California’s fractured approach violates one of the core principals of risk communication in an emergency: “Deliver a clear and consistent message by trusted messengers,” said Lori Peek, director of the Natural Hazards Center at the University of Colorado at Boulder.

Instead, the messages were mixed, and confusing. Masks were useless, then essential. The virus might kill you, or do nothing. Children are safe, except sometimes; schools are dangerous, depending. And the messengers were politicized: Republicans said the threat was exaggerated, killing jobs and undermining the president. Democrats warned of disaster, urging more spending.

Early on, we had a clear goal: Don’t overwhelm hospitals. But that shifted, said Dr. Rajiv Bhatia, clinical assistant professor of primary care and population health at Stanford.

Source: Coronavirus: Six months after sheltering, why things fell apart

Public health messaging has been inconsistent, contradictory, and frequently incoherent throughout this entire event.

A 2006 paper by four epidemiologists explained that most public health mitigation steps do not actually work. First, the measures that might work – like lockdowns – are understood to be infeasible over wide regions or countries and not sustainable. Many other steps that seem intuitive do not work well or do not work at all, or have no evidence to support that they work.

The authors note that all pandemics eventually end – usually through herd effects, vaccines, virus mutates to less virulent form – or people eventually just get on with life. Hard lock downs and long term restrictions are not sustainable. The effect is that most public health mitigations eventually stop working.

The result is literal random effects from measures, followed by erroneous correlations that measure A worked. Many correlations neglect the time dimension. For example, if we compared China in February/March to the U.S., we could conclude that U.S. measures were working – until they were not. Similarly, Hawaii did “everything right” – lock downs, mandatory face masks since April, very strict travel restrictions, mandatory quarantines for travel to, from or between the islands – and arresting violators. It all worked until it didn’t.

In May and June, Hawaii was the prime example that public health mitigation measures work! Until August when it was obvious they failed. Numerous peer reviewed published papers ignored the time dimension – and neglected to recognize that places doing well tend to do poorly later (Hawaii), and those that did poorly tend to do much better later on (New York, Sweden).

When we look at these epicurves we see similar trends across the country. California did everything right – until it wasn’t right. Mandatory face masks in all populated counties since April. In the Spring, media widely reported that California did everything right compared to New York. And it was “everything right” until it wasn’t right anymore – like Hawaii, California’s cases exploded in August.

When presented the data, public health messaging and mitigation steps have been a disaster.

There is no evidence that places that did or are doing well are doing so because of these mitigation factors.

The problem is we look at this as a single variable problem when multiple variables are in play – such as the virus itself does what it does, and that we now understand that some demographics or geographic areas may have some degree of pre-existing protection against Covid-19. But we seek to simply to a single variable problem (one or more public health mitigations explain success or failure, or blame one or more politicians for not doing the right things) and ignore the time dimension as we seek patterns and correlations that are, in fact, meaningless.

Couple this with literally failed public health messaging and ultimately, the virus merely does what viruses do and will eventually end as past pandemics have ended.