Pneumococcal Vaccine Effectiveness
- Sage has completed two major projects on the effectiveness of Pfizer’s polyconjugate pneumococcal vaccines--PCV7, added to the infant vaccine schedule in 2000, and PCV13 an improved version added in 2010. Both were independent efforts funded by Pfizer.
In the first we used State Inpatient Database data to show that PCV7 very effectively cut hospitalizations and deaths from pneumococcal disease not only in kids, but their parents and grandparents as well; we published these results in mBio in 2011.
In the second study, we used very timely hospitalization data from a private vendor to get an early population-level look at whether the new vaccine, which covered six new pneumococcus serotypes, was working as hoped. Happily, it was. This study appeared in Lancet Respiratory Medicine in 2014.
A New Method to Track Vaccine Uptake
- As part of Sage’s work on pneumococcal vaccines, we invented a new way to track, in near-real time, uptake of newly introduced vaccines using readily available pharmaceutical sales data. We used the method to assess PCV7 and PCV13 vaccine coverage in the years after they first appeared on the market. But the method could be applied to any vaccine, including one introduced in response to an emerging infectious disease outbreak.
Global Mortality Burden of the 2009 Influenza Pandemic
- In collaboration with the Netherlands Institute for Health Services Research (NIVEL), Sage assembled a global research network to assess the global and regional numbers of deaths caused by the 2009 H1N1 influenza pandemic. The team solicited mortality data from collaborators in 26 countries, and used standard modeling techniques to assess pandemic mortality in 21 of these. We then applied a completely new statistical method to project single-country results to each WHO region and the world as a whole. The work, funded by the WHO, was published in PLoS Medicine in late 2013.
Burden of RSV and Influenza in the UK and US
- Under contract from GSK, Sage analyzed used “big data” to assess the burden of two important respiratory viruses, RSV and influenza, in the UK and US. We analyzed a total of six large datasets to estimate how many office visits, hospitalization, and deaths each of these viruses cause every year in these two countries. Papers from this effort will be published later in 2014.
Modeling Benefits of Quadrivalent Influenza Vaccine
- As part of our respiratory virus burden work for GSK, Sage created an Excel-based model of the potential benefits of going from a trivalent influenza vaccine, containing only one of two common influenza B strains to a quadrivalent vaccine, containing both. The model allowed easy exploration of various scenarios of coverage, effectiveness, and virus severity. Results will be published later in 2014.
Supporting a Web-based Portal for Influenza Data
- From 2009 through 2013, Sage was part of the team creating the Influenza Research Database. The project, funded by the National Institute for Allergy and Infectious Diseases (NIAID), collects all data related to influenza that it can and makes it easy for researchers to access and use. Sage supplied expertise in influenza epidemiology, conducted outreach to users, designed web tools for data analysis and improved site usability.
A cost/benefit model for pandemic planning
- Sage developed a computer-based model that allowed businesses to explore the costs and benefits of stockpiling antiviral drugs. The model, created for an international pharmaceutical company, incorporated both epidemiology and health economics modules. Sage Analytica devised the analytical strategy and programmed the model with an Excel-based user interface, then prepared a report explaining in lay language what the model does and how corporate executives should use it. The pharmaceutical client uses the model to encourage large corporations to improve their pandemic preparedness.
Training Software to Read Medical Text
- Sage Analytica partnered with Lockheed Martin in two efforts to train computer software to "read" and interpret medical text records. Lockheed supplied the computer experts and the text processing software, while Sage supplied subject matter expertise in statistical analysis, medicine, and medical communication; Sage also developed analytic strategies to optimize sensitivity and specificity of the system output.
- In the first project the team created a system to identify medical diagnoses of conditions, such as cardiovascular disease and congestive heart failure that are frequent co-morbidities of obesity, by analyzing jargon-filled hospital discharge reports. The system tied for first place in an academic contest in which more than 40 teams competed. The Lockheed/ Sage team was among the few participants invited to submit a paper to the Journal of the Association for Medical Informatics (JAMIA).
- The second project applied an improved system to an exhaustive library of discharge reports, progress reports, radiology reports, nurses’ notes and other text-based records for 33 men who died of prostate cancer. We tuned the system to map each patient’s experience of pain over the course of his illness, providing the first systematic look at the this important quality of life issue in cancer patients. The work, which included innovative ways to visualize complex longitudinal data, was published in JAMIA in 2013.