Can risk models tackle Scotland’s drugs crisis? 

Innovative ideas are required to lower grim drug death rate

Drug-related deaths in Scotland are continuing to spiral out of control. The latest government statistics make for grim reading: Scotland has by far the worst drug death rate in Europe and around three-and-a-half-times that of England and Wales.

The government has publicly acknowledged its failings in this area: the First Minister said that the figures were “completely unacceptable” and “indefensible” and admitted that there was “much to do to sort this out”.

The question is, what exactly can or should be done? Drug addiction, drug crime and drug-related deaths are difficult, interrelated issues with no obvious solution. But we can still be open to new ideas and solutions for mitigating the problem.

Drug deaths rarely arise out of the blue. Habitual users often have complex needs and interact with a multitude of public services – including health and social care, education, benefits, and law enforcement.

The problem is that these interactions are often only known to the individual agencies that deal with them, which makes it difficult to get an overall view of whether someone is in danger.

This is an area where government could learn from the financial services sector. The 2008-09 financial crisis, and its aftermath, compelled banks and insurers to invest in sophisticated statistical analysis to better assess risks, make predictions, and recommend actions, very often at the customer-specific level. They have utilised technology that is extremely well understood and battle-tested to do this.

But how can we apply citizen-specific risk modelling to public sector problems such as drug deaths? A recent collaboration between SAS and the Spanish government, the VioGen project, shows the way forward.

Spain’s VioGen system brings together relevant information from various public agencies involved
in cases of gender violence and uses risk modelling techniques to calculate the risk that an individual will suffer further violence in the future.

When the risk is determined to be high, the system automatically alerts the relevant agencies so that they can intervene quickly to protect victims.

The results of VioGen have been startling. For individuals in the high-risk category, the rate of repetitive violence has reduced by 50 per cent.

Scotland could apply the same techniques to help mitigate its drug death crisis, creating a real-time assessment and alert system that gives health and social care agencies the insight they need to intervene sooner, and potentially save hundreds of lives each year.

Drug deaths cut deep into the heart of Scottish society, and it’s an issue that seems to be getting worse, not better. But with the right risk models – combined with a data governance framework that ensures privacy and fairness – I’m confident we can get Scotland moving in the right direction.

Hugo D’ulisse is Director Scotland at SAS UK

Partner content in association with SAS