The MIDAS Webinar Series features research by MIDAS members, and is open to the public.
Date: Friday, June 17th, 2022
Time: 12:00pm – 1:00pm Eastern (USA)
Topic: Introduction to the Center for Forecasting and Outbreak Analytics (CFA)
Speaker: Dr. Marc Lipsitch, Director of Science, CFA
Professor, Harvard T.H. Chan School of Public Health
Marc Lipsitch, DPhil, serves as the director of science in CDC’s Center for Forecasting and Outbreak Analytics (CFA). Dr. Lipsitch received his BA from Yale University and DPhil from the University of Oxford. Dr. Lipsitch is seconded to CDC from the Harvard T.H. Chan School of Public Health, where he is professor of epidemiology and director of the Center for Communicable Disease Dynamics. He is an internationally recognized expert in infectious disease transmission modeling and has been a leading scientific authority in scientific research and public communication during the COVID-19 pandemic. His COVID-19 research has been in areas including the ethics of human challenge COVID-19 studies, transmission dynamics, basic epidemiology, clinical severity and sequelae, vaccine allocation, vaccine effectiveness, and equity. That work builds on his earlier research focus, which included antimicrobial resistance, epidemiologic methods, mathematical modeling of infectious disease transmission, bacterial and human population genetics and evolution, and molecular genetics of Streptococcus pneumoniae infection and immunity. He is an elected member of the American Academy of Microbiology and the National Academy of Medicine.
Follow-up Answers to Questions posted during the Webinar:
Question 1: Are there any opportunities for trainees (PhD to Post-doc level ) to engage? What about with existing fellowship programs at the CDC like EEP. Also, I’m curious about antibiotic resistance modeling with increased ability to attribute risk factor to specific sub-lineages (particularly of enteric bacteria) with risk factors associated with resistance – seems like an area where complex modeling could be utilized to assess resistance in particular pathogens.
Answer1: Yes for trainees. Post-PhD there is the Public Health Analytics and Modeling Fellowship https://www.cdc.gov/pef/analytics-and-modeling/index.html . There will be regular positions at CFA going up on USAJobs.gov soon for people at the GS9 to GS12 levels, which typically include master’s and recent PhDs. We are working on plans for rotators from academia and private sector to spend a period with CFA, but these don’t exist yet. Stay tuned.
On the resistance modeling issue – this is indeed an interesting area of research. CFA will be focusing for the moment on acute events (so far we have engaged with COVID-19, MPX, and acute pediatric hepatitis). It is possible that we may become involved in some AMR-related topics but primarily these are addressed by other parts of CDC and of course other agencies.
Question 2: How flexible are the job start dates? Especially coming from academia, we might have obligations that last to say, the end of fall semester.
Answer 2: They will be somewhat flexible. Sooner is better but we will be able to adapt to some extent to people’s schedules.
Question 3: Will the CFA provide prescriptive decision-analytic advice to local decision makers and help coordinate policy decisions? For instance, I would love to see CDC guidance be accompanied by cost-effectiveness analyses supporting important decisions.
Answer 3: As discussed briefly in the Q&A decision-analytic approaches such as cost-effectiveness will not be CFA’s focus; rather disease dynamic modeling for forecasting, scenarios, and specific decision problems but with a focus on the health outcomes. I believe it is likely we will move in that direction over time, but at present we are focused in this way.
Question 4: Are there any plans for integrating CFA roles with academic programs (i.e. PhD students/post-docs working for CFA)?
Answer 4: As noted under #1 this is something we are working on and consider valuable but have not yet formulated an approach.
Question 5: Are economic factors typically incorporated in the comparison of intervention effectiveness in your models?
Answer 5: So far not explicitly. See above.
Question 6: What types of outputs does the CFA plan to generate? Reports? Dashboard/apps? Journal publications? All of the above?
