Number crunching: the search for potential therapies

R&D in the pharmaceutical industry is having a tough time. As blockbuster drugs fall foul of the so-called ‘patent cliff ’, and the cost of healthcare continues to rise, many big pharma companies, including Merck, Novartis and AstraZeneca, have resorted to laying off R&D staff to cut costs.

Pushing pipeline drugs through to the approval stage as quickly as possible is as crucial as ever, but the business of bringing new therapies to market from discovery to bathroom cabinets is an expensive and risky endeavour.

While it’s difficult to calculate the full cost of bringing a new drug to market, Forbes estimates that given the low success rate of drug candidates making it to market (95 percent of candidates fail to show their safety and efficacy, it says) on average, pharmaceutical companies will spend about $5 billion on each successful medicine.

Big data

Harnessing what is known as ‘big’ data could be the answer. Business analytics firm the SAS Institute defines big data as the exponential growth and availability of data, which can be both structured or unstructured. It’s an allencompassing term that includes information generated digitally as well as that from traditional sources.

Recently called upon by retail and marketing firms looking to better understand their customers, big data can allow pharmaceutical companies to identify potential drug candidates and develop them into effective medicines in a more efficient way.

One of big data’s advantages is its abundance. Vast amounts of information are generated by pharmaceutical companies in the R&D process and clinical studies. This information, catalogued and analysed, has the potential to improve the efficiency of research and clinical drug trials, which can bring therapies to market far quicker.

In a process it calls de-risking, California-based technology company NuMedii ‘translates’ its big data technology to create drugs with a higher probability of therapeutic success.

By working with existing drugs, the drug development timeline from discovery to approval can be shortened from 15 years to seven or eight years, says Gini Deshpande, chief executive and founder of NuMedii.

NuMedii analyses data from raw human, biological, pharmacological and clinical data points to find drug candidates and biomarkers that predict their efficacy for treating diseases.

Deshpande says that some existing drugs have the potential to have applications in different disease models that NuMedii may identify. Big data gives NuMedii a window on how the drugs are likely to behave in a human before clinical trials for the new indication have even begun.

“It’s not trial and error—we’re in a better position to begin with,” she says. “We’re building technology to increase the likelihood that the drugs will work in relevant populations. It’s sometimes hard to justify investment in a particular drug, unless the drug shows activity in small patient studies.”

Deshpande says that NuMedii can also optimise the formulation of a drug and how to administer it after it has been identified.

According to the McKinsey Global Institute, the business and economics research arm of management consulting firm McKinsey & Company, using big data strategies and analytics creatively to inform decisions could generate up to $100 billion in value a year in the US healthcare system, and reduce expenditure in the system by about 8 percent.

“It’s not trial and error—we’re in a better position to begin with.”

Some countries are already embracing the potential of big data analytics. The UK’s National Institute for Health and Clinical Excellence uses large clinical datasets to investigate the clinical and cost-effectiveness of new drugs, while the Italian Medicines Agency collects and analyses clinical data on expensive new drugs as part of a national programme focused on cost-effectiveness.

NuMedii uses publicly available data to discover connections between drugs and diseases. The technology was originally developed in Atul Butte’s lab at Stanford University in the US, and has been exclusively licensed to the company.

“We face a challenge with the public domain data—it needs to be cleaned up and normalised before we can use it,” Deshpande says.

The big data comes in information silos, which can be difficult to work with. Deshpande says that one of the biggest challenges for companies in the big data space is normalising and integrating the data so it may be easily used. Stanford has developed some methods to clean up and normalise data.

The company’s big data technology is made of billions of comprehensive disease, pharmacological and clinical data points. As well as partnering with specialty pharmaceutical firms, NuMedii works with companies that may hold patient data, such as medical centres.

It’s a work in progress: the company uses the data by pulling it on to the platform alongside information it already has, creating a more proprietary platform, Deshpande says.

“The platform is dynamic, enabling additional data types to be incorporated as they become available.”

The platform lends itself well to possibilities for expansion. On November 4, the company announced a strategic initiative with the IP & Science business of Thomson Reuters in which the latter’s databases in combination with NuMedii’s own proprietary data will discover FDA-approved drugs as well as discontinued development compounds that can be repurposed.

NuMedii also works with specialty pharmaceutical company Aptalis Pharma to discover and develop new methods of treatment for some gastrointestinal disorders and cystic fibrosis.

Finding the evidence

Deshpande says that pharmaceutical companies are not necessarily looking for additional tools but for drugs that work in patient populations: “Pharmaceutical companies want evidence that the drug will benefit the patient population—it’s about finding a proof of concept.”

In addition, removing risky drugs from the pipeline at an earlier stage will allow for more clinical trials to go ahead.

Licensing

NuMedii owns an IP portfolio of method of use and formulation patents that it licenses to its pharmaceutical partners. The database initially licensed to NuMedii from Stanford is a combination of patents and trade secrets.

“We’re currently working with specialty pharmaceutical companies, but will welcome partnerships from traditional pharma looking to revitalise their pipelines,” Deshpande says.

From a commercial perspective, the NuMedii’s big data technology creates the possibility of capturing new IP, as the company builds new method of use patents around drugs that have already been approved for one indication.

“We’re looking a new uses for things that have not been looked at before, things that could be commercially viable,” Deshpande says. “We can apply technology to drugs that have been shelved, or did not show efficacy.”

Researchers have suggested that the use of big data in the pharmaceutical industry will lower the cost of healthcare while patients enjoy better access to information about the treatments they use.

However, whether this will work all depends on how easily information may be accessed, and as some countries have regulations that protect personal medical data, it may become necessary for policy makers to re-evaluate the laws to allow for a more open world of data, while preserving individual patients’ privacy.

The success of companies such as NuMedii streamlining the drug development process may eventually become a vital factor in making it a reality.