Public health is a vast and complex ecosystem – as the events of the past year have made explicitly clear. Maintaining the health of millions of citizens requires multiple strands of policy, law, education, research – and, increasingly – technology and data.
Indeed, as governments become increasingly able to collect, analyse and harness big data pertaining to the health and wellbeing of their citizens, so they are able to develop more sophisticated and effective approaches to public health. In particular, they can develop so-called precision public health – highly targeted and intelligent approaches to public health, tailored to the needs of different groups and even individuals. But whilst the theory might sound simple, putting it into practice is rather more complex. So how can governments best harness the promise of big data when it comes to the health of their populations?
What is precision public health?
As outlined in this government blog, precision medicine ‘has been defined as a set of prevention and treatment strategies that take individual variability into account’. That is, precision medicine aims to deliver a highly bespoke approach to each individual patient, based on an in-depth understanding of the particular risks, expected reactions and outcomes associated with that patient.
As the same government blog underlines, there are two broad forms of precision public health. One form focuses on organising citizens into groups based on particular factors which are determined to have a substantial impact on health. This has long been an important part of public health – older people, for example, clearly require different healthcare interventions and approaches to children. However, in an era of big data collection and analysis, it is becoming possible to do this categorisation in far more targeted ways. Risk scores for particular illnesses, for example, can be intelligently generated based on multiple different datasets.
The other form of precision medicine – and a more expansive one – focuses on using ‘data and analytical techniques to design and implement interventions that benefit whole populations’. Rather than merely predicting where a particular individual fits in a series of groups, this form of precision medicine considers the population as a whole, balancing myriad datasets to determine which combination of strategies will be optimal for improving health. In a sense, this form of precision public health is more proactive – though both can be hugely beneficial.
How can governments harness the promise of big data in precision public health?
The very term ‘precision public health’ implies a laser-like focus on the individual. Nevertheless, it is important to underline that a starting point for any government undertaking a precision public health programme should be a focus on structural underlying causes of health problems, above and beyond individual risk. Reducing risk to individuals – for example, by encouraging an individual to follow a healthier diet – can be hugely beneficial for that individual. However, the broader societal benefits come from understanding why vast groups of people are tending to follow less healthy diets, and developing policies to tackle those factors en masse. The sugar tax, for example, or healthier school meals, could be thought about in these terms.
By thinking in terms of structural causes, governments can recognise the importance of drawing together vast and disparate datasets when developing precision health initiatives – which is precisely where big data analytics come in. From there, the principles of running an effective programme are the same as for any other big data project. The data in question must be clean, and it must be complete – or as near to complete as possible. Risks of bias must be carefully analysed and mitigated. The data must be properly stored and secured; it must be sense-checked.
Big data analytics have transformed myriad business sectors over recent years, and they have extraordinary potential in the field of health. But whilst precision public health might ultimately be about individualisinghealthcare, there are some general data hygiene points to get right if that potential is to be realised.