Powering Drug Commercialisation through Small Data
Over the past decade, several technological advances have taken place in the healthcare sector globally. One of the largest advances is that pharmaceutical companies, insurers and providers have been digitising their research and development data, claim data, and patient records (Cattell, Chilukuri & Levy, 2013). These advances coupled with rising global healthcare costs (Aon Hewitt, 2016) led experts to predict a “big-data” revolution in healthcare. The hope was that companies would use their vast swaths of data to achieve extensive cost savings. However, companies soon realised that big data alone would not provide them with a competitive advantage. There were several insights they needed that could only be derived from small data, and small data in combination with big data.
In order to assess the role of small data in drug development, it’s important to first define small data. For these purposes, small data includes datasets of less than a petabyte. These datasets are used to track specific attributes, and explain events that are happening now. Smart labels on medicine bottles can illustrate the use of small data. Small data from smart labels can be used to determine storage temperatures, the shelf life of the medicine, and whether the seal is broken (Kavis, 2015). Small data can then be combined with other datasets often yielding big data, datasets larger than a petabyte. This data is often used to derive future predictions (Kavis, 2015). In this case, small data from smart labels can be used in combination with other supply chain data to predict and prevent patterns that lead to spoilage of medicine supply.
Some experts believe that small data may be the new big data, citing that 60 – 65% of the top 100 biggest innovations have been based on small data (Lindstrom, 2016). However, for companies to derive new insights and to create and maintain a competitive advantage, there is a critical need for both small and big data. Small data cannot replace big data, nor the reverse. The two can be used powerful in combination. To understand how this applies to drug commercialisation, below is a brief explanation of the drug development process, and how small data and small data in conjunction with big data can be useful in the context of this process.
The Drug Development Lifecycle
According to the U.S. Food and Drug Administration (FDA), the drug development process consists of the following 5 steps (FDA, 2015):
1. Discovery and development phase: This phase involves extensive research in the lab.
2. Preclinical research: Drugs are tested in the lab, often on animals, to explore their safety.
3. Clinical research: Drugs are tested on humans to ensure both safety and effectiveness.
4. FDA Review: The FDA reviews all data from the previous three stages, which informs their decision on whether the drug is approved.
5. FDA Post-Market Safety Monitoring: The FDA continuously monitors drug safety after it is on the market.
Understanding this framework can help us to better understand how small data can be used to improve the process.
The Role of Small Data in Improving Drug Development
The big data revolution has brought much attention to small data hurdles that need to be ad-dressed (Feigenbaum, 2013). One such challenge is the lack of integrated data across the several different divisions within a pharmaceutical company. For example, large pharmaceutical companies have started searching for drug candidates earlier in the development phase. Investing in drugs during these early stages is risky, so it’s important that companies can access integrated data on the competitive landscape, results from clinical trials and other critical data in one place. Currently, it’s quite difficult to access this integrated small data that equates to just a few gigabytes. Having access to this small data would facilitate decision-making and improve a company’s competitive advantage (Feigenbaum, 2013).
Another such challenge is that out-of-pocket prescription costs have increased by 250% in the past 5 years (PwC Health Research Institute, 2013). Due to customers’ increased expenses, their demands have also increased. There is now pressure on pharmaceutical companies to pro-duce an experience for the patient, not just a drug. Experts predict that the companies that can pull insights from customer feedback, a form of small data, will hold a competitive advantage in the market (PwC Health Research Institute, 2013). This illustrates two of many examples in which small data insights can improve the drug development cycle.
Combining Small Data with Big Data to Improve Drug Development
To garner big data insights, small data must often be combined with large and complementary data. Genentech, the biotechnology company, for example, uses clinical trial data – small data – alongside real-time patient data in electronic health records (EHRs) – big data. Clinical trials often exclude certain groups of people who are unhealthy. The combination of these datasets allows companies to ascertain how effective the drug might be for these omitted groups (Copping and Li, 2016). In another example, companies might combine genomic data with real-world (small) data to assess potential unexplored pathways for a particular disease (Berger, Axelsen & Subedi, 2014).
In a fast-changing world, it’s important to clearly define what small data is first. In doing so, it is clear that small data will not replace big data. Rather, each holds a unique importance in the drug development process in ensuring a company derives the latest insights and remains competitive. The power of small data is to help bring light to customer insights, to understand the here and now, and to help the many departments of a company make better decisions. Small data can also be combined with big data to help pharmaceutical companies understand clinical outcomes in populations they haven’t yet studied, and to make future predictions that can save the company and patients health and monetary costs. A company must understand this and hire people with up-to-date skills in order to maximise the potential of small data in their company.