Applications of Digital Tools for Precision Public Health in the COVID-19 Era: Where Are We?

Two recent systematic scoping reviews explore the use and limitations of digital tools in public health surveillance and their applications to the pandemic response

Digital health tools offer increasing potential for substantial benefits to medicine and public health. In clinical practice, digital applications include personal wearable devices, devices within the body, and sensors that can be used to identify health status and aid disease diagnosis and management. Wearable devices can continuously measure and track health data for individuals, such as their pulse rate and physical activity level. In a recent blog, we discussed how most emerging digital tools, while promising, are not yet ready for routine clinical applications and need more rigorous evaluation.

For public health practice, digital tools—including internet-based platforms, social media, machine learning and predictive analytics—also have the potential to enhance the conduct of public health surveillance and response. Digital information can allow for more refined stratification of population health-related data by time, place, and personal characteristics—thus potentially allowing greater “precision” in public health response. As internet connectivity and social media are now part of our daily lives, harnessing digital data may provide deeper insights into disease occurrence and trends and even help us track the impact of health-related guidelines and policies. Digital applications relevant to using geospatial tracking of populations during disease outbreaks, providing local and near real-time information (e.g. contract tracing) to help recognize the emergence of an outbreak, and population-based clustering of behavioral risk factors such as physical inactivity, substance use, and poor diet. Most recently, digital tools and technologies have been applied to the public health response to the COVID-19 pandemic. In many cases, the technology has been deployed rapidly in response to the great need to obtain more precise and deidentified population level data regarding occurrence, spread, risk factors, morbidity, mortality, mobility, contact tracing, effectiveness of lockdowns, vaccine distribution and adverse outcomes, as well as differentiating subpopulations with increased susceptibility to infection and poor outcomes.

So what is the current status of research and evaluation of the use of digital tools in public health?

In the first paper, Abad et al. conducted a systematic scoping review on digital public health surveillance (DPHS). The authors assessed 755 articles that were published over the last 15 years. Studies were conducted in 54 countries and used 26 digital platforms related to 16 themes in public health surveillance. The most common focus area of the studies was infectious diseases and, less frequently, behavioral risk factors. The review noted the dearth of longitudinal studies and numerous methodological and practical limitations that can interfere with the integration of DPHS in traditional public health activities. The authors note limitations in information on internet users’ demographic characteristics, and that many studies did not stratify by geographic region. In total, they describe nine areas where the studies tended to have important limitations. Of the included studies, 40% had limited sample size and scope, 80% may have had content bias, none assessed health issues in vulnerable populations, 41% used biased or unrepresentative populations, and 36% did not segment results by geographic locations. Finally, only 6 studies showed how DPHS has been implemented in public health programs. The review uncovered many questions around validity and generalizability of social media and internet data. The authors noted that user-generated content often contains a large amount of noise, making automatic information extraction and classification of free-text data challenging. They also raised concerns about the accuracy of health-related digital data and the potential impact of misinformation on the public’s health (such as disease or vaccine specific information) .

In the second paper, Gunasekeran et al. conducted a systematic scoping review on digital applications in the first 6 months of the public health response to the COVID-19 pandemic. Applications included artificial intelligence (AI), telehealth, and other relevant digital health solutions. Their analysis included 247 studies. Major findings included identification of a relatively large body of scientific literature on applications of telehealth (n = 99 reports), AI (n = 111), and predictive analytics (n = 89). Potential gaps and challenges in public health were inferred from less frequently reported applications, such as the internet of things (n = 5 reports), digital platforms for communication (n = 27), and digital screening (n = 22). The authors note many shortcomings in study design that could limit generalizability and translation to health care settings and public health programs.

The two reviews provide valuable insights into the current limitations of digital health technologies in public health. Significant challenges remain in adoption, scale-up, and integration into healthcare systems and public health programs. Other considerations include equity issues in deployment of such technologies among disproportionately affected groups and in the potential for introducing measurement bias, especially for racial and ethnic minority groups– bias due to unvalidated AI algorithms. The reviews highlight the urgent need for more investigations and evaluation of the validity and utility of digital platforms including AI and predictive analytics methodologies. Finally, there are only a limited number of studies investigating patient and healthcare provider acceptance of digital tools, illustrating the need for more participatory research for more sustained adoption of these tools beyond the pandemic.

As a service to clinical and public health researchers and practitioners, our office continues to track the field of digital applications in public health. Our online searchable database (PHGKB) is updated daily and features curated articles in genomics and precision health, and highlights the most recent publications in this rapidly growing field. In addition, our COVID-19 GPH portal provides a ready resource to help keep up to date on emerging genomics and precision health information relevant to the pandemic. We encourage our readers to check our resources and provide feedback on the applications of digital tools in precision public health.

Please submit your comments here.