Why Pandemic Preparedness Failed to Predict COVID-19 Deaths

Why Pandemic Preparedness Failed to Predict COVID-19 Deaths

In early 2020, the world had a checklist. Countries had spent billions building pandemic preparedness scorecards, ranking nations by their ability to detect, contain, and survive an outbreak. The Global Health Security Index, published in 2019, gave the United States a score of 83.5 out of 100. The United Kingdom scored 77.9. These were the gold standards of readiness. Then COVID-19 arrived, and the gold standard turned out to be fool’s gold.

By September 2021, the U.S. had recorded over 700,000 deaths. The U.K. had over 130,000. Meanwhile, Vietnam, which scored 49.1 on the same index, had fewer than 1,000 deaths. The checklist had failed. But the question is not just why it failed. The question is what the checklist was actually measuring.

A 2022 study in The Lancet by Thomas Bollyky and colleagues at the Institute for Health Metrics and Evaluation took a hard look at 177 countries and asked: What actually predicted COVID-19 infection and death rates? The answer is not what anyone expected. The preparedness indices were not just unhelpful. They were essentially irrelevant.

The Numbers That Didn’t Add Up

Bollyky’s team did something most pandemic preparedness analysts had not. They standardized the data. They accounted for the obvious variables that skew raw numbers: a country’s age structure, its obesity rates, its air pollution, its GDP per capita, even the seasonality of pneumonia. This is the kind of statistical housekeeping that separates real insight from noise. Once they had cleaned the data, they tested whether any of the 12 major pandemic preparedness indices actually correlated with how countries performed.

The result was stark. “Pandemic preparedness indices, which aim to measure health security capacity, were not meaningfully associated with standardized infection rates or IFRs,” the authors wrote (Bollyky et al., 2022). Not weakly associated. Not modestly associated. Not at all.

The indices were supposed to measure things like laboratory capacity, emergency response systems, and border controls. These are the things that sound right. They are the things that governments spend money on. But in the real world, they predicted nothing.

What Actually Drove Deaths

If the checklists were useless, what mattered? Bollyky and colleagues found that the single biggest predictor of a country’s infection fatality ratio (IFR) was its age profile. This explained 46.7% of the variation between countries. That is not surprising. Older populations die more. But the next biggest factor was GDP per capita, which explained only 3.1% of variation. National mean body mass index explained 1.1%. Together, these three factors accounted for just over half of the variation in IFR. The other 44.4% remained unexplained (Bollyky et al., 2022).

For infection rates, the story was even more mysterious. The proportion of the population living below 100 meters of elevation explained 5.4% of variation. GDP per capita explained 4.2%. Seasonality of pneumonia explained 2.1%. The rest was a black box. “Most cross country variation in cumulative infection rates could not be explained,” the authors admitted (Bollyky et al., 2022).

This is the uncomfortable truth the preparedness industry does not want to face. They built models that assumed we knew the levers. We did not.

The Trust Variable That Outperformed Everything

The most provocative finding in the Bollyky paper is not about what failed. It is about what succeeded. The researchers tested something that almost never appears on preparedness scorecards: trust.

They measured trust in government and interpersonal trust using standardized survey data. They also measured government corruption. Then they asked whether these social variables predicted infection rates better than the official indices.

They did. By a lot.

Countries with high levels of trust in government had lower standardized infection rates. Countries with high interpersonal trust had lower infection rates. Countries with less corruption had lower infection rates. These associations were statistically significant, and they held even after controlling for GDP, age, and other confounders (Bollyky et al., 2022).

The size of the effect is what makes you stop. The authors ran a counterfactual simulation. If every country in the world could achieve the level of trust that Denmark has (the 75th percentile), global infections might have dropped by 12.9% for government trust and 40.3% for interpersonal trust (Bollyky et al., 2022). Forty percent. That is not a rounding error. That is a different pandemic.

How Trust Changed Behavior

The Bollyky paper does not just show a correlation. It traces a mechanism. The researchers looked at mobility data and vaccination rates. They found that trust and low corruption were associated with greater reductions in mobility during lockdowns, and with higher COVID-19 vaccine coverage in middle income and high income countries where vaccines were available (Bollyky et al., 2022).

Think about what this means. A country can have the best hospital system in the world. It can have stockpiles of ventilators and a national response plan. But if its citizens do not trust the government, they will not stay home. They will not get vaccinated. They will not follow guidance. The infrastructure is irrelevant if the social contract is broken.

This is the kind of finding that should keep health ministers up at night. You cannot buy trust. You cannot stockpile it. You cannot deploy it in a crisis. You have to build it over years, through consistent governance, low corruption, and genuine community engagement. And most countries have not done that.

