Why does India’s health data need a booster?

The search was in vain. And even today not much has changed. As we grapple with the third wave of the COVID-19 pandemic and the Omron edition, the gap is just as disappointing.

Today, statistics on health indicators come from many sources. There are National Sample Survey Organisation, National Family Health Survey (NFHS), Sample Registration System, Disease Registration, Surveillance Reports, Annual Health Survey in select states and few others. However, this plethora of sources does not give us the data we are looking for in most cases. We have many opaque data systems that operate in watertight compartments, which makes their interoperability difficult. These datasets are also not shared between ministries and are certainly not open source for use by analysts and commentators.

Consider this: The pandemic has disproportionately affected doctors, nurses and policemen – they are the ones fighting the ongoing battle against the virus. The Health Ministry made a shocking announcement that it has no data on how many of our frontline health workers were killed. During the second wave the oxygen supply ran out and the whole country went through a sad time. The answer, once again, was sadly the same. By the government’s own admission, there was no information on those who died in intensive care units and hospital beds waiting to be supplied with oxygen.

The lack of information has many implications. First, health officials are unable to determine the prevalence of COVID-19 infection across different age groups, and therefore are not able to focus on the most vulnerable. Secondly, it is not possible to get age-wise or gender-wise numbers, for example, looking at death figures. This has turned out to be a completely avoidable controversy. While the government announced 400,000 deaths due to COVID-19, in July 2021 some very credible analysts including former chief economic adviser, Arvind Subramaniam, announced that the deaths could be 10 times that number. Other consequences are drug stockouts, vaccine supply shocks, overused ventilators and overcrowding of COVID centres.

Why are we facing a lack of reliable, transparent and unified data? It is ironic that given the changing nature of governance, we rely more and more on evidence to formulate new policies. While this is true for all sectors, nowhere is the evidence more significant than in health care.

The exact numbers would have helped with the management of supplies, morgues and even cremation facilities during the second wave. In this third wave, where we are all expecting far fewer deaths, reliable data will ease the pressure on healthcare workers and frontline workers, while better preparing our medical facilities when they hit a peak .

a design problem

Efforts have been made on data collection; Money and resources are actually allocated. The problem is more a matter of trust in the design, transparency and release of the information collected.

Aarogya Setu is a great example of technology that would have given us all the data we needed to trace, track and monitor the spread of COVID-19. When asked, the National Informatics Center first replied that it had no information about who created the app, undermining the credibility of what can be a great resource for epidemiologists. The government later clarified that it was built on a public-private-partnership and was indeed well designed and preserved. Even now, the data it collects is not available to government departments, even as a whole.

The government has announced its policy of using open-source software and announced a knowledge sharing protocol. However, all these systems are almost always inflexible apart from being proprietary and expensive with their complex architectures.

Then, there’s the Heath Management Information System, introduced 13 years ago. It collects a huge amount of information but uses only one tenth of it to generate health indicators. In almost all sheets, half of the fields are blank or marked as ‘not applicable’.

Even after collecting this data, major problems arise because very often the denominators, such as age and gender, are not available. For example, it is not enough to simply provide the number of people who have tested positive – such as 10 out of 500 tested. We need to know what populations are at risk and how many of them are positive. If a whopping 500 is young and vaccinated, 10 is a seriously high number.

Inaccurate data is another serious concern, especially with data collected through surveys or administrative means. If a data set shows a higher proportion of older people in an area with a population of youth, this needs to be corrected. However, in most cases, these errors are corrected at a central level and not at the point of data collection, leading to drastic changes in the results. The numbers should be corrected in the hospital itself or by the enumerator who is collecting field data and not by an analyst looking at millions of data points in Delhi.

The issue of collecting lots of irrelevant data gets complicated with the same data being collected more than once. When the same data is collected multiple times on different platforms, it confuses healthcare workers, data collection agents and the surveyed population itself. Different sources give different numbers—allocating the budget becomes too confusing. The best example is India’s TB data set. There are many organizations that collect TB data. While there are 10 million cases in India according to one estimate, the number has been pegged at 3 million in another study.

