• Big Data
  • Mobile phone data can track rubella in Kenya

    Social media and Big Data’s value to health and healthcare is being applied in Kenya. A study published in the Proceedings of the National Academy of Sciences shows that a research team from Harvard and Princeton universities have used about 15 million anonymised records of mobile phone calls and text messages in Kenya to track rubella.

    The team compared the data with long-term data about Kenya’s rubella transmission and outbreaks. It found that the two sets of data matched. It included showing the peaks in September, February and May. The team says the “Data represent a potentially valuable previously unidentified resource for the direct observation of seasonal population dynamics on refined spatial scales that underlie disease transmission. Unlike other static measures of travel such as road networks, small-scale surveys, travel time surfaces, or census data, mobile phone data can provide a dynamic picture of mobility and population travel and changes over large geographic scales.”

    There’s a caveat. “Their relevance to infectious disease dynamics has yet to be formally assessed, requiring parallel longitudinal estimates of phone use and transmission rates.”

    Despite this, the team’s clear that “Our analysis of population fluxes from mobile phone data and rubella transmission dynamics shows that mobile phone data may be used to capture seasonal human movement patterns relevant for understanding childhood infection dynamics and significantly outperform previous proxies.”

    The next goal for the team’s to use its methodology to track cholera, malaria and measles. Could this study be a big breakthrough for Big Data in Africa’s health and healthcare?

  • Exploiting Big Data and analytics needs sophistication

    The potential, possibly the hype, around Big Data and analytics in healthcare is rolling on. A report by Greg Nelson, Founder and CEO of ThotWave, a company specialising in big data warehouses, reporting, and-analysis, sets out the need for a sophisticated approach to maximising the benefits.

    Published by Health Data Management and Source Media, a main proposition of Big Data and Analytics is that having more data and technology aren’t necessarily the challenges, but defining an effective strategy is. It should include:

    Focus, to know what should be retained and integrated, with a methodology for continuous ranking of its utility Governance, to manage the processes for data flow, enrichment, definitions, quality assurance, and security applied to data sharing Human capital and talent development because it’s critical to find, attract, train, nurture and motivate talent who understand data, analytics, and healthcare’s delivery ecosystem.

    A strategy’s also needed to retain the recruited talent. This is very demanding for Africa’s health systems.

    From these, a big question then is, how will healthcare organisations make the most of all this data? It is important because healthcare organisations are accumulating data, but it’s unstructured and not integrated into a cohesive whole. This makes it difficult to use it effectively, so converting data into actionable intelligence is the challenge for healthcare leaders. Three recommended actions are:

    Establish the data exchange value proposition Build trust and community data competence Build community data infrastructures.

    These fit into five stages of analytics maturity:

    Analytically Impaired, where organisations aren’t data driven, but take decisions intuitively Localised analytics, mainly using reporting and analytics in silos Analytical aspirations, where organisations see the value of analytics, but struggle to become more analytical Analytical organisations that are good at analytics, highly data oriented, have analytical tools, use analytics extensively, but aren’t relying on analytic strategies Analytical competitors, where organisations experience analytical nirvana, and use analytics broadly and deeply across the enterprise as a competitive differentiator.

    It’s easy to see where African health systems are, and how they can move through the five stages. How health systems do it, and how quickly, depends on the value they can derive from Big Data and analytics. It’s a complex set of learning curves that need good leadership.


    Image from www.kdnuggets.com

  • Big Data's on its way to track Ebola

    Tracking and mapping Ebola’s spread is a manual endeavour. Infection Control Today (ICT) has a report on a project at Florida Atlantic University’s (FAU) College of Engineering and Computer Science that aims to develop an innovative model of Ebola’s spread that uses Big Data analytics techniques and tools. It’s financed by a Rapid Response Grant (RAPID) from the USA’s National Science Foundation.

    The programme will use data from many sources including Twitter feeds, Facebook and Google. It’ll be entered into an mHealth decision support system that models Ebola’s spread pattern, then create dynamic graphs and predictive models on the outcome and impact on individuals and communities. These will improve the precision on predictions.

    Computer power for the project is a combination of FAU’s Cloud system and LexisNexis HPCC Systems®, an open-source platform for Big Data. LexisNexis will provide the research team with the Big Data needed to develop and model the program for Ebola spread patterns and their expertise in Big Data analytics. An aim is to track the movement of people and their contacts in locations affected by Ebola and comply with privacy laws.

    The USA and the UK have had Ebola patients, so the project can benefit their health systems. If it succeeds, the project can transform data management and healthcare’s response in African countries too.

  • Analytics identifies patients with diabetes

    About 10% of patients may have confirmed or undiagnosed diabetes. Of these 54% may be undiagnosed. These are some of the findings from a large scale analytics study published in CMAJ Open. The team from the USA and UK used data from 11,540,454 EHRs of patients from over 9,000 USA primary care clinics. The team used algorithms to pick out the 62,620 patients with undiagnosed diabetes.

    That about 5% of people have undiagnosed diabetes may not be the main finding. There’ll always be people with undiagnosed diabetes. The method and tools used to identify them as a cohort and be able to follow up the findings may be the big step forward.

