Jianqi Liu,Jiafu Wan,Shenghua He,and Yanlin Zhang

（1.School of Information Engineering，Guangdong Mechanical and Electrical College，Guangdong 510515，China; 2.School of Electrical Engineering and Automation，Jiangxi University of Science and Technology，Ganzhou 341000，China)

E-Healthcare Supported by Big Data

Jianqi Liu1,Jiafu Wan2,Shenghua He1,and Yanlin Zhang1

（1.School of Information Engineering，Guangdong Mechanical and Electrical College，Guangdong 510515，China; 2.School of Electrical Engineering and Automation，Jiangxi University of Science and Technology，Ganzhou 341000，China)

The era of open information in healthcare has arrived.E-healthcare supported by big data supports the move toward greater trans?parency in healthcare by making decades of stored health data searchable and usable.This paper gives an overview the e?health?care architecture.We discuss the four layers of the architecture—data collection，data transport，data storage，and data analysis—as well as the challenges of data security，data privacy，real?time delivery，and open standard interface.We discuss the necessity of establishing an impeccably secure access mechanism and of enacting strong laws to protect patient privacy.

healthcare;wireless body network;big data;disease prediction;remote monitoring;medical data

1 Background

I n China，more than 10%of the population is over six?ty，and by 2015，the number of elderly in China will be 200 million.A rapidly aging population is one of the long?term effects of the one?child policy intro?duced in the 1970s.This trend could lead to a huge labor short?age by 2050 as well as increased demand for more paramedics and medical services.Many experts warn that healthcare for the elderly is an urgent issue.In America，healthcare expenses have been rapidly increasing over the past two decades;high medical costs is not a new phenomenon.In 2009，medical costs in the United States were 17.6%of GDP，nearly$600 bil?lion more than expected.Escalating costs are starting to alter provider reimbursement trends.Risk?sharing models have started to replace many fee?for?service models in order to curb expenses and encourage judicious use of resources.The pa?tient is shouldering more of the cost of medical services than ever before.However，in the United States，the elderly are less likely to be able to pay for these services.This problem is even more serious in China，where many older people so not have health insurance.If they fall chronically ill，their families will not be able to afford the bill.Also，lack of real?time moni?toring is an issue.Real?time physiological data is very impor?tant for treating heart disease in particular，but collecting such data is not easy because heart disease is usually characterized by fleeting symptoms.Similar other diseases require real?time monitoring.There are some issues in the Chinese healthcare system，such as slow response and isolated clinical data;how?ever，advances in wireless body?area networking(WBAN)[1]， [2]and big data[3]introduce new ways to address these is?sues.A modern e?healthcare system takes full advantage of big data[4]and heterogeneous networks，which include WBAN，3G，and 4G technologies.An e?healthcare system collects data on the physical and mental health of a patient and also collects data from clinical records.It then uses data?mining algorithms to provide well?reasoned，useful information for patient care.E?healthcare has three main benefits:

?It eliminates redundant treatment by providing the doctor with a more comprehensive medical profile of the patient. This helps the doctor precisely target their treatment.

?It provides continuous，timely monitoring and diagnosis.A WBAN can monitor the vital signs of patient and transmit this information in real time to a data center via cellular net?work or the Internet.This enables timely diagnosis.

?It reduces costs.The patient may be eligible for discount medicines by authorizing the medical data to be used by a pharmaceutical company.

Here we consider two e?healthcare scenarios.In scenario one，(Fig.1)，an elderly person with health problems is sleep?ing at home.Sensors on the body continuously monitor the per?son’s vital signs，motor activity，social interaction，sleep pat?terns，and other health indicators.The data is sent to a big?da?ta storage server through the hybrid network so that a doctor can assess the data and make an accurate diagnosis[5]，[6].

In scenario two(Fig.2)，an elderly person takes a walk in the garden.Sensors are deployed on the balcony and in the yard to detect sunlight and weather conditions outdoors.The collected data is so large and complex that it cannot be pro?cessed using conventional database management and data?pro?cessing tools.Big?data applications are therefore introduced.

▲Figure 1.Body sensors collect vital signs in an indoor home environment.

▲Figure 2.An elderly person takes a walk in the garden.Sensors collect vital signs and environmental information and transmit this to a big?da?ta storage server via a hybrid network.

