ICTworks

Digitization has revolutionized range of services, especially data systems in Low-Middle Income Countries. As highlighted in USAID’s report, Artificial Intelligence in Global Health, presents the many promises of digitization from tools to improve data collection at the point of care to application of artificial intelligence (AI) in global health.

The commentary in the Lancet Digital Health on Big data and health by Snyder and Zhou), highlights that improving individual health is complex and requires data from multiple aspects of individuals life that can then be used for construction of personalized health models and with personal health dashboards.

Termed as the “data revolution”, multiple articles by, Steinhubl SR, Muse ED, Topol EJ; Li X, Dunn J, Salins D, et al. and Dunn J, Runge R, Snyder M, discuss that with rapid expansion in use of direct-to-consumer devices such as smart watches and blood pressure cuffs and other wearable devices, more and more health data is being collected and use in discreet systems.

World Health Organization’s blog on the determinants of health, discusses that to a large extent, factors such as where we live, the state of our environment, genetics, our income and education level, and our relationships with friends and family all have considerable impacts on health.

The question is, how can we design future data systems in the digital age that leverage of ability to do such powerful analysis and design better programmes?

Consider a potential future scenario

A child is born. As the child grows, s/he needs a national ID, receive immunizations, nutrition monitoring and action, education, other health services, voting registration (at appropriate age), employment, get married (or not), move from one place to another, access to social protection, insurance, and many other services throughout the entire life-time.  Eventually get old and die (at some-point) for which knowing the cause of death is essential to plan for improvement.

Currently, these groups of services are provided by different sectors or Ministries. Every sector has its own data recording and management system that are designed on a mix of software applications. When individuals move from one service to another there is duplication in collecting basic information and at the end the data is available in an analytic-ready format. At the end it is difficult to track the individual from birth to death and across the continuum of services during their course of life.

Existing ID schemes are often limited to a sector or program area. On one hand this results in increased burden of information collection, as basic individual’s data eg, name, age, address, family information is collected multiple times. On the other, the potential to support care at multiple points of service by connecting disparate care episodes into a longitudinal, person-centric shared health record is not achieved. This can also improve efficiencies in information flows and reduce the magnitude of the steps in our cascade of care.

How do we benefit from the data revolution?

Building on The World Health Organization’s: Framework on integrated, people-centered health services (sixty-ninth World Health Assembly, 15 April 2016), a person-centered data system is a vision that aims to build upon distributed data system with appropriate linkages (interoperability) across them. The rapidly expanding use of digital tools to support implementation of Civil Registration and Vital Statistic (CRVS), health services including Electronic Medical Records at hospitals and health facility level, electronic tools at community level, education systems, national IDs, and others are great opportunities to start planning and building such linkages.

A proposed Future Vision

As country’s are working towards achieving Sustainable Development Goals (SDGs) and Universal Health Coverage (UHC), a multi-faceted approach to health and well-being is required. From the health perspective, how can we ensure that data that is collected and used for multiple purposes outside of the health domain that may affect individual health behavior and status, is used for better analysis.

Similarly, health data maybe useful for improving educational messages and programmes, guide on improved agricultural crops that should be harvested, etc. Harnessing opportunities to link multiple data systems would greatly improve the availability and quality of statistics that can then be used to build better predictive models and used for developing improvement plans across multiple sectors.

Linking data across sectors and systems such as health, education and social welfare, will provide valuable opportunities to generate better information and lead to a more in-depth understanding of the complex behavioral, social and economic aspects of health. Longitudinal data collected on risk patterns, lifestyle and behavioral can be very powerful in building prediction models using AI.

The known challenges

Linking of records across multiple distributed data systems that are owned and managed by multiple Ministries or sectors has some challenges. These challenges can be technical as well as political. On the technical side, discreet but overlapping business processes result in duplicate and limited linkages across systems. No usage of data standards that facilitate data exchange, limited human capacities and skills and developing ICT infrastructure are some other technical challenges.

Challenges on the political side are more complex. Lack of common legal and governance frameworks across multiple Ministries of sectors that restrict data sharing and interoperability across systems; concerns about data privacy and security and client confidentiality are some political issues that need be addressed to provide a strong foundation.

