Modelling the COVID-19 pandemic

The bitter dispute over India’s pandemic mortality (with Aashish Gupta, in The Hindu, 12/05/22). After publication of the WHO’s estimates of India’s pandemic mortality, government rebuttals got more strident. These rebuttals include wild claims about civil registration coverage which do not stand up to scrutiny. (The article is paywalled, but an image is available here.)

Estimating Covid-19 Fatalities in India. (The India Forum, 05/05/22). What is the basis for Indian government objections to estimates of pandemic mortality in India? And do these objections hold water? This piece is also in a more abbreviated form here.

Estimating potential excess mortality during the third (Omicron) wave of India’s COVID-19 epidemic. (Murad Banaji, technical document, 24/01/22).

Why the IIT Kanpur Report on UP’s COVID-19 Crisis Was Dishonest (Murad Banaji, The Wire, 20/10/21). Some highly dubious and unscientific work on the pandemic and its handling. This piece is focussed on the questions the IIT report failed to ask about mortality.

Many say Covid-19 will transition ‘from epidemic to endemic’. But what does this mean? (Murad Banaji, Scroll, 19/10/21). Reflections on the future of COVID-19, some common misconceptions, and what a transition to endemicity might mean in practice.

With COVID-19, India Experienced Its Greatest Mortality Crisis Since Independence (With Aashish Gupta, The Wire, 05/10/2021). Summary of the key points in our preprint on pandemic excess mortality in India. Contextualising the numbers and seeing how they align with the progress of the epidemic.

Estimates of pandemic excess mortality in India based on civil registration data (preprint on medrXiv with Aashish Gupta, 01/10/2021). We use death registration data from twelve states comprising around 60% of the national population to estimate the scale of the pandemic mortality surge in India.

Covid-19: What data about excess deaths reveals about Mumbai’s class divide (Murad Banaji, Scroll, 15/09/2021). All-cause mortality data indicates that the bulk of Mumbai’s “uncounted” deaths were probably amongst slum-dwellers. It tells a story, probably repeated across the country, of how official data risks understating the toll of the pandemic on marginalised communities. Based on the updated preprint on medrxiv here.

Estimating COVID-19 infection fatality rate in Mumbai during 2020 (Murad Banaji, 09/09/2021). Preprint on medrxiv updated to take into consideration extra insights from monthly all causes mortality data.

Two talks:
1) The COVID-19 pandemic in India: data, stories and myths (to the Kerala Council for Historical Research, July 22, 2021)
2) Lessons from India’s pandemic data (to the Centre for the Study of Developing Societies, September 3, 2021)

Lessons from India’s all-cause mortality data. (With Aashish Gupta, The Hindu, 20/08/2021. Article behind a paywall, but image available here.) Some estimates on excess mortality in India during the pandemic. We discuss the issues with death registration data, why making excess mortality estimates is complicated, and why the numbers are likely to rise. Based on a technical document: Estimates of pandemic excess mortality in India based on civil registration data.

Survey evidence of excess mortality in Bihar in the second COVID-19 surge. (30/07/2021, preprint, with Apurva Bamezai, Aashish Gupta, Shivani Pandey, Sharan MR, Kanika Sharma and Chanchal Kumar Singh). A small survey in Bihar found a huge surge in mortality during the second COVID-19 wave: around four times normal (95% CI: 2.2x to 6.4x normal).

COVID-19 in Kerala: Is the Story Very Different From the Rest of the Country? (Murad Banaji, The Wire, 22/07/2021). Kerala has managed and surveilled its epidemic better than many states in the country. Mortality is lower than some states, but still significant.

What does mortality data tell us about Bihar’s first Covid-19 wave? (Murad Banaji, Scroll, 11/07/2021). Bihar’s all cause mortality data reveals just how badly it was hit in the first COVID-19 wave.

Making Sense of Madhya Pradesh’s Shocking Deaths Increase in April and May (Murad Banaji, The Wire, 14/06/2021). An analysis of the alarming excess mortality data obtained from the state of Madhya Pradesh. In a single month there were five times as many deaths as normal.