The e?healthcare scenarios described here benefit the pa?tient as well as the provider，medical institution，pharmaceuti?cal company，and government.E?healthcare has four main re?quirements:usability and comfort，data acquisition，real?time communication，and big?data processing.The key design pa?rameters for an e?healthcare system are:

?small，light，wearable sensor modules

?perception of both indoor and outdoor environments by the WSN

?low power consumption(to avoid frequent changing of batter?ies)

?real?time transmission

?data?analysis capabilities

?disease prediction model.

E?healthcare is designed to be discrete，usable，comfortable，and intelligent.

2 Architecture

In the health sciences，scientific method is based on experimentation or clinical data，but the limiting factor is lack of relevant data that either supports or refutes the initial hypothesis.Patients cannot benefit from their historical medical and environmental data.With the development big?da?ta techniques，huge amounts of data are derived from patients，doctors，research institutions，and pharmaceutical companies.The e?healthcare sys?tem needs to take advantage of these massive amounts of data and provide the right intervention to the patient at the right time.It should offer per?sonalized care to the patient.According to Peter Groves et al.，an e?healthcare system should achieve the follow?ing goals[7]:

?Right lifestyle.Patients can maintain a positive attitude in their treatment，including complying with the doctor's treat?ment guidelines and taking active measure to prevent dis?ease.With advice given by e?healthcare，people can make proper lifestyle choices that help them remain healthy.

?Right medical care.Patients should be offered the most rea?sonable treatments when sick.In China，some doctors tend to over?treat or use expensive drugs in order to obtain a higher income.E?healthcare can provide treatment exam?ples for the same disease.Remote monitoring and diagnosis provided by e?healthcare is a particularly good choice for outpatients.If a patient is in crisis，e?healthcare sends a warning to the doctor for quick intervention.

?Right medical workers.Patients should always be treated by high?performing professionals that are best matched to treat the disease.However，patients often cannot get de?tailed information about a doctor and are unable to make ap?propriate choices.The e?healthcare system should provide comprehensive information about medical staff so that pa?tients can make informed decisions.

?Right value.Medical workers，payers and other institutions will continuously enhance e?healthcare value while preserv?ing or improving on quality.E?healthcare can offer advice about the best treatments;it can avoid overtreatment;it of?fers real?time monitoring;and it involves multiple measures for ensuring cost?effectiveness.It can eliminate fraud，waste，and abuse in the system.

An e?healthcare architecture supported by big data has four layers:data collection，data transport，data storage，and data analysis(Fig.3).The data?collection layer includes wireless sensors，for collecting environmental data such as temperature and humidity，and a GPS receiver，for determining the location of the patient.The data?transport layer converges raw datafrom the wireless sensor network(WSN)，WBAN to router，and sends this data to a data center via traditional Internet or LTE cellular network.In the data storage and analysis layer，the pa?tient’s health information is collected and a prediction is made on possible signs of disease.

▲Figure 3.E?healthcare architecture.

3 Data Collection

3.1 Data Classification

E?healthcare’s innovation mostly depends on the patients’medical data.Some research institutions use private data de?rives from medical experiments.But many hospitals still cap?ture patients’medical data to make further research.Accord?ing to the data sources，the data can be categorized as clinical trial，genetic experimental，private health，mental health，and other according to the source of the data.

Clinical trial data is derived from public papers or electron?ic medical records[8].In clinical trials，the quality of the data collected depends first and foremost on the quality of instru?ments used.No matter how much time and effort go into con?ducting a trial，if the correct data points are not collected，it may not be possible for an e?healthcare system to do a mean?ingful analysis.Therefore，the design，development，and quali?ty assurance of medical instrumentation must be given the ut?most attention.Clinical trial data is structured and very easy to put into big?data storage.As well as structured electronic med?ical records，there are also some unstructured clinical notes and medical images.These are also important data sources for an e?healthcare system.

Experimental data described here is private or proprietary data，usually derived from R&D con?ducted by pharmaceutical companies or govern?ment institutions[9].This data is usually confiden?tial and not publically available.It includes genet?ic data(DNA sequences).If the system is secure，limited sharing of data will improve the universali?ty and veracity of the data system.

