Tom Frieden was director of the Centers for Disease Control and Prevention from 2009 to 2017. He is president and chief executive of Resolve to Save Lives, part of the public health organization Vital Strategies, and senior fellow for global health at the Council on Foreign Relations.

Although it’s wonderful to see widened interest in epidemiological principles that just a few months ago were obscure, it’s alarming to see the exponential rise of not only the novel coronavirus but also of clueless opinions about how to track and halt the spread of disease. This endangers our efforts to get the epidemic under control while we reopen our economy. Here are six of the most egregious amateur epidemiology errors and five places we should focus our attention instead.

Cases. Obsession with case counts is misleading; we estimate that only about 10 to 15 percent of U.S. infections are diagnosed. Attempting to predict trends from this small fraction of cases without considering the distribution of cases within a community, who gets tested and how intensively testing is done is pointless.

Tests. Tracking the number of tests done also provides little useful information. It’s more useful to track the percent of tests that are positive and more useful still to monitor trends in test numbers and positivity rates. But most important is whether testing is done the right way: soon after patients feel sick; intensively in nursing homes and other congregate facilities; and followed by prompt isolation, contact tracing and quarantine.

Models. The many published models of how covid-19 might progress are based on varied assumptions and can change radically. Models can goad leaders into action and steer specific responses, but the appropriate use is to change the future — such as how many people will die — not predict it.

Reproductive number. The basic reproductive rate is a deceptively simple concept — how many people each case infects — and it can suggest whether control measures are working. But it is a rough estimate, based on untestable assumptions, and lags by at least a week; it is of limited utility for day-to-day monitoring or action.

Shifts in recommendations. When experts change their advice, they draw criticism. Although some changes reflect errors, many are responses to new, better information. Wearing masks is an example. As evidence of asymptomatic spread emerged, it became clear that infections can be reduced if people wear masks when they are within six feet of one another, particularly indoors. The changed recommendation was progress, not correction of a mistake.

Number of staff doing contact tracing. Tracing the contacts of infected people is crucial to stopping spread; focus on the number of contact tracers needed has become a distraction. I accept some blame for this: To indicate the scale of effort needed, I noted that for the United States to have, proportionally, the same tracing force as Wuhan, we would need up to 300,000. But far more than the number of staff, it’s the quality of the program that matters.

Here are five of the most important things we do need to track closely to understand the pandemic and improve our control measures.

Number of unlinked infections. These are rarely reported in the United States; countries with effective programs track them closely. Tracking the number of infections without an identified source case or event reveals the effectiveness of the contact tracing process. Areas with unlinked infections can control the virus by improving contact tracing and physical distancing.

Speed of isolating infected people. Testing the right people, getting results fast and finding and isolating patients immediately halts spread. There should be no more than three days from symptom onset to isolation.

Proportion of cases arising among quarantined contacts. This is the fundamental outcome indicator of a contact tracing program. If all new cases arise from among known, quarantined contacts, spread of disease stops.

Number of health-care worker infections. In the United States, more than 72,000 health-care workers have been infected and 400 have died. We must track and reduce this number to improve care of covid-19 and other health problems and to protect the people willing to risk their own health for the health of others.

Trend in excess mortality. Information on total deaths, compared weekly with historical trends and analyzed by age, race and ethnicity, gives essential information on what’s happening with both coronavirus (including undetected cases) and non-coronavirus health problems and helps target interventions.

The art and science of field epidemiology identifies where and how the virus is spreading and how to stop it. Overburdened public health staff have been distracted by having to generate numbers that have little meaning and less utility. During a recent conversation, one public health leader commented about indicators such as the proportion of cases arising from quarantined contacts, “If we reported those, it would be zero every day.” And that’s exactly the point. If public health is allowed to focus on doing the hard, meticulous work of field epidemiology and tracking meaningful indicators such as the five above, we will better understand and stop the virus. That will save lives and will restore livelihoods faster.

Watch Opinions videos:

Read more: