Model to project, schedule O2 demand

MUMBAI: A team of researchers at the Tata Institute of Fundamental Research (TIFR), Mumbai, have developed a digital tool to route medical oxygen for Covid-19 patients to avoid the shortage experienced during the second wave of the Covid pandemic.

The oxygen supply crisis peaked in April in many parts of the country as hospitals ran out of medical oxygen that was needed to be supplied to severe patients of the coronavirus disease, forcing the Centre to first reduce and then reroute industrial oxygen all together to various states to manage the demand.

The customisable planning tool, developed by the School of Technology and Computer Science (STCS) at TIFR, will help public health officials project the oxygen demand in metric tonnes for each district for a few days or weeks. Based on these projections and suppliers’ capacities and locations, they can use the tool to schedule their oxygen supply to different districts.

For example, in Maharashtra, according to June 3 data of 15,229 Covid-19 cases, the model projects that the oxygen requirement for the state will be 887MT on Saturday, and reduce gradually to 725MT by next Thursday (June 10). In Mumbai, the projection is that oxygen requirement will drop from 41MT on Saturday to 37 MT by next Thursday. For Delhi, the projections are 155MT on Saturday. This will reduce to 95MT by next Wednesday (June 9).

These are default projections made based on guidelines by the Indian Council of Medical Research (ICMR), public health officials from different states can input these details to get a customised projections and schedule for their state and districts.

“The default assumptions are that 3% of all detected Covid-19 patients will become critical and require substantial oxygen support for about 18 days, as per ICMR guidelines. However, if a state is experiencing a more severe outbreak, they can customise the consumption pattern to better match their situation easily on the tool,” said Sandeep

Juneja, professor and dean of STCS, who was part of the project.

For each of these criteria, a range of variables are considered such as patients with severe disease are first admitted in nonICU wards; the period between detection of case and start of non-ICU hospitalisation is same as that for patients with moderate disease (three days); after some time, these patients are transferred to ICU wards where the period between case detection and start of intensive care is 11 days and period of intensive care is 18 days. The model also takes into account the time taken to get oxygen tanks into each district from the suppliers. All these can be changed to suit the conditions on the ground.

For the projections, they use a linear programming model, a mathematical optimisation to find the best possible solution for a problem. “It is a very simple model to build. We have made it available on our website for use and are updating the features to better meet user needs. States can input details and download the schedule easily.

We have also provided basic guidelines on how to easily input data to suit your needs. The idea is to improve it and get more people to use it, especially as we prepare for a potential third wave,” said Juneja.

Maharashtra chief secretary Sitaram Kunte said the model was promising and could potentially streamline the oxygen supply chain. “I have seen their model. It is based on operations research concepts and has the potential to streamline the huge logistics of oxygen transportation. Very promising,” said Kunte.

Dr Shashank Joshi, a member of Maharashtra’s Covid taskforce, said, “The need for oxygen changes every minute for severe and critically ill patients. The demand for oxygen depends on various factors and mathematical models may not necessarily work. The Supreme Court appointed national task force is also looking at this problem. In the third wave, our focus should be on zero deaths and preventing severe conditions, thus reducing the demand for oxygen altogether.”