Private health data includes data on vital signs and environment collected by sensors on the pa?tient’s body and the personal data of a patient col?lected from a hospital.

Mental health may include an individual’s abili?ty to enjoy life，balance life activities，and achieve psychological resilience.The state of a person’s mental health can be collected by a trained psy?chologist.Statistics from the World Health Organi?zation show that nearly half the world's population is affected by mental illness，which affects people’s self?esteem，relationships and ability to function in everyday life.Long?term psychological illness impacts physical health.Therefore，mental health data is important in an e?healthcare system[10].

There is also other data，such as epidemiology data，environ?mental data and location data，which affects the health of a pa?tient[11].

3.2 Sensors

資本主義社會(huì)與合理化進(jìn)程有著密不可分的聯(lián)系。馬克斯·韋伯寫道：“資本主義精神的發(fā)展似乎最好理解為合理主義整體發(fā)展的一部分，并且似應(yīng)能夠從合理主義對(duì)生活基本問(wèn)題的原則立場(chǎng)中推演出來(lái)?！保?2］51他還將宗教社會(huì)學(xué)與社會(huì)經(jīng)濟(jì)關(guān)系聯(lián)系起來(lái)，確認(rèn)由新教倫理生發(fā)的勤奮勞動(dòng)、誠(chéng)實(shí)經(jīng)營(yíng)刺激著理性的工具化擴(kuò)張，構(gòu)成理性工具化的文化根源。借助于天職的概念，倫理從宗教信仰轉(zhuǎn)向了世俗信念，系統(tǒng)而合理的追逐利潤(rùn)的價(jià)值觀構(gòu)成了資本主義之精神。理性成就了世俗信念的倫理觀，卻瓦解了理性超越與批判力量，價(jià)值合理性被遮蔽。

A sensor node mainly comprises the physiological and envi?ronmental signal sensor and the radio platform that sensors are connected to.The general function of a sensor is to collect ana?log signals that correspond to the human physiological condi?tion.The analog signals are received by a corresponding radio?equipped board，where they are then digitized.The digital sig?nals are then forwarded by the router to the data center，where they are stored.

A blood pressure sensor is a non?invasive sensor for measur?ing systolic and diastolic blood pressures.The sensor automati?cally inflates an upper arm cuff at customizable time intervals to acquire readings and transmit the data to the storage server. The sensor was developed by Johns Hopkins University Ap?plied Physics Laboratory and is based on the NIBP module from SunTech Medical.It takes a reading every five minutes [12].It has a battery pack comprising four 9 V lithium batter?ies and can operate for ten hours.It shows a pulse rate accu?rate to±3 beats per minute and blood pressure accurate to±3 mmHg[13].

A CO2sensor measures gaseous carbon dioxide level and ox?ygen concentration in human respiration.In[14]，a wireless，passive carbon?nanotube gas sensor is introduced.Multiwall carbon nanotubes(MWNTs)remotely receive data on carbon dioxide，oxygen，and ammonia based on the measured changes in MWNT permittivity and conductivity with gas exposure.

An electrocardiographic(ECG)sensor records the heart’s electrical activity.Healthcare providers use it to help diagnose heart disease.In order to obtain an ECG signal，several elec?trodes are attached at specific sites on the skin，and the differ?ence in electrical potential between these electrodes is mea?sured[15].

A humidity and temperature sensor measures the tempera?ture of the human body as well as the humidity and tempera?ture of the surrounding environment[16].

A pulse oximetry(SpO2)sensor measures the oxygen satura?tion and heart rate through a non?invasive probe.The SpO2 sensor board developed by Smiths Medical is used in ambu?lances and is very accurate.According to the manufacturer’s specifications，it has an SpO2 accuracy of±2%and heart rate accuracy of±2 bpm.The sensor board was chosen for the AID?N project.

A camera sensor may be a complementary metal?oxide?semi?conductor(CMOS)active?pixel sensor，such as OV9650，that is embedded in a PDA，cellphone，or other device.Video re?cordings have been used to treat Parkinson’s disease.An ex?pert examines the video recordings and provides clinical cores representing the severity of the tremor，dyskinesia，and brady?kinesia.

