How data science can help India's frontline workers in the fight against COVID-19

Healthcare workers, police officers, municipal workers and many others; standing in the line of duty, those working on the COVID-19 frontline have risked not only their lives, but their family members. Are these COVID-19 warriors being adequately taken care of? Are they diagnosed or monitored regularly? What happens if frontline workers get infected, and can this be proactively identified? What responsive measures have been put in place? And are these measures sensitive to how this crisis might be taking a toll on these workers' emotional, social and physical wellbeing?

Nations today are struggling to treat their citizens who are infected with COVID-19, regardless of the robustness of their healthcare systems.

In countries without sufficient healthcare facilities, lockdowns have given officials time and space to plan, design and develop makeshift health centres. As the battle against the pandemic has unfolded, we have seen the importance of building temporary facilities at full tilt, while ensuring enough healthcare staff are available onsite. These staff also need to be monitored daily, as they are more prone to contracting the virus while on duty.

Widespread testing is mooted as the best way to understand and arrest the spread of the coronavirus - but in a country like India, with a population of more than 1.3 billion, it is difficult to get every citizen tested. There is an alternative model, however - and it is based on the collection and analysis of data.

‘A stitch in time saves nine’: this old phrase is relevant to everyone fighting COVID-19. In addition to making healthcare supplies and personal protective equipment (PPE) available to healthcare workers, we are working to develop a continuous health assessment programme based on collecting and analysing data gathered from frontline workers.

To achieve this, a precise and brief digital survey form could capture vital details under three key categories: demographic data, co-morbidities, and symptoms and signs. Demographic data encompasses the name, age, gender, and other particulars that give an idea about the geography of the health care worker. Co-morbidities - for example hypertension, diabetes or cardiac problems - must be be captured, as those with a pre-medical history are more susceptible to infection. A prompt reporting by individuals of any symptoms of fever, dry cough or respiratory problems would be highlighted and sent for immediate diagnosis, along with those who have come in contact with a COVID-19 patient, so that potential carriers can be quarantined. This exercise should be mandatory for health workers and their immediate families, so that timely assistance and action can be taken.

Continuous health self-reporting coupled with social distancing can help countries and their patrons contain the pandemic. In this way, data received on a timely basis can help cease the spread of the infectious virus.

The data thus collected through the digital survey would be consolidated, cleansed, and then categorised for detailed investigation. As a result of this investigation, patterns can be generated to display information about respective categories in real time. This would show the impact on different units to which frontline workers belong; their respective organizations will then be empowered, with actionable insights delivered to help them prioritize their efforts. These findings will become the building blocks of a future national data landscape, which will have big implications for the model currently under development. After analysing any outliers, the normalized data will be processed and analysed with the help of a statistical algorithm, which will produce further actionable intelligence. This becomes the core cycle of the work.

The possibilities and combinations of data on co-morbidities and symptoms are innumerable. Personnel who are exposed to an infected patient are categorized as ‘critical’; those with co-morbidities are termed as ‘at risk’; and those who do not fall under any of the above categories can be termed ‘healthy’.

Visualized data would be submitted for review by a panel of doctors who would further assess its accuracy and the quality of outcomes. Post-data analysis, critical information patterns and insights will be reported in the form of infographic charts to the respective department heads, who can then act as required; this will create edges within the data that link the individual to their segments to the region which helps in the creation of data traction.

For example, such reports might help police departments categorise their employees by risk level; managers could then immediately call back individuals from their duty or have them checked by the doctor and advise next steps. It does not stop there; those who are in the ‘at risk’ category will be allocated work in areas which are less prone to - or are free from - COVID-19.

If executed well, the aforesaid approach can become a healthcare paradigm not just during this crisis but in entirety, leading towards a proactive health awareness programme. A data science-led, expeditious health audit and reporting programme focused on frontline workers could become a new norm, which would not only protect public health and save lives, but also generate confidence and trust among the greater population.