For players in the biopharmaceutical space, it is becoming increasingly clear that advanced analytics can be of enormous assistance in solving many of the unique challenges the industry faces. To understand the extent of the impact that advanced analytics can make, it’s first necessary to examine how healthcare in the US has undergone a major transformation over the past decade.
First, there’s the presence of managed care. It puts pressure on pharmaceutical companies to provide stronger evidence of efficacy and safety, reduce costs of drug development and healthcare in general, and provide personalized care by targeting patient groups that are most likely to benefit from treatments and least likely to suffer adverse events.
That said, managed care is only a part of the transformation of healthcare. We’re also seeing a new breed of patients. These patients have access to far more information than they used to, as well as access to their own healthcare data, and are now taking an active role in researching treatment options and in monitoring their own diseases and treatments.
Another trend we’re seeing in biopharma is the expansion of data into big data, even more so than in most other industries. New sources of data in biopharma include:
- Patient-level data collected through electronic health/medical records
- Social media information capturing customer feedback, patient disease monitoring, device assistance, etc.
- Medical devices and sensors providing real-time data on disease progression
- Pharmacy records indicating both what was prescribed and what was actually filled by the patient, leading to more accurate views on patient adherence and adverse events
- Genomic sequencing data
- Medical imaging data
- Epidemiological datasets like those provided by WHO and NIH following the spread of diseases
And this is all in addition to the plethora of data being collected in clinical trials and outcomes research. That’s a lot to dig through to get usable insights, which is why advanced analytics can make such a large difference here.
Unique Challenges Faced by Biopharma
As regulators, payers, and patients all demand more personalized care in a data-rich environment, they are creating the perfect conditions for machine learning and AI use cases. Of course, advanced analytics aren’t new to biopharma and they are certainly on the rise as data scientists look to machine learning to address a number of questions. However, in general, biopharma as an industry faces unique challenges that have inhibited the application of advanced analytics. Whereas other industries have jumped on the big data bandwagon, biopharma, with its ‘massive data,’ has lagged behind a bit.
One reason for this is that even with all the data available in the industry, it’s actually very difficult to get your hands on it. Privacy policies successfully protect patients’ confidentiality, but by doing so, they also make it nearly impossible to access the patient-level data that’s required for many analyses. And who do you go to in order to gain access? Data ownership is often unclear or proprietary. Even patients don’t own their own EMR/EHR data. On top of that, once data is accessed, it is often disparate, unstructured, and/or incomplete. For example, the FDA reports that only 1 percent of adverse effects from treatments are reported.
On the other hand, logistics may be the barrier. Genomic sequencing and medical imaging data may make up the vast majority of big data in biopharma, but before even considering how to tackle mining and analyzing it, practical matters of data storage and management need to be addressed. For example, processing of a full genome for a single patient requires 500GB.
In addition to data quality issues and access to data, the biopharma marketplace is far more multifaceted than most, adding complexity to many advanced analytics use cases. Rather than simply a buyer and seller, there’s a dynamic market of payers, regulators, patients, and physicians. The actions of each one will impact the choices of the others. Understanding and mapping this marketplace is challenging, and ignoring one player may lead to misguided insights.
Where Biopharma Can Benefit from Advanced Analytics
Despite all the challenges mentioned, there are real gains to be made and serious healthcare problems to be solved by utilizing machine learning and other advanced analytics methodologies in the biopharma space. Plus, the addition of machine learning algorithms into your analytics mix establishes a cycle of continuous improvement that compares predictions against actuals as you go, which can help you refine and focus your data over time.
Issues of data storage and management are often considered within the scope of the advanced analytic effort and can be tackled as part of the overall solution. Tools such as Spark are available to cope with the sheer size of big data and can process 10x faster than Hadoop. Before engaging directly with advanced analytics, organizations often take an assessment of their overall data management, governance, and architecture to create a solution capable of supporting such data and analytics.
Biopharma Use Case Examples
Consider the following use cases that span across research and development, pharmacovigilance, health economics, real-world outcomes, and market research. The pursuit of these and other advanced analytics solutions can make major impacts on efforts to reduce healthcare costs, reduce adverse events, reduce the time to bring drugs to market, and provide personalized care.
Improve the process for selecting candidate molecules for development using data mining methodologies to detect molecules with the right combination of attributes.
Select patients to be included in clinical trials by integrating multiple data sources such as social media data, pharmacy data, and electronic medical records. In doing so, isolate the patients most likely to benefit and thus reduce costs and timelines of trials.
Monitor patients in clinical trials in real time to detect causes for concern earlier and take action immediately, again reducing costs, adverse events, and delayed timelines.
Use simulation designs to detect safety issues and early fails prior to engaging in human trials.
Use historical data and lookalike analysis to forecast clinical trial timelines.
Use segmentation to optimize your sales force by targeting the right physicians who are willing to prescribe and use a given treatment, and who see the appropriate patient population. This is accomplished by reaching your target audience with a compelling message that’s delivered via their preferred marketing channel.
Predict adverse events before they occur by mining for potential drug interactions and patient types not likely to respond to treatment.
Mine patient data for hidden and rare drug interactions associated with harmful outcomes.
Track and predict real world outcomes more accurately by incorporating multiple sources such as pharmacy data, social media, text analytics, and EMR data.
Mine historical patient-level data to uncover patterns associated with later disease or behavioral problems (e.g., drug dependence).
Predict trends in disease prevalence over time and across geodemographic regions.
The bottom line is that the data might be complicated, but the results of advanced analytics in biopharma are well worth the effort of investigating. Each biopharma company that uses advanced analytics is not only taking a significant step toward maximizing the efficacy of its respective organization, but also toward pursuing the commonly-shared goal of improving overall healthcare quality and outcomes.
Ironside was founded in 1999 as an enterprise data and analytics solution provider and system integrator. Our clients hire us to acquire, enrich and measure their data so they can make smarter, better decisions about their business. No matter your industry or specific business challenges, Ironside has the experience, perspective and agility to help transform your analytic environment.