Int. Journal of Electrical & Electronics Engg. Vol. 2, Spl. Issue 1 (2015) e-ISSN: 1694-2310 | p-ISSN: 1694-2426

NITTTR, Chandigarh EDIT -2015 81

Dr. Gagandeep Jagdev , Sukhpreet Singh1Dept. of Comp. Science, Punjabi University Guru Kashi College, Damdama Sahib, Bathinda

2M.Phil (Comp. Sc.), Punjabi University, Patiala(PB)1gagans137@yahoo.co.in

Abstract- Big Data can be combined with new technologyto bring about positive conversion in the health caresegment. A technology aimed at making Big Dataanalytics a certainty will act as a key element intransforming the way the health care industry operatestoday. The study and analysis of Big Data can be usedfor tracking and managing population health careeffectively and efficiently. In ten years, eighty percent ofthe work people do in medicine will be replaced bytechnology. And medicine will not look anything likewhat it does today. Healthcare will change enormously asit becomes a data-driven industry. But the magnitude ofthe data, the speed at which it’s growing and the threat itcould pose to individual privacy mean mastering "bigdata" is one of biomedicine's most pressing challenges.Hiding within those mounds of data is knowledge thatcould change the life of a patient, or change the world.This also plays a vital role in delivering preventive care.Health care will change a great deal as it becomes a data-

driven industry. But the size of the data, the speed atwhich it’s growing and the threat it could cause toindividual privacy mean mastering it is one ofbiomedicine's most critical challenges. In this researchpaper we will discuss problems faced by big data,obstacles in using big data in the health industry, how bigData analytics can take health care to a new level byenhancing the overall quality of patient care.

Keywords- Big Data, framework, medical science, EMR,HIS

INTRODUCTION

Big data [1, 2] are rapidly all over the place. Everyoneseems to be collecting, analyzing, and making moneyfrom it. No matter whether we are talking aboutanalyzing zillions of Google search queries to predictflu outbreaks, or zillions of phone records to detectsigns of terrorist activity, or zillions of airline stats tofind the best time to buy plane tickets, big data are onthe case. By combining the power of moderncomputing with the enormous data of the digital era, itpromises to solve virtually any problem like crime,public health, the evolution of grammar, etc.The goal of big data management is to ensure a highlevel of data quality and accessibility for businessintelligence and big data analytics applications.Corporations, government agencies and otherorganizations employ big data management strategies tohelp them contend with fast-growing pools of data,typically involving many terabytes or even petabytes of

information saved in a variety of file formats. Effective bigdata management helps companies locate valuableinformation in large sets of unstructured data and semi-structured data from a variety of sources, including calldetail records, system logs and social media sites.Most big data environments go beyond relationaldatabases and traditional data warehouse platforms toincorporate technologies that are suited to processing andstoring non-transactional forms of data. The increasingfocus on collecting and analyzing big data is shaping newplatforms that combine the traditional data warehouse withbig data systems in a logical data warehousing architecture.As part of the process, it must be decided what data must bekept for compliance reasons, what data can be disposed ofand what data should be kept and analyzed in order toimprove current business processes or provide a businesswith a competitive advantage. This process requirescareful data classification so that ultimately, smaller sets ofdata can be analyzed quickly and productively.

ISSUES RELATED WITH BIG DATACHARACTERISTICS

Data Volume – With the increase in volume, the worth ofdifferent data records will decrease in proportion to age,type, richness, and quantity among other factors.Data Velocity – Data is being generated at tremendousspeed with each minute passing. The velocity at which thisdata is being generated is beyond the handling power oftraditional systems.Data Variety - Mismatched data formats, non-aligned datastructures, and inconsistent data semantics representssignificant challenges that can lead to analytic collapse.Data Value – Often it is witnessed that there is a huge gap inbetween the business leaders and the IT professionals. Themain concern of business leaders is to just add value to theirbusiness and to maximize their profit. On the other hand, ITleaders deal with technicalities of the storage andprocessing.Data Complexity - Data scientists have to link, match,cleanse and transform data across systems coming fromvarious sources. It is also necessary to connect and correlaterelationships, hierarchies and multiple data linkages or datacan quickly spiral out of control [3].Data Veracity - Veracity refers to the messiness ortrustworthiness of the data. With many forms of big dataquality and accuracy are less controllable. [4 , 10].

