Data Are Not Enough to Reimagine Public Health

COVID-19 has revealed the weaknesses of health information systems worldwide. In many countries, included the United states, data were not missing. We have been rather overwhelmed with data, which have been highly accessible through open access repositories, like never before. From these data emerged many statistical analyses and predictions as well as comments in the media and social networks. Hence, on top of the viral epidemic, we have experienced a digital epidemic of information—reliable or false—resulting in an “infodemic.”¹ However, little information useful for decision making has been produced. Why is that? We believe it is because gathering and analyzing data cannot replace a true public health surveillance system.

In the current issue of AJPH, Brownson et al. (p. 1605) create a grim picture of what this crisis has uncovered about the state of public health. They highlight several major failures, including that our surveillance capabilities are insufficient and that public health actions are increasingly countered by the rapid spread of misinformation. “Surveillance is the foundation of assessment in the public health system,” the authors say, but the COVID-19 pandemic has revealed a “lack of investments in preparedness for surveillance” (p. 1606). We would like to stress that the failure of surveillance goes far beyond insufficient “capabilities” and “preparedness”; it is a system failure.

At the heart of the problem is the confusion between public health surveillance and health data science. Public health surveillance is effective when it produces information useful for decision making and action.² It does not consist merely in gathering data, conducting analyses, and making them available. Key is to make this information useful for decision making. Of course, we need to improve the quality, scope, and completeness of data, but this would be insufficient. Many health data scientists believe that properly analyzing data will produce useful information. They are, however, not connected with people who produce the data and are not trained to make their analyses useful to the people who need this information. There is a cultural gap between the world of health data science and the world of policymaking and applied public health. How models and predictions have misled us in this crisis is the result of this gap.³

As Brownson et al. remark, we do indeed need to “reinvent our public health systems” (p. 1605), notably by strengthening surveillance employing “21st century data sciences” (p. 1608). Data, however, will not be enough. Improving our public health surveillance systems requires policymakers and health data scientists to work together; they have to develop a common culture and agree on surveillance goals. Policymakers must be trained in surveillance principles and methods in this age of data science.² Health data scientists, beyond providing sound data analyses, must understand the constraints of the health care system and of public health, the needs of policymakers, and the population risk perception as well as the socioeconomic and political implications of their analyses.

We must have stronger public health surveillance systems if we are to succeed in the fight against COVID-19 and to manage future epidemics.