A vital signs sensor benefits from the development of wire?less communications;in particular，miniaturization and energy efficiency in embedded computing has improved significantly. The sensor node is therefore becoming much smaller.

4 Data Transport

In e?healthcare，data is transmitted via heterogeneous net?works.Physiological and mental health data is relayed via the WBAN;environmental data is relayed via wireless sensor net?works;and experimental or clinical data is relayed via tradi?tional Internet.In situations where long?distance transmission is required and there is no wired network，LTE should be de?ployed.

WBAN is one of the best technologies for building unobtru?sive，scalable，robust wearable health monitoring systems.A WBAN for health monitoring comprises multiple sensor nodes. Each node is typically capable of sensing and processing one or more physiological signals，caching them，and transmitting the data to a big?data storage server.

A WSN comprising spatially distributed autonomous sensors monitors environmental conditions and passes the data via a wireless network to the storage server.Modern networks are bi?directional，meaning that sensor activity can be controlled.To?day such networks are mainly used in industrial and consumer applications，such as monitoring and control of industrial pro?cesses and machine health.

LTE is 4G communication technology that enables peak download rates of up to 299.6 Mbit/s and peak upload rates of up to 75.4 Mbit/s depending on the user equipment category (with 4×4 antennas and 20 MHz of spectrum).In the case of a moving patient，LTE can easily transmit the vital signs to the server and overcome the defects of a traditional network[17].

5 Data Storage

Mass medical data requires an appropriate storage mecha?nism.Google products are stable solutions and can be classi?fied into three levels.From the bottom up these are:basic file system，such as Google file system(GFS)or Colossus;manage?ment database，such as BigTable;and programming model，such as MapReduce[18].

Google’s Colossus is second?generation GFS and an expand?able distributed file system that supports large?scale，distribut?ed，data?intensive applications.It overcomes problems such as a single point failure and poor performance with small files.Co?lossus is the foundation of upper?level applications.In addi?tion，open?source file systems，such as HDFS and Kosmosfs can be acquired[19].

BigTable is a column?oriented，distributed，structured data storage system designed to process large?scale(petabyte)data of thousands of commercial servers.The basic data structure of BigTable is multidimensional sequenced mapping with sparse，distributed，persistent storage.BigTable is based on many fun?damental components of Google，including Colossus，cluster management system，SSTable file format，and Chubby.

MapReduce[20]is a simple but powerful programming mod?el for large?scale computing.It uses a large number of commer?cial PC clusters for automatic parallel processing and distribu?tion.The MapReduce computing model only has two func?tions—Map and Reduce—both of which are programmed by the user.The Map function processes input key?value pairs and generates intermediate key?value pairs.MapReduce then combines all the intermediate values related to the same key and transmits them to the Reduce function，which further com?presses the value set into a smaller value set.

6 Data Analysis

In the healthcare industry，medical data is important role predicting diseases，and some researchers use the genome data (DNA)for this purpose.When medical experts identify genes in DNA that are markers for disease，a person can make appro?priate lifestyle or other changes to lower the risk of disease.Ge?nome data has used to predict heart disease and brain disor?ders.For inherited diseases，identifying a parent who is a carri?er but does not express the disease can help parents make in?formed choices regarding pregnancy.Some researchers use da?ta?mining to predict diabetes and breast cancer according to the health profile of the individual.Some researchers use envi?ronmental data or information on epidemics to predict diseases such as the flu.In the big?data era，people are concerned about how to rapidly extract important information from mass data inorder to benefit both enterprises and patients.At present，the main methods for processing big data are[18]:bloom filter，hashing，indexing，triel，and parallel computing.

A bloom filter comprises a series of hash functions.The prin?ciple of a bloom filter is to store the hash values of data，rather than the data itself，by using a bit array.This is，in essence，a bitmap index that uses hash functions for lossy data compres?sion and storage.It is highly efficient in terms of query speed and space?saving，but it also has some disadvantages，such as misrecognition and deletion.

Hashing is a method is a method for transforming data into shorter fixed?length numerical values(index values).It has the advantages of rapid reading，writing，and query speed，but it is difficult to find a sound hash function.