BIG DATA IN MEDICAL SCIENCE

Application of Big Data in Medical Sciencebrings revolution in managing health

care of humans

Int. Journal of Electrical & Electronics Engg. Vol. 2, Spl. Issue 1 (2015) e-ISSN: 1694-2310 | p-ISSN: 1694-2426

NITTTR, Chandigarh EDIT -2015 82

Big data in health care is different from big data inmarketing and product development because ofregulation, medical ethics, privacy and the diversity ofdata sources. Goals differ as well. Understanding thesedifferences will be important to unlocking its hiddenvalue. Health data volume is expected to growdramatically in the years ahead. Although profit is notand should not be a primary motivator, it is vitallyimportant for healthcare organizations to acquire theavailable tools, infrastructure, and techniques toleverage big data effectively or else risk losingpotentially millions of dollars in revenue and profitsHealth care is one of the top social and economic issuesin many countries, such as the India, the UK, SouthKorea, The United States and even middle-incomecountries. In India, health care sector suffers fromunderfunding and bad governance. No doubt, India hasmade huge improvements since independence; majority(70%) of the effort has been led by the private sector.Still India accounts for 21% of the world’s burden ofdisease. The term big data [1, 2] refers to the collectionof data sets so large and complex that it becomesdifficult to process using readily available databasemanagement tools. In actual, big data refers to thesituation where more and more portions and objects ofeveryday life are available in digital form, like personalprofiles or company profiles, social network and blogpostings, health records etc. through which hugeamount of data gets dynamically produced particularlyon the Internet and on the Web.Unstructured data forms close to 80% of information inthe healthcare industry and is growing exponentially.Getting access to this unstructured data such as outputfrom medical devices, doctor’s notes, lab results,imaging reports, medical correspondence, clinical data,and financial data is an invaluable resource forimproving patient care and increasing efficiency.In the last few years there has been a move towardevidence-based medicine, which involves making useof all clinical data available and factoring that intoclinical and advanced analytics. The outcomes of thismovement include improved ability to detect anddiagnose diseases in their early stages, assigning moreeffective therapies based on a patient’s genetic makeup,and adjusting drug doses to minimize side effects andimprove effectiveness.In recent years the Indian government has increasedspending in the health care industry. The governmentplans to increase it even further by 2.5% of the GDP inthe 12th five year plan. As compared to other emergingeconomies the amount of public funding that Indiainvests in health care is very small. India ranks amongthe last 5 countries with 6% of GDP expenditure onhealth care. Hospital bed density in India has beenstagnating at 0.9 per 1000 population since 2005 andfalls significantly short of WHO laid guidelines of3.511 per 1000 patients’ population. Moreover, there isa huge disproportion in utilization of facilities at thevillage, district and state levels with state level facilitiesremaining the most tensed. India is currently known tohave approximately 600,000 doctors and 1.6 millionnurses. This interprets into one doctor for every 1,800people. The recommended WHO guidelines suggest

that there should be 1 doctor for every 600 people. Thistranslates into a resource gap of approximately 1.4 milliondoctors and 2.8 million nurses. There is also a cleardisproportion in the man power present in the rural andurban areas [7, 8].

