. This is particularly important for sales and marketing analysts who attend conferences and meetings in varied geographical locations. Pharma groups can integrate their cloud solutions across company systems to combine commercial data and business intelligence tools. Cloud platforms are customizable and can display in-house and external news, sales and pricing news and competitor profiles through accessible modes like dashboards and layouts. Combining valuable data analytics insights, AI and cloud-based platforms streamlines market data for CI teams and strategy groups.

Cloud computing platforms in addition to AI-powered tools can harvest data from millions of online websites to draw out detailed insights by reading a competitor’s entire digital footprint both on and off the competitor site.

Natural Language Processing can analyze the sentiment that competitor brands face on social media and read reviews from social networking platforms to convert it into usable data and actionable intelligence.

Finally, using a good visualization engine would help CI users to create insights in a structured and standard manner that can be distributed to key stakeholders through a push reporting system, as against the traditional method of distributing information through PPT or emails. Information in PPT or emails is easily lost and relevance of it also is lost over time. However, through a visualization platform, the change over time can easily be monitored to derive better insights. Implementing a good visualization engine also eliminates the need for CI Teams to spend time in preparing presentations. These result in huge time savings, allowing CI teams to spend more time on insight generation than in data collection and preparation of actionable intelligence.

Pharma companies in the last decade or so have increasingly recognized the importance of social media posts as a valuable source of patient views, symptoms, and outcomes of health research. These have a potential to reveal important information about patient needs, drug availability and key market conditions apart from the general use of competitive products.

Drug major Roche understood the importance of analyzing social media posts and content for understanding the voice of the patient in critical drug delivery and development. It used text mining with NLP technology to unlock the value of social media chatter. This enabled the company to glean in sights that influenced the drug development and marketing process. For a developmental drug to treat Parkinson’s Disease, Roche tried to understand the Parkinson’s conceptual disease model on social media. The researchers at Roche established a series of NLP-based text mining queries to understand patient talk on Parkinson’s across several social platforms, patient forums, websites, and blogs. This was observational research with data being downloaded from open health social communities and sites. NLP helped them to use open linguistic patterns to understand and extract unknown symptoms and impacts by looking at key words and phrases.

As a result of this endeavor, Roche researchers gained two major insights. They added two impacts to the original disease model and identified several additional symptoms. This NLP-enabled social media listening project also helped Roche enhance its understanding of symptoms that impact Parkinson’s patients and gain awareness to improve the design of future clinical trials.

In another AI and Machine Learning (ML) initiative by California-based Roche subsidiary Genentech, researchers partnered with GNS Healthcare to use its AI platform to analyze its oncology therapeutics and its reach within the patient community. Genentech aimed to use ML to convert high volumes of patient data on cancer afflicted individuals into computer models and use these to identify new targets for cancer therapy. It also applied AI to develop research into better diagnostics and biomarkers and develop new drug targets and design better drugs for marketing.