Indexing is an effective way of reducing disk reading and writing and improving insertion，deletion，modification，and query speeds in both traditional relational databases，which manage structured data，and other technologies that manage semi?structured and unstructured data.However，with index?ing there is additional cost for storing index files，which need to be maintained dynamically when data is updated.

Triel(also called trie tree)is a variant of hash tree.It is mainly used for rapid retrieval and word frequency statistics. The main principle of triel is to use common prefixes of charac?ter strings to reduce comparison of character strings to the greatest possible extent and increase query efficiency.

Parallel computing，unlike traditional serial computing，in?volves the simultaneous use of several computing resources to complete a computing task.The basic principle of parallel com?puting is to decompose a problem and assign the parts of the problem to several separate processes to be independently dealt with.This is also called co?processing.

We can envision a thorough healthcare system in which medical devices，wearables，diagnostic tools，and analytics em?power patients and their families to better care for themselves. A disease?prediction model is essential to the e?healthcare sys?tem.Real?time data alerts enable the doctor or healthcare pro?vider to intervene when necessary.A developed disease?predic?tion model first and foremost needs to provide accurate，inter?nally and externally validated probabilities of specific health conditions or outcomes in a patient.Such models must guide the doctor’s decision?making process and the patient’s behav?ior.This will improve the outcome for the patient and reduce the cost of care.

6.1 Breast Cancer Prediction Model

Researchers，clinicians，and the public are increasingly in?terested in statistical models designed to predict the occur?rence of cancer[21].Susan M.Domchek et al.use a cancer risk prediction model to estimate the risk of breast cancer for women[21].The model estimates the likelihood of breast can?cer risk due to genetic susceptibility，such as BRCA1 or BRCA2 mutations.Recent developments have reinforced the clinical importance of breast cancer risk assessment.Tamoxi?fen chemoprevention and studies such as the Study of Tamoxi?fen and Raloxifene are available to women who are at in?creased risk of breast cancer[26].In addition，specific manage?ment strategies are now defined for BRCA1 and BRCA2 muta?tion carriers.Risk?assessment can be used to determine the likelihood that a woman will develop breast cancer，or prior probability models can be used to determine the likelihood of BRCA1 or BRCA2 mutation.Several models may be needed to give the woman optimal counseling，provide the woman and her family with accurate，useful information，and make a sound clinical judgment.

6.2 Coronary Heart Disease Prediction Model

A computer simulation model was developed to project the future mortality，morbidity，and cost of coronary heart disease (CHD)in the United States[22].The model contains a demo?graphic?epidemiologic submodel that simulates the distribu?tion of coronary risk factors and the conditional incidence of CHD in a demographically evolving population.It also con?tains a“bridge”submodel that determines the outcome of the initial CHD event as well as a disease history submodel that simulates subsequent events in a person who has previously had a CHD event.The effects of either preventive or therapeu?tic intervention on mortality，morbidity，and cost can be simu?lated for up of to a 30 years.The baseline projection is based on no changes in risk factors or efficacy of therapies after 1980.It shows how the aging of the population，especially the baby?boomer generation，naturally increases the annual inci?dence，mortality，and costs of CHD by about 40-50%by 2010. Unprecedented reduction of risk factors would be required to offset these demographic effects on the absolute incidence of CHD.However，forecasts could be inaccurate because of mis?placed assumptions or poorly estimated baseline data，and the model awaits validation using actual future data.

7 Applications

Some healthcare leaders are already gaining value from healthcare aided by big data.The following two examples show how mass medical data can be used to reduce the medi?cal costs.

7.1 HealthConnect

In 2002，Kaiser Permanente ceased construction of its own clinical information system and ultimately sought out another vendor，Epic Systems，to undertake this construction.The new system，called HealthConnect[23]，[24]，is one of the largest private electronic health systems in the world.The system inte?grates more than 611 medical offices and 37 hospitals，linking patients to their healthcare teams and private health data.This system has served 9.1 million members and helps patients re?fill 1.2 million prescriptions monthly.HealthConnect facili?tates communication between members and doctors to help make getting well and staying healthy easy and convenient. HealthConnect share the latest findings and best practices that a comprehensive electronic health record can increase consum?er convenience and satisfaction and provider efficiency，while maintaining clinical quality.

