In the pharmaceutical industry, pipeline analysis plays a vital role in managing and optimizing the drug portfolio. The goal is to provide comprehensive insights into the lifecycle of drug candidates, from discovery through clinical trials to market launch. Traditional approaches to portfolio management can offer some significant benefits, but there are also notable shortcomings, particularly as the volume of pharma data rapidly expands. Pharmaceutical companies have typically relied on interdisciplinary teams of senior experts and support staff to manually analyze the numerous factors involved in capital allocation, including clinical endpoints, FDA regulatory pathways, disease biology, Chemistry Manufacturing and Controls (CMC), toxicology, and more. Many companies hire individual experts or teams to perform these analyses to optimize their portfolios; however, the resulting assessment produces a siloed analysis that’s both based on limited data and highly dependent on an individual expert’s specific knowledge of the subject. Some pharmaceutical companies don’t conduct drug portfolio optimization through pipeline analysis at all. Other pharmaceutical companies conduct pipeline analysis in-house, but internal teams also struggle to incorporate all material information into their risk assessments and decision-making process, resulting in low return on investment. Whether in-house or outsourced, human teams inevitably encounter difficulty in terms of:

• Capturing the complex risks and rewards behind a program
• Gaining a deep understanding of the biology and landscape of drug programs in a space
• Ensuring they are leveraging the complete historical scientific learnings in an indication

Adding to these challenges, it’s difficult to collaborate and coordinate with other functional leaders and disease experts to make sense of all this information and leverage it effectively. It’s time to address the shortcomings of these time-consuming, manual, and subjective approaches to make them truly effective in identifying the most viable drug candidates to pursue.

The High Cost of Drug Development

The cost of bringing new drugs to market is a significant problem in the pharmaceutical industry, with studies estimating R&D cost for a new drug ranging from $314 million to $4.46 billion. Cost estimates vary widely depending on the therapeutic area (for example, nervous system agents vs. antineoplastic agents), company size and experience, the type of drug (such as orphan drugs or first-in-class drugs), and modeling assumptions. These estimates are just that — estimates — and the true costs vary widely depending on the specific drug being developed and which company is bringing it to market. Multiple factors contribute to the high costs of drug development. One serious issue is that approximately 90% of drugs fail after the start of clinical development, which significantly increases overall costs. Clinical development accounts for the lion’s share of R&D costs at 50% to 58% per new drug. Alongside R&D is the significant cost of capital, which ranges from 33% to 51% of the total costs per new drug. Process development and manufacturing contribute about 13% to 17% of the R&D budget from pre-clinical trials to approval. And, of course, government policies and regulations affect development costs, though that percentage is more difficult to pinpoint. Pipeline analysis platforms have the potential to bring down these costs by more comprehensively analyzing all of the data available, not just for their pipeline but everything in market and in development, a task that is all but impossible for human analysts to do manually. There are large and growing datasets of corporate presentations to scientific publications, as well as conference posters or slides and press releases, that are difficult for individual teams to effectively incorporate into their decision-making processes but have the potential to increase the success rate for bringing drugs to market. By streamlining research processes and improving an internal team’s ability to access and understand relevant data, pipeline analysis platforms can lead to significant time and cost savings for pharmaceutical companies.

Creating a Pipeline Analysis Platform

Technology can address many of the existing challenges in pipeline analysis by creating a software solution that helps organizations aggregate, manage, and analyze data related to drug development projects across various stages. In order to meet the needs of the pharmaceutical industry, a pipeline analysis platform for drug development must include:

