• Analytics
  • Predictive analytics needed for better infectious disease tracking

    When outbreaks of new diseases emerge, public health’s impact inevitably follows events. Eventually, it catches up. The US Government Accountability Office (GAO) report in May 2017, Emerging Infectious Diseases Actions Needed To Address the Challenges of Responding to Zika Virus Disease Outbreaks, found that the Zika virus case counts from the national disease surveillance system underestimate the total number infections because: 

    Infected people may not seek medical care because they have only mild or no symptoms, or other reasons,May not be diagnosed because of limitations in Zika virus diagnostic testingIncomplete surveillance reporting.                                                                                                                                   

    Better, prompt and accurate information’s still needed. Three Congress representatives have written to the GAO boss, the Comptroller General, suggesting a subsequent study into predictive models and systematically integrating modelling into outbreak responses. They think the US can respond more effectively bt learning if:

    Federal agencies use predictive modelling to inform planning for emerging infectious diseases?How federal agencies use models to inform regulatory decisions about infectious disease outbreaks?Do medical product sponsors use predictive modelling?What funding’s available for infectious disease predictive modelling?What challenges do predictive modellers face?If and how, federal agencies validate models’ predictions? 

    With predictive modelling and analytics expanding their potential, these seem like a good set of questions that all countries should ask. Answers can inform eHealth strategies and strengthen the role and impact of public health professionals.

  • Machine learning use cases for health points to the future

    Machine learning (ML) and artificial intelligence (AI) have quickly rocketed to the top of the industry’s buzzword list, driven partly by heightened interest in big data analytics amongst healthcare providers and vendors

    The allure of intelligent algorithms to mine masses of structured and unstructured data for innovative insights get’s health planners pretty excited. However, a fragmented health ICT landscape and sluggish analytics development have thus far kept that Holy Grail beyond reach.

    Regardless, ML is already making a difference.  Here are some examples;

    Imaging analytics and pathology

    ML can supplement the skills of human radiologists by identifying subtler changes in imaging scans more quickly and potentially leading to earlier and more accurate diagnoses.  At Stanford University, ML tools performed better than human pathologists when distinguishing between two types of lung cancer.  The computer also bested its human counterparts at predicting patient survival times.

    Natural language processing and free text data

    Using natural language processing (NLP), ML algorithms can turn images of text into editable documents, extract semantic meaning from those documents, or process search queries written in plain text to return accurate results.  Anne Arundel Medical Center is using a natural language interface, similar to any of the widely known internet search engines, to allow users to access data and receive trustworthy results.

    Clinical decision support and predictive analytics

    Identifying and addressing risks quickly can significantly improve outcomes for patients with any number of serious conditions, both clinical and behavioral. The University of California San Francisco’s Center for Digital Health Innovation (CDHI) and GE Healthcare are creating a library of predictive analytics algorithms for trauma patients in an attempt to speed up the delivery of critical care.

    Cybersecurity and ransomware

    At the end of 2016, IBM Watson launched its Cyber Security Program.  Watson’s ML and cognitive computing skills are used to flag cyber threats and check for suspicious activity against known malware or cyber crime campaigns.  This helps IT staff take better decisions based on known characteristics of malware.

    ML and AI are the keys to addressing health care inadequacies.  These technologies can help predict and control disease, expand and augment service delivery, and address several persistent social inequities. Ubiquitous health tech is by no means inevitable.  Successful rollout will entail an immense amount of concerted effort, capital, labor, and partnership.

  • Is machine learning the new buzzword for healthcare?

    By now, it’s old news that big data will transform healthcare. Electronic health records and health information systems have arrived, data flows, and there’s a lot of it. However, all this data, is only useful when it has been analyzed, interpreted and acted on. So will it be algorithms that perform this analysis that will really transform healthcare?

    Access to lab results via a mobile app, has helped clinicians diagnose and treat patients faster. Imagine how much more useful these results would be if they also showed the patient’s risk for cardiovascular disease or renal failure, based on the last several years of the patient’s lab reports. This is where machine learning might help physicians to make better decisions at point of patient care.

