Dean Samet Archive

Another good week for Colorado. All indicators for the COVID-19 pandemic continue to move downward as are those for influenza and respiratory syncytial virus (RSV). And, by my own highly subjective observations, Coloradans are mostly behaving as though the COVID-19 pandemic had left the state. It hasn’t. Last week, there were still 185 Coloradans hospitalized, a substantial and non-zero number. In contrast, that figure was almost 10-fold higher a year ago.
 
The evolving picture of the COVID-19 pandemic has led to a change in the Food and Drug Administration (FDA)’s recommendations for vaccination against SARS-CoV-2. Last week, the agency’s Vaccines and Related Biological Products Advisory Committee (VRBPAC) met to consider a single question:
 
Vaccine composition: Does the committee recommend harmonizing the vaccine strain composition of primary series and booster doses used in the U.S. to a single composition, e.g., the composition for all vaccines administered currently would be a bivalent vaccine (Original plus Omicron BA.4/BA.5)?
 
The intent is to simplify the now confusing vaccination landscape, which includes the original series of two doses against the initial strain and the subsequent boosters, now including the bivalent booster directed against the original strain and the Omicron variant. The committee’s vote was unanimous and affirmative for the question. The FDA also presented a process for selecting future strains to be targeted by vaccines. The shift signals the routinization of immunization against SARS-CoV-2, reflecting what may be the start of an era of endemicity.
 
Here is another indication of that shift. In March 2020, the Colorado School of Public Health established the Colorado COVID-19 Modeling Group to track the pandemic in the state as a basis for supporting decision-making by the Colorado Department of Public Health and Environment (CDPHE) and the governor’s office. The team drew from the Colorado School of Public Health, the University of Colorado School of Medicine, the University of Colorado Denver and Boulder, and Colorado State University. We used a standard type of epidemic model (SEIR—where S=susceptible, E=exposed, I=infected, and R=recovered) that we tailored to Colorado. With the model, we projected the course of the pandemic and explored the impacts of policy measures, such as mask mandates.
 
Quantitative infectious disease modeling has a long history, dating back at least to the early 20th century. Early compartment models, segmenting the population into susceptible, infected, and recovered, were developed by Kermack and McKendrick, and by Reed and Frost. They have long been applied to influenza epidemics and more recently to emerging infections. The field of infectious disease modeling has evolved greatly, propelled by the recent wave of emerging infections, including SARS-CoV-2. Modeling capacity has been accelerated by networks of investigators, e.g., MIDAS, and by growing computing capacity. Once again, the George Box quote remains relevant: “All models are wrong, but some are useful.” During the many months of challenges posed by a rising and falling COVID-19 epidemic curve, model results were useful for guiding decision-making. With the decline of COVID-19 in Colorado, the modeling effort has been halted—encouraging news.
 
Prompted by a positive review in Science, I just finished Escape from Model Land by Erica Thompson. The title gives away the message. Modelers create an artifice that may have little relation to the “real world.” This is not news for those of us who are denizens of “model land.” Modelers of any sort may find this book’s theme to be obvious. It did remind me that there are feedback loops between “model land” and reality. Model findings influence decision-making and models are re-tuned to the changes that resulted from model outputs. And, models are not value-free: as their creators should be aware of how they present the world; how the population is characterized in the model; and how findings in different directions may influence decision-making. For me, the chapter on modeling financial markets was informative and discouraging, as are my descending retirement accounts. This book may be of interest to non-modelers looking for a general, but opinionated, overview of modeling.

And, temporary good news: Representative George Santos has deferred taking House committee positions, including his appointment to the House Science, Space, & Technology Committee.
 
Stay well and vaccinated.

Jonathan Samet, MD, MS
Dean, Colorado School of Public Health

Another good week for Colorado. All indicators for the COVID-19 pandemic continue to move downward as are those for influenza and respiratory syncytial virus (RSV). And, by my own highly subjective observations, Coloradans are mostly behaving as though the COVID-19 pandemic had left the state. It hasn’t. Last week, there were still 185 Coloradans hospitalized, a substantial and non-zero number. In contrast, that figure was almost 10-fold higher a year ago.
 
The evolving picture of the COVID-19 pandemic has led to a change in the Food and Drug Administration (FDA)’s recommendations for vaccination against SARS-CoV-2. Last week, the agency’s Vaccines and Related Biological Products Advisory Committee (VRBPAC) met to consider a single question:
 
Vaccine composition: Does the committee recommend harmonizing the vaccine strain composition of primary series and booster doses used in the U.S. to a single composition, e.g., the composition for all vaccines administered currently would be a bivalent vaccine (Original plus Omicron BA.4/BA.5)?
 
The intent is to simplify the now confusing vaccination landscape, which includes the original series of two doses against the initial strain and the subsequent boosters, now including the bivalent booster directed against the original strain and the Omicron variant. The committee’s vote was unanimous and affirmative for the question. The FDA also presented a process for selecting future strains to be targeted by vaccines. The shift signals the routinization of immunization against SARS-CoV-2, reflecting what may be the start of an era of endemicity.
 
Here is another indication of that shift. In March 2020, the Colorado School of Public Health established the Colorado COVID-19 Modeling Group to track the pandemic in the state as a basis for supporting decision-making by the Colorado Department of Public Health and Environment (CDPHE) and the governor’s office. The team drew from the Colorado School of Public Health, the University of Colorado School of Medicine, the University of Colorado Denver and Boulder, and Colorado State University. We used a standard type of epidemic model (SEIR—where S=susceptible, E=exposed, I=infected, and R=recovered) that we tailored to Colorado. With the model, we projected the course of the pandemic and explored the impacts of policy measures, such as mask mandates.
 
Quantitative infectious disease modeling has a long history, dating back at least to the early 20th century. Early compartment models, segmenting the population into susceptible, infected, and recovered, were developed by Kermack and McKendrick, and by Reed and Frost. They have long been applied to influenza epidemics and more recently to emerging infections. The field of infectious disease modeling has evolved greatly, propelled by the recent wave of emerging infections, including SARS-CoV-2. Modeling capacity has been accelerated by networks of investigators, e.g., MIDAS, and by growing computing capacity. Once again, the George Box quote remains relevant: “All models are wrong, but some are useful.” During the many months of challenges posed by a rising and falling COVID-19 epidemic curve, model results were useful for guiding decision-making. With the decline of COVID-19 in Colorado, the modeling effort has been halted—encouraging news.
 
Prompted by a positive review in Science, I just finished Escape from Model Land by Erica Thompson. The title gives away the message. Modelers create an artifice that may have little relation to the “real world.” This is not news for those of us who are denizens of “model land.” Modelers of any sort may find this book’s theme to be obvious. It did remind me that there are feedback loops between “model land” and reality. Model findings influence decision-making and models are re-tuned to the changes that resulted from model outputs. And, models are not value-free: as their creators should be aware of how they present the world; how the population is characterized in the model; and how findings in different directions may influence decision-making. For me, the chapter on modeling financial markets was informative and discouraging, as are my descending retirement accounts. This book may be of interest to non-modelers looking for a general, but opinionated, overview of modeling.

And, temporary good news: Representative George Santos has deferred taking House committee positions, including his appointment to the House Science, Space, & Technology Committee.
 
Stay well and vaccinated.

Jonathan Samet, MD, MS
Dean, Colorado School of Public Health