Big Data (27)

As a relatively new part of eHealth, Big Data has a negligible effect on Africa’s health systems and eHealth programmes. Big Data insights are in Chapter 8 of the WHO and Global Observatory for eHealth (GOe) publication eHealth Report of the third global survey on eHealth Global diffusion of eHealth: Making universal health coverage achievable. WHO Global Survey 2015 provides the data source for the report.

It hasn’t taken off globally yet. Fewer than a fifth of countries say they have a national policy or strategy for regulating Big Data in health and healthcare. In Africa, it’s about 2%. This Big Data deficit isn’t much of a cause for concern. As the eHNA posts about WHO’s report show, Africa’s health systems have many other eHealth priorities. One that wasn’t included in WHO’s survey is stepping up cyber-security. Acfee’s report from its African eHealth Forum (AeF) our priorities include  cyber-security and others, such as Interoperability (IOp), cloud computing, eHealth governance, regulation and capacity building are well ahead of Big Data.

WHO found that a lack of integration, privacy and security are major barriers to Big Data adoption. It’s constructive that Africa’s health systems are focusing on these as part of their expanded eHealth initiatives. Acfee’s activities in 2017 will support them.

Africa’s not yet investing much in Big Data compared to the global trend. It’s revealed in the WHO Global eHealth Survey 2015. An article in Healthcare Informatics, by Joe Mario, founder of Healthcare Integration Strategies, sets out two requirements for success that Africa’s health systems should adopt:

  1. Capture useful data in the proper format needed
  2. Have the analytic tools needed to extract meaningful information.

He also implies that healthcare organisation’s need to set up a data mining unit. He could’ve added a fourth requirement, having people with the analytic skills to design and run the data capture and use the analytic tools effectively. All four Big Data components are in short supply across Africa.

He sets out two organisational options. One is a centralised Big Data unit of analytic specialists supporting healthcare teams. The other’s decentralised, where each healthcare provider has its own team. For Africa, a decentralised model seems unrealistic until enough analytic specialists are available and affordable.

Setting up a centralised model needs a long-term plan to recruit, develop and retain analytic specialists. Success depends on health ministries working closely with universities to ensure a steady stream to build up a steady stream of health analytics specialists. The pace and scale needs to match the equivalent endeavour for health informatics specialists and the availability of data sources such as EHRs.

There are many other competing demands for these eHealth resources and tight affordability constraints. These combine to create a scenario where Africa’s Big Data programmes are likely to be relatively small, highly focused and provided by small, centralised teams. A big step forward’s including something like this in the next round of eHealth strategies.

Africa’s health systems will always need to adopt modern eHealth for disease surveillance, especially epidemics like Ebola and Zika. The recent WHO eHealth survey reported in eHNA found that Africa’s health systems are well behind in their Big Data and analytics strategies. Most countries have no Big Data strategy. A report on the Information Age shows why they have to catch up.

It says intelligent uses of Big Data in healthcare are helping to predict epidemics, find cures for disease, improve quality of life and avoid preventable deaths. These are big impacts, and even more important as Africa’s population increases and ages. Big Data can help to develop new and changing models of healthcare treatment and delivery.

Big Data blood supply, its data, is expanding rapidly. More data’s been created in the last two years than in the entire previous history of the human race. It’s not stopping. By 2020, about 1.7 megabytes of new information will be created every second for every person on the planet. By then, data’s accumulated digital universe will’ve grown from 4.4 zettabytes today to around 44 zettabytes, or 44 trillion gigabytes, a ten-fold increase in just four years.

Whether it’s the megas, zettas or trillios is less important than adopting, developing and using Big Data for its benefits. Understanding more about individual patients as early in their life as possible offers opportunities to find warning signs of serious illness early enough so treatments can be easier, simpler, cheaper and more successful than later diagnoses.

Big Data’s not a simple solution. It has several challenges. Storing and managing needs big archives to preserve large volumes of data for future research. Security and compliance challenges run alongside this. Ben Rossi, the article’s author suggests that pathology services offer a good way for healthcare organisations to deal with this.

Pathology services are regularly adopting new digital technologies that offer new Big Data opportunities. eHealth’s helping pathology labs to benefit from digital workflows that foster innovation that can transform patient care. A modest pathology lab with one small slide scanner will generate over 15 terabytes of data a year. 

It could be a way into Big Data for some African health systems. When will all African countries start on Big Data?

Once patients are back home, complying with medication regimes seems an continuing intractable challenge for many of them. SMSs have played a valuable role. Now, analytics has emerged on the scene.

