You are here: Home / Internet Governance / Blog / National Health Stack: Data For Data’s Sake, A Manmade Health Hazard

National Health Stack: Data For Data’s Sake, A Manmade Health Hazard

Posted by Murali Neelakantan, Swaraj Barooah, Swagam Dasgupta and Torsha Sarkar at Aug 17, 2018 12:40 AM |
On Oct. 5, 2017, an HIV positive woman was denied admission in Hyderabad’s Osmania General Hospital even though she was entitled to free treatment under India’s National AIDS Control Organisation programme. Another incident around the same time witnessed a 24-year-old pregnant woman at Tikamgarh district hospital in Madhya Pradesh being denied treatment by hospital doctors once she tested positive for HIV. The patient reportedly delivered the twins outside the maternity ward after she was turned away by the hospital, but her newborn twin girls died soon after.

The op-ed was published in Bloomberg Quint on August 14, 2018.

Apart from facing the severity of their condition, patients afflicted with diseases such as HIV, tuberculosis, and mental illnesses, are often subject to social stigma, sometimes even leading to the denial of medical treatment. Given this grim reality would patients want their full medical history in a database?

The ‘National Health Stack’ as described by the NITI Aayog in its consultation paper, is an ambitious attempt to build a digital infrastructure with a “deep understanding of the incentive structures prevalent in the Indian healthcare ecosystem”. If the government is to create a database of individuals’ health records, then it should appreciate the differential impact that it could have on the patients.

The collection of health data, without sensitisation and accountability, has the potential to deny healthcare to the vulnerable.

We have innumerable instances of denial of services due to Aadhaar and there is a real risk that another database will lead to more denial of access to the most vulnerable.

Earlier, we had outlined some key aspects of the NHS, the ‘world’s largest’ government-funded national healthcare scheme. Here we discuss some of the core technical issues surrounding the question of data collection, updating, quality, and utilisation.

Resting On A Flimsy Foundation: The Unique Health ID

The National Health Stack envisages the creation of a unique ID for registered beneficiaries in the system — a ‘Digital Health ID’. Upon the submission of a ‘national identifier’ and completion of the Know Your Customer process, the patient would be registered in the system, and a unique health ID generated.

This seemingly straightforward process rests on a very flimsy foundation. The base entry in the beneficiary registry would be linked to a ‘strong foundational ID’. Extreme care needs to be taken to ensure that this is not limited to an Aadhaar number. Currently, the unavailability of Aadhaar would not be a ground for denial of treatment to a patient only for their first visit; the patient must provide Aadhaar or an Aadhaar enrolment slip to avail treatment thereafter. This suggests that the national healthcare infrastructure will be geared towards increasing Aadhaar enrollment, with the unstated implication that healthcare is a benefit or subsidy — a largess of government, and not, as the courts have confirmed, a fundamental right.

Not only is this project using government-funded infrastructure to deny its citizens the fundamental right to healthcare, it is using the desperate need of the vulnerable for healthcare to push the ‘Aadhaar’ agenda.

Any pretence that Aadhaar is voluntary is slowly fading with the government mandating it at every step of our lives.

Aadhaar Seva kendra. (Source: Aadhaar Official Account/Facebook) 

Aadhaar Seva kendra. (Source: Aadhaar Official Account/Facebook

Is The Health ID An Effective And Unique Identifier?

Even if we choose to look past the fact that the validity of Aadhaar is still pending the test of legality before the apex court, a foundational ID would mean that the data contained within that ID is unique, accurate, incorruptible, and cannot be misused. These principles, unfortunately, have been compromised by the UIDAI in the Aadhaar project with its lack of uniqueness of identity (i.e, fake IDs and duplicity), failure to authenticate identity, numerous alleged data leaks (‘alleged’ because UIDAI maintains that there haven’t been any leaks), lack of connectivity to be able to authenticate identity and numerous instances of inaccurate information which cannot be corrected.

Linking something as crucial and basic as healthcare data with such a database is a potential disaster.

There is a real risk that incorrect linking could cause deaths or inappropriate medical care.

The High Risk Of Poor Quality Data

The NITI Aayog paper envisages several expansive databases that are capable of being updated by different entities. It includes enrollment and updating processes but seems to assume that all these extra steps will be taken by all the relevant stakeholders and does not explain the motivation for stakeholders to do so.

In a country where government doctors, hospitals, wellness centres, etc are overburdened and understaffed, this reliance is simply not credible. For instance, all attributes within the registries are to be digitally signed by an authorised updater, there must be an audit trail for all changes made to the registries, and surveyors will be tasked with visiting providers in person to validate the data. Identifying these precautions as measures to assure accurate data is a great step towards building a national health database, but this seems an impossible task.

Who are these actors and what will incentivise them to ensure the accuracy and integrity of data?