How can this be achieved?

Imagine a Family Tree that shows your ancestry and maps you, your parents, grand-parents and all your brother and sisters. Now think of that in relation to an individual’s lifetime and that you are linked with your family and kin like a “distributed data system”. They are all linked to each other and at the same time maintain individual identity. Either they have their last name or use a hyphenated one or it is used their previous name. The key is that they can be searched using a common-key i.e. a family name.

Now, in this case of using distributed systems across multiple sectors, we need to make sure that there is common searchable key that is available in all the systems. This can then be used to link (or interoperate) these systems. It does not need to be the same key in all the data systems but linked. In the digital age this can be done much more easily.

The World Bank and DIAL joint report The Role of Digital Identification for Healthcare: The Emerging Use Cases, shares examples from 5 countries on use existing foundational identification systems, such as population registers, unique identification numbers (UINs) or national ID (NID) cards, as the basis for patient identification, verification, and authentication. Leveraging a foundational system in this way may create additional benefits beyond those offered by a functional system.

Samuel Mills, et al. article Unique health identifiers for universal health coverage discusses examples from South Korean and Thailand in using the national unique identification number as the unique health identifier. In other countries, a unique health identifier is created in addition to the national unique identification number, but the two numbers are linked; Slovenia offers an example of this arrangement.

An Analysis of unique patient identifier options by the US Department of health and human services presents (6) components that should be integral parts of the Unique Patient Identifier: (1) An Identifier (numeric, alphanumeric, etc.) Scheme; (2) Identification Information; (3) Index; (4) Mechanism to hide or encrypt the Identifier; (5) Technology infrastructure to search, identify, match, encrypt, etc.; and, (6) Administrative infrastructure including the Central Governing Authority.

Developing and Using Individual Identifiers for the Provision of Health Services including HIV, a report by UNAIDS and PEPFAR highlights the development and use of unique identifiers will assist countries in the process of developing HIV services within the context of universal access to HIV prevention, treatment, care and support services, in addition to strengthening the country’s healthcare system as a whole.

Looking at the broader data eco-system, USAID’s report on Identity in a digital age: Infrastructure for inclusive development, highlights the benefits of creating digital identification (DID) systems and focuses on five technology trends that are poised to have near-term impact on the DID ecosystem. USAID’s How To: Create Digital ID for Inclusive Development describes two key shifts.

First, creating more sustainable ID systems requires a shift from instrumental design that leads to isolated, single-application ID systems to infrastructural design of systems that can be repurposed for similar projects and are compatible with existing local systems.

This first shift, which is the focus of this guide, will help support a second, longer-term shift from functional systems that serve only one purpose or program to more foundational systems that can serve as a public good and underlie multiple functional purposes.

Important considerations for this an initiative

Overcoming the challenges to establishing a person-centered data system that is interoperable across multiple distributed data systems will require high level, multi-sectoral coordination structures between ministries of health, civil registration authorities, education, national statistics office, and others to identify linkages and design of a shared architecture.

Similar to Health Insurance Portability and Accountability Act of 1996 (HIPAA) is United States legislation that provides data privacy and security provisions for safeguarding medical information and the General Data Protection Regulation (GDPR) 2016/679 is a regulation in EU law on data protection and privacy for all individual citizens of the European Union and the European Economic Area, countries need to define such legal frameworks for the protection, privacy and security of individuals data.

A strong data governance framework and adoption of common data standards that support interoperability and common language across various sectors is key. This would facilitate smooth exchange of data across multiple systems. Most low- and middle-income countries need to develop cadres, capacities and skills of data stewards and managers skilled and empowered to collect, analyze, curate and use data across the continuum of care in the digital age.

Lastly, development and implementation of a connected person-centered data system is complicated by the fact that there are multiple knowledge gaps, many partners and stakeholders who need to be involved, a complex external environment, and wide diversity of digital systems. Implementation research effort are needed to learn and continuously improve implementation and use of such systems.

Vikas Dwivedi is a Sr Health Information Systems Advisor at Palladium