ROLE PLAYED BY BIG DATA IN MEDICALSCIENCE

These are some example use cases that illustrate how bigdata is being used in healthcare, helping to increaseefficiency and improve patient care.Personalized Treatment PlanningPersonalized treatment planning is a way to customizetreatment for a patient to continuously monitor the effects ofmedication. The dose can be modified or the medication canbe changed based on how the medication is working for thatparticular individual.Assisted DiagnosisBig data being able to access a broad combination ofknowledge across multiple data sources aids in the accuracyof diagnosing patient conditions. Assisted diagnosis isaccomplished using expert systems that contain detailedknowledge of conditions, symptoms, medications and sideeffects.Fraud DetectionHealthcare organizations need to be able to detect fraudbased on analysis of anomalies in billing data, proceduralbenchmark data or patient records. For example, they cananalyze patient records and billing to detect anomalies suchas a hospital’s over utilization of services in short timeperiods, patients receiving healthcare services from differenthospitals in different locations simultaneously, or identicalprescriptions for the same patient filled in multiplelocations.Monitor Patient Vital SignsHealthcare facilities are looking to provide more proactivecare to their patients by constantly monitoring patient vitalsigns. The data from these various monitors can be used inreal time and send alerts to nurses or care providers so theyknow instantly about changes in a patient’s condition.

Digitization of DataTill date most of the data in the health care industry arestored in the form of hard copy, but the current trend istoward rapid digitization of these large amounts of data.The medical community has accepted big data as a research

tool and beliefs that the society’s enormous amount ofdiverse health information has the potential to help solvesome of medicine’s most troublesome problems. Bydiscovering relations and understanding patterns within thedata, big data analytics has the potential to improve care,save lives and lower costs [5, 11, 13].

TECHNOLOGY USED BY BIG DATA

Big Data needs a framework for running applications onlarge clusters of commodity hardware which produces hugedata and to process it. One such framework is Hadoop.Hadoop includes two main components. First one is HDFS(Hadoop Distributed File System) and the second one isMap/Reduce technology.The process starts with a userrequest to run a MapReduce program and continues until theresults are written back to the HDFS . As MapReduce is an

Int. Journal of Electrical & Electronics Engg. Vol. 2, Spl. Issue 1 (2015) e-ISSN: 1694-2310 | p-ISSN: 1694-2426

83 NITTTR, Chandigarh EDIT-2015

algorithm, it can be written in any programminglanguage.Hadoop map reduce works in three stages:First Stage: mapping: In this stage, a list of elementsis provided to a ‘mapper’ function to get it transferredinto pairs. The mapper function does not modify theinput data, but simply returns a new output list.

Intermediate stages: Shuffling and Sorting: After themapping stage, the program exchanges the intermediateoutputs from the mapping stage to different ‘reducers’.This process is called shuffling.

Final Stage: Reducing: In the final reducing stage, aninstance of a user-provided code is called for each keyin the partition assigned to a reducer. In particular, wehave one output file per executed reduce task.

Fig.1 Working of Map Reduce

IMPACT OF BIG DATA ON HEALTH CARESYSTEM

RIGHT LIVING - The right-living pathway focuses onencouraging patients to make lifestyle choices that helpthem remain healthy.CORRECT CARE - It involves ensuring that patientsget the timely and appropriate treatment available.ACCURATE PROVIDER - It proposes that patientsshould always be treated by high-performingprofessionals that are best coordinated to the task andwill achieve the best outcome.PRECISE VALUE - To fulfill the goals of precisevalue, providers will continuously enhance healthcarevalue while preserving or improving its quality [6].RIGHT INNOVATION - It involves the identificationof new therapies and approaches to delivering care,across all aspects of the system, and improving theinnovation engines themselves.

SECURITY ISSUES IN BIG DATA

A major challenge to healthcare cloud is the securitythreats including tampering or leakage of sensitivepatient’s data on the cloud, loss of privacy of patient’sinformation, and the unauthorized use of thisinformation. The main security and privacyrequirements for healthcare clouds are discussed belowAuthentication: In a healthcare cloud, both health careinformation offered by CSPs (cloud service providers)and identities of users should be verified at the entry ofevery access using user names and passwords assignedto users by CSPs.

Authorization: It is a crucial security requirement that isused to control access priorities, permissions and resourceownerships of the users on the cloud.

Non-repudiation: It implies that one party of a transactioncannot deny having received a transaction nor can the otherparty deny having sent a transaction.

Integrity and Confidentiality: Integrity means preserving theprecision and consistency of data. In the healthcare system,it refers to the fact that EHRs (electronic health records)have not been tampered by unauthorized use.

Availability: For any EHR system to serve its purpose, theinformation must be available when it is needed. Highavailability systems aim to remain available at all times,preventing service disruptions due to power outages,hardware failures, and system upgrades.