    Of the approximate 1.1 million EHRs records indicating diagnosed diabetes, about 62% had a diagnostic code. Of the approximate 10.4 million records for other patients, 0.4%, 40,359, had at least two abnormal fasting or random blood glucose values. About 0.2%, 23,261 of the remaining records had at least one documented glycated haemoglobin (HbA1c) value of 6.5% or more.

    The team also found that electronic coding of the diagnosis was associated with improved quality of care, in turn, reinforcing the value of eDiabetes registers and patient cohort management. This healthcare mode is underused in US primary care, and other countries.. It’s reasonably well developed for some types of patients in the England’s NHS, who had better quality of care than diabetes patients in the USA. This comparison confirms the new opportunities offered by registries and patient cohort management for a systematic, structured proactive approach to providing proactive healthcare.

    EHRs are not enough. How they’re used is more important, and is where the benefits are. Steadily, two streams of benefits are emerging:

    Sharing clinical and health information in clinical settings Using analytics to provide information for proactive management of patients and their conditions.

    Both can help to transform health and healthcare. African countries have an opportunity to learn from studies like these to invest in realising benefits promptly.

  • Social media and Big Data help to develop HIV services

    Social media as a source for Big Data can provide valuable insights about people’s behaviours and their likelihood of engaging in high-risk activities that can lead to contracting HIV. This is a conclusion of Sean D. Young from the Center for Digital Behavior, Department of Family Medicine, University of California (UCLA) in his post in Trends in Microbiology.

    He sets out how social media data can contribute to Big Data science and current approaches to using social media to monitor and predict health behaviours and disease outbreaks. From these, he recommends tools and approaches needed.

    He sees Big Data as a combination of relational and structured data, such as medical and genetics datasets and unstructured such as publically available free text from social media conversations. Some of the conversations contain large volumes of personal information, and it’s feasible to analyse these to collect a range of psychological information about attitudes and behaviours affecting health. Some of the data shows that people who discussed HIV-prevention on social media are more likely to ask for an HIV test. The data can provide part of a forecast healthcare demand and feed into epidemiological studies that monitor risk behaviours and predict disease outbreaks and progressions.

    Using social media for Big Data is not just an analytical activity. It needs:

    A multi-disciplinary team and approach Availability of large and frequently updated datasets An understanding of its limitations, such as data validity levels, missing data, observational data and samples’ representativeness.

    These offer a good model for African countries to adopt for their initiatives. It may be worth adding that it’s advantageous to start small with Big Data.

  • Does Big Data need IoT to succeed?

    The promise of Big Data and analytics in health care tends to rise when there’s an epidemic. Now, it seems that Big Data’s not big enough to stand on its own feet. It needs the help of the Internet of Things (IoT). Together, they’re the Industrial Internet. This is what a report from GE and Accenture seems to show. It’s based on a survey across several sectors.

    The survey shows an increasing urgency for organizations to use Big Data analytics to advance their strategies. The belief is that Big Data analytics can change industries’ competitive landscape over the next year, so companies are investing in it. The security, information silos and systems integration challenges seem outweighed by the operational, strategic and competitive advantages. The main findings are:

    93% say that new market entrants leverage Big Data as a critical differentiation strategy 89% say that companies that don’t adopt a Big Data strategy in the next year risk losing market share and momentum 85% say that their companies have Big Data in their top three priorities 84% say that Big Data can transform their companies’ competitive landscape over the next year 65% focus on monitoring assets to identify operating issues 60% expect Big Data to improve their profitability 31% of healthcare organizations say they’re significantly ahead in analytics About 51% of healthcare organizations are investing between 10% and 20% of their technology budgets on Big Data 49% say they’re planning Big Data initiatives for new business opportunities for extra revenue streams with their Big Data strategy, while 60%expect to increase their profitability by using the information to improve their resource management.

    That’s the main exciting Big Data possibilities, so now for the constraints:

    36% say system barriers between departments prevent data collection and correlation 29% say it’s difficult to consolidate disparate data and use the resulting data repository 44% say security and cyber-crime is an issue.

    A balanced solution is:

    Use experts to assess risks and consequences and understand the vulnerabilities, then invest in effective security Break down the barriers to data integration Acquire and develop Big Data and analytics talent Develop new business models Actively manage regulatory risk Leverage mobile technology to deliver analytic insights.

    These form a nucleus of a Big Data strategy for healthcare organisations in Africa. The challenge is fitting into constrained budgets that aren’t enough for all the eHealth possibilities, but this is the nature of eHealth leadership.

  • IBM's Watson to help fight Ebola

    There have been several calls and proposals for using technology, ICT, Big Data and analytics to help in the fight against Ebola. IBM has joined fray. The Guardian has a post on IBM’s plan to use Watson, it’s super-computer to help to identify undetected victims, many of whom try to treat themselves or rely on traditional cures. Karen Jacobsen, the IBM specialist leading the team, says that smartphones should play a greater role in combating the disease. eHNA posted a commentary on an article in The Economist saying the same.

