top of page

Why MPI and National IDs cannot work as Patient ID in LMICs

After years of research into centralized Patient Identification, I have concluded that neither a national ID alone nor a central Master Patient Index (MPI) is suitable for developing countries. Yes, I know; this goes against all logic. I was a strong advocate of centralized Patient ID management until the recent past. An ideal Patient ID scheme needs to meet to a large extent all of the following - unique, Inexpensive, simple, uncontroversial, unchanging, uncomplicated.

Since no central ID schemes can ideally meet all these, the friction has meant little adoption of centralized ID schemes in healthcare. The concept of Decentralized IDs (DID) is ripe for adoption in healthcare; I am excited at the opportunities they present. Of course, because they are censorship-resistant, it will be tough for healthcare regulators. World-wide-web consortium (W3C) has since published a draft standard in this regard.

How does this work? Let me describe this with an adapted self-sovereign ID use case from W3C. In traditional ID, every new software solution (eg. EMR) will need authentication-infrastructure, network-connection and meet a host of other requirements. To my knowledge, I do not know of any developing country that uses this scheme. This not practicable for most Low and Middle-Income Countries (LMICs)

I must now stop and acknowledge that Estonia is a shining light for centralized Patient ID use. In 2019, Estonia published that it achieved 98% ePrescriptions. But they are factors beyond the reach of developing countries that made such inevitable, like 100% electricity infrastructure and 100% internet availability in all health facilities. These health facilities have the proper hardware to support biometric authentication at service points. In contrast, Countries with less than 30% broadband coverage whose health facility hardware implementation consists mainly of tablets and whose majority of health facilities lack connection to the national electricity grid will only continue to dream of replicating the Estonian-dream.

A sovereign ID can be cryptographically proved by any software application using a set of the Patients' characteristics (eg. Age, age, phone number, or Name) pre-agreed and a set of mathematical rules (or algorithms). This will allow for care coordination and collaboration and ensure that software vendors and developers design for collaboration from the start. This proposal deviates slightly from the W3C approach that envisages the use of blockchain. This model will mean that with a Patient's characteristics (or combination there-off), a consistent ID can be generated or proved across 'disconnected' health institutions. This model can be truly game-changing, though far from ideal.


bottom of page