Back

A Novel Approach to Data Driven Pandemic Recovery: The Pandemic Recovery Acceleration Model

Gold, J. P.; Wichman, C.; Bayles, K.; Khan, A. S.; Kratochvil, C.; Lawler, J. V.; Lowe, J. M.; Schwedhelm, S.; Grimm, B.

2020-05-22 public and global health
10.1101/2020.05.17.20104695 medRxiv
Show abstract

A data driven approach to guide the global, regional and local pandemic recovery planning is key to the safety, efficacy and sustainability of all pandemic recovery efforts. The Pandemic Recovery Acceleration Model (PRAM) analytic tool was developed and implemented state wide in Nebraska to allow health officials, public officials, industry leaders and community leaders to capture a real time snapshot of how the COVID-19 pandemic is affecting their local community, a region or the state and use this novel lens to aid in making key mitigation and recovery decisions. This is done by using six commonly available metrics that are monitored daily across the state describing the pandemic impact: number of new cases, percent positive tests, deaths, occupied hospital beds, occupied intensive care beds and utilized ventilators, all directly related to confirmed COVID-19 patients. Nebraska is separated into six Health Care Coalitions based on geography, public health and medical care systems. The PRAM aggregates the data for each of these geographic regions based on disease prevalence acceleration and health care resource utilization acceleration, producing real time analysis of the acceleration of change for each metric individually and also combined into a single weighted index, the PRAM Recovery Index. These indices are then shared daily with the state leadership, coalition leaders and public health directors and also tracked over time, aiding in real time regional and statewide decisions of resource allocation and the extent of use of comprehensive non-pharmacologic interventions.

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.