    Machine learning is an area of artificial intelligence (AI) that is starting to attract interest in healthcare. It is a set of algorithms that help a system to automatically learn and predict outcomes, after continuous exposure to variable datasets. The value of machine learning in healthcare is its ability to process huge amounts of clinical information, beyond that of human capability, and then reliably convert analysis of that data into clinical insights. This will help physicians plan better and ultimately lead to better outcomes, lower costs of care and increased patient satisfaction.

    Machine learning is already making headlines in healthcare. Google has developed a machine learning algorithm to help identify cancerous tumors on mammograms. Stanford is using a machine learning algorithm to identify skin cancer. A JAMA article, last year, reported the results of a deep machine-learning algorithm that was able to diagnose diabetic retinopathy in retinal images. Others, like Philips, are transforming TB screening, using machine learning algorithms that can offer an objective opinion to improve efficiency, reliability, and accuracy.

    Machine learning puts a new arrow in the quiver of clinical decision-making.

  • EU’s BigData@Heart aims to improve heart disease treatments

    A report on Cardiovascular Disease (CVD) published by Springer says CVD prevalence in sub-Saharan Africa’s increasing. Limited access to prevention and continuing care are seen as constraints to improvements. The EU’s BigData@Heart project may contribute to the development of treatments for heart disease patients. It has lessons for Africa’s health systems.

    It’s a large-scale, five-year, €19 million project. Its aim’s to use data and advanced analytics to develop a translational research platform of phenotypical resolution to improve patient outcomes and reduce societal burdens of atrial fibrillation (AF), heart failure (HF) and acute coronary syndrome (ACS). Data sources include real-world evidence, best-practices in drug development and personalised medicines.

    Four outputs are:

    ·         New universal, computable, definitions of diseases and outcomes relevant for patients, clinicians, industry and regulators

    ·         Informatics platforms linking, visualising and harmonising data sources, completeness and structures

    ·         Data science techniques to develop new definitions of disease, identify new phenotypes, and construct personalised predictive models

    ·         Guidelines that allow for cross-border use of big data sources acknowledging ethical and legal constraints and data security.

    It can have considerable value for Africa. The continents health systems and cardiologists could move their CVD services ahead in the wake of BigData@Heart’s progress. 

  • EU’s BigData@Heart aims to improve heart disease treatments

    A report on Cardiovascular Disease (CVD) published by Springer says CVD prevalence in sub-Saharan Africa’s increasing. Limited access to prevention and continuing care are seen as constraints to improvements. The EU’s BigData@Heart project may contribute to the development of treatments for heart disease patients. It has lessons for Africa’s health systems. 

    It’s a large-scale, five-year, €19 million project. Its aim’s to use data and advanced analytics to develop a translational research platform of phenotypical resolution to improve patient outcomes and reduce societal burdens of atrial fibrillation (AF), heart failure (HF) and acute coronary syndrome (ACS). Data sources include real-world evidence, best-practices in drug development and personalised medicines. 

    Four outputs are:

    New universal, computable, definitions of diseases and outcomes relevant for patients, clinicians, industry and regulatorsInformatics platforms linking, visualising and harmonising data sources, completeness and structuresData science techniques to develop new definitions of disease, identify new phenotypes, and construct personalised predictive modelsGuidelines that allow for cross-border use of big data sources acknowledging ethical and legal constraints and data security.

    It can have considerable value for Africa. The continents health systems and cardiologists could move their CVD services ahead in the wake of BigData@Heart’s progress.

  • Rothman Index predicts patients’ increasing risks

    It’s often reassuring to hear that hospital patients are in a stable condition.  When they’re deteriorating, it’s not so good. It’s even worse when the deterioration’s a set of marginal steps that are challenging to find, but lead to catastrophic states. This was the motivation for the Rothman Index (RI), a predictive health analytics tool.

