Sanofi's big data alliance to predict diabetes medicine adherence

Sanofi has unveiled a new alliance with Duke University and the Massachusetts General Hospital to mine big data to try to predict which patients will adhere to their medication – and which won't.

Alliances based on exploiting 'real world' healthcare data are all the rage in pharma at the moment, and Sanofi's move is aimed at helping it build a more holistic 'beyond the pill' offering to patients and healthcare systems.

Poor compliance with medicines in people with type 2 diabetes is a massive problem, with between 25% and 50% not taking their medicines properly, or not at all.

Now Sanofi's US division has teamed up with Duke Clinical Research Institute (DCRI) and, separately, with the Center for Assessment Technology and Continuous Health (CATCH) at Massachusetts General Hospital to create new tools to help understand and predict which patients might have adherence problems.

This big data crunching is called 'predictive analytics'.

"The DCRI's collaboration with Sanofi has the potential to transform chronic disease population management by analysing how predictive analytics might forecast medication adherence and result in more personalised patient adherence programmes," said Michael Pencina, PhD, director of biostatistics at the DCRI.

The alliance aims to improve patient health outcomes, drug development, clinical trial design and quality of care. CATCH and the DCRI say they are using novel machine learning methods to extract meaningful patient insights, based on large scale use of anonymised individual patient level data.

Patient privacy is always an issue in these studies, and the partners say their work will adhere to Sanofi's own strict privacy standards.

They say the insights gleaned from the analytics can help doctors tailor treatment strategies to each patient based on their patterns of behaviour.

The researchers will focus on using non-traditional data sources, including whether or not patients fill their prescription, socio-geographic and behavioral patterns. For example, if a community in one geographic area shows a higher than average level of low medicines adherence, this could flag the need for greater patient engagement.

This data can then help to better predict patient-specific drug adherence, guide future clinical trial designs and ultimately improve outcomes for patients.

‎Peter Juhn, Vice President, Sanofi Global Diabetes Integrated Care says the initiative will help build 'ever more personalised engagement programmes', practical tools and services for people with diabetes.

The initiative has been launched as Sanofi struggles with falling revenues from its diabetes business, which have been hit particularly hard in the US.

Diabetes drug sales account for about 20% of the company's revenue, but fell 13% to 1.9 billion euros in the fourth quarter of 2015, hurt by lower sales of its top-selling drug, insulin analogue Lantus.

The product lost patent protection last year, with Lilly launching its biosimilar Basaglar in several European markets.

The biosimilar won't be launched in the US until December this year, but Sanofi is being forced to offer larger discounts to the government, insurers and health-care providers across its diabetes portfolio.

Meanwhile, rivals Novo Nordisk continue to win a greater share of the global diabetes treatment market. It has also launched its own initiatives to exploit big data and technology, including a collaboration with IBM Watson Healthcare to create a 'virtual doctor'.