• Data Collection and Visualization: collecting data from various sources, such as clinical trials, regulatory databases, and company pipelines, and creating visual dashboards and reports on drug pipelines.
• Aggregated drug information: displaying information about drugs from discovery through clinical trials to launch, including details on indications, mechanisms of action, the status of clinical trials, and so on.
• Tracking and updates: providing updates on drug development progress, clinical trial results, and regulatory decisions to enable monitoring of competitor pipelines and industry trends.
• Analytics and forecasting: offering analytics on the likelihood of a drug being approved as well as market forecasts and analysis of the drug’s commercial potential.
• Customization and filtering: allowing users to filter data by therapeutic areas, companies, development stages, and so on, as well as creating customized views and reports based on user needs.
• Convert unstructured data into structured data: bringing in PDFs, presentations, and Word documents and converting them into structured data enables a more complete analysis of information.
• Integration with other data: linking pipeline data with patent information, sales forecasts, epidemiology data, and more to provide a holistic view of the drug landscape.

To truly help pharma companies, a pipeline analysis platform should also integrate commentary and insights from industry analysts to help in-house teams interpret key industry events and trends. With these capabilities in place, a pipeline analysis platform could help support decision-making for R&D strategy, business development, and investment by enabling organizations to assess opportunities, analyze risk, and gather and analyze competitive intelligence more quickly and easily than ever before.

A Pipeline Platform with an AI Analyst

Although a pipeline analysis platform is vital to making data-driven decisions quickly, the sheer volume of data generated today can still overwhelm users, who may find it difficult to extract actionable insights. To maximize the capabilities of a pipeline analysis platform, pharmaceutical companies need artificial intelligence to clean, process, and analyze vast amounts of data quickly and accurately. Here are four important capabilities of an AI-powered pipeline analysis platform that enable biopharma companies to make strategic decisions involving disease areas or drugs:

• Ingestion: upload all the available data for a disease area, commercial drugs, development programs, and failed studies. Include public, private, unstructured (such as a PDF or Word document), or structured (for example, tables and databases) data to ensure analysis encompasses scientific articles, FDA filings, data rooms, intellectual property, and more.
• Data Preparation: validate, transform, and enrich the ingested data to ensure data is standardized, organized, accurate, semantically mapped, and able to be queried. This cleaned data can be mapped and securely stored in a private data store for each organization’s unique analysis needs.
• Customized Query Creation: define analysis objectives and priorities, relevant drug attributes (including efficacy endpoints and results, safety, dosing regimens, mechanism of action, and more), inclusion and exclusion criteria, and other important inputs that form each query to extract the critical information required to generate a comprehensive and useful report.
• Analysis: use custom-built AI models that use information derived from custom queries to generate reports that include customized analysis, risk assessment, scorecards, recommendations, and other insights. These reports enable each organization to make strategic decisions about disease areas to pursue, specific drugs to fill pipeline gaps, and the likelihood of success for drugs in the pipeline. Ensure that outputs from the pipeline analysis platform are traceable to the original source data, enabling each organization to understand and explain exactly how each insight was derived.

With an AI Analyst embedded into the pipeline analysis platform, making informed decisions about the pipeline becomes far simpler.

AI-enabled pharma experts

Any AI analysis is only as good as the expertise behind it. Vibe Bio was founded and built by experienced pharmaceutical experts who work side-by-side with each organization to understand its objectives, verify input data, and track the analysis through Vibe Bio’s pipeline analysis platform, VibeOne. Pharmaceutical companies can trust the output from VibeOne because Vibe Bio’s experts verify all the analyses and reports generated and present the results to internal teams, ensuring that they come away with actionable insights from any conclusions and recommendations in the pipeline analysis platform. Drug development is rapidly advancing, accelerated both by necessity (to address rare and emerging diseases) and by new technologies and discoveries. This rapid evolution is compelling pharmaceutical companies and investors to make decisions faster about which disease or drug to pursue next. Driven to reduce development costs and timelines and quickly react to external advancements, teams of human experts cannot move quickly enough to stay competitive without the aid of AI analysis. VibeOne applies artificial intelligence and large language models to huge volumes of unstructured data, allowing organizations to identify new indications, choose high-potential drugs for development, eliminate human biases, and allocate capital intelligently to deliver new cures, faster.