Understanding What Went Wrong

Besides the failure of too few coronavirus tests, why has America’s public health response been so ineffective? Two recent simulations provide important clues. The first examined the gold standard response to emerging outbreaks: case identification, contact tracing, and quarantine of cases. It estimated how effective contact tracing would have to be to stop travel-related cases from causing local transmission given various viral, public health, and outbreak factors. Based on an estimated viral transmissibility of R0 = 2.5, contact tracing would have needed to be nearly flawless (80% of all contacts traced) to effectively squelch outbreaks. However, if at 10% of cases were asymptomatic, then even perfectly executed contact tracing would have failed to stop 35% outbreaks (Figure 3). Emerging evidence suggests that 20% of COVID-19 infections could be asymptomatic. If accurate, these models suggest contact tracing wouldn’t have necessarily stopped the US outbreak unless combined with more restrictive social distancing measures and travel restrictions.

These results suggest that the US’s initial strategy of case identification and quarantine alone was unlikely to succeed because SARS-CoV-2 is highly transmissible and asymptomatic cases are prevalent. Perhaps these limitations could have been overcome had more tests been available, but this is not certain. Current anecdotal reports (here and here) suggest that PCR testing is only 70% sensitive meaning 30 out of every 100 patients with COVID-19 could be missed when tested. This also helps explain why patients with classic symptoms, chest X-ray, and CT findings may need to be re-tested if their initial test is negative.

The second model examined the effectiveness of a 2-week bout of social distancing on an established outbreak where local transmission is occurring. Social distancing measures included: quarantining cases and family members, quarantining cases and family members plus closing schools, quarantining cases and family members plus closing non-essential businesses or combining all 3 measures. Assuming viral transmissibility was R₀ = 2.5 and 22.7% of infections were asymptomatic, simple contract tracing and quarantine of family members would only reduce 13% of future transmissions. Adding school closures yielded minimal additional reductions. Adding work closures increased social distancing’s effectiveness and reduced 22% of future infections. Because school closures had minimal impact, combining all 3 measures did not achieve significant reductions over quarantine plus work closures (Figure 4). The finding that school closures had little impact bears further consideration given that attack rates appear to be similar in adults in children.

This second simulation reinforces the notion that COVID-19 is difficult to control because of its ease of transmission and its high number of asymptomatic cases. For these reasons, sustained social distancing over several months may be required to contain this outbreak; otherwise, our sacrifices will not “flatten the curve” and preserve our ability to treat all who need it.