    The process aims to collect data where citizens use SMS or voice calls with data on locations. This is then analysed to identify correlations and highlight challenges that need addressing.

    Watson is a $100m project to use analytics and data to help solve some of Africa’s problems and challenges. Access to Watson could help enable poorer parts of Africa to leapfrog development stages similar to the take-off for mobile phones on the continent where landlines were limited.

  • Ebola: how can mobile phones and Big Data help?

    Like all types of new information, Big Data has a lot to offer, but limitations too. It’s been suggested that eHealth can help health professionals respond to the tragic Ebola outbreak, and Big Data can have a specific role. An article in The Economist sets out how.

    A main role is in tracing people’s locations, journeys and places of their contacts. The data is phone companies’ Call Data Records (CDR). It’s been used historically before in this role, such as:

    Karolinska Institute used CDRs to locate people after the Haiti earthquake and cholera outbreak in 2010 Carnegie Mellon University and, Harvard used CDRs to track malaria and people’s mobility, and reported the study in Science A Telefonica Research team found that in Mexico’s 2009 swine flu epidemic, alerts didn’t reduced mobility, but closing workplaces did.

    It seems that these types of initiatives aren’t being used to combat Ebola in West Africa. Regulation and privacy are inhibitors. Groupe Speciale Mobile Association (GSMA) has developed standards and legal codes to facilitate access to and use of CDRs. The initiative hasn’t been converted into action by governments requiring mobile operators to provide their CDR data to researchers. It’s a bit surprising because The Economist says that Orange has done it in Senegal and Ivory Coast for years as part of the Data for Development initiative.

    Just accesing CDRs isn’t enough. It needs analysing by highly skilled analytics staff to see what the CDRs reveal, rather than just by researchers: a common capacity constraint for most types of eHealth and analytics in African countries.  In the meantime, it’d be good to see some progress.

  • What is Big Data?

    Like eHealth, Big Data is an expression that seems to crop up in several guises. Is there a common definition, or is it being misused?

    Wikipedia’s a start point. It defines it as “a blanket term for any collection of data sets so large and complex that it becomes difficult to process using on-hand data management tools or traditional data processing applications.”

    The MIT Technology Review, a Massachusetts Institute of Technology website, says it’s a conundrum shown by the survey by Ward and Barker at the University of St Andrews in Scotland. They identified six definitions:

    Gartner: an initial three-fold definition was three Vs: Volume, Velocity and Variety with a fourth subsequently added for Veracity reflect trust and uncertainty. Oracle: derivation of value from traditional relational database-driven business decision making, augmented with new sources of unstructured data. Intel:  where organisations generate a median of 300 terabytes of data a week, with the most common forms of data being business transactions stored in relational databases, then documents, e-mail, sensor data, blogs, and social media. Microsoft: the process of applying considerable computing power that included modern machine learning and artificial intelligence, to massive volumes and often highly complex sets of information. The Method for an Integrated Knowledge Environment (MIKE): the open-source project says it’s not a function of the size of a data set but its complexity, so it’s a high degree of permutations and interactions within a data set. The USA’s National Institute of Standards and Technology (NIST): data that needs more than the capacity or capability of current or conventional methods and systems.

    Ward and Barker says Big Data describes the storage and analysis of large and or complex data sets using a series of techniques including, NoSQL, MapReduce and machine learning.

    Forbes has a report by Arthur saying that Big Data is a collection of data from traditional and digital sources inside and outside an organisation’s as a source for continuous discovery and analysis. She distinguishes unstructured and multi-structured data. She also says that an organisation can “add its own individual tweaks here or there.”

    That means that there’ll be no agreed definition for some time to come. In the meantime, African countries can carry on with their own ideas.

  • Google to healthcare: we know what we're good at

    I wonder what Google’s healthcare enthusiasts make of the recent comments by Google co-founder Sergy Brin. He said, “Generally, health is just so heavily regulated. It’s just a painful business to be in. It’s just not necessarily how I want to spend my time. Even though we do have some health projects, and we’ll be doing that to a certain extent. But I think the regulatory burden in the U.S. is so high that think it would dissuade a lot of entrepreneurs,” according to David Shaywitz’s Forbes article, aptly titled “Google Co-Founders to Healthcare: We’re Just Not That Into You.”

    Fortunately the world is bigger than the US and its regulations. It’s different in other countries, and it can change, and there is plenty of evidence that Google is interested in healthcare as a market for its technology.

    Based on a small sample of eHNA articles we know that Google’s passionate about fitness, is the heartbeat of some wearables, has developed a glucose-testing contact lens for diabetics, owns robots that could help healthcare, is using Google Glass in surgical procedures, is also in emergency rooms with Google Glass, and has an EHR for Google Glass. Google tracks influenza outbreaks, has joined a research team to tackle autism and has setup Calico, a new company to focus specifically on health.

    It’s an impressive list of health innovation by anyone’s standards. It sounds like a company that’s interested in healthcare, but not in providing it. Perhaps it’s an opportunity for African countries to attract the Google projects that are already doing so much eHealth and invariably want to do more.