In other words, what incentive and accountability structures will ensure that data entry and updating is accurate, and not approached from a more ‘jugaad’ ‘let’s just get this done for the sake of it’ attitude that permeates much of the country. How will patients have access to the database to be able to check its accuracy? Is it possible for a patient (who will presumably be ill) to gain easy access to an updater to change their data? If so, how? It is worth noting that the patient’s ‘right’ to check her data assumes that they have access to a computer that is connected to the internet as well as a good level of digital literacy, which is not the case in India for a significant section of the population. Even data portability loses its potential benefits if the quality of data on these registries is not reliable. In this case, healthcare providers will need to verify their patients’ health history using physical records instead, rendering the stack redundant.

Who will be liable to the patient for misdiagnosis based on the database?

A sonographic image is displayed on a monitor as a patient undergoes an ultrasound scan in Bikaner, Rajasthan, India. (Photographer: Prashanth Vishwanathan/Bloomberg)

A sonographic image is displayed on a monitor as a patient undergoes an ultrasound scan in Bikaner, Rajasthan, India. (Photographer: Prashanth Vishwanathan/Bloomberg)

Leaving the question of accountability vague opens updaters to the possibility of facing dangerous and unnecessarily punitive measures in the future. The NITI Aayog paper fails to address this key issue which arose recently. Despite being a notifiable disease, there are reports that numerous doctors from the private sector failed to notify or update TB cases to the Ministry of Health and Family Welfare ostensibly on the grounds that they did not receive consent from their patients to share their information with the government. This was met with a harsh response from the government which stated that clinical establishment that failed to notify tuberculosis patients would face jail time. According to a few doctors, the government’s new move would coerce patients to go to ‘underground clinics’ to receive treatment discreetly and hence, would not solve the issue of TB.

The document also offers no specific recommended procedures regarding how inaccurate entries will be corrected or deleted.

It is then perhaps not a stretch to imagine that these scenarios would affect the quality of the data stored; defeating NITI Aayog’s objective of researchers using the stack for high-quality medical data.

The reason why the quality and integrity of data is at the head of the table is that all the proposed applications of the NHS (analytics, fraud detection etc.) assume a high quality, accurate dataset. At the same time, the enrolment process, updating process and disclosed measures to ensure data quality will effectively lead to poor quality data. If this is the case, then applications derived from the NHS dataset should assume an imperfect data, rather than an accurate dataset, which should make one wonder if no data is better than data that is certainly inaccurate.

Lack Of Data Utilisation Guidelines

Issues with data quality are exacerbated depending on how and where it is used, and who uses it. The paper has identified some users to be health-sector stakeholders such as healthcare providers (hospitals, clinics, labs etc), beneficiaries, doctors, insurers and accredited social health activists but misses laying down utilisation guidelines. The foresight to create a dataset that can be utilised by multiple actors for numerous applications is commendable, but potentially problematic -- especially if guidelines on how this data is to be used by stakeholders (especially the private sector) are ignored.

In order to bridge this knowledge gap, India has the opportunity to learn from the legal precedent set by foreign institutions. As an example, one could examine the Health Information Technology for Economic and Clinical Health Act (HITECH) and the Health Insurance Portability and Accountability Act (HIPAA) in the U.S. which sets out strict guidelines for how businesses are to handle sensitive health data in order to maintain the individual’s privacy and security. It goes one step further to also lay down incentive and accountability structures in order that business associates necessarily report security breaches to their respective covered entities.

If we do not take necessary precautions now, we not only run the risk of poor security and breach of privacy but of inaccurate data that renders the national health data repository a health risk for the whole patient population.

There’s also the lack of clarity on who is meant to benefit from using such a database or whether the benefits are equal to all stakeholders, but more on that in a subsequent piece.

A medical team uses a glucometer to check the blood glucose level of a patient at a mobile clinic in Pancharala, on the outskirts of Bengaluru, India. (Photographer: Dhiraj Singh/Bloomberg)

A medical team uses a glucometer to check the blood glucose level of a patient at a mobile clinic in Pancharala, on the outskirts of Bengaluru, India. (Photographer: Dhiraj Singh/Bloomberg)

It’s Your Recipe, You Try It First!

If the NITI Aayog and the government are sure that there is a need for a national healthcare database, perhaps they can start using the Central Government Health Scheme (which includes all current and retired government employees and their families) as a pilot scheme for this. Once the software, database and the various apps built on it are found to be good value for money and patients benefit from excellent treatment all over the country, it could be expanded to those who use the Employees’ State Insurance system, and then perhaps to the armed forces. After all, these three groups already have a unique identifier and would benefit from the portability of healthcare records since they are likely to be transferred and posted all over the country. If, and only if, it works for these groups and the claimed benefits are observed, then perhaps it can be expanded to the rest of the country’s healthcare systems.

Murali Neelakantan is an expert in healthcare laws. Swaraj Barooah is Policy Director at The Centre for Internet and Society. Swagam Dasgupta and Torsha Sarkar are interns at The Centre for Internet and Society.