A report in Health Analytics says Sanofi has partnered with Duke Clinical Research Institute (DCRI), Massachusetts General Hospital and its Center for Assessment Technology and Continuous Health (CATCH) to improve medication adherence rates for Type 2 diabetes patients. The team’ll use Big Data analytics tools developed by DCRI and CATCH. 

Using all available data, including demonstrated behaviour, is seen as an innovation to direct personalised care and engagement programmes and practical tools and services so people with diabetes can engage more proactively with their treatment. Better outcomes are the planned benefits. They’ll be achieved by healthcare provider anticipating medication adherence rates for individual patients, then intervening as needed with appropriate actions. 

The project’s part of a trend. Medical researchers are relying increasingly on predictive analytics to improve chronic disease management, especially medication compliance. Failure to comply has a high costs. In the USA, it’s estimated that it cost the healthcare industry approximately US$337b in 2013. It’s also associated with unsatisfactory population health management programmes.

A data challenge’s that healthcare providers are reluctant to integrate data about medication adherence into EHRs, mainly due to data overload. There’s also some dissatisfaction with value of patients’ data from pharmacies and health insurers. 

A solution’s seen as semantic data analytics tools. Healthcare providers and Big Data scientists are relying more on semantic databases for comprehensive insights on population health management. They draw from different data sets, such as patient information, socio-economic data and pharmacy information to produce more accurate and reliable predictions about patients’ outcomes. These are from connections of different concepts rather than displayed rows of specific information.

As US healthcare strives to move Big Data ahead, Africa’s health systems can begin to develop their Big Data strategies. WHO’s latest eHealth survey, reported on eHNA, showed a marked lack of progress.

Big Data offers big promises. Even if it’s not true, Big Data does offer new opportunities. eHNA has reported that WHO’s 2015 eHealth Survey found that out of 33 African countries, only one’s started to deal with Big Data.

IBM Watson has released its white paper Social and behavioural determinants of health A look at fundamental drivers of health and disease to help improve population health and reduce costs. It has valuable constructs that can help Africa’s health systems build their Big Data and analytics strategies, plans and services.

It proposes a shift in the approach to Population Health Management (PHM) from a context of patient populations to communities where people live, work and play. Many factors comprising Social Determinants of Health (SDH), such as employment, healthy food and people’s physical environment are beyond doctors’ and hospital’s control, but healthcare organisations can work with communities to help patients deal with SDH factors that affect their health and access to healthcare. A current limitation to progress is knowing how to do it.

Explaining SDH

The white paper explains:

  • What SDHs are
  • How they influence people’s health
  • What healthcare organisations can do to improve population health
  • How behavioural health affects physical health
  • Why it should be integrated with primary care
  • A new concept of service delivery integrating healthcare with social services
  • The kind of eHealth needed to support these.

PHM models 

Five are set out as:

  • Model 1: targeting health behaviours where doctors and healthcare teams try to induce
  • patients to modify their health behaviour
  • Model 2: referral to community services, especially community health centres care for people o lower incomes
  • Model 3: limited social support in healthcare frameworks
  • Model 4: patient-centred medical homes that emphasise SDH
  • Model 5: holistic care management that integrates healthcare and social services.

Advantages of integrating healthcare and social care

It’s shown to produce measurable benefits. They include:

  • Twice as effective at treating depression
  • Improves physical and social functioning and quality of life
  • When behavioural health is engaged, healthcare organisations seeking to address SDH must work with social services and other community resources
  • Reduces overall healthcare costs.

Information sources

An integrated service needs a wider range of information than segmented services. This’s where Big Data’s needed. The range of sources is considerable, and includes:

  • Social services adult and child teams
  • Functional assessment surveys, even though they’re currently uncommon in mainstream healthcare
  • Health insurance databases
  • EHRs and administrative systems
  • Places of residence
  • Administrative records
  • Distances to nearest clinics, health centres and food stores from geographical information systems
  • Living situations, such as living alone, from social service agencies and public records
  • Physical disabilities, from EHRs and disability insurance records
  • Employment status from unemployment insurance and social security records
  • Environmental hazards, such as poor housing from building reports
  • Air quality from weather reports
  • Diet and exercise from health risk assessments
  • Medication compliance, from prescription records and mHealth reminders.

As African countries are still at the Big Data starting gate, planning to achieve all these isn’t realistic. A more modest approach’s needed to make a start, and communities with potentially high levels of SDH or poor access to healthcare may offer the best returns. These could be urban populations living in poor housing and remote rural communities. From a small start, Africa’s health systems and can learn and grow their Big Data skills, priorities and benefits.

