Living in the world of data science and analytics, I often find myself being asked by customers and colleagues “what are your predictions for the pharma/biotech industry and what steps should we take to truly make AI/ML operational from a multichannel marketing and real-world evidence standpoint?”
It’s a BIG question, and there are definite trends emerging and actions companies can take to stay ahead of their competition.
Pharma/Biotech’s Changing Landscape
I’ve noticed four key emerging trends in the space. First, manufacturers are moving away from blockbuster drug development and toward personalized medicine. With an ability to leverage genomics, biomarkers and other technologies companies are spending more energy on rare disease/specialty medicines. Second, is increased pricing pressures. Companies are finding themselves lowering drug prices to either gain access/use or protect share and margins. Third, increased pressure on patents are challenging the lifetime value of pharma/biotech therapies. The advent of biosimilars/generics and innovative legal strategies are resulting in increased competition and pressure on market share and price points. Fourth, access to healthcare professionals, and attention spans, continue to decrease.
These four key trends demand we create efficiencies, not only in the way drugs are being developed, but how they are marketed and managed throughout their lifecycle.
Burgeoning Demand for Healthcare Analytics
Pharma/biotech companies and the analytics providers that serve them are investing huge sums of money to try and get ahead of the curve. According to a Google Markets & Markets report, healthcare analytics is expected to reach $50B by 2024 (at a CAGR of 28%). If you’re thinking that hospitals and healthcare systems are driving the majority of growth, you’d be wrong. The pharma/biotech industry is expected to drive 66% of the total business in 2024.
3 Critical Ingredients Healthcare Analytics Needs to Succeed
To bake a great cake, you need the perfect mix of all the right ingredients. If you’re missing one, you can adjust the oven temperature, adjust the baking time, but no matter what you do, the cake simply won’t be great. To truly deliver superior healthcare analytics, you need a combination of three key ingredients:
- Domain Expertise
Healthcare ecosystems today are incredibly complex, with multiple players and interdependencies. Stringent regulatory and legal frameworks require knowledge of laws and processes. HIPAA and privacy laws demand expertise in merging and interpreting data sets.
- High Touch + Relationships/Trust
While the industry is changing and transforming, tenured professionals value long-term deep relationships where trust has been built. Further, traditional organizational silos still exist, requiring high-touch hand-holding to break down silos and operate in a cross-functional way.
- AI/ML Analytical Expertise
Advanced analytical skills and expertise are required to drive powerful real-time insights from data – to make them actionable and deliver results. Powered by a cutting-edge platform, AI/ML can be leveraged to mine large data sets and generate scalability and efficiency.
Selecting the Right Tool/Platform
I’ve seen time and again, the best companies and teams fail because their tools, platforms and partnerships are too complex, inefficient, or inflexible. Many healthcare analytics providers, especially the large ones, create agreements or platforms that operate in a “black box” allowing pharma/biotech companies the ability to visualize only the end result. If they want to test another model or run another simulation, they have to go back to the solutions provider (and ring the cash register). If they need additional analysis or deeper level reporting, they have to go back again and incur additional costs for human resources.
I’m frequently asked by pharma/biotech companies what to look for when evaluating tools and platforms. My advice is to use the following criteria to select state-of-the-art technology:
- User-friendly (no data science or business analytics exerience required)
- No “Body Shop” dependency (platform efficiencies should reduce labor needs)
- No “Black Box” of the data
- Ability to runs millions of models in a parallel environement to find the best model
- Platform validated via thousands of models and iterations across hundreds of brands, therapeutic categories and countries/regions
- Norms and benchmarks for ROI across therapeutic categories
The ideal platform needs to be truly customizable and allow for a multitude of data sets (open claims, closed claims, EMR, online/offline, social, etc.). It needs to continually iterate over the data to improve the accuracy of machine learning, identify, and predict the effectiveness of each initiative on a real-time basis.
Based on our own research, validated through thousands of models and iterations across more than 100 brands, a dozen therapeutic categories and more than 40 countries/regions, when done correctly you can expect to see significant results
- Up to 20% gain in incremental sales from bottom up optimization
- Up to 30% improvement in cost efficiency from profit max optimization
- >50% gain in productivity (cut time for analysis and optimizations in half)
- Identify patients likely to be non-compliant up to 2 months ahead of disruption of therapy
- Proactively find emerging patients in rare disease
Simply put, you can achieve astonishing results.