Why Digital Healthcare Cannot Substitute Foundational Infrastructure

After COVID-19, digital healthcare saw an increasing global interest. Digital technologies can be instrumental for tracking epidemics, limiting transmission, developing new drugs and diagnostics, and advancing medical research to improve well-being. The World Health Organisation (WHO) issued guidelines to help countries implement technology in healthcare on a large scale. Countries like UK and US have already adopted digital technologies in health, including electronic health and medical records to improve healthcare. These aim to streamline and expand access to healthcare services by digitising patient data.

India is following a similar trend, with initiatives like Ayushman Bharat Digital Mission (ABDM) aiming to address long-standing issues in India’s healthcare access, quality, and affordability, factors that push 60 million people into poverty each year. While the use of technology in healthcare sounds promising, collecting electronic medical and health records for 1.4 billion citizens is complex, and may face challenges such as discrepancies in data collection, resources, and competing interests of stakeholders. Even during COVID-19, there was no clear data on deaths or shortages in services like oxygen, hospital beds, and medicines. While this would affect everyone, those facing socio-economic injustice are at greater risk.

Digitising healthcare, for example, requires resources, which service providers may be reluctant to invest in. Large healthcare providers may face disruption, while smaller platforms will struggle with the resource-intensive nature of digitisation. This has led to resistance to the ABDM, particularly from stakeholders within the medical community. A major factor is the shortage of medical doctors, with the current number just exceeding one-fourth of the WHO’s recommended threshold. 

Doctors are expected to upload patient data themselves, yet the average Indian doctor is understood to be seeing 40-60 patients a day, leaving little time for this. With consultations lasting only two minutes, healthcare professionals may overlook digital records even if available.

In USA, a study found that patients are often dissatisfied with digitisation, and doctors, whose days are increasingly filled with brief encounters, find it “downright deadening.” The study observed, “You are sitting in front of a patient, with so many tasks and limited time, seven to 11 minutes, probably, so when do you really listen?”

Discrepancies

Decisions made by doctors and patients when engaging with technology-led healthcare, such as what data to collect, how much, and its accuracy, will affect ABDM’s functionality and effectiveness. In the US, for example, a lawsuit was filed against an EHR system that failed to display prescribed medications correctly, showed discontinued drugs as current, and even mixed up patient profiles and notes, leading to misdiagnosis and incorrect prescriptions. In India, combining digital records and paperwork may create confusion and a tiring process as doctors would like to verify digital records against physical ones, undermining the technology’s effectiveness. 

Even in the Pradhan Mantri Jan Arogya Yojana (PMJAY) scheme, audits revealed multiple discrepancies. Adults over 18 were sometimes treated under “paediatric specialty packages,” and 45,846 cases recorded discharge dates before admission dates. There were also instances where a single patient was listed as hospitalised in multiple hospitals. It raises the important question of how healthcare providers can act based on such inaccuracies in the datasets. While these may seem like minor or reasonable trade-offs, any error in health data can directly cost lives. In PMJAY, thousands of beneficiary cards were also cancelled, and despite requirements for unique IDs after verification, the audit uncovered over 1.57 million duplicate IDs. During the ABDM pilot, people struggled to remember their IDs and passwords, leading to duplicate IDs and inaccurate data. Even many doctors lack understanding of Health ID and its connection to electronic health records. 

Moreover, the involvement of multiple intermediaries, such as state agencies, insurance companies, and pharmaceutical firms, each with their own competing interests, increases the risk of fragmentation in healthcare data collection. This poses challenges in achieving interoperability. Evidence suggests that achieving interoperability in healthcare remains a challenge due to competing interests of different stakeholders. UK’s NHS still struggles with interoperability due to implementation challenges, limited stakeholder engagement, and data sharing issues. Similarly, USA has not been able to achieve interoperability, despite legislation against “information blocking.” 

UHID concerns

The rollout of Health ID has also raised concerns around consent. During COVID-19, individuals registering for vaccination via Aadhaar on CoWIN had their UHID created, often without consent. The ABDM strategy document differentiates between Health ID creation methods: using Aadhaar for those seeking government subsidies and other forms of identity (such as email or phone number) for others. It encourages linking Aadhaar to Health ID, thereby connecting personal health records to Aadhaar details, such as receiving  benefits from PMJAY.

Similar to the Aadhaar project’s early phase, Aadhaar became nearly indispensable between its launch and the 2016 law enactment. Mandating Aadhaar for accessing welfare programmes like PMJAY creates risks of exclusion, as seen in other welfare schemes such as food ration. 

Data as the means to an end

In healthcare, treating data as mere exhaust rather than infrastructure can lead to misinformed decisions, compromised patient care, and missed opportunities for innovation. The potential data gaps in the ABDM, including data inaccuracies, exclusion, lack of diversity, and poor interoperability, can negatively impact the training of AI models in healthcare. For example, without effective participation from all relevant stakeholders, data from certain areas could skew policies as well as outputs produced by AI systems.

While urban doctors may attribute heart disease to cardiovascular issues, rural doctors may see it linked to rheumatic heart disease, leading to biases in data sets and AI’s recommendations. A similar issue is seen in the US healthcare system, where algorithmic bias led to underestimating the healthcare needs of Black patients compared to White patients with similar risk levels. 

While digital technologies are important, they alone cannot solve India’s healthcare challenges. Strengthening foundational healthcare, especially at the grassroots level, is crucial. This includes empowering primary health centres, village-level health workers, and implementing preventive measures, which can be complemented by technological interventions to ensure equitable access.

Asheef Iqubbal is technology policy researcher at CUTS International.

The Wire is now on WhatsApp. Follow our channel for sharp analysis and opinions on the latest developments.