In addition，the mass data in HealthConnect helps clinicians with their research.Kaiser Permanente operates one of the larg?est non?university research programs in the United States.Kai?ser Permanente clinicians and researchers have approximately 2000 studies in progress at any given time and publish 900 to 1000 articles annually.Accessing fully digitized health re?cords is much cheaper and far less time consuming than ac?cessing paper records when conducting research.With this mass?digitized data，Kaiser Permanente researchers have been able to explore a number of areas，including how vaccines，medications，and lifestyle affect the whole population.The re?search outcome is used to improve the healthcare.

7.2 IMS Health Disease Analyzer

The Disease Analyzer patient database contains data on di?agnoses，prescriptions，risk factors，and laboratory results for approximately 5 million patients per year in Germany[25]. The database also contains data from various groups of special?ist doctors and general practitioners.Data from approximately 3200 office?based doctors form the basis of this investigation. Data are delivered with OLAP based software application(Dis?ease Analyzer Software)or data are used for ad?hoc studies which are sold with analytic services to customers.

8 Challenges

With the more widespread release of personal health infor?mation，the government，leading companies，and research insti?tutions need to consider regulations about its use，as well as privacy protections and data security.To encourage data shar?ing and streamline the repetitive nature of granting waivers and data?rights administration，it may be better for data approv?als to follow the patient，not the procedure.Further，data shar?ing could be made the default，rather than the exception.It is important to note，however，that as data liquidity increases，physicians and manufacturers will be subject to increased scru?tiny，which could result in lawsuits or other adverse conse?quences.

8.1 Data Privacy

Personally identifiable information or other sensitive infor?mation including healthcare records，biological traits，such as genetic material is possible to be divulged in e?healthcare.The challenge in data privacy is to exchange data while protecting personally identifiable information.As heterogeneous informa?tion systems with differing privacy rules are interconnected and information is shared，policy appliances will be required to reconcile，enforce and monitor an increasing amount of pri?vacy policy rules(and laws).There are two categories of tech?nology to address privacy protection in commercial IT systems: communication and enforcement.As a user，you also need to protect yourself data privacy，on the internet you almost always give away a lot of information about yourself:Unencrypted e?mails can be read by the administrators of the e?mail server，if the connection is not encrypted(no https)，and also the internet service provider and other parties sniffing the traffic of that connection are able to know the contents.

8.2 Data Security

The e?healthcare system can realize benefits from democra?tizing big data access.The researchers can more easily collabo?rate，engage in peer review and eliminate duplication of efforts. Researchers will also be able to more readily identify opportu?nities where they can contribute and collaborate.The system makes exposing and sharing data easy and relatively inexpen?sive.However，significant security concerns exist.A creden?tialing process could facilitate and automate this access，but there are complexities and challenges.Since providers，pa?tients and other interested parties such as researchers need se?cure access;data access should be controlled by group，role and function.Finally，the security of the data once it leaves the cloud also needs to be assured.Big data can be used to identi?fy patterns and irregularities indicating and preventing securi?ty threats，as well as other types of fraud.

8.3 Real-Time Communication

The healthcare is performance?critical applications，which require bounded delay latency.Whether the heterogeneous net?work has capability of providing bounded delay guarantees on packet delivery is very important to monitor.Since WSN and WBAN deal with real world，it is often necessary for communi?cation to meet real?time constraints.In surveillance systems，for example，communication delays within sensing and actuat?ing loops directly affect the quality of tracking.To date，few re?sults exist for WSNs that adequately address real?time require?ments.Tian He and John A Stankovic proposed a real time pro?tocol SPEED to meet requirement.The redundancy of nodes is a solution for wireless network.The presence of multiple dis?joint paths between nodes makes them robust to link and node failures.

8.4 Data Format and Standard Interface

The data of e?healthcare are derived from different institu?tions，the formats are different.These heterogeneous systems need to adapt to different formats and interfaces，affect the ex?change of data.So we need an open standard for data formats along with open standardized interfaces，and the healthcare system can access each other more easily with various health?care providers，hospital，and pharmaceutical company.This is a challenge for e?healthcare，the government or guild has a re?sponsibility to build a specification for data exchange.