The BMI Factor That No One Wants to Talk About

There is another uncomfortable finding in this paper. National mean body mass index explained a small but significant portion of the variation in IFR: 1.1% (Bollyky et al., 2022). That might sound trivial, but the authors ran another counterfactual. If all countries had a mean BMI at or below the 25th percentile (essentially, a population with a healthy average weight), global IFR would have dropped by 11.1% (Bollyky et al., 2022).

This is a politically charged finding. Obesity is linked to COVID-19 severity through inflammation, metabolic dysfunction, and impaired immune response. But the Bollyky paper is not a moral judgment. It is a statistical observation with a policy implication. The authors suggest that “increasing health promotion for key modifiable risks is associated with a reduction of fatalities in such a scenario” (Bollyky et al., 2022). In plain English: if you want to survive the next pandemic, invest in public health now. Not just in hospital beds. In the underlying health of your population.

What the Preparedness Industry Got Wrong

The Global Health Security Index, the Joint External Evaluation tool from the World Health Organization, and other preparedness indices were not designed to be predictors of pandemic outcomes. They were designed to measure capacity. But capacity is not the same as performance. A country can have a state of the art laboratory and a population that refuses to use it.

The Bollyky paper makes this distinction explicit. The indices “measure capacity and not the likelihood that a country will effectively use that capacity during a crisis” (Bollyky et al., 2022). It is the difference between owning a fire truck and knowing how to get people out of a burning building.

The authors also note that these indices are heavily weighted toward technical capabilities that are easier to measure: number of labs, number of hospital beds, existence of a national plan. They do not measure the social and political factors that determine whether those capabilities actually get used. Trust is hard to quantify. Corruption is hard to admit. Community engagement is hard to score. So the indices ignore them.

This is not just an academic problem. Governments use these indices to allocate funding. The World Bank, the Gates Foundation, and national aid agencies have poured billions into preparedness based on these metrics. If the metrics are measuring the wrong things, that money is being wasted.

The Limits of This Study

The Bollyky paper is an observational analysis. It cannot prove causation. The authors are careful to say that their findings “should not be interpreted as causal” (Bollyky et al., 2022). The association between trust and lower infection rates could be driven by other factors. Maybe countries with high trust also have better public health systems. Maybe they have more homogenous populations. Maybe they got lucky.

The study also covers only the first 21 months of the pandemic, through September 2021. It does not account for the Delta and Omicron variants that followed, which changed the dynamics of transmission and severity. And the data on infections is notoriously unreliable, especially in low income countries where testing was limited. The authors standardized for this, but standardization is not the same as accuracy.

Perhaps most importantly, the study cannot tell us how to build trust. It can tell us that trust matters. It cannot tell us how to create it in a country where trust has been destroyed by decades of corruption, inequality, or political polarization. That is a question for political science, sociology, and history. The Bollyky paper is a diagnosis, not a prescription.

What This Actually Means

The Bollyky paper is not a call to abandon pandemic preparedness. It is a call to redefine it. Here is what the evidence suggests we should actually do.

• ▸Stop funding preparedness indices that measure capacity without measuring trust. The Global Health Security Index and similar tools should be revised to include social and political variables, or replaced entirely. A country’s score should reflect not just how many labs it has, but whether its citizens would trust a government recommendation to use them.

• ▸Invest in community engagement before the crisis. The finding that interpersonal trust had a larger association with lower infection rates than government trust suggests that social cohesion matters more than institutional authority. Governments should fund community based organizations, local health workers, and communication strategies that build trust over time. This is not a public relations campaign. It is a long term investment in social infrastructure.

• ▸Address modifiable health risks now. The BMI finding is a reminder that the health of a population before a pandemic determines how many people die during it. Countries with high obesity rates should treat this as a national security issue. That means investing in nutrition, physical activity, and healthcare access. It is not about shaming individuals. It is about changing the conditions that make people vulnerable.

• ▸Measure what matters, not what is easy. The Bollyky paper found that 44.4% of variation in IFR and most of the variation in infection rates remained unexplained. That is a huge gap. It means we do not understand the most important drivers of pandemic outcomes. Researchers should be funded to investigate the factors we are missing: political polarization, media ecosystems, housing density, occupational risks, and yes, trust.

• ▸Prepare for the next pandemic by strengthening democracy. This is the uncomfortable implication of the trust finding. Countries with less corruption and higher trust did better. Those are features of functioning democratic institutions. The Bollyky paper does not say that democracy is a pandemic preparedness tool, but the data points in that direction. If you want to survive the next outbreak, you might need more than a stockpile of masks. You might need a government that people believe in.

The pandemic preparedness industry sold the world a checklist. The checklist did not work. The Bollyky paper is not the final word on why, but it is the most honest accounting so far. It tells us that the things we measured were not the things that mattered. And it tells us that the things that mattered trust, social cohesion, underlying health are the hardest things to build. They are also the only things that might save us.