1) Ask yourself if a number is necessary

DO THIS

A report on influenza observes that “Children’s Hospital in New Orleans is nearly full with children sick from the flu.”

DON’T DO THIS

An article on mpox in Massachusetts states that “the count of 15 new monkeypox cases in the Bay State was down from 17 infections during the previous week.”

DISCUSSION

Sometimes, it is best not to use numbers at all. Knowing that there were 17 cases of mpox one week and 15 the next does not really help readers understand the meaning of this data. Instead of these numbers, using textual descriptions — how these cases are affecting hospital capacity or how quickly they are spreading in an individual community — might more effectively communicate how widespread a given disease is.

2) When reporting disease counts, don’t just give the official number. Highlight that this is always going to be an undercount, and explain why. (Say what we don’t know.)

DO THIS

A report on mpox includes the following paragraph: “Today, more than 10,000 people in virtually every state in the U.S. have been infected with MPXV, according to the latest data from the Centers for Disease Control and Prevention. Confirmed cases in the U.S. alone are doubling every eight days, said Dr. Theresa Chapple, an epidemiologist and public health expert in Illinois. Epidemiologists say that estimate very likely undercounts actual infection rates.”

DON’T DO THIS

A report on mpox mentions that state health authorities “reported 15 new monkeypox cases,” and adds that the “total number” in Massachusetts is 396.

DISCUSSION

Reports that focus only on confirmed cases often paint a very incomplete picture of a disease’s true prevalence. This is because official figures are almost guaranteed to be undercounts. COVID has made this problem more obvious, but it is true for almost all infectious diseases. This is why it’s important to highlight gaps in morbidity and mortality data, and to provide estimates as to how many cases have likely gone unnoticed or unreported.

3) When reporting disease counts, look for ways to talk about testing rates as well

DO THIS

A report on COVID-19 notes that tracking the emergence of new variants is becoming more difficult, because “although some Americans are still getting PCR tests…reported testing levels are now at their lowest point in eight months, with numbers dropping by nearly 75% since the beginning of the year.”

DON’T DO THIS

A report states that COVID-19 case rates across the Americas have fallen “by 31% in the last week, marking a sixth successive week of declines” but offers no context about testing rates and how they could affect the data at hand.

DISCUSSION

Simply indicating the number of positive cases is not enough, because case counts are a function of testing (and in the case of COVID-19, a function of the availability of tests and people’s willingness to report on the results of self-testing). If testing itself is falling, then cases will seem to decline as well. This is why it’s as important to report on testing itself. Doing this can help people understand why case counts are always underestimated.

4) Make sure to acknowledge or explain the many different tests scientists use to measure how widespread a disease is

DO THIS

This reporting on COVID-19 uses multiple forms of data (including wastewater analyses, hospitalization rates, and case counts) to paint a picture of the current state of the epidemic. Its main findings are supported by multiple pieces of data (as can be seen in phrases like “which also reflects the increasing trend in COVID in wastewater in that region”).

DON’T DO THIS

An article refers to PCR tests as “the gold standard of testing,” without explaining what makes them superior to other means of assessing COVID-19’s incidence or prevalence.

DISCUSSION

There are many different ways to estimate how widespread a disease is. To help readers or viewers get a true picture of a disease’s prevalence, it’s best to talk about more than just the number of officially diagnosed cases. Look for ways to bring results from multiple kinds of testing into your reporting. If you are pressed for space or time and cannot mention all the other tests, then make sure to reference — even as briefly as in a clause — why the test you are citing is particularly revealing, effective or important.

5) Draw attention to the many different things that contribute to health disparities, and highlight potentially overlooked causes or alternative explanations

DO THIS

To show how differences in health outcomes among racial and ethnic groups are due to long-term structural racism, not biological or personal traits, a story profiles a Houston woman who, because of a variety of circumstances (low wages, high cost of housing, lack of health insurance) became homeless when she was diagnosed with breast cancer.

DON’T DO THIS

An article on COVID-19 death rates among African Americans and Latinos in the U.S. quantifies racial health disparities, but does not mention any of the structural variables (for example, the impact of racism on access to health care, education, employment) that drive these disparities.

DISCUSSION

Media reports discussing differences in mortality and morbidity rates between different groups sometimes fail to convey how structural forces (including racism, sexism, ableism, and other forms of systemic discrimination) play into these differences. Failure to do this can contribute to the mistaken impression that differences in infection or case fatality rates are purely a result of differences in behavior or “lifestyle” or (even worse) presumed biological differences. Of course, there are instances in which behaviors and biological factors influence health outcomes, but these are often a product of historical and social processes. For example, though Tay-Sachs disease is more prevalent among Ashkenazi Jews, the genetic cause is rooted in the historical isolation of this population group, not some innate, primordial biological factor. To avoid this implication, it is important to draw attention to the socioeconomic determinants of health–even if quantitative data on this is lacking.

Provide feedback.