What local news reports tell us about Covid-19 mortality in rural areas of North and Central India (Murad Banaji, Leena Kumarappan and Aashish Gupta, Scroll, 10/06/21). A study of COVID-19 mortality in 61 villages or clusters, based on newspaper reports. Rapid surges, and high median excess mortality (0.29%).

Mumbai’s COVID Situation Is Improving – but Is the City Safe From More Surges? (Murad Banaji, The Wire, 26 May, 2021). A summary of what we know and don’t know about variants, how many were infected during the latest surge, and why it wound down. Includes some speculation about the role of prior infections and vaccination in mortality trends.

Scientific uncertainties, weak institutions, and nationalism led to COVID-19 devastation (Murad Banaji, The Caravan, 14/05/2021). An attempt to understand how scientific narratives and political considerations got entangled, leading to the complacency which preceded India’s devastating second COVID-19 wave.

Estimating COVID-19 fatalities in India (Murad Banaji, The India Forum, 10/05/2021). Why it’s hard, why it’s important, and how we can try – a detailed overview.

Bengaluru is reaching a Covid-19 crisis point. How did it get there? (Murad Banaji, Scroll, 07/05/2021). How did the city get to a point where it had the highest active case-load of any Indian city at any point in the pandemic?

Are COVID-19 Vaccines Saving Lives in Mumbai? Here’s the Math (Murad Banaji, The Quint, 26/04/2021). Although the kind of data needed to conclusively demonstrate a vaccine effect is missing, the city’s fatality rates and age-distribution of COVID deaths give strong hints that vaccination is saving lives. The scale and timing of the effect match what we would expect in terms of strong protection against severe disease a few weeks after first dose vaccination.

What could be driving Delhi’s devastating fourth Covid-19 surge? (Murad Banaji, Scroll, 22/04/2021). Why is disease spreading so fast in the national capital, and why are deaths so high?

COVID-19 Is Surging In India – but Will There Be Fewer Deaths This Time? (Murad Banaji, The Wire, 16/04/2021). Why there is little cause for optimism about fatalities being lower during this wave.

What could be driving Mumbai’s surge? (Murad Banaji, 30/03/2021, Article behind a paywall, but image available here.) Trying to make sense of the brutal new surge in already-hard-hit Mumbai. What role could uneven spread in earlier waves, and new variants be playing?

Has a More Deadly Novel Coronavirus Variant Increased Fatalities in Punjab? (Murad Banaji, The Wire, 27/03/2021). The variant B.1.1.7 (first found in the UK) has reportedly become widespread in Punjab. And we can see the effects of B.1.1.7’s greater lethality if we carefully track cases and fatalities in the state.

As Mumbai battles its third Covid-19 surge, could vaccination stop the wave? (Murad Banaji, Scroll, 16/03/2021). How bad could Mumbai’s third surge be, and what could bring it under control?

How the Economic Survey Used Dubious Analysis to Spin India’s COVID Response (Murad Banaji, The Wire, 08/03/2021). Incomplete data and flawed analysis were used to create narratives about India’s COVID-19 epidemic and response which had little substance.

India’s Dwindling COVID-19 Cases – Sign of Effective Control or Increasing Immunity? (Murad Banaji, The Wire, 20/02/2021). An attempt to unravel what lies behind the wind-down in cases and deaths nationally, and go beyond some of the simple narratives in the media and social media.

What lessons do India’s Covid-19 serosurveys hold? (Murad Banaji, Scroll, 06/02/2021). India’s serosurveys give us vital clues about the Indian COVID-19 epidemic. But sometimes they tell us less about the disease and more about the interaction between science, media and politics.

COVID-19: What the Third National Sero-Survey Result Does and Doesn’t Tell Us (Murad Banaji, The Wire, 05/02/2021). The survey indicates that India is unlikely to be near herd immunity, and that disease has moved increasingly to rural areas where detection is poorer.