    Florence Rothman was diagnosed with aortic stenosis, narrowing of the exit of her heart’s left ventricle. She had a low-risk surgical procedure and seemed to be recovering, then became weaker and started a slow, steady, subtle decline that detected when her condition became critical. Relevant data was recorded in her EHR, but the trends weren’t easy to see, so wasn’t used by the skilled and caring healthcare professionals. She was discharged, and four days later, collapsed and died in the ER.

    Michael and Steven, Florence’s sons, one an engineer, the other a scientist, both skilled in data analysis, were inspired reveal EHRs’ crucial insights to improve healthcare. They created the Rothman Index (RI), a statistically validated patient acuity score across all diseases and conditions. It presents patients’ real time conditions and can be trended and visualised, alerting doctors and nurses of deterioration before it’s critical.

    Pera Health, formerly Rothman Healthcare Corporation until 2012, provides RI. It includes graphical user interfaces so healthcare professionals can visualise trends in health status from patients’ data in their EHRs. Regularly updated health scores are derived from vital signs, nursing assessments and lab results. The model transforms each input into a common representation of univariate risk, enabling heterogeneous data to be summed, solving the data fusion problem. Outputs are continuous measures of patients’ conditions integrated into their EHRs. Trends enable deteriorating and vulnerable patients to be identified, often with less than24 hours warning, and with minimal false alerts.

    The company says RI correlates well with:

    24 hour mortalityUnplanned transfers to ICUsICU readmissionsCode Blue events to for cardiac or respiratory arrests.Readmissions within 30 days of dischargeLengths of stay.

    RI can be a part of Africa’s EHR programmes to build predictive health analytics into hospitals; routines. It’ll help to maximise their EHRs’ benefits.

  • eHealth’s wide data range can benefit from embedded analytics

    Assembling data from eHealth’s increasing range is a considerable challenge. Business Intelligence (BI) and its analytics can help. It’s also moving on. An article in Business Intelligence Best Practices  by Wayne Eckerson describes how. It’s a trend that Africa’s health systems need to follow.

    He’s a BI thought leader, author and founder of Eckerson Group, a research and consulting company specialising in BI, analytics and Big Data. Two of his books are Performance Dashboards: Measuring, Monitoring, and Managing Your Business, published by Wiley, and The Secrets of Analytical Leaders: Insights from Information Insiders, published by Technics Publications.

    His view of BI’s next wave’s the enhancement towards embedded analytics. It’s part of a requirement that BI tools must be easier to use if they’re to fulfil their potential and provide real time, constructive insights into organisations’ activities business events. Embedding analytics into the operational applications and processess’ the best way to simplify and operationalise it. So far, most embedded analytics have barely scratched the surface its possibilities.

    The concept’s not new. Java and .NET developers sometimes create new reporting capabilities when they build custom applications. Portals can display charts and tables generated by reporting tools inside portal windows. Microsoft Office maintains live links to reports on BI servers. Online applications embed predictive models to score customer transactions in real time to detect fraud, cross-sell products, evaluate risk and assess credit worthiness.

    These analytics aren’t sophisticated. They display canned views of existing reports with limited ability to drill down, publish views in other formats or compare outputs with other data. Compelling analytically-driven applications are used by rich leading-edge companies that have big teams of skilled developers who can code, debug, and test monolithic applications that are time-consuming and expensive to build and more costly to modify.

    BI’s future’s leading to a blurred distinction between analytical and operational applications. Embedding analytics’ techniques will make it as easy as dragging and dropping objects onto a workbench. Instead of stand-alone BI toolsets, users will leverage embedded BI functionality as an integral part of applications. It’ll also be used in reverse, to launch operational processes and tasks. 

    Two types of workbenches available for the analytically minded developer are:

    Traditional Interactive Development Environments (IDEs), such as IBM Rational, Microsoft Visual Studio.NET and SAP Visual ComposerAdvanced Development Environments (ADEs) are similar to IDEs, but often enhance development so users and developers, can rapidly prototype and build applications. 

    Africa’s eHealth strategies need to find a place for these trends. It’ll include a combination of tools, skills, recruitment and retention.