Corporates are increasingly harnessing the power of Big Data and analytics to improve productivity, gain market share and a competitive edge. While many have embraced Big Data and its benefits, there are sectors that lag behind by not fully capitalising on its potential to produce actionable insights says an article in IT-Online. Healthcare still has along way to go. 

According to Yudhvir Seetharam, head of analytics at First National Bank (FNB) Business, global business challenges and developments put more pressure on sectors that are sensitive to economic cycles. As a result, Big Data’s insights can help solve some of the biggest business challenges. 

He says there are five sectors in South Africa that should consider incorporating Big Data analytics into their operations. They’re:

  • Agriculture
  • Financial services
  • Healthcare
  • Retail
  • Franchising.

Healthcare’s always looking for new ways to increase accessibility and affordability while maintaining costs. Big Data can play a big role in making that happen. It can be used to determine the best geographic location for new hospitals, and provide insight into trends and potential solutions in medical research. Identifying and predicting epidemics, help in finding cures for diseases and an overall increase in the quality of care can be extracted from Big Data and its analytics.

Big Data uptake in many African countries has been slow for healthcare. It’s not that governments or hospitals don’t relealise the benefits but the fact they have a long list of challenges to address before they can use it. Big Data and analytics are also competing for very limited resources and many countries struggle to justify the cost when they desperately need the money to buy more drugs or hire more staff.

Although several implementation challenges remain for African countries, including limited resources and basic infrastructure shortfalls, healthcare may be coming to the point where it is not longer a matter of if, but rather when. The potential for using Big Data seems too big to miss out on. The challenge is integrating into a long list of other eHealth priorities, especially the long list of long-standing challenges identified in Advancing eHealth in Africa.

Big Data’s still a young methodology. It’s reasonable to expect that its initial potential will convert into wide-scale benefits.

A recent post in the Wall Street Journal (WSJ) highlighted Big Data’s potential to bring about revolutionary advances in finding effective cancer treatments. It describes analysing vast genetic and clinical data from hospitals and doctors to personalise cancer prevention and treatment based on the genetic characteristics of patients’ tumors, family history and other unique characteristics. The result is to minimise unwanted treatments and their often serious side effects.

In Africa, an article in Fortune highlights using Big Data in Mozambique to improve HIV testing services for infants to allow timely initiation of life saving anti-retroviral therapy. Investigators found that by analysing enough data from shipments between clinics and laboratories, they could reorganise testing facilities and their locations to speed up HIV test results. The result’s increased efficiency in processing tests.

Big Data has considerable potential to improve other population and patient outcomes, so ensure more healthy Africans. Effectively mining the vast insights hidden in Big Data to improve our health remains an exciting frontier to follow closely. Initiatives such as IBM's Watson fighting Ebola or mobile phone data to track rubella in Kenya are making a start. African countries can build on the early gains to move into a position where they can exploit Big Data for their own health transformation.

Ever heard of a quintillion? No, neither had I until I came across this piece by IBM. It says that every day we create approximately 2.5 quintillion bytes of data. IBM believes it’s a lot. I looked up quintillion to get a feel for how much we’re talking about.

By the UK definition a quintillion is a million to the power five, or a number with 30 zeros after it. It’s also used in English literature to mean a number so large as to exceed normal description. I guess that was before IBM did its number crunching. For American readers, there is a difference in the definition when one crosses the Atlantic. The US regards a quintillion as a number followed by 18 zeros.

It’s why there is a Big in Big Data and looks like a reasonable judgment call regardless of which definition you chose.

IBM doesn’t clarify which definition they’ve used, which is a pity, but they do have a more practical concept to illustrate the accelerating speed of data production. Apparently, “90% of the data in the world today has been created in the last two years”.

This data comes from all kinds of places, particularly anything online and everywhere you could imagine you might put a sensor. That includes social media, purchasing transactions, digital pictures and videos, feeds from devices of all kinds, including mobile phones, and for everything, it includes their GPS signals.

The Big Data concept includes structured and unstructured data and implies such an astronomical size that it’s beyond the ability of commonly used data processing software tools to make sense of it. It’s high volume, high velocity, highly varied and can be complex to transform into meaningful insights.

One definition of Big Data adds that it needs “a set of techniques and technologies with new forms of integration to reveal insights from datasets that are diverse, complex and of a massive scale”. Thankfully that is changing, with a growing number of tools, such as Apache Hadoop, machine-learning Spark, NoSQL database tools and others, available to tame Big Data to allow big utility.  