COVID-19 in India: Where Is It at and What Does the Future Have in Store? (Murad Banaji, The Wire, 06/12/20). What has driven the national slowdown in cases and deaths, will there be another surge, and with highly variable detection will we even know if there is?

Why is Covid-19 surging in Delhi eight months after the pandemic started? (Murad Banaji, Scroll, 15/11/20). I used some Monte Carlo simulation to try to understand the origins of the third wave in Delhi, and to estimate levels of infection, detection changing fatality rate, and so forth. The technical version is here.

Two pieces on Bihar’s COVID-19 epidemic: Bihar’s COVID-19 Epidemic Is Not Over – but Where Is It, Exactly? (Murad Banaji, The Wire, 27/10/2020) and Does Bihar’s COVID-19 Data Seem Too Good To Be True? (Murad Banaji, The Wire, 02/11/2020). Technical notes for the first piece here.

Herd Immunity for COVID-19 – a Straightforward Explainer (Murad Banaji, The Wire, 11/10/20). What is herd immunity all about, and what is a “herd immunity approach” to COVID-19? And is it a good idea? In Q and A form.

What Early Reports From Mumbai’s Second Seroprevalence Survey Tell Us (Murad Banaji, The Wire, 06/10/20). A smaller fraction of people were found with COVID-19 antibodies compared to the last survey (early July). Was this good news, bad news – or about waning antibodies? (Accompanying calculations on the effects of waning test sensitivity.)

What Is Driving Mumbai’s COVID-19 Surge? And How Bad Is It? (Murad Banaji, The Wire, 28/09/20). Despite high COVID-19 prevalence Mumbai saw a surge in cases in September recording its highest daily figures of the epidemic so far. What drove this surge? (Accompanying technical material focussed on different levels of infection detection in slum and non-slum areas of Mumbai.)

Covid-19: Why it’s impossible to draw firm conclusions from the first national serosurvey (Murad Banaji, Scroll, 23/09/20). The first national serosurvey in India was riddled with flaws, and yet implausible narratives built on its tenuous data abound. (A detailed peer review of the first serosurvey paper in IJMR.)

Mumbai Had 13k ‘Excess’ Deaths in March-July. How Does Its COVID Story Change? (Murad Banaji, The Wire, 10/09/20). The Indian Express obtained all cause mortality data for Mumbai which showed a sharp rise in excess deaths during March-July over and above the recorded COVID-19 deaths. The city’s COVID story is gradually becoming clearer. (A more fully linked version of this piece with technical detail here.)

Are Health Experts Right To Focus on Cutting COVID-19 Deaths, Not Containment? (Murad Banaji, The Wire, 03/09/20). Some thoughts on recommendations by public health bodies that “the government’s focus should be to prevent deaths from COVID-19 and no longer on containing the infection.”

A city divided: Covid-19 reveals Mumbai’s faultlines (Murad Banaji, Scroll, 20/08/20). Mumbai’s COVID-19 story is only gradually starting to become clear. With rapid spread in the slums, and a more long drawn out epidemic in middle class areas, Mumbai has seen two interlinked epidemics; housing poverty is the key dividing line. A fully referenced version of this piece here.

What Pune’s Sero-Survey Does and Doesn’t Tell Us About Its COVID-19 Epidemic (Murad Banaji, The Wire, 18/08/20). A quick look at the serosurvey results from Pune with some astonishingly high seroprevalence values. Supporting material here.

A quick look at Ahmedabad’s serosurvey results.

What Do the Delhi and Mumbai Sero-Survey Results Tell Us About COVID-19 in India? (Murad Banaji, The Wire, 30/07/2020). A first quick analysis of seroprevalence results from two major Indian cities, and what they tell us about fatality rates and prevalence of COVID-19 in India. Supporting material for Delhi here and Mumbai here.

COVID-19: Has Delhi Passed its Peak or Is its Test Strategy Confusing the Picture? (Murad Banaji, The Wire, 22/07/2020). Delhi ramped up COVID-19 testing, and then gradually switched from sensitive RT-PCR tests, to less sensitive rapid antigen tests. This created a sharp peak and rapid descent in daily case data which hid a more ambiguous reality.