Answer 6: All of the above. To an extent this depends on the stage of the event we are addressing. Early on, qualitative and partly-quantitative technical reports and risk assessments, eg https://www.cdc.gov/ncird/investigation/hepatitis-unknown-cause/technical-report.html. Later, modeling and forecasting products (we have not yet taken on that function in a regular way; it is in progress). Our primary role is to support decision making by officials at various levels of government and eventually for the public, and our scientific activities are in the service of that mission. However, we have been involved in several publications so far (see the list on our webpage https://www.cdc.gov/forecast-outbreak-analysis/) . It is my personal belief that we will be most effective at our mission of decision support by being a scientifically vibrant organization whose members are deeply involved and play leading roles in the relevant scientific and technical communities, as is the case for many of the CDC’s disease-specific programs.
Question 7: Increasingly, it appears that ensemble approaches provide better forecast and projections. Will CFA compromise the quality of its work if it does not actively involve the diversity of approaches from the academic modeling community?
Answer 7: It is our intent to do so. We will bring an ensemble capability within CDC and will continue to engage with academic modelers. Some academic modelers have expressed a desire to reduce their effort in generating weekly forecasts but a willingness to share their approaches. Others who wish to are of course encouraged to continue their work, and we will be developing means of integrating these approaches.
Question 8: What preparation would you recommend for an undergrad looking to get into the field?
Answer 8: A combination of biology/public health (to understand how data are generated), coding/computation, modeling and its mathematical basis (calculus, linear algebra, probability and statistics). Also useful for some purposes: causal inference, communications/writing/dataviz, bioinformatics/evolution/phylogenetics.
Question 9: How will novel methods developed by the center, including model code, be disseminated to the public?
Answer 9: We are working on this. It will be, but the plans are still evolving.
Question 10: Can you comment on the time that CFA employees will have to prepare publications? If their primary work is generating rapid analyses for decision makers, that would seem to present a trade-off with the slower process of publishing?
Answer 10: See answer to #6 above.
Question 11: What does CFA consider the role of industry in data modernization and public health response?
Answer 11: CFA included an Industry Day in its April 2022 launch (https://www.youtube.com/watch?v=LznbzggUb3Q) to recognize the value of collaboration with industry and accelerate a dialogue we had already begun with industry. Clearly industry is a locus of innovation, technological change, and capabilities that are of great value to public health data modernization.
Question 12: Will the center confine its work to infectious diseases? How about other epidemics with contagion-like properties, such as drug use, or gun violence?
Answer 12: We are beginning with infectious disease emergencies but are expected and expecting to broaden the mandate as we staff up to include some noninfectious diseases.
Question 13: What is the approximate timeframe from application close date to hire date? I am a late-stage postdoc and wanted to apply to the most recent job posting, but ultimately didn’t because my current position ends in six months?
Answer 13: As noted in response to an earlier question, it is somewhat flexible. I would advise applying if these positions reopen.
Question 14: How will and other CFA employees manage conflicts of interest in decisions about hiring, awarding research funding, establishing collaborations, etc.?
Answer 14: CDC has procedures to manage this issue which will apply to CFA as to the rest of the agency.
Question 15: Can you talk more about the equity component?
Answer 15: As noted the commitment to design and present analyses in ways that enhance health equity is core to CDC and to CFA. We are working now to enumerate the ways in which we will fulfill this commitment. A still-incomplete list includes:
- Working with colleagues in incident responses and programs to enhance the collection of demographic data such as age, race/ethnicity, disability etc. to improve measurement of health disparities
- Designing studies that include diverse participants to ensure that data are gathered on groups that are often underrepresented.
- Designing analyses with a clear health equity focus: to identify the causes of disparate outcomes in order to inform effective measures to improve the health of those at greatest risk. The work of the UK Race Disparity Unit on COVID-19 (Four reports during the pandemic to date https://www.gov.uk/government/publications/final-report-on-progress-to-address-covid-19-health-inequalities) is exemplary in this regard.
- Compiling data in advance of future events on the cross-tabs of potential risk factors for infection and / or severity of outcome such as age, race/ethnicity, geography, and underlying conditions in order to inform vaccine and other prevention planning efforts.
Question 16: How can we get additional information about funding opportunities through the NSF and DoE?
Answer 16: For DoE the opportunities are internal to the US Government. From NSF, the RAPID grants need to be invited. Our ability to invite may vary over time, but you can inquire at email@example.com. Put NSF RAPID Inquiry in the subject line.