BLOG

Why Data Science is the Key to Improving Medical Research Processes

Jo Ann Stadtmueller

July 2021

Blog

Leveraging Data Science Insights to Drive Research Innovation

Over the past ten years, the increase in data collection and computing power has transformed the way that industries operate and provide value to their clients and constituents. The pharmaceutical industry is no exception. Increasingly, drug companies are leveraging data science insights to drive innovations in research and development, especially in improving existing research processes. In fact, McKinsey identified this back in 2017 as the...

“$100 Billion opportunity”

This integration of data is unlocking more efficient ways of identifying the most promising compounds and molecules to use for drug development. With Artificial Intelligence (AI) and Machine Learning (ML) based tools, data scientists can use predictive modeling to accelerate drug discovery, optimize clinical trial design, and target specific patient populations for new drugs. Here’s why data science is key to improving medical research processes:

Data Science Accelerates Drug Discovery

Only about 12 percent of drugs entering clinical trials ultimately gain FDA approval. Recent studies also found that the average R&D cost of a new drug ranges from $1 billion to more than $2 billion. Optimizing these processes without sacrificing the quality of the research is the holy grail to safer medicines that are accessible to the public on a shorter timeline.

During drug discovery, pharmaceutical companies spend millions of dollars screening compounds to test in pre-clinical trials. Until recently, this process could take a decade or longer due to incorporating many manual processes that were time-consuming.

However, using data science and machine learning algorithms can drastically shorten the time that researchers and scientists need in order to achieve results.

Data analytics can be helpful during all parts of drug discovery, from the initial screening of drug compounds to forecasting the drug’s success rate based on patients’ biological factors. Text mining algorithms can sift through past clinical trial data for drugs with similar molecules and compounds or medications created to treat the same illness. Predictive modeling can run through thousands of simulated trials to identify the ones that are most likely to produce the best results.

Thanks to increased computing power, researchers can generate new insights based on enormous datasets of drugs and assets undergoing preclinical testing using data science. These insights allow organizations to prioritize which experiments to run and help them understand the potential impact of a new drug. In addition to enhancing drug discovery, data analytics is becoming crucial to avoiding adverse outcomes and potential disasters from patient risk factors.

Data Science Enables Better Clinical Trial Planning

Drug companies must file an Investigation New Drug (IND) application with the FDA before any clinical trials begin. This application includes vital information, like the candidate drug’s molecular structure, the results of pre-clinical work, predicting how the experimental drug works in the body, and listing any potential side effects. An IND also needs to provide a clinical trial plan that describes how, where and by whom the studies will be conducted.

The IND application process can be complex, especially when data is collected at multiple locations that involve many investigators. However, data science initiatives can help organizations break down data silos and leverage that data to design more optimized clinical trials. For example, data scientists can use an AI software solution to analyze patient profiles along with their medical histories to pick patients who will respond best to a drug being tested. The result is valuable time saved when trying to find suitable patients for clinical trials.

Through predictive analytics, pharmaceutical companies can use their past trials or a database of trial data from other companies to find and establish best practices for upcoming clinical trials. Having information about procedures, clinical trial operations, and the trial’s relative success rate will help researchers plan future trials with ideal conditions and avoid past mistakes.

Craig Lipset,
former head of clinical innovation at Pfizer

Modern data science… will accelerate the creation of smarter studies, with fewer protocol amendments and greater confidence that trials will more efficiently understand the efficacy and safety of new medicines.

Data Science Targets Specific Patient Populations

With any disease or illness, different patients will respond differently to treatments for various reasons. But regardless of the condition, there are massive amounts of patient data being collected. Electronic medical records, medical sensor data, and genomic sequencing data are all becoming widely available. Combining all the data from these different sources is greatly simplified by modern data analysis software. Novartis developed the “Nerve Live” program, which greatly accelerated their researchers’ access to valuable clinical databases and also encouraged collaboration across their departments. While they’re still evaluating the impact of this transformational system, Novartis has already seen faster enrollment periods, optimized cost planning, and more efficient resource allocation.

Trained data scientists can use the latest data analytic tools to spot trends and patterns that enable drug companies to create more targeted medications for patients with common features. Giant pharmaceutical corporations, such as Pfizer, are combining data from a patient’s sequenced genome, clinical trials, and electronic medical records to find more treatment opportunities for specific patient populations.

Taking an analytical approach to testing compounds allows companies to identify a subset of patients with a specific gene mutation who lack treatment options. Cancer patients with specific gene mutations, for example, are one subset of patients who have benefited from this targeted approach to drug development and research.

Victor Lobanov,
VP, informatics solution development at Covance

Starting with protocol design, we can better utilize observational, epidemiological, and safety data to identify the right patient population, inclusion and exclusion criteria, and sample size.

Leveraging Data Science Requires Tools and Training

Across the numerous organizations that are incorporating these data practices into their operations, the ones that have seen the most success have emphasized data literacy and data governance across teams and departments. Typical research can be siloed and stored in disparate Excel spreadsheets, which leads to duplicated efforts and more time expended. Building a database that is accessible to researchers (provided that the appropriate HIPAA and privacy regulations are in place) is the foundation to a strong analytics approach.

It’s important to note that tools can only do so much without the proper training and guidelines. No matter what platforms or databases a company builds, they will never see the real value from them without ensuring that their staff is trained on best practices for data analytics and data science as it pertains to pharmaceutical research. While SAS may have been the preferred tool for many, Burtch Works’ recent study has demonstrated a massive increase in preference for Python (40%) and R programming (34%) over SAS (25%). This trend has been consistent over the past five years. The open source nature of these languages is compelling due to their no-cost implementation and robust community of millions of users, academics, and researchers.

Given these insights, data science training can be one of the best decisions a drug company can make for its long-term success. By investing in data science initiatives, organizations can shorten the biopharmaceutical research and development process, which typically requires billions of dollars in investment and many years of testing. From drug discovery to pre-clinical tests to clinical trials, data science can speed up the research and development of new medicines.

Conclusion

From a quicker drug discovery process to better optimized clinical trials, data science is driving industry-wide innovations to medical research. Thanks to accurate predictive modeling and machine learning algorithms, researchers can efficiently develop new drugs under time pressure and identify new uses for existing drugs. However, these benefits can only be leveraged with the appropriate data infrastructure, as well as professional development that focuses on best practices of data collection, storage, and visualization.

Training can prepare your organization to harness critical data, creating substantial competitive advantages during the research process and beyond.

Toward a Healthcare Data Revolution

Whitepaper

With more data science applications in healthcare around the globe, it is clear the industry's transformation has already begun. However, achieving the data maturity required to leverage these capabilities effectively does not come without significant infrastructure, culture, and education challenges.

We'll explore how to start a data science movement that impacts

Hospital Operations
Personalized Medicine
Patient Self-Care
Public Health

Download and View Full Report

Subscribe to our newsletter