  • Healthcare organisations need several approaches to analytics

    Data needed for business and healthcare analytics can be complex. A white paper from Sisense Business Analytics and the Data Complexity Matrix shows four components:

    Data size

     

     

    Large    

    BIG

    COMPLEX

    Small

    SIMPLE 

    DIVERSIFIED

     

    Few        

    Many

    Number of Sources as Data Tables

    Analytics for each of the four types have different requirements. Using simple data can rely on access to databases and visualisation tools. Big data needs data warehouse and appliance tools to step up analytics to this next level. It requires resources for data preparation and aggregation, and to configure, maintain, and support the tools and their integration with other parts of the analytics map.

    Diversified data adds to the sophistication. Extract, Transform Load (ETL) become more complicated and may need ICT teams input to set up and structure data and analytical tools before generating analytic output. The workload may be so laborious that specialised ETL tools are needed to automate data preparation.

    Complex data needs specialised skills and resources. Costs are high, comprising manpower costs that increase with each data preparation or ETL activity and the need for sophisticated tools.

    As Africa’s health systems move towards more analytics, they’ll need to set up increasingly sophisticated methodologies and develop people with specialised skills. Sisense has a free test drive of its analytics tool that provides an insight to a health analytics journey.

  • Smart dashboards are essential for eHealth benefits

    Realising benefits depends extensively on maximising the number of users. It also depends on them using the data effectively. This, in turn depends on meeting their requirements. There are two main parts to this, the information they need and having in a format, style and presentation that they can use for faster decision taking.

    Tableau, a dashboard supplier, has a white paper saying there are four main ways to use data to improve healthcare:

    Using analytics for better  population health managementUsing real-time analytics to increase productivityAggregating and blending data to reveal and fix supply chain inefficienciesAutomating ad hoc visual analysis for better revenue cycle management.

    Providing more data doesn’t always help. The first step’s to simplify data that’s already available. It might easier said than done. In a hospital organisation, there can be a thousand or more health workers. Common themes for simplification include: 

    Use data visualisation so users can quickly automate processes rapidlyEnable users to visualise and assimilate data the way their minds workHelps users see and understand their healthcare data no matter how big it is, or how many systems it is stored inConnect quickly to any data, analyse it and share insights to reveal opportunities to benefit patients, health workers and healthcare organisations.

    As Africa’s eHealth expands, it’s vital that these concepts are in place too. Maximising eHealth’s benefits depend on it. 

  • Keep health analytics simple, for now

    Claims for Big Data’s and analytics’ benefits are considerable for commercial enterprises. For health and healthcare, it seems a bit more complicated, and sometimes bewildering. Where and how to start is a challenge.

    In an article in Health Analytics, Dr Danyal Ibrahim, emergency physician and Chief Data and Analytics Officer at Saint Francis Care in Connecticut, invokes the wisdom of the ancient Chinese philosopher Lao Tzu, who famously said “A journey of a thousand miles begins with a single step.”

    He sees the first step towards health analytics as a daunting leap of faith. Succeeding depends on:

    Clinical leadership and championsThe right partnershipsA strong, diverse Big Data analytics teamCreate an analytics infrastructure that delivers actionable, meaningful data to points of careIdentify and break down data silos and develop a Big Data roadmap.

    The last point holds the key to creating a streamlined data analytics infrastructure that reveals comprehensive and meaningful patients’ stories. These lead on to help make better and timely decisions. While healthcare data’s supposed to be connected around individual patients, it isn’t. It ends up in different siloes, creating a big barrier to using data to improve care.

    From the streamlined data, the next steps are:

    Bring all the data stewards togetherRedesign the analytics team to focus on value for better care, more cost effective care and better patients’ experiences for patients

    Effective analytics teams comprise a wide range of skills. They include, clinical, financial analytics, SQL development, data warehousing and data science, including statistical modelling and natural language processing. For Africa’s health systems, all these skills and knowledge may not be readily available. Analytics teams will need structured, integrated and technical development, so a sustainable training budget.

    Lao Tzu’s proposed first step of analytics costs has a second step of producing valuable outputs that healthcare teams can use for measurable, proven patient benefits. Analytics will then justify itself.