In healthcare this helps improve disease surveillance, outbreak modelling and predictions, predictive disease management, medication adherence management, prescription refill management, population health analytics, optimising medical insurance and planning emergency services.

For those of us in African countries it’s an opportunity to move into a position where we can exploit Big Data for our own health systems strengthening and transformation.

Now that we’ve reached a few quintillion, I’ve been wondering what’s next. Fortunately there’s no need to worry. We’ve lots more big, fun numbers to learn, such as a “Googol”, which is ten-to-the-power-of-one-hundred, and many more beyond.

As management gurus go, Tom Peters, the American author of the famed book In Search of Excellence, is in the premier league. He said that “The whole secret to our success is being able to con ourselves into believing that we're going to change the world because statistically we are unlikely to do it.” Where does this leave Big Data users in healthcare?

Dr Eric Schadt is the founding director of the Icahn Institute for Genomics and Multiscale Biology at New York’s Mount Sinai Health System. He’s been interviewed about Big Data by Sastry Chilukuri, a principal in McKinsey’s New Jersey office. His view is that medicine can build better health profiles and better predictive models for individual patients to gain better diagnose and treatments.

It’ll help to overcome a major limitation with medicine and the pharmaceutical industry: the understanding of disease biology. “Big data comes into play around aggregating more and more information around multiple scales for what constitutes a disease—from the DNA, proteins, and metabolites to cells, tissues, organs, organisms, and ecosystems. Those are the scales of the biology that we need to be modelling by integrating big data.”

It won’t happen overnight. Models will have to evolve, and eventually, they’ll become more predictive for given individuals. mHealth is contributing to the trend.

Schadt sees wearables and engagement through mHealth as the future, and not just for research into diseases, but medical practices. Information stored in medical records is not extensive enough, so wearables are starting to offer ways to monitor patient’s longitudinally and across individual’s numerous health states. The result’ll be better and more accurate health profiles, measurements of any deviations from baselines that may predict a disease state, or an emerging state, enabling healthcare professionals to intervene sooner.

Achieving this needs wearables to continue their development trends that are changing the balance from recreation to research and clinical roles. This will be in parallel to the possibilities for Big Data and predictive analytics.

It offers an added, parallel track to Africa’s health systems current trends in mHealth expansion. This combination may be Big Data’s big step forward for the continent.

eHealth can help health systems be more proactive in improving health and populations’ healthcare. A report from Caradigm, a global company specialising in Population Health Management (PHM) and analytics, and available from Health Data Management, sets out several examples from a team of contributors of how health ICT and analytics can help. It provides a valuable checklist for Africa’s health systems and their PHM initiatives. 

Examples of how ICT can support PHM:

  • Aggregating community-wide information in real time
  • Stratifying a defined population by risk using predictive modelling
  • Co-ordinating and managing patient care
  • Delivering information needed by clinician workflows
  • Enable personalised healthcare
  • Identify courses of action and interventions
  • Right care for the right person
  • Changing behaviour to lead to better health
  • Better, more informedions
  • Improved patient engagement leading individual health gains
  • A shared, interoperable, virtual EHR’s an essential foundation for PHM
  • Big Data and analytics capabilities to profile populations and evaluate their financial and clinical risks
  • Facilitate benchmark measurement.

How analytics will play a part in PHM: 

  • Distinguish between predictive analytics that identifies what will happen, when and why, from prescriptive analytics that define what we do and what the effects are
  • Develop a better understanding of healthcare enterprises and the patients they serve
  • Reduce unintended variations in healthcare and promote best practices
  • Identify high-cost, high-risk individuals needing short and long-term interventions, and their receptivity to interventions
  • Pinpoint where programmes can have the most impact and can inform the development of incentive programmes
  • Leverage predictive modelling to understand where patients are on their care pathways, where they may be going, and what behaviours and interventions could help
  • Create reports that can help organisations understand where their high cost services are, what interventions are effective, who’s participating and overall success
  • Use analytics to identify opportunities and drive appropriate action that has the greatest impact, such as improving quality or minimising costs
  • Analytical strategies need three elements:
    • Highly accurate data powered by integrated clinical and claims data across the continuum of care
    • The ability to predict at-risk patients to focus on which can be appropriately managed to reduce preventable costs
    • Comparative clinical benchmarks
  • Better strategies based on analyses of patients’ current clinical status, healthcare needs, future needs and the costs per unit of clinical values created of a population or individual patient allows healthcare.

Adopting eHealth initiatives that contribute to these types of goals enables Africa’s health systems to move rapidly beyond the more traditional roles ascribed to eHealth. Going further expands eHealth benefits.