Medical big data – ‘the totality of data related to patient healthcare and wellbeing’[1] has grown exponentially in recent years as new technologies enable the capture of patient data at an unprecedented volume and rate. In medicine, analysis of big data has helped cut the crippling cost of healthcare[2] and improved patients’ quality of life.
Despite these benefits the reality is that big data is not always good data - and without good governance it can conversely lead to inefficient spend, and even poor medical decisions. So how can healthcare professionals navigate the data minefield and be sure their data is robust enough to help improve healthcare in the real world?
Sources of Medical Big Data
Sources for obtaining medical data are increasing daily. In addition to traditional collection methods, such as patient and doctor-reported data, insurance claims, prescriptions and clinical trials, huge volumes of data are now collected through technology. It is increasingly common for consumers to own a personal health tracker such as a FitBit or Apple Watch for example and according to Juniper Research, activity trackers are going to be used by more than 1 in 5 Americans by 2021. Web searches, social media and mobile devices all generate real time, personal information and mean healthcare data is now estimated to make up 30% of the world’s daily data production.[3]
Benefits of Big Data
There are many benefits of data on this size and scale. In public health, big data collected in real time, such as that through wearable technology, enables healthcare professionals to track a patients’ symptoms over time, giving a clearer picture of their overall health and improving their ability to treat them.
Big Data improves our ability to anticipate and treat illness, supporting the move towards personalised medicine where medical decisions and treatments are tailored to the individual patient, helping to improve outcomes.
In disease surveillance, big data analysis is used to spot and report disease outbreaks. In 2014, HealthMap, a sophisticated online mapping tool run by Boston Children’s Hospital, noticed a mystery fever spreading in Guinea - nine days before the World Health Organisation announced the outbreak of Ebola.
Big data can also accelerate and enhance research and development. It improves clinical trials as it helps to identify and expedite insight into drug efficacy and development into approved, reimbursed medicines, saving both money and time
Challenges
The collection, verification and analysis of data on such a vast scale presents significant challenges.
Data is not static, and rates of change need to be understood, with data updated to ensure veracity.
There is a lack of standardisation with medical data regarding how it is collected, processed and analysed. The lack of a uniform system to capture and measure this data means healthcare data can be of poor quality; inconsistent, unstable and hard-to-access. Disparate databases, which are rarely compatible with each other, prevent easy and effective analysis. Furthermore, the sheer volume of information can lead to instability.
What makes Big Data Good data?
While big data is often characterised by the 3Vs – Volume, Variety and Velocity, the key to good data is Validity; the information must be factually sound and reliable, and Value; the data must be meaningful and facilitate decisions about healthcare in the real world.
The validity of big data becomes critical once healthcare professionals have decided that they wish to use it, or subsets of it, for decision-making. Ensuring validity, that the data is correct and accurate for the intended use, is a prerequisite of applying the data in an operational context. To reflect the importance of this, it is estimated that data scientists spend 60% of their time cleaning and organising data. [4]Pharmaceutical and healthcare companies must therefore adopt good data governance practices to ensure the underlying data they wish to use is valid and of value.
How to ensure high quality data
Unsurprisingly perhaps, technology is providing many of the solutions to the challenge of ensuring the validity and value of data captured - and getting it right first time. Innovations in online products and digital tools mean that companies can avoid outdated collection methods that are inaccessible to others and incompatible with multiple systems.
In Pharmacovigilance, a data-driven process whose success relies on both the quality and breadth of data, key information on medicines and associated adverse event information can now be captured in real time, enabling a faster response and better interventions from pharma companies. Reportum, for example, is a cloud-based, fully secure, multi-platform safety data capture tool which enables the standardised capture of adverse event data at source, which optimises its quality. This gives pharma companies far more accurate data on which to base their assessment and analysis to help produce more refined and effective drugs.
Healthcare companies which are uncompromising in their insistence on ‘good’ big data and which are willing to fully embrace technology’s potential to deliver, analyse and apply it, will be the ones that achieve the greatest successes in the coming decade.
[1] Raghupathi2014
[2] Mckinsey and Company estimates in the future big data will save Americans between $300 billion to $450 billion in reduced healthcare spending annually - ‘The ‘big data’ revolution in healthcare: Accelerating value and innovation ‘
[3] Fortune magazine, April 2018
[4] Forbes, March 2016 - https://www.forbes.com/sites/gilpress/2016/03/23/data-preparation-most-time-consuming-least-enjoyable-data-science-task-survey-says/#775511b36f63
