Turning Medical Data Into Information
“In God we trust. All others must bring data.”– W. Edwards Deming
On a daily basis, this world is full of events and happenings which are seemingly uncorrelated to each other. Everything around us is changing and this change is so rampant that we can say the only constant thing in this world is ‘change’ itself. At first sight, these changes appear totally haphazard and unpredictable but a closer look and analysis can identify some patterns, trends or rules which bring about these changes. For example, census of population has been a very effective method used by governments to study the change in demographics and to have an understanding of the characteristics of a population before devising policies.
Interestingly, in this age of digital information, one does not need to have a huge government machinery to conduct such data acquisitions and analyses. Everything that we do today is somehow contributing towards generating a pool of data which might seem useless to us but it can be very precious to some groups who have understood the power of data and information.
Although it sounds a bit scary, our activities on internet, smart phones and computers generate data which has the power of defining us by giving a glimpse of psychology behind our choices. The study of this data to find out the inherent patterns and trends is called ‘data mining’ and it has been used effectively in a range of fields such as business, finance, marketing and research.
With the efforts about digitising medical health records being done already, it is not a surprise that big pools of medical data are being generated. How well we read this data or how well we ‘mine’ this data for precious insights wields the power to change the landscape of whole medical healthcare industry. One of the most natural outcome of this data mining is prediction based on existing patterns. Given the family medical history and genetics of a person, it might be possible to predict the threat of a disease based on similar cases. Given a certain portion of demographic living in a particular part of country, it can be predicted that what kind of infections, viruses or diseases they are prone to. World Health Organisation (WHO) is already using such data acquisitions from developing countries to monitor the outbreak of various diseases and viruses.
One of the factors that is often cited in medical surveys is the cost of readmission of patients. Such costs can be greatly reduced if the patient data is being monitored by health care provider remotely and necessary actions are advised before the situation becomes worse. Having a huge database of general symptoms for mild wellness and corresponding medical advice available to clients can significantly cut down one on one visits to clinics and hospitals. Medical insurance companies can use this medical information to devise better insurance policies. Companies who provide medical insurance for their employees can send notifications to them if they are lagging behind on some key health indices. It can also prevent frauds and detect any embezzlement in insurance policies. Data mining can also help practitioners by providing them with a comparison of different treatments applied to a disease and their relative effectiveness.
The above prospects sound very enticing and some work has already been done in this regard yet data mining has not realised its true potential in digital healthcare realm. One of the hindrances is the unstructured nature of medical data. In order to study data, it is important to have it in a standard form so that key attributes and values can be identified. Collaborative efforts need to be done to develop standards and protocols for medical data gathering and sharing. Another very important factor is what we do with the information provided by data mining. All these deep insights and preemptive guidelines would be rendered useless if there is no proper infrastructure to make sure that they are acted upon. Development of analytics tools is also a need of time as it would provide a certain level of encryption to physicians and patients by showing data in a more presentable tables and graphs. Last but not least, there would be genuine concerns over the privacy and these should be handled by developing strict confidentiality measures and ensuring fair use of data.