Meanwhile, the private sector plays a major role in healthcare today—about 75% of all diseases are treated in the private sector, both in rural and urban centres. The same is true of all outpatient care. Even for inpatient care, the proportion of the private sector is close to 70%. This means that at least two thirds of all data is held by non-state actors. However, none of this data, except for a few minor exceptions, is ever notified or reported.

old demand new

How are our surveys conducted? Surveys in a large population suffer from sampling issues, and the NFHS-4 (2015-16) is a great example of the survey that the sample used in some states was too small. Same is the case with National Sample Survey.

We need to collect and disseminate regular administrative data, a cheaper and more reliable form of information gathering, near the point of data collection. Imagine what we could have achieved if we had a regular set of data provided by the Ministry of Health. This is the data that is available in most countries.

The first is the old demand. The birth weight of all children must be recorded and recorded in the birth certificate. This will allow us to see what happens as our children grow up and will reduce our under-five mortality, which in developing countries is among the highest in the world today. Similarly, the cause of death should be clearly mentioned in all death certificates.

In the context of COVID-19, most researchers, and now, most citizens, would like to know the periodic and routine results of genomic sequencing, which is being done in a small number of cases. We should be able to get daily updates on tests done, hospitalization rates. We also need daily information about infections and reinfections in hospitals. All of this will allow us to increase the number of beds needed and the hospitalizations required for next week and more.

way forward

The roll-out of the National Digital Health Mission (NDHM) in September 2021 was indeed a step in the right direction- NDHM started with a vision to improve efficiency, effectiveness and transparency of healthcare delivery. This could enable a unified digital database for health care in India; Disseminated data can allow public policy to be shaped. However, NDHM will be successful only if the system allows transparent collection and distribution of the data collected. This needs to include private care and community based hospital services – they will provide most of the information given their reach.

All stakeholders should be aware that data is being collected to be used for policy purposes. This was the secret behind the success of data collection when it comes to the National Rural Employment Guarantee Act (NREGA), which aims to guarantee the ‘right to work’. All users felt that it would be used for budgeting and monitoring performance. Collection and dissemination became real-time, and the data base became accessible to all – researchers, panchayat heads, state governments and the central government.

Similarly, health data should be viewed as useful. Health Management Information System (HMIS), a web-based information system launched by the Ministry of Health and Family Welfare, captures service delivery data (reproductive, maternal and child health related, immunization, family planning, etc.) on a monthly basis . However, for data enumerators, the end result is unclear and hence the collection is done carelessly. It is important that the entire health system sees this as a useful exercise that helps with decision making. If the data remains hidden behind various firewalls and is inaccessible, the usability will be questioned.

Another approach is to decentralize the ownership of data. State governments should take pride in collecting and disseminating information – as is done with casual workers and card holders under NREGA. The Panchayat also then takes pride in keeping the data ready.

Meanwhile, private hospitals and diagnostic centers should be encouraged and encouraged to share information, register cases and report infections. Some private data aggregators can also be used. This data should be available in the public domain, open to researchers and everyone who is interested in the subject.

A great example of this is the way the NFHS-3 (2005-06) data has been openly disseminated, yet it has brought embarrassment to the present government. NFHS-3 showed how India’s dazzling economic growth led to malnutrition in large numbers. The data then forced the central and state governments to take nutrition planning seriously and correct the many mistakes made in food policy.

Now, we need a new data policy that enables our publicly funded information to be easily accessed. We have ample examples of this from Israel, the UK and most of Europe, where real-time data helped avert deaths. This policy should also ensure that data privacy is respected and that theft is not easily tolerated. Here, it is important to mention that the Digital Information Security in Health Care Act (DISHA) has been passed and needs to be tightened further.

Finally, robust health data collection mechanisms are now possible with easy availability of technology and spread of bandwidth across the country. Real time data monitored by GPS tools can be collected and verified instantly. This can pave the way for quick decision making. For example, in India’s procurement policy, it will ensure that we never run short of medicine supplies, oxygen cylinders and personal protective equipment, or PPE kits.

(Aamir Ullah Khan teaches at MCRHRDI and the Indian School of Business, and Salima Razvi is a senior economist at the Copenhagen Consensus Center.)