Covid Fast Fax: A system for real-time triage of Covid-19 case report faxes

  1. Adam Lavertu
  2. Alison Stribling
  3. Matt White
  4. Greg McInnes
  5. Russ B. Altman
  6. Rajiv Pramanik
  7. Amit Kaushal

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

The scale and speed of the COVID-19 pandemic has strained many parts of the national healthcare infrastructure, including communicable disease monitoring and prevention. Many local health departments now receive hundreds or thousands of COVID-19 case reports a day. Many arrive via faxed handwritten forms, often intermingled with other faxes sent to a general fax line, making it difficult to rapidly identify the highest priority cases for outreach and monitoring. We present an AI-based system capable of real-time identification and triage of handwritten faxed COVID-19 forms. The system relies on two models: one model to identify which received pages correspond to case report forms, and a second model to extract information from the set of identified case reports. We evaluated the system on a set of 1,224 faxes received by a local health department over a two-week period. For the 88% of faxes of sufficient quality, the system detects COVID-19 reports with high precision, 0.98, and high recall, 0.91. Among all received COVID-19 faxes, the system identifies high priority cases with a specificity of 0.87, a precision of 0.46 and recall of 0.83. Our system can be adapted to new forms, after a brief training period. Covid Fast Fax can support local health departments in their efforts to control the spread of COVID-19 and limit its impact on the community. The tool is freely available.

Article activity feed

  1. ScreenIT

    SciScore for 10.1101/2020.12.15.20248256: (What is this?)

    Please note, not all rigor criteria are appropriate for all manuscripts.

    Table 1: Rigor

    NIH rigor criteria are not applicable to paper type.

    Table 2: Resources

    Software and Algorithms
    SentencesResources
    We used PyTorch 1.4.0 to construct and train several small convolutional neural networks8,9.
    PyTorch
    suggested: (PyTorch, RRID:SCR_018536)
    Pages that are predicted to be a relevant CMR are then registered to the predicted form template using the pyStackReg (https://github.com/glichtner/pystackreg) and imreg_dft (https://github.com/matejak/imreg_dft) Python packages.
    Python
    suggested: (IPython, RRID:SCR_001658)

    Results from OddPub: Thank you for sharing your code.

    Results from LimitationRecognizer: We detected the following sentences addressing limitations in the study:
    We recognize the system as it currently exists has several limitations. We are limited by the quality of the faxes being received and have no effective means to process low-quality faxes. This issue could be addressed through coordination between the LHD and regional healthcare providers to improve the quality of the faxes being sent. Additionally, image registration takes up a large amount of the current compute time of the system, while using a neural network to quickly identify which images should be registered provides a large improvement in system processing times, the need to retrain the neural network to identify new templates limits the ability of our system to generalize to other LHD in other states or counties out of the box. However, template detection and image registration are upstream of all other tasks and therefore it’s crucial they maintain high performance. Additionally, in the future medical efforts where a high volume of faxes may be received and AI-based assistance in triaging is desired, the usage of a standardized form that is designed for machine reading would further improve the performance of systems like the one outlined here. Key components would be isolating key information, such as checkboxes that are tied to high priority cases, away from other parts of the form would facilitate machine-based extraction. The inclusion of optical character recognition oriented written fields could make the deployment of text extraction easier, as model performanc...

    Results from TrialIdentifier: No clinical trial numbers were referenced.

    Results from Barzooka: We did not find any issues relating to the usage of bar graphs.

    Results from JetFighter: We did not find any issues relating to colormaps.

    Results from rtransparent:
    • Thank you for including a conflict of interest statement. Authors are encouraged to include this statement when submitting to a journal.
    • Thank you for including a funding statement. Authors are encouraged to include this statement when submitting to a journal.
    • No protocol registration statement was detected.

    About SciScore

    SciScore is an automated tool that is designed to assist expert reviewers by finding and presenting formulaic information scattered throughout a paper in a standard, easy to digest format. SciScore checks for the presence and correctness of RRIDs (research resource identifiers), and for rigor criteria such as sex and investigator blinding. For details on the theoretical underpinning of rigor criteria and the tools shown here, including references cited, please follow this link.

  2. Version published to 10.1101/2020.12.15.20248256v1 on medRxiv