CONCLUSION

The real issue is not that we are acquiring large amounts ofdata. It's what you do with the data that counts. Today, asignificant proportion of the cost and time spent in the drugdevelopment process is attributable to unsuccessfulformulations. By enabling researchers to identifycompounds with a higher likelihood of success, Big Datacan help reduce the cost and the time to market for newdrugs. Also, by integrating learning from medical data in theearly stages of development, researchers will now be able tocustomize drugs to suit aggregated patient profiles.Currently, information privacy concerns are the singlebiggest obstacle to Big Data adoption in health care.Another is the absence of an analytics solution powerfulenough to gather massive volumes of largely unstructuredhealth data, perform complex analyses quickly, and triggermeaningful solution, for instance, gather all the data fromICU monitors, which today goes un-stored, put it on theCloud, decipher significant medical patterns that are yetundiscovered, and trigger a medical action instead of merelyan alarm.By providing an overview of the current state of big dataapplications in the healthcare environment, this researchpaper has explored the existing challenges that governmentsand healthcare stakeholders are facing. All big data projectsin leading countries and healthcare industries have similargeneral common goals, such as the provision of easy andequal access to public services, better citizens' healthcareservices, and the improvement of medical-related concerns.However, each government or healthcare stakeholder has itsown priorities, opportunities, and threats, based on itscountry's unique environment (e.g., healthcare expendituresin the United States, the inefficient and wasteful healthcaresystem in Japan, regional disparities in the healthcareresources in India, etc.) which big data projects mustaddress.Second, for medical data that cuts across departmentalboundaries, a top-down approach is needed to effectivelymanage and integrate big data. Governments and healthcarestakeholders should establish big data control towers tointegrate accumulated datasets, whether structured orunstructured, from each silo. In addition to this,governments and healthcare stakeholders need to establish

Int. Journal of Electrical & Electronics Engg. Vol. 2, Spl. Issue 1 (2015) e-ISSN: 1694-2310 | p-ISSN: 1694-2426

NITTTR, Chandigarh EDIT -2015 84

an advanced analytics agency, which will be taskedwith developing strategies on how big data could bebest managed through new technology platforms andanalytics as well as how to secure skilled professionalstaff to use the new tools and techniques.Third, real-time analysis of in-motion big data shouldbe carried out, while protecting privacy and security.Thus, governments and healthcare stakeholders shouldexplore new technological playgrounds, such as cloudcomputing, advanced analytics, security technologies,legislation, etc.Fourth, leading big data governments appear to havedifferent goals and priorities; therefore, they usedifferent sets of data management systems,technologies, and analytics. While such information isnot readily available in the literature, the main concernswith big data applications among these countries andcompanies converge on the following: security, speed,interoperability, analytics capabilities, and the lack ofcompetent professionals [11, 12].

Finally, this study is limited in that the practical applicationsof big data for investigating healthcare issues have not yetbeen fully demonstrated due to the dearth of practice. Withregard to future study, practitioners and researchers shouldcarefully look at and accumulate information with regard tothe practical applications of big data in order to determinethe best ways of using big data in healthcare issues.

REFERENCEShttp://radar.oreilly.com/2012/01/what-is-big-data.htmlwww.forbes.com/sites/lisaarthur/2013/08/15/what-is-big-data/http://www.business2community.com/digital-marketing/4-vs-big-data-digital-marketing-0914845#!bgCHyQhttp://dashburst.com/infographic/big-data-volume-variety-velocity/http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3848064/http://www.informationweek.in/informationweek/cio-blog/175124/health care-compulsion-choicehttp://www.ncbi.nlm.nih.gov/pmc/articles/PMC3717441/http://www.datanami.com/2013/07/19/big_data_emerges_in_indian_health_care/http://eandt.theiet.org/magazine/2013/03/journey-to-the-centre-of-big-data.cfmhttp://itbussinessconsulting.jimdo.com/2014/08/11/how-big-data-and-analytics-can-reshape-healthcare/http://www.hissjournal.com/content/2/1/3