How many people in India have had Covid-19? (Murad Banaji, Scroll, 09/07/2020). The tale of the ICMR serosurvey which led to some wild claims, but whose results have never been released. Supporting documentation: Data-driven estimation of prevalence and fatality undercounting for COVID-19 in India and Analysis of ICMR’s estimates of COVID-19 prevalence and infection fatality rate.

Why case fatality rate itself is almost never the story – but is often made to be. (Murad Banaji, The Wire, 27/06/2020). Case fatality rate (CFR) is perhaps the most misused of all statistics in discussions about COVID-19. A quick summary of what CFR does not tell us, how CFR is misused, and how CFR could still provide some useful hints about the trajectory of COVID-19 if analysed carefully.

Maharashtra’s missing COVID-19 deaths: an update. (Murad Banaji, The Wire, 18/06/2020) + details of the simulations here. It was acknowledged by the state government that a number of COVID-19 deaths had not been added to the official toll, vindicating an earlier piece; but even after the update, the data suggests that a lot of COVID-19 deaths in Maharashtra are missing.

How will the Covid-19 pandemic end? (Murad Banaji, Scroll, 12/06/2020) + Supplementary material: How do COVID-19 epidemics end? On the various possible futures of COVID-19 epidemics and the dangers of being simplistic about them. Modelling and data analysis.

Missing COVID-19 deaths in Russia – how many? (Murad Banaji, 06/06/2020). I wanted to see if modelling supported claims of fatality underreporting in Russia and, if so, gave any indication of the scale.

A brief analysis of Madhya Pradesh’s COVID-19 data. (Murad Banaji, 02/06/2020). Does the data suggest that there are missing COVID-19 fatalities in Madhya Pradesh?

A brief analysis of West Bengal’s COVID-19 data. (Murad Banaji, 01/06/2020). The story of missing fatalities, a correction, and a case fatality rate which declined, rose, and declined again…

The Similarities and Differences in COVID-19 Data From Delhi and Tamil Nadu. (Murad Banaji, The Wire, 29/05/2020) + Supplementary material: Notes on the evolution of COVID-19 in Delhi and Tamil Nadu. These two regions have superficially similar COVID-19 data-sets, but quite different COVID-19 trajectories.

What effects has the lockdown had on the evolution of Covid-19 in India? (Murad Banaji,, 27th May 2020) + Supplementary material: Notes on lockdown and its effects on the evolution of COVID-19 in India.

The Delhi COVID-19 epidemic so far: missing fatalities (Murad Banaji, 18th May 2020). A brief model-based exploration to study the trajectory of the COVID-19 outbreak in Delhi, and see whether there is evidence in the data for claims that COVID-19 deaths are being underreported in Delhi.

Brief notes on the London COVID-19 epidemic so far (Murad Banaji, 16th May 2020). These notes respond to modelling described in an Evening Standard article from May 15th which appears to provide support for lifting the lockdown in London.

The Tricky Issue With Knowing How Many Are Dying From COVID-19 in India (Murad Banaji, The Wire, 15/05/2020) + Supplementary material: Notes on the COVID-19 epidemic in Maharashtra, a change in protocol, and “missing” fatalities. I’ve added an update on Mumbai’s COVID-19 data anomaly here (June 14th).

Brief notes on the German COVID-19 epidemic so far (Murad Banaji, 09/05/2020).

Individual based modelling of the COVID-19 pandemic (Latest description of the modelling, Murad Banaji, 02/05/2020). This is the latest version of this document and supercedes previous documents – it may not, however, be completely up-to-date. The accompanying code is at

The Problem With Spinning Simple Stories Based on India’s COVID-19 Numbers (Murad Banaji, The Wire, 22/04/2020)

Modelling the COVID-19 pandemic – how it can be done and why we should be cautious. A summary for the non-mathematician of different methodologies, and why COVID presents particular challenges. (Murad Banaji, 03/04/2020)