Applying Big-Data in Public Sector Healthcare

Big Data is one thing that everyone talks about, but only a few know what to do with it? The Narendra Modi led NDA Government has promised to set up an institute of Big data and analytics (BJP Election Manifesto). It is time for us to look at what Big data can do to healthcare.

Build the &ldquo Big Data&rdquo as an &ldquoinfrastructure&rdquo for healthcare stakeholders
Take a holistic prospective to healthcare and wellbeing that enables deeper understanding of administrative, clinical, research, economic and population prospective. Extending into key influencers to individual and population health like lifestyle, vector or organism based information weather information shall be a key to make the volume of data relevant and comprehensive.

 

Be proactive in taking advantage of automation initiatives in public sector
There are several initiatives at state level in automation of care delivery at hospitals, clinics and last mile ASHA/ANM workers. Number of these initiatives is executed through NHM (National Health Mission) and/or World Bank funding. The Blueprint architecture must consider integration and interoperability with the Big Data platform. This approach shall achieve harmonization across initiatives and reduce the cost of building insightful data.

 

Consolidate existing data assets and generate quick wins
Big-Data holds many promises for Healthcare ecosystem.

Various government and non-government agencies collect pertinent data which is used for primary use in their day-to-day business. However the same data can be put to secondary use when corroborated with other related care domains. A typical example of such an opportunity is targeted epidemiology studies by ing relevant cohorts that combines direct health data points with indirect healthcare data points like lifestyle, environment, weather data etc. &ldquoBig Data&rdquo infrastructure can enable epidemiologists in planning and executing epidemiology studies in real life scenarios enabling direct applicability of these study results in driving policies, procedures and interventions for Indian population.

 

Indian healthcare system has a unique opportunity to leverage its maturing Healthcare IT to leverage &ldquoBig Data&rdquo as an integral part of the ecosystem however definite steps shall be taken to align technology infrastructure to business objectives and illustrate quick wins. Consolidation of existing data in healthcare research organizations, public and private hospital clinical data combined with public and private payer can be a good starting point for a Big Data initiative for Indian Healthcare economy.

 

Big Data Analytics at massive scale for Healthcare applications

1) Collection and amalgamation of enormous data varieties for Healthcare, 2) its use in extending clinical and scientific knowledge, and 3) to generate machine learning rule base and knowledge for computational engines in future. A framework needs to be built which can accept data from different sources and makes sense out of the combined data. Highlighted below, a few Big Data Healthcare applications that is the need of the hour at Pan-India level:

Prescription analysis &ndash Analyze all prescriptions at a given time point in India to ascertain which are
the most common symptoms/ medicines being prescribed. Include traditional medicine practitioners in data collection as they outnumber allopathic practitioners. This will help in giving advance notice even before disease epidemics emerge.

Pathology reports analysis &ndash Generate new clinical knowledge thresholds on new normal values, for ex: BMI, Hb, various cell counts, etc. Identify and define new bounds based upon ethnicities,
geographical location and existing health conditions.

Medical records analysis &ndash Generate new hypothesis like disease co-morbidities for functional and
clinical validation. We may use temporal patterns in medical records to discern adverse drug
events from indications. This could be data of the same patients/ families from different time
points and diseases context. As of now, the nature of clinical trials doesn&rsquot allow the detection of all serious side effects and drug interactions before approval.

Easy to understand health records &ndash Patients will soon have access to their own health records over the cloud. There will be a growing need for online facilities that can help patients without medical knowledge to access and receive their records and relevant information in the health records in a generally understandable form.

Patient stratification and data driven medicine &ndash Patient stratification is needed for therapy
planning in hospitals and clinical trials for pharma cos. It will be extremely important to identify the new comparison metrics at both clinical and population level for accurate stratification.

Molecular pathology reports &ndash We can integrate National Cancer Registry Program to include molecular
tests for identifying mutations involved in Cancer. This can be later extended to other disease registries.

Advanced database querying &ndash Can we take a subset of data, say 5-10% and give results in real-time,
this could be the screening test and the more advanced diagnostic test takes 100% of data into account for providing answers in real time.

Crowdsourcing &ndash Different dictionaries may be designed and populated for medicines, dosages, symptoms, etc. so that crowdsourcing of epidemiology and healthcare data can be put easily through mobile phones. This will involve developing applications and interfaces for &ldquocitizen scientists&rdquo for collecting annotations leading to effective learning from crowd-annotated or crowd-augmented datasets.

If we summarize the value of technology in transformation of a health system it has to be &ldquoTransformation from a Reactive to Proactive model&rdquo. With the daunting task of providing quality of care at minimum costs, only a proactive approach will rescue health economies like India to meet surging demand of health in coming years. This applies to the health ecosystem as a whole while business intelligence has successfully provided us with an ability to convert data into information, we need transformational approach to enable valuable insights supporting tactical and strategic decision making for stakeholders in a health economy. Driving valuable insights creates a paradigm shift on how we define healthcare system and extend the health system into a &ldquocitizen centric&rdquo ecosystem that influences individual and population health span directly or indirectly, and it cannot happen without Big Data.