9 Conclusion

In order to achieve these goals，including right living，right care，right provider，right value，and right innovation，the big data technique have been introduced into e?healthcare.With the support of big data，the e?healthcare can resolve massive data storage，data management，and data analysis.If e?health?care can protect the personal data security and privacy，the e?healthcare can reduce the healthcare cost by a large margin. The disease prediction can give patient appropriate strategies to lower the risk of disease.

Acknowledgement

The authors would like to thank the Natural Science Founda?tion of Guangdong Province，China(No.9151009001000021)，the Ministry of Education of Guangdong Province Special Fund Funded Projects through the Cooperative of China(No. 2009B090300341)，the National Natural Science Foundation of China(No.61262013)，the Open Fund of Guangdong Prov?ince Key Laboratory of Precision Equipment and Manufactur?ing Technology(No.PEMT1303)，and the Higher Vocational Education Teaching Reform Project of Guangdong Province (No.20130301011)for their support in this research.

[1]M.Chen，S.Gonzalez，V.Leung，and Q.Zhang et al.，“A 2G?RFID?based e?healthcare system，”IEEE Wireless Communications,vol.17，pp.37-43，2010. doi:10.1109/MWC.2010.5416348.

[2]J.F.Wan，C.F.Zou，and S.Ulah et al.，“Cloud?enabled wireless body area net?works for pervasive healthcare，”IEEE Network，vol.27，no.5，pp.56-61，2013. doi:10.1109/MNET.2013.6616116.

[3]S.LaValle，E.Lesser,and R.Shockley et al.，“Big data，analytics and the path from insights to value，”MIT Sloan Management Review,vol.21，2014

[4]H.Cao，V.Leung,and C.Chow et al.，“Enabling technologies for wireless body area networks:A survey and outlook，”IEEE Communications Magazine,vol.47，pp.84-93，2009.doi:10.1109/MCOM.2009.5350373.

[5]J.Liu，Q.R.Wang,and J.F.Wan et al.，“Towards key issues of disaster aid based on wireless body area networks，”KSII Transactions on Internet and Infor?mation Systems,vol.7，pp.1014-1035，2013.

[6]F.H.Messerli，B.D.Nunez,and H.O.Ventura et al.，“Overweight and sudden death:increased ventricular ectopy in cardiopathy of obesity，”Archives of Inter?nal Medicine,vol.147，pp.1725，1987.

[7]P.Groves，B.Kayyali,and D.Knott et al.，“The‘big data’revolution in health?care，”McKinsey Quarterly,2013.

[8]R.H.Miller and I.Sim，“Physicians’use of electronic medical records:barriers and solutions，”Health Affairs,vol.23，pp.116-126，2004.

[9]B.S.Weir，Genetic Data Analysis，U.S.:Sinauer Associates，Inc.，1990.

[10]K.Demyttenaere，R.Bruffaerts and J.Posada?Villa et al.，“Prevalence，severi?ty，and unmet need for treatment of mental disorders in the World Health Orga?nization World Mental Health Surveys，”JAMA:The Journal of the American Medical Association,vol.291，pp.2581-2590，2004.

[11]J.Q.Liu，Q.R.Wang,and J.F.Wan et al.，“Towards real?time indoor localiza?tion in wireless sensor networks，”in IEEE 12th International Conf.on Comput. and Inform.Technology，Cheng du，China，2012，pp.877-884.doi:10.1109/ CIT.2012.182.

[12]T.Gao，T.Massey,and L.Selavo et al.，“The advanced health and disaster aid network:A light?weight wireless medical system for triage，”IEEE Transactions on Biomedical Circuits and Systems,vol.1，no.3，pp.203-216，2007.doi: 10.1109/TBCAS.2007.910901.

[13]A.Georgiades，A.Sherwood,and E.C.Gullette et al.，“Effects of exercise and weight loss on mental stress-induced cardiovascular responses in individuals with high blood pressure，”Hypertension,vol.36，pp.171-176，2000.

[14]K.G.Ong，K.F.Zeng，and C.A.Grimes，“A wireless，passive carbon nanotube?based gas sensor，”IEEE Sensors Journal,vol.2，no.2，pp.82-88，2002.doi: 10.1109/JSEN.2002.1000247.

[15]J.Welch，F(xiàn).Guilak，and S.D.Baker，“A wireless ECG smart sensor for broad application in life threatening event detection，”in 26th Annual Int.Conf. IEEE Engineering in Medicine and Biology Society(IEMBS’04)，2004，pp. 3447-3449.

[16]T.Polk，W.Walker,and A.Hande et al.，“Wireless telemetry for oxygen satura?tion measurements，”in IEEE Biomedical Circuits and Systems Conference(Bio?CAS 2006)，2006，pp.174-177.

[17]S.Sesia，I.Toufik，and M.Baker，LTE:the UMTS Long Term Evolution，Wiley Online Library，2009.

[18]M.Chen，S.W.Mao，and Y.H.Liu，“Big data:a survey，”Mobile Networks and Applications,vol.19，no.2，pp.1-39，2014.doi:10.1007/s11036?013?0489?0.

[19]B.Peng，B.Cui，and X.M.Li，“Implementation issues of a cloud computing platform，”IEEE Data Eng.Bull.,vol.32，pp.59-66，2009.

[20]J.Dean and S.Ghemawat，“MapReduce:simplified data processing on large clusters，”Communications of the ACM,vol.51，pp.107-113，2008.

[21]S.M.Domchek，A.Eisen,and K.Calzone et al.，“Application of breast cancer risk prediction models in clinical practice，”Journal of Clinical Oncology,vol. 21，no.4，pp.593-601，2003.

[22]P.W.F.Wilson，R.B.D’Agostino,and D.Levy et al.，“Prediction of coronary heart disease using risk factor categories，”Circulation,vol.97，no.18，pp. 1837-1847，1998.

[23]K.D.Mandl and I.S.Kohane，“Healthconnect:clinical grade patient?physician communication，”in Proc.AMIA Symposium,1999，p.849.

[24]P.Bolton，S.Ganon，and D.Aro，“HealthConnect:a trial of an after?hours tele?phone triage service，”Australian Health Review,vol.25，pp.95-103，2002.

[25]L.Annemans，E.Spaepen,and M.Gaskin et al.，“Gout in the UK and Germa?ny:prevalence，comorbidities and management in general practice 2000-2005，”Annals of the Rheumatic Diseases,vol.67，no.7，pp.960-966，2008.

[26]V.G.Vogel，J.P.Costantino，and D.L.Wickerham et al.，“Update of the na?tional surgical adjuvant breast and bowel project study of tamoxifen and raloxi?fene(STAR)P?2 trial:preventing breast cancer，”Cancer Prevention Research，vol.3，pp.696?706，2010.

Biographiesphies

Jianqi Liu(liujianqi@ieee.org)is a lecturer at Guangdong Mechanical and Electri?cal College，China.He received his MS degree in computer application technology from Guangdong University of Technology in 2009.He is currently pursuing his PhD degree at GDUT.His research interests include body area networks，embedded systems，internet of vehicle，and cyber?physical systems.He is a member of the IEEE.

Jiafu Wan(jiafuwan_76@163.com)is an associate professor in the School of Electri?cal Engineering and Automation，JiangXi University of Science and Technology，China.He received his PhD degree in mechatronic engineering from SCUT in 2008. He has authored or co?authored one book and more than 60 scientific papers.His re?search interests include cyber?physical systems，internet of things，machine?to?ma?chine communication，mobile cloud computing，and embedded systems.He is a CCF senior member and a member of IEEE，IEEE Communications Society，IEEE Control Systems Society，and ACM.

Shenghua He(2001hsh@21cn.com)is an associate professor at Guangdong Mechan?ical and Electrical College.He received his MS degree in mechatronic engineering from Guangdong University of Technology in 2004.His research interests include management information system，automatic control system，energy management sys?tem，etc.

Yanlin Zhang(ylzhxx@qq.com)is a professor at Guangdong Mechanical and Electri?cal College.He received his MS degree from the Institute of Optics and Electronics，Chinese Academy of Sciences，in 1998.His research interests include mobile net?works，enterprise service bus，and energy?saving technologies in large public build?ings.

t

2014?04?19

10.3939/j.issn.1673-5188.2014.03.006

http://www.cnki.net/kcms/detail/34.1294.TN.20140818.1654.001.html,published online 18 August,2014