Q: What do we know about the long-term consequences of Covid-19?
A: As of early December, Covid-19 infections have topped 64.66 million worldwide and an estimated 13.93 million in the US. We know that many infected people have minimal or no symptoms, some have a flu-like illness, and others may progress to a more severe illness that requires hospitalization, intensive unit care, and may die of complications such as lung disease, stroke or heart attacks.
We also know that Covid-19 infections span all age groups but that younger people have milder illness and the very old and those with underlying medical conditions such as diabetes, lung disease, or obesity are most vulnerable to becoming seriously ill.
However, we do not have as good a handle on what happens to infected people who have minimal or no symptoms and aren’t hospitalized, or the long-term complications of those who leave the hospital after a Covid-related illness. What is the incidence of heart attacks, strokes, and venous thromboembolism in this population?
There is accumulating evidence that post-hospitalization people may experience complications and symptoms including long-term fatigue, headaches, cognitive issues, and inflammatory conditions such as systemic lupus and arthritis.
There is also data suggesting that children and young adults may also experience unusual complications related to activation of the body’s immune and coagulation systems. While these complications appear to be rare, the accumulating evidence as the prevalence of COVID-19 infection increases, is that we will see more unusual near- and long-term complications.
Given the size of the pandemic, these will likely have important consequences for health care systems and policy makers.
Q: What are some of the data and analytics challenges?
A: A challenge in identifying and following Covid-19 infected patients who recover but are not hospitalized is the non-specificity of testing. In the acute period of infection, testing for the virus is a specific way of determining whether a person is infected. Following the resolution of the acute infection, which is generally within a few weeks after exposure, the body’s immune system develops antibodies which can be measured to determine if infection occurred. These antibodies may not last a long time, hampering identification of people who had a Covid-19 infection but were not tested with a direct viral test. This would clearly lead to an underestimate of persons at risk for post Covid-19 complications.
Much of the understanding we have of Covid-19 related illness, its complications, and outcomes, is based on real world data (RWD). This evidence may come from data collected by Johns Hopkins University, national and international agencies, and healthcare data from insurers and providers.
Q: Which types of questions can be addressed by RWD and analytics?
A: Health claims data collected by insurers, and aggregated by companies into large data sets inform trends of illness and complications across the United States. Data from Medicare and Medicaid beneficiaries is also useful in assessing outcomes in this vulnerable population. Data from electronic health records also may be helpful in confirming the diagnosis of Covid-19 and measuring clinical events.
There are many unanswered questions that real world data will be able to be used to inform us in the future. These questions include: What are the long-term consequences of post-Covid infections? Will there be illnesses similar to post-influenza Parkinson’s disease that will be found years after the infection? How will these illnesses affect the health care system in terms of demands for care, need for hospitalizations, specialized physician services, and long term care? What are the budgetary and policy implications? How would a vaccine affect these outcomes in various populations? How would better treatment affect these outcomes?
In addition to important policy issues and economic issues related to health care that post-Covid-19 infection complications raise, another area is how to value potential prevention strategies and treatments? For example, if a treatment with a drug or prevention with a vaccine proves to be effective in a certain populations with Covid-19 and/or prevents complications or hospitalization, then what is the long-term value of this drug?
Data science is a means of using real-world data to glean new insights into predictors of disease severity and associated outcomes. These data can inform about who is vulnerable to contract Covid-19, be hospitalized, admitted to an intensive care unit, placed on a ventilator, or survives. RWD can further be analyzed to the efficacy of interventions during and after hospitalization.
The tools that data scientists use include advanced platforms that rapidly analyze large data sets and apply a variety of statistical methods, as well as creative thinking, to generate and test hypotheses.
The potential for real world evidence and data science to inform policy makers, health care administrators, and researchers, as well as clinicians and patients, is limitless.
Q: What types of outcomes need to be collected directly from Covid-19 patients?
A: Some RWD on key patient outcomes may be more limited in common sources, such as claims databases and electronic health records in tracking the course of recovery in persons who are infected with the Covid-19 virus, but are either not hospitalized or the long term outcomes in post- hospitalized patients. For example, how debilitating are lingering symptoms of Covid-19, such as fatigue and shortness of breath? Has the patient been able to return to normal functioning, such as able to return to work? What is the patient’s quality of life?
To understand the full societal burden of Covid-19, it will be important to collect RWD directly from patients who are recovery from this devastating illness. The collaboration between private and public researchers and increased availability of rich RWD sources will lead to actionable insights and reduce human suffering from Covid-19 complications.
Mark Friedman is VP of Medical Affairs for Panalgo. He is an internist with over 20 years of clinical and pharmacoeconomics consulting experience. He has management experience in the formulary and managed care implementation processes and also plays a major role in the design and implementation of clinically-oriented outcomes research studies, including evidence-based medicine studies, clinical-trial–based evaluations, prospective outcomes studies, and economic models. He is a member of the American Board of Internal Medicine and the American College of Physicians. Dr. Friedman is a graduate of Columbia University and attended the Harvard School of Public Health.