In recent news, scientists have expressed concerns over the government-backed SUTRA (Susceptible, Undetected, Tested (positive), and Removed Approach) model. This model had played an instrumental role in creating the perception that a second wave of Covid was unlikely in India; however, since April 2021, the devastating second wave has claimed thousands of lives. Let’s delve deeper into the intricacies involved.
The Application and Initial Predictions of SUTRA Model
Scientists from the Indian Institutes of Technology (IITs) in Kanpur and Hyderabad used the SUTRA Model to predict the Covid-19 trajectory in India. Notably, the model gained public attention in October 2020 when an expert member announced, according to the model’s predictions, that India was “past its peak”. The forecasting mechanism of the SUTRA Model depends upon three primary parameters: Beta or contact rate, Reach or exposure level of the population to the pandemic, and Epsilon, which denotes the ratio of detected to undetected cases.
The Core Components of SUTRA Model’s Prediction System
Naming them elaborately, ‘Beta’ signifies how many individuals a person infected with the virus can infect per day. It corresponds closely to the R0 value, quantifying the number of people an infected person can spread the virus to during their infection period. ‘Reach’ reflects the community’s overall exposure level to the pandemic. Lastly, ‘Epsilon’ represents the proportion between detected and undetected cases.
Identifying Problems with the SUTRA Model
However, the efficacy of the SUTRA Model in predicting the pandemic’s course has been brought under scrutiny due to several factors. Here are some noted issues:
Variability in SUTRA’s Forecasts
The model’s forecasts have often deviated significantly from the actual caseload, reflecting high degrees of variability. As a result, using the SUTRA model for guiding government policies has been questioned.
Dependence on Too Many Parameters
Critics argue that the SUTRA model heavily relies on a plethora of parameters. The model also recalibrates these parameters whenever its predictions falter, leading to potential overfitting. With three or four parameters, any curve can be adjusted over a short period.
Neglecting Virus Behaviour and Population Strata
The SUTRA model has also been called out for neglecting crucial factors like the virus behaviour and population strata. For instance, it fails to acknowledge that certain individuals (such as barbers or receptionists) are more likely to transmit the virus compared to people working from home. Furthermore, it does not take into account the social or geographic heterogeneity and lacks stratification of the population by age, ignoring contacts between different age groups.
Disregarding the Reasons Behind Changes
New variants appearing in the SUTRA model are represented as an increase in ‘beta’ value (which estimates contact rate). However, the model merely observes changes in parameter values without considering the reasons behind these changes, which can lead to misleading interpretations. In conclusion, while the SUTRA model’s efforts in predicting the Covid graph are noteworthy, it is crucial to address these highlighted issues to